"Pharmaceutical Sciences, Faculty of"@en . "DSpace"@en . "UBCV"@en . "Zhang, Tingting"@en . "2013-07-30T09:15:36Z"@en . "2013"@en . "Doctor of Philosophy - PhD"@en . "University of British Columbia"@en . "Background: Drug therapy is the mainstay medical treatment for asthma patients. Many asthma patients (up to 70%) receive suboptimal drug therapy. Inadequate use of inhaled corticosteroids (ICS) has been associated with increased emergency department (ED) visits and hospital admissions for asthma. To understand patients\u00E2\u0080\u0099 asthma drug use in British Columbia (B.C.) and improve health outcomes, this study describes the burden of asthma, identifies patients who received suboptimal asthma drug regimens according to asthma clinical practice guidelines, and examines the link between regimen optimality and health services utilization for asthma in an entire population with treated asthma in BC from 1996 to 2009.\nMethods: A cohort of 336,901 asthma patients between 5\u00E2\u0080\u009355 years of age was identified using provincial health services utilization data from 1996 to 2009. Annual patient medication dispensings of short-acting bronchodilators (SABA) with or without ICS were categorized into optimal or suboptimal regimens based on the asthma clinical practice guidelines. The association between regimen optimality and health services utilization was examined in one-year, as well as during a 14-year study period, using logistic regression models and Cox Proportional regression models, respectively.\nResults: The prevalence (~2%) and incidence (0.7%) of asthma was stable in patients 5-55 years of age in B.C. from 1996 to 2009. Asthma-related specialist visits, ED visits and hospital admissions declined by over 50% during the study period. In 2009, patients with suboptimal regimens had significantly greater risk of using health services than patients with optimal regimens of SABA and/or ICS. Over time, switching from a suboptimal to an optimal drug regimen was associated with a 30% reduction in the use of hospital services for asthma\niii\n(hazard ratio (HR) 0.71; 95% CI 0.54 \u00E2\u0080\u0093 0.93), and a 50% reduction in the use of ED services for asthma (HR 0.49; 95% CI 0.33 \u00E2\u0080\u0093 0.73).\nConclusions: Much of the healthcare burden associated with asthma is preventable by optimizing drug therapy, in particular, with improved ICS adherence. Identifying patients with suboptimal asthma management practices is a critical step in reducing the burden of asthma on the healthcare system and ultimately improving the quality of life of asthma patients."@en . "https://circle.library.ubc.ca/rest/handle/2429/44726?expand=metadata"@en . "ASTHMA DRUG REGIMEN OPTIMALITY AND HEALTH SERVICES UTILIZATION: A POPULATION-BASED ANALYSIS IN BRITISH COLUMBIA by TINGTING ZHANG B.Sc, Capital University of Medical Sciences, 2005 M.Sc, Simon Fraser University, 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Pharmaceutical Sciences) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) July 2013 \u00C2\u00A9 Tingting Zhang, 2013 ii ABSTRACT Background: Drug therapy is the mainstay medical treatment for asthma patients. Many asthma patients (up to 70%) receive suboptimal drug therapy. Inadequate use of inhaled corticosteroids (ICS) has been associated with increased emergency department (ED) visits and hospital admissions for asthma. To understand patients\u00E2\u0080\u0099 asthma drug use in British Columbia (B.C.) and improve health outcomes, this study describes the burden of asthma, identifies patients who received suboptimal asthma drug regimens according to asthma clinical practice guidelines, and examines the link between regimen optimality and health services utilization for asthma in an entire population with treated asthma in BC from 1996 to 2009. Methods: A cohort of 336,901 asthma patients between 5\u00E2\u0080\u009355 years of age was identified using provincial health services utilization data from 1996 to 2009. Annual patient medication dispensings of short-acting bronchodilators (SABA) with or without ICS were categorized into optimal or suboptimal regimens based on the asthma clinical practice guidelines. The association between regimen optimality and health services utilization was examined in one-year, as well as during a 14-year study period, using logistic regression models and Cox Proportional regression models, respectively. Results: The prevalence (~2%) and incidence (0.7%) of asthma was stable in patients 5-55 years of age in B.C. from 1996 to 2009. Asthma-related specialist visits, ED visits and hospital admissions declined by over 50% during the study period. In 2009, patients with suboptimal regimens had significantly greater risk of using health services than patients with optimal regimens of SABA and/or ICS. Over time, switching from a suboptimal to an optimal drug regimen was associated with a 30% reduction in the use of hospital services for asthma iii (hazard ratio (HR) 0.71; 95% CI 0.54 \u00E2\u0080\u0093 0.93), and a 50% reduction in the use of ED services for asthma (HR 0.49; 95% CI 0.33 \u00E2\u0080\u0093 0.73). Conclusions: Much of the healthcare burden associated with asthma is preventable by optimizing drug therapy, in particular, with improved ICS adherence. Identifying patients with suboptimal asthma management practices is a critical step in reducing the burden of asthma on the healthcare system and ultimately improving the quality of life of asthma patients. iv PREFACE This dissertation is an original intellectual product of the author, Tingting Zhang. A version of Chapter 3 [1] has been published [Zhang T, Smith A, Camp P and Carleton B. Asthma drug regimen optimality and health service utilization: a population-based study in British Columbia. Pharmacoepidemiology & Drug Safety. 2013 Apr 5. doi: 10.1002/pds.3444]. I am the first author, responsible for the study design, literature review, data clean and statistical analyses, results preparation and interpretation, manuscript writing and revisions. Ms. Anne Smith contributed to study design and interpretation of results. Dr. Pat Camp contributed to study design and interpretation of results. Dr. Bruce Carleton was the supervisory author on this project and was involved throughout the project in study design, interpretation of results, and extensive review of the manuscript. This research was approved by The University of British Columbia Clinical Research Ethics Board. The UBC CREB number is H04-80927. Chapter 1. Figure 1.1 \u00E2\u0080\u009CMechanisms of airway inflammation in patients with asthma\u00E2\u0080\u009D is used with permission from the McGraw-Hill Education. Chapter 1. Figure 1.2 and Figure 1.3 \u00E2\u0080\u009CStepwise approach to therapy for managing asthma\u00E2\u0080\u009D are used with permission from the National Institutes of Health of the United States. Permission was obtained from the National Heart, Lung and Blood Institute Health Information Center. v TABLE OF CONTENTS ABSTRACT ............................................................................................................................... ii PREFACE ................................................................................................................................. iv TABLE OF CONTENTS ......................................................................................................... v LIST OF TABLES .................................................................................................................viii LIST OF FIGURES .................................................................................................................. x LIST OF ABBREVIATIONS ...............................................................................................xiii ACKNOWLEDGEMENTS .................................................................................................. xvi CHAPTER 1: INTRODUCTION ............................................................................................ 1 1.1 DEFINITION OF ASTHMA ........................................................................................ 2 1.2 PATHOPHYSIOLOGY ............................................................................................... 2 1.2.1 Bronchoconstriction\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...2 1.2.2 Airway inflammation\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.3 1.2.3 Airway remodeling\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.6 1.3 ETIOLOGY OF ASTHMA .......................................................................................... 6 1.3.1 Host factors\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A67 1.3.2 Environmental factors\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A69 1.4 EPIDEMIOLOGY OF ASTHMA .............................................................................. 10 1.4.1 Challenges in asthma burden estimates\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A611 1.4.2 Limitations in current asthma case definitions\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.12 1.5 DIAGNOSIS OF ASTHMA ....................................................................................... 14 1.6 GOAL OF ASTHMA TREATMENT ........................................................................ 16 1.7 TREATMENT OF ASTHMA .................................................................................... 16 1.7.1 Inhaled corticosteroids\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.17 1.7.2 Leukotriene receptor antagonists (LTRAs)\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..23 1.7.3 Mast cells stabilizers\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.25 1.7.4 Methylxanthines\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A626 1.7.5 Anti-IgE agents\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.26 1.7.6 Long-acting \u00CE\u00B2 agonists (LABA)\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...26 1.7.7 Systemic corticosteroids\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...27 1.7.8 Short-acting bronchodilators\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A628 1.7.9 Anti-cholinergic medications\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A629 1.8 ASSESSMENT OF ASTHMA CONTROL ............................................................... 30 1.9 CLINICAL MANAGEMENT OF ASTHMA ............................................................ 36 1.9.1 Long-term control\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.36 1.9.2 Managing exacerbations of asthma\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..40 vi 1.10 ADHERENCE TO ASTHMA DRUG REGIMENS .................................................. 43 1.10.1 Regimen adherence\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..43 1.10.2 Reasons for non-adherence\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..47 1.11 PROGNOSIS OF ASTHMA ...................................................................................... 51 1.12 RESEARCH GOALS, HYPOTHESIS AND OBJECTIVES ..................................... 53 CHAPTER 2: BURDEN OF ASTHMA IN BRITISH COLUMBIA ................................. 55 2.1 SYNOPSIS ................................................................................................................. 56 2.2 METHODOLOGY ..................................................................................................... 57 2.2.1 Data Sources\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.57 2.2.2 Quality of data\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..62 2.2.3 Data preparation\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...64 2.2.4 Cohort definition\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..65 2.2.5 Validity of asthma case definitions\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..67 2.2.6 ICD-9 and ICD-10 diagnostic codes for asthma\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..70 2.2.7 Estimation of asthma burden in B.C. 1996-2009\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..71 2.3 RESULTS................................................................................................................... 77 2.3.1 Identification of patients with treated asthma\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..77 2.3.2 Overall burden of asthma\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.81 2.3.3 Burden of asthma by age group and gender\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.86 2.3.4 Burden of asthma by HSDA\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.93 CHAPTER 3: ASTHMA REGIMEN OPTIMALITY AND HEALTH SERVICES UTILIZATION\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6. ...................................................................................................... 96 3.1 SYNOPSIS ................................................................................................................. 97 3.2 METHODOLOGY ..................................................................................................... 98 3.2.1 Cross-sectional study design\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A698 3.2.2 Retrospective cohort study design\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..105 3.3 RESULTS................................................................................................................. 114 3.3.1 Results based on the cross-sectional study design\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.114 3.3.2 Results based on the cohort study design\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...119 CHAPTER 4: CHANGING FROM SUBOPTIMAL TO OPTIMAL ASTHMA DRUG REGIMENS: SIGNIFICANT IMPROVEMENTS IN HEALTH SERVICES UTILIZATION OUTCOMES ............................................................................................. 128 4.1 SYNOPSIS ............................................................................................................... 129 4.2 METHODOLOGY ................................................................................................... 131 4.2.1 Study design and study patients\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.131 4.2.2 Exposure assessment-measure of changes in regimen optimality\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..131 4.2.3 Research outcomes assessment\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..134 vii 4.2.4 Statistical analysis\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..134 4.2.5 Sensitivity analyses\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6136 4.3 RESULTS................................................................................................................. 137 4.3.1 Baseline characteristics\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..137 4.3.2 Switching from suboptimal to optimal regimens\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...138 4.3.3 Health services use associated with switching from suboptimal to optimal regimens...........................................................................................................................140 4.3.4 Results from sensitivity analyses\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...147 CHAPTER 5: REGULAR VS. INTERMITTENT USE OF INHALED CORTICOSTEROIDS ......................................................................................................... 149 5.1 SYNOPSIS ............................................................................................................... 150 5.2 METHODOLOGY ................................................................................................... 151 5.2.1 Study design and study patients\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.151 5.2.2 Study measures\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...153 5.2.3 Data analysis\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..156 5.2.4 Sensitivity analysis\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.164 5.3 RESULTS................................................................................................................. 165 5.3.1 Baseline characteristics\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..165 5.3.2 Comparison between regular and intermittent use of ICS\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6.169 5.3.3 Sensitivity analyses\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6174 CHAPTER 6: CONCLUDING CHAPTER ....................................................................... 179 6.1 KEY FINDINGS ...................................................................................................... 180 6.1.1 Key findings in the asthma burden estimate project\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..180 6.1.2 Key findings in regimen optimality and health services utilization project...189 6.1.3 Key findings in the comparison between regular & intermittent use of ICS.196 6.2 STUDY STRENGTHS ............................................................................................. 197 6.3 STUDY LIMITATIONS .......................................................................................... 199 6.3.1 Selection biases\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..199 6.3.2 Information biases\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6..199 6.3.3 Confounding\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6\u00E2\u0080\u00A6...202 6.4 FURTHER RESEARCH .......................................................................................... 203 6.5 POLICY IMPLICATIONS ...................................................................................... 206 6.6 CONCLUSIONS ...................................................................................................... 208 BIBLIOGRAPHY ................................................................................................................. 210 APPENDIX ............................................................................................................................ 228 APPENDIX A CHAPTER 2 ADDITIONAL TABLES .................................................................... 229 viii LIST OF TABLES Table 1.1 Factors affecting asthma control level .................................................................... 33 Table 2.1 Data sources of the present research ....................................................................... 58 Table 2.2 ICD-9 and ICD-10 codes for respiratory conditions tracked in the asthma cohort 60 Table 2.3 Prescription medications for the treatment of asthma by class ............................... 66 Table 2.4 Prevalence of asthma in British Columbia aged 5-55 years of age, in 1996, 2003 and 2009......................................................................................................................................... 83 Table 2.5 Patients with newly diagnosed asthma in British Columbia aged 5-55 years of age, in 2001, 2005 and 2009 ........................................................................................................... 84 Table 2.6 Use of family physician services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 ......................................................................................... 88 Table 2.7 Use of specialist services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 ......................................................................................... 90 Table 2.8 Use of hospital services for in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 ............................................................................................................... 91 Table 2.9 Use of ED services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 ............................................................................................................... 92 Table 3.1 Regimen optimality classification for asthma patients aged 12 years or older .... 103 Table 3.2 Propensity score variables and association with suboptimal use of asthma regimens ............................................................................................................................................... 109 Table 3.3 Characteristics of study patients ........................................................................... 115 Table 3.4 Comparison of health service utilization between patients with suboptimal and optimal regimens ................................................................................................................... 117 Table 3.5 Unadjusted and LABA use-adjusted ORs in the comparison of health service utilization between patients with suboptimal and optimal regimens .................................... 118 Table 3.6 Comparison of health service utilization between patients with suboptimal and optimal regimens in each age group ..................................................................................... 118 Table 3.7 Characteristics of patients before and after propensity score matching ............... 121 Table 3.8 Adjusted hazard ratios for ED and hospital admissions associated with suboptimal and optimal use of asthma regimens ..................................................................................... 124 ix Table 3.9 Adjusted hazard ratios for ED and hospital admissions associated with suboptimal and optimal use of asthma regimens by age group ............................................................... 127 Table 4.1 Characteristics of study patients (n=4,059) .......................................................... 138 Table 4.2 Characteristics of patients over the one year prior to their first change from suboptimal to optimal asthma drug regimens during follow-up ........................................... 140 Table 4.3 Unadjusted and adjusted hazard ratios and 95% confidence intervals for risk of ED visits for asthma exacerbations ............................................................................................. 143 Table 4.4 Unadjusted and adjusted hazard ratios and 95% confidence intervals for risk of hospital admissions for asthma exacerbations ...................................................................... 144 Table 5.1 Propensity score variables and association with using ICS regularly .................. 159 Table 5.2 Characteristics of patient groups before and after propensity score matching (based on the 90-day definition) ....................................................................................................... 167 Table 5.3 Adjusted hazard ratios for ED and hospital admissions associated with regular (90 days definition) and intermittent use of ICS ......................................................................... 170 Table 5.4 Characteristics of patient groups before and after propensity score matching (based on the 120-day definition) ..................................................................................................... 176 Table 5.5 Adjusted hazard ratios for ED and hospital admissions associated with regular (120 days definition) and intermittent use of ICS ......................................................................... 178 x LIST OF FIGURES Figure 1.1 Mechanisms of airway inflammation in patients with asthma (copied from Goodman & Gilman\u00E2\u0080\u0099s The Pharmaceutical Basis of Therapeutics, 12th Edition with permission) [2]. ........................................................................................................................... 4 Figure 1.2 Stepwise approach to therapy for managing asthma in patients 5-11 years of age (copied from the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma, Expert Panel Report-3 with permission) [3] .................................... 38 Figure 1.3 Stepwise approach to therapy for managing asthma in patients 12 years or older (copied from the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma, Expert Panel Report-3 with permission) [3] .................................... 39 Figure 2.1 Form of study cohort ............................................................................................... 67 Figure 2.2 Number of patients that met each case definition criteria in B.C. 5-55 years of age, 1996 - 2009 ............................................................................................................................... 79 Figure 2.3 Prevalence of asthma based on different asthma case definitions in B.C aged 5-55 years, 1996 \u00E2\u0080\u0093 2009 .................................................................................................................... 80 Figure 2.4 Incidence of asthma based on different asthma case definitions in B.C. aged 5-55 years, 1996 - 2009 ..................................................................................................................... 80 Figure 2.5 Trends of using health services for asthma in patients with treated asthma in B.C. aged 5-55 years, 1996 - 2009 .................................................................................................... 82 Figure 2.6 Percentage of patients dispensed asthma medications in B.C. aged 5-55 years, 1996 \u00E2\u0080\u0093 2009........................................................................................................................................ 86 Figure 2.7 Use of family physician services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 .......................................................................................................... 94 xi Figure 2.8 Use of specialist services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 ........................................................................................................................ 94 Figure 2.9 Use of hospital services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 ............................................................................................................................... 95 Figure 2.10 Use of ED services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 ............................................................................................................................... 95 Figure 3.1 Asthma regimen optimality classification (patient ages 12-55 years) ................... 101 Figure 3.2 Asthma regimen optimality classification (patient ages 5-11 years) ..................... 102 Figure 3.3 Study population in the comparison of health services utilization between patients with suboptimal and optimal regimens ................................................................................... 106 Figure 3.4 Standardized difference in the distribution of each confounding variable before and after propensity score matching .............................................................................................. 110 Figure 3.5 QQ plots \u00E2\u0080\u0093 assessment of propensity score matching ........................................... 111 Figure 3.6 Log minus log plot to assess Cox proportional hazard model assumption ........... 114 Figure 3.7 Cumulative incidence for asthma-related ED visits for propensity score matched cohort ...................................................................................................................................... 125 Figure 3.8 Cumulative incidence for asthma-related hospital admissions for propensity score matched cohort ........................................................................................................................ 125 Figure 4.1 Examples of study scenarios ................................................................................. 133 Figure 4.2 Association between the number of optimal regimen years and the likelihood of hospital admission for asthma ................................................................................................. 146 Figure 4.3 Association between the number of optimal regimen years and the likelihood of ED visits for asthma ...................................................................................................................... 146 xii Figure 5.1 Form of study patients in the comparison of health services utilization between patients with regular and intermittent use of ICS ................................................................... 153 Figure 5.2 Distribution of ICS dispensings in patients who used ICS intermittently during the index year ................................................................................................................................ 155 Figure 5.3 Standardized difference in the distribution of each confounding variable before and after propensity score matching .............................................................................................. 160 Figure 5.4 Quantile-quantile plots: evaluation of propensity score matching ........................ 161 Figure 5.5 Log minus log plot to asses Cox model hazard function assumption ................... 164 Figure 5.6 Cumulative incidence curves of hospital admissions for unmatched cohort ........ 171 Figure 5.7 Cumulative incidence curves of hospital admissions for matched cohort ............ 172 Figure 5.8 Cumulative incidence curves of emergency department visits for unmatched cohort ................................................................................................................................................. 173 Figure 5.9 Cumulative incidence curves of emergency department visits for matched cohort ................................................................................................................................................. 174 xiii LIST OF ABBREVIATIONS ACG Adjusted Clinical Group ADAM A Disintegrin and Metalloproteinase 33 ADRs Adverse Drug Reactions AMP Adenosine Monophosphate APP Alternative Payment Program ATP Adenosine Triphosphate BC British Columbia BDP Beclomethasone Dipropionate CI Confidence Interval CHSPR Centre of Health Services and Policy Research CIHI Canadian Institute for Health Information COPD Chronic Obstructive Pulmonary Disease DAD Discharge Abstract Database DPP10 Dipeptidyl Peptidase 10 ED Emergency Department EDC Expanded Diagnosis Cluster EPR Expert Panel Report FDA Food and Drug Administration FEV1 Forced Expiratory Volume in one second FFS Fee for Service FP Family Physicians xiv FVC Forced Vital Capacity GRPA G-Protein-Coupled Receptor for Asthma Susceptibility GI Gastrointestinal HA Health Authorities HR Hazard Ratio HSDA Health Service Delivery Areas ICD-9 International Classification of Diseases \u00E2\u0080\u0093 version 9 ICD-10 International Classification of Diseases \u00E2\u0080\u0093 version 10 ICS Inhaled Corticosteroids ICU Intensive Care Unit IgE Immunoglobulin E IL Interleukin IUC Inhaler Use Checklist LABA Long Acting Beta Agonist LCM Latent Class Modeling LHA Local Health Areas LTRA Leukotriene Rceptor Antagonists MDI Metered Dose Inhaler MPR Medication Possession Ratio MSP Medical Services Plan NCHS National Center for Health Statistics NHLBI National Health Heart, Lung and Blood Institute xv OR Odds Ratio PEOPLE 34 Population Extrapolation for Organization Planning with Less Error Version 34 PHF11 Plant Homeodomain Finger protein 11 POPi Pharmaceutical Outcomes Programme PPDC Proportion of Prescribed Days Covered QQ plot Quantile-Quantile Plot RCT Randomized Controlled Trial RSV Respiratory Syncytial Virus SABA Short-Acting Bronchodilators SD Standard Deviation SES Socioeconomic Status SPINK5 Serine Peptidase Inhibitor Kazal Type 5 SPSS Statistical Package for the Social Sciences Th T Cell Helper US United States UK United Kindom URTI Upper Respiratory Tract Infections VC Vital Capacity VHC Valved Holding Chambers xvi ACKNOWLEDGEMENTS My research and thesis would not have been possible without continuous guidance, support, encouragement from my supervisory committee members, colleagues, family members and friends. My sincerest gratitude to my supervisory committee \u00E2\u0080\u0093 Drs Bruce Carleton, Ujendra Kumar, Ran Goldman and Craig Mitton \u00E2\u0080\u0093 your guidance, suggestions, extreme support and expertise has enriched my training. I thank the Drug Safety and Effectiveness Cross-Disciplinary Training (DSECT) program and the Canadian Institutes of Health Research for funding support for my PhD projects. The Pharmaceutical Outcomes Programme (POPi) has seen me through both academic and life successes and challenges, and I deeply appreciate everyone in this great group for helping me grow, not just as a researcher, but also as a person. To Dr. Bruce Carleton for your supervision and support to my academic and personal life, your encouragement and support during my difficult time in my PhD study period; Ms. Anne Smith and Ms. Gabriella Groeneweg for making everything at POPi work so all I have to do is my research, Dr. Ricardo Jimenez-Mendez, Dr. Benji Heran and Dr. Mahyar Etminan for methodological, career and life advice, Hao Luo for statistical assistance with more complicated statistical models and use of statistical software, Claudette Hildebrand, Ursula Amstutz, Kaitlyn Shaw, Adrienne Borrie, Shevaun Hughes for support and presentation skills. I thank Dr. Stuart MacLeod for great support in my national and international network building, career and life advice, Ms. Ruth Milner for statistical expertise, career and life advice. Finally, support from my family through my years of studies has been essential for me to xvii achieve this important milestone. To my parents, thank you for your life-long support. My child, Cassie Lei was born and grew up while I was working on my degree. Thank you for being such a nice kid. To my husband, Chao Lei, who continued providing me support, thank you. 1 CHAPTER 1 : INTRODUCTION 2 1.1 DEFINITION OF ASTHMA Asthma is a complex and highly variable condition defined by its physiological, clinical and pathological characteristics as a chronic inflammatory disease of the airways. In general, asthma is characterized by airway obstruction, airway hyperresponsiveness, inflammation, and other recurrent symptoms including shortness of breath, chest tightness, as well as wheezing and coughing, particularly at night and in the early morning that can result in premature awakening and sleep disturbances[4, 5]. While a number of pharmacological agents provide effective symptom relief, several longitudinal studies indicate that significant proportions of patients experience symptom control followed by a return of symptoms[6]. Moreover, asthma symptoms can vary significantly from mild to severe, from person to person, and from one episode to the next. Because of this variability, there are challenges associated with asthma diagnosis and with adherence to therapeutic regimens. Asthma exacerbations are increases in symptom intensity and/or duration above the level a patient experiences when their condition is stable[7]. Severe exacerbations can last up to several weeks and may require ED visits or hospitalization[7, 8]. Recurrent exacerbations are a major cause of morbidity and health care expenditures for asthma patients [9]. 1.2 PATHOPHYSIOLOGY The three primary pathophysiological factors contributing to asthma are bronchoconstriction, airway inflammation, and airway modeling, which will be discussed separately below. 1.2.1 Bronchoconstriction Bronchoconstriction, or narrowing of the airways, is the primary contributor to 3 clinical symptoms of asthma. During acute exacerbations \u00CE\u00B22 adrenergic receptor-mediated, smooth muscle contractions in the bronchia narrow the airways in response to asthma triggers such as inhaled allergens or viruses[10]. Beta adrenergic receptors belong to the superfamily of G protein coupled receptors and include subtypes \u00CE\u00B21, \u00CE\u00B22 and \u00CE\u00B23. \u00CE\u00B21 receptors express primarily in heart and kidney, where they function to increase heart rate and force of contraction. \u00CE\u00B23 receptors are known to express in adipose tissue, but their physiological role in humans is poorly understood. \u00CE\u00B22 receptors are highly expressed in the smooth muscle of the airways and uterus[10]. In the airways of asthma patients, inhibition of \u00CE\u00B22 adrenergic receptor activity reduces the conversion of adenosine triphosphate (ATP) into 3\u00E2\u0080\u0099,5\u00E2\u0080\u0099 adenosine monophosphate (cAMP). As a direct result, the entry of calcium\u00E2\u0080\u0094an important signaling molecule\u00E2\u0080\u0094into smooth muscle cells is impaired, which in turn, limits activation of myosin light-chain kinase and causes smooth muscle contractions that lead to airway constriction[10]. \u00CE\u00B22 receptor activation promotes bronchodilation, this can be enhanced by inhaled \u00CE\u00B22 agonists used in the treatment of asthma (see Section 1.7.8). 1.2.2 Airway inflammation Airway inflammation is another important defining characteristic of asthma and involves several types of cells including mast cells, structural cells like endothelial and epithelial cells, as well as eosinophils, neutrophils, lymphocytes, and dendritic cells[2]. Inflammation is initiated when asthma triggers activate these cells to produce inflammatory mediators or drive inflammatory responses in the airway. Figure 1.1 demonstrates the mechanisms of airway inflammation in patients with asthma. 4 Figure 1.1 Mechanisms of airway inflammation in patients with asthma (copied from Goodman & Gilman\u00E2\u0080\u0099s The Pharmaceutical Basis of Therapeutics, 12th Edition with permission)[2]. Respiratory viral infection-induced inflammation is a result of damage to the airway epithelium or an initiation of inflammatory immune responses to the infection sites[11]. Respiratory viruses, such as influenza and respiratory syncytial virus (a common flu virus in infants and young children) cause asthma symptoms mainly through damaging airway epithelium[11]. These damages disturb the integrity of the epithelial layer, and in turn increase mucosal permeability and epithelial shedding, together with mucus production, 5 cause airway obstruction. The damage to airway epithelium also increases exposure of inflammatory and antigen cells to allergens. Rhinovirus, a common virus in older children and adults, do not damage the airway epithelium, in contrast, causes asthma symptoms mainly through initiating inflammatory responses. These inflammatory responses include increased inflammatory cells recruitment and secretion of a wide variety of chemokines (e.g., IL-6, IL-8 and IL-11)[11]. In asthma triggered by inhaled allergens, an increased number of mast cells is characteristic of asthma; these cells contain high-affinity immunoglobulin E (IgE) receptors that, upon activation, trigger release of histamine, cytokines, interleukins, leukotrienes, and other inflammatory mediators[12, 13] that increase airway microvascular leakage and mucus secretion. Likewise, structural cells produce inflammatory mediators that, together with those produced by mast cells, initiate an inflammation cascade by recruiting additional inflammatory cells including dendritic cells, eosinophils, T-lymphocytes, and neutrophils to the airways. These, in turn, are activated to release additional inflammatory mediators, which contributes to a positive feedback mechanism that results in additional inflammatory cell recruitment and constriction, edema, and mucus secretion. In addition, airway epithelial cells produce arachidonic acid, which contributes to bronchoconstriction as well as nitric oxide that increases edema[2]. Moreover, epithelial cells can slough off and cause blockages within large and small airways, forming mucous plugs which further block airways[2, 10]. Structural cells are a major source of inflammatory mediators contributing to chronic inflammation and thus are the target cells of ICSs used as controller medications. Activation of dendritic cells by the binding of asthma triggers leads to subsequent 6 activation T-helper type 2 (Th2) cells, which produce interleukins including IL-4 and IL-5 that recruit and activate eosinophils. Activated eosinophils have been associated with the development of airway hyperresponsiveness, and releasing and producing other inflammatory mediators such as leukotrienes and histamine to further injure airways[14]. Leukotrienes produced by eosinophils are actively involved in smooth muscle contraction, vascular permeability, increased mucus secretion, and the recruitment and activation of inflammatory cells, and their pharmacological inhibition reduces airway inflammation[2, 10]. 1.2.3 Airway remodeling When asthma symptoms are frequent and airway inflammation becomes chronic, structural changes occur, these include: hyperplasia, which results in increased airway thickness (can range from 10% to 300% of normal[15]); hypertrophy of mucous glands, increased mucus production; and increased blood supply. These structural changes may lead to a significant and permanent reduction in airways efficacy[12]. In summary, airway constriction, airway inflammation and airway remodeling are the three main pathophysiological mechanisms of asthma. Constriction and inflammation are the two main target of asthma drug therapy, which will be described in detail in Section 1.7. 1.3 ETIOLOGY OF ASTHMA The etiology of asthma involves complicated mechanisms and remains poorly understood, but it is widely accepted that the development of asthma results from an interaction between host factors including immunity and genetics, and environmental factors. 7 1.3.1 Host factors 1.3.1.1 Innate immunity and asthma Innate immunity is critical for the inflammatory response[16], which is essential to the development of asthma symptoms. Specifically, an imbalance in the expression of two types of T-helper lymphocytes (Th1 and Th2) appears to play a major role in the development of asthma[16, 17]. Th1 cells produce IL-2 and interferon-\u00CE\u00B3, which are a primary defense against intracellular pathogens. In contrast, Th2 generates interleukins (IL-4, -5, -6, -9 and -13), which mediate allergic inflammation. It has been suggested that asthma is characterized by either over-expression of Th2, under-expression of Th1 cytokines, or a combination of the two[3]. 1.3.1.2 Genetics and asthma There is a strong correlation between genetics and asthma. One study of 344 families revealed that among families with neither parent having asthma, only 6% of children developed asthma; while among families with both parents having asthma, the proportion of asthma development in their children reached 60%[18]. Another study examined 325 pairs of twins[19]. Among these, 84 pairs of twins where one twin had asthma were identified. Of these, 39 were identical and 55 were non-identical. 59% (23 pairs) of the 39 pairs of identical twins both had asthma, while 24% (13 pairs) of the 55 non-identical twins both had asthma. The strong familial clustering of asthma has encouraged research into the genetic predisposition to disease, which is important for the development of new asthma treatments. Five genes or gene complexes have now been associated with asthma development, including a disintegrin and metalloproteinase 33 (ADAM) on chromosome 20p; plant 8 homeodomain finger protein 11 (PHF11) on chromosome 13q12, dipeptidyl peptidase 10 (DPP10) on chromosome 2q14, G-protein-coupled receptor for asthma susceptibility (GRPA) on chromosome 7p14 and serine peptidase inhibitor kazal type 5 (SPINK5) on chromosome 5q33.1 [20-24]. The expression of DPP10, GRPA and SPINK5 has been studied comprehensively, and studies suggest that these gene products may play a role in the response of airway epithelium to triggers and damage caused by environmental factors[25], or in the production of mucus, perhaps via IL-3 which is known to regulate mucus production[25]. 1.3.1.3 Gender and asthma Gender is another critical factor in asthma. Asthma incidence in young children is greater among boys than girls (details in Section 1.4)[26-28]. However, after puberty, incidence becomes higher in females than in males[28-30]. Multiple theories have been put forth to explain the age and gender interaction in asthma prevalence and incidence. The first postulates that airway hyperresponsiveness is more common and severe among boys in childhood[31], but that it increases in adolescent females[32, 33]. By adulthood, hyperresponsiveness is more common and severe among adult women[34-36]. The second theory suggests that atopy (the production of IgE in response to allergens) is more common in males before age 13 years [37], and becomes more prevalent among females during adolescence and is equal between males and female in adulthood [38, 39]. The third theory suggests that hormones may have an impact on airway inflammation and airway smooth muscle function. Indeed, the fluctuation of estrogen levels due to female menstrual cycles has been reported to activate proteins that produce an inflammatory response[40]. Increased estrogen levels have been associated with reduced airway constriction via regulation of 9 calcium-dependent potassium channels, whereas low estrogen levels have been found to increase risk of asthma exacerbations[41]. 1.3.2 Environmental factors Allergens and viral respiratory infections are regarded as the two most important environmental factors in the development and persistence of asthma[42]. 1.3.2.1 Allergens Exposure to indoor allergen like dust mites, pet dander, cockroaches, rodents, and mold have been found to trigger asthma episodes and exacerbations [43, 44], while other studies demonstrate increased risk of sensitization and subsequent development of asthma in individuals who are exposed to dust mites and cockroaches [45-48]. 1.3.2.2 Respiratory infections Respiratory infections during infancy and early childhood have been associated with asthma development in later life. The Tucson Children\u00E2\u0080\u0099s Respiratory Study group evaluated 880 children to determine whether development of lower respiratory tract infections before 3 years of age is associated with physician-diagnosed asthma and/or wheezing at ages 6 and 11 years. The investigators reported that respiratory syncytial virus (RSV)-induced respiratory infections increased the risk of infrequent wheezing (3 episodes per year) by 3-fold and increased the risk of frequent wheezing (>3 episodes per year) by 4-fold[49]. Correspondingly, Sigurs and colleagues found that children who were hospitalized for RSV bronchiolitis in the first year of life had significantly higher risk of developing asthma up to age 7 compared to children who did not have RSV bronchiolitis during infancy[50]. Viral 10 respiratory infections cause damage to the epithelium, inducing edema, shedding, and mucus production, which can cause airway obstruction and wheezing. Viral respiratory infections also cause proinflammatory immune responses, which produce vast numbers of inflammatory cells and contribute to inflammation and obstruction[51, 52]. 1.3.2.3 Other environmental exposures Air pollution and tobacco smoke have also been associated with an increased risk of developing asthma. Reports detailing the role of pollution in asthma development are mixed, but the mechanism may be related to increased airway sensitization among patients exposed to heavy air pollution. Studies examining the association of tobacco smoke and asthma reported an increased risk of wheezing in infants whose mothers were exposed to tobacco smoking during pregnancy[53, 54]. Maternal smoking during pregnancy may affect the utero-placental flow, resulting in a reduction in foetal nutrition and consequent retardation in foetal lung development[55-57]. Moreover, in asthmatic adults, evidence indicates that smoking decreases responsiveness to ICSs[53]. 1.4 EPIDEMIOLOGY OF ASTHMA The World Health Organization estimates that asthma affects over 235 million people worldwide [58]. In Canada, asthma is one of the most common respiratory diseases, affecting an estimated 2.7 million Canadians[59], and Canada ranked the 10 th in asthma prevalence among more than 100 countries[60]. Analysis of Canadian Community Health Surveys showed that 10% of females and 7% of males aged 12 years or older were physician- diagnosed with asthma each year from 2001 to 2010[61], with the highest prevalence being in children (13%)[61]. In Ontario alone, asthma was responsible for 864,868 physician visits, 11 57,458 ED visits and 14,737 hospital admissions in 2006[61]. In B.C., between 1996 and 2000, the most recent evidence based on population-based data available on asthma burden estimates, asthma affects 2.5% of B.C residents 0-64 years of age. The asthma incidence, or the rate of newly diagnosed asthma was highest in children aged between 0 and 14 years (1%)[62]. These data are out of date, suggesting a need of updating the evidence on asthma burden estimates in B.C. The most recent data available on the economic burden of asthma shows that the costs of asthma-related hospital admissions, ED visits and prescription drug dispensings were estimated to be $41,858,610 between 1996 and 2000 ($311 per patient-year) in B.C. Of special interest and relevance to this study is that during this same period in B.C., 64% of patients had poorly controlled asthma and these same patients accounted for 95% of health care expenditures[63]. 1.4.1 Challenges in asthma burden estimates Epidemiological statistics for asthma are commonly obtained using questionnaires and healthcare administrative data, which present several challenges. For example, questionnaire-based studies usually include the question \u00E2\u0080\u009Cdid patients receive any diagnosis of asthma in the previous year\u00E2\u0080\u009D or \u00E2\u0080\u009Cin their lifetime\u00E2\u0080\u009D. Since these analyses include one-time physician-diagnosed asthma questionnaire-based studies tend to overestimate the true prevalence of asthma in a given population. Health services utilization and prescription drug dispensing databases is widely used for burden estimation studies since these databases are inexpensive to maintain and provide information for large populations over a long period of time, but they are also problematic. For example, asthma cases are commonly defined by physician visit, hospital admission and/or prescription dispensing criteria from population-based databases, and their validity 12 depends on both the accuracy of diagnosis and the accuracy of transcription from medical charts to electronic database. Asthma prevalence and incidence data can vary significantly depending on which asthma case definition has been used and different researchers frequently base their studies on different case definitions, which can lead to disparate and potentially inaccurate estimates of the burden of asthma. For example, Lix et al. used 28 different asthma case definitions to produce prevalence estimates for patients 12 years or older during the same time period (1998-2002), and the resulting prevalence estimates ranged from 2% to 18%[64]. As another example of reported variability, Van Wonderen et al. found 60 different definitions of asthma from 122 papers; these produced astonishingly variable prevalence estimates ranging from 15.1% to 51.1%[65]. Thus, researchers should be aware the limitations of using different case definitions to estimate the burden of asthma and be cautious in the interpretation of results. 1.4.2 Limitations in current asthma case definitions A variety of asthma case definitions have been used in previous asthma epidemiological studies. Several previous studies [66-68] used a single criterion of 1 or more physician visits during a 12-month period to define asthma cases. For example, using this criterion, Senthilselvan et al. reported an average asthma prevalence of 7% in children 0-4 years of age, 5.5% in patients 5-14 years of age, 2.5% in patients 15-34 years of age and 2.2% in patients 35-64 years of age between 1991 and 1998[66]. A major limitation arising from the use of this criterion is that the study cohort captures patients who have only a single visit for asthma which may in fact just be a presumed diagnosis and no future visits may occur. These patients do not have asthma or future visits would occur given the chronic nature of this disease. 13 In 2004, Kozyrskyj et al. developed an asthma case definition: 1or more physician visits, or 1 or more hospital admissions for asthma, or 1or more ICS/ketotifen/cromone prescription dispensings, or 2 or more SABA or oral beta-agonist dispensings using Manitoba data[69]. In addition to the limitation associated with the use of 1 or more physician visits discussed above, study cohorts based on a criterion of either 1 or more or 2 or more prescription dispensings are likely to include patients who only received a single set of medications to help make a diagnosis, or for the treatment of wheezing in patients with respiratory infections who presented with wheezing as a chief symptom. Other studies used a criteria of 1 or more hospital admissions, or 2 or more physician visits, or 2 or more prescription drug dispensings to identify patients with asthma [70-72]. In B.C., there is a 100-day reimbursement policy for prescription refills [72], which means that patients with treated asthma need to refill their asthma prescription medications every 3 months. When used with B.C. data, case definitions using 2 or more prescription drug dispensings will be limited by containing patients who only received a single set of medications to help develop a diagnosis or a group of patients with very mild asthma (i.e., patients who require SABA no more than 3 doses per week for symptom treatment, and do not require daily ICS for asthma management). Thus, this case definition is not suitable to limit the capture of patients who have chronic symptoms requiring regular drug use to manage symptoms. To address these limitations, the present research developed a superior case definition (i.e., 1 or more hospital admissions, or 2 or more physician visits or 3 or more asthma prescription drug dispensings during a 12-month period), with the aim of capturing the entire population with treated asthma in B.C. between 1996 and 2009. 14 1.5 DIAGNOSIS OF ASTHMA Proper diagnosis is the critical in order to facilitate optimal drug therapy and asthma management. Key components in diagnosis include a detailed medical history, family history, physical exam and proper testing of pulmonary (lung) function. Physical examinations include identification of allergic reactions such as skin rashes and/or abnormal sounds in the chest in response to various potential allergens. Hallmark symptoms include wheezing, shortness of breath, cough, nocturnal symptoms that lead to awakening and sleep disturbances, and a worsening of symptoms in the presence of triggers[13]. Testing of pulmonary function is important as it provides objective values to estimate the extent of airway limitation and its reversibility. However, it requires patient cooperation, making difficult to apply in young children, in patients who are experiencing asthma exacerbations, or in patients with poor health. The diagnosis of asthma is not always straightforward particularly since wheezing, coughing or chest tightness can arise from other conditions, such as bronchitis, colds, or pneumonia in children and adults and from chronic obstructive pulmonary disease (COPD) in adults. Thus, when patients are presenting with these conditions are admitted to clinics or EDs, diagnosis of asthma could be delayed or missed altogether. Further complications arise since other respiratory diseases can be misdiagnosed as asthma. On the other hand, some asthma symptoms like cough and/or wheeze are common and recurrent in many conditions, it is possible that other respiratory diseases are misdiagnosed as asthma in some patients. For example, one study examined whether asthma was overdiagnosed in patients from eight Canadian cities in 2008[73]. The investigators defined patients as not having asthma if they did not display evidence of acute worsening of symptoms, reversible airflow obstruction or 15 bronchial hyperresponsiveness while being weaned off asthma medications. Based on these criteria, this study identified that more than 30% of the study population with physician- diagnosed asthma did not actually have asthma. Use of spirometry, a pulmonary function test, has been recommended to establish an asthma diagnosis[3]. Spirometry measures the amount and speed of air that can be inhaled and exhaled. The most common parameters measured in spirometry are vital capacity (VC; the maximum volume of air a person can expel from the lungs after maximal inhalation), forced vital capacity (FVC; the volume of air that can be forced to exhale after full inhalation), and forced expiratory volume (FEVX; volume of air exhaled at time point X during forced expiration). Spirometry is generally available only to children who are 5 years or older because of difficulties in application in young children. To aid diagnosis, asthma clinical practice guidelines has recommended that physicians measure FEV1, FEV6, FVC, and the FEV1/FVC ratio before and after the patient inhales a short-acting bronchodilator. Reduced values for FEV1 and the FEV1/FVC ratio (or FEV1/FEV6 ratio) compared to predicted values are indicators of airway obstruction. An FEV1 increase of >200 mL combined with \u00E2\u0089\u00A512% from baseline or an \u00E2\u0089\u00A510% increase in FEV1 value after SABA administration indicates clinically significant airway reversibility. FEV1 and FEV1/FVC values have also been used to measure degree of asthma control, which is described in Section 1.8. Although spirometry provides objective values in pulmonary function, it does have limitations. In particular, test results may not always correlate directly with symptom severity; For example, in one study, Stout and colleagues reported that one third of children with moderate-to-severe asthma were reclassified to a more severe category when FEV1 values 16 were considered in addition to symptom frequency[74]. In contrast, another study by Bacharier and colleagues found that 68% of children with mild-to-moderate asthma classified by symptoms had normal FEV1[75]. In summary, several challenges exist in clinically diagnosing asthma, and these challenges produce variable results (e.g. in quantifying prevalence and incidence) that can result in errors in studies estimating the epidemiology of asthma. Degree of accuracy in asthma diagnosis also affects the validity of health service utilization database, which is discussed in detail in Section 1.5. 1.6 GOAL OF ASTHMA TREATMENT The ultimate goals of asthma drug therapy are to enable a patient to live without asthma symptoms and associated functional limitations, to improve quality of life, and to reduce the risk of adverse side effects of medications[3]. Because of significant variations in asthma symptomology over time and in patients\u00E2\u0080\u0099 response to drug therapy, regular monitoring of the effectiveness of asthma control is necessary to achieve these goals. Research indicates that ongoing monitoring can help achieve disease management through routine follow-up visits with patients to review compliance with management plans, and self- management skills including the use of inhalers, spacers, and peak-flow meters[13]. 1.7 TREATMENT OF ASTHMA Asthma is predominately managed by pharmacologic therapy, which can prevent and control symptoms, reverse airflow obstruction, and improve quality of life. There are two main classes of drug that are used to manage asthma: quick-relief medications that are taken to provide prompt reversal of acute airflow obstruction and relief of bronchoconstriction and 17 long-term controller medications that are taken daily over a longer period of time to achieve and maintain control. Long-term controller medications include: corticosteroids, mast cells stabilizers (i.e., cromolyn sodium and nedocromil), anti-IgE agents, leukotriene receptor antagonists (LTRA), long-acting bronchodilators (LABA) and methylxanthines. 1.7.1 Inhaled corticosteroids Inhaled corticosteroids (ICSs) are the cornerstone of asthma drug therapy, and are by far the most effective and safest controller medication in asthma management. ICS were initially developed as a replacement therapy for oral steroids, which cause systemic adverse drug reactions (ADRs). Oral steroids were found to be effective at treating asthma and were widely used in the 1950s. However, potential ADRs, including growth retardation in children, osteoporosis, and metabolic disturbances soon became major concerns and suggested the need for development of an inhalation formulation designed to localize therapy at the target site (the lungs) and therefore reduce systemic ADRs associated with oral therapy. In the 1970s, beclomethasone dipropionate (BDP) became the first corticosteroid developed for inhalation. Studies comparing the effectiveness and safety of ICSs in patients who transferred their drug therapy from oral steroids to ICSs[76-79], reported a 99% success rate in symptom control, along with relief of ADRs (especially Cushingoid features). Currently there are six different ICSs on the Canadian market, including beclomethasone, budesonide, fluticasone, flunisolide, mometasone and ciclesonide, and these share similar pharmacological characteristics. Clinical effects of ICSs are achieved mainly through inhibition of inflammatory cells via modulation of gene transcription, as well as prevention of inflammatory cell recruitment into the airways. The general mechanism of the 18 therapeutic action of corticosteroids is to increase transcription of anti-inflammatory genes and inhibit transcription of inflammatory genes, particularly in airway epithelial cells, the latter is the most important action of ICSs in inhibiting inflammation in asthma patients. In addition, ICSs reduce the number and activation of inflammatory cells in the airways and lungs, inhibit production of inflammatory mediators (e.g., cytokines and chemokines) by T- lymphocytes, macrophages, and mast cells[13], and decrease vascular permeability, aiding in the reduction of airway edema. While they do have numerous therapeutic actions, ICSs are not able to reverse airway remodeling caused by chronic inflammation in the airways. 1.7.1.1 ICS efficacy in asthma treatment The therapeutic efficacies of ICSs relative to LTRA and cromolyn sodium have been established in a number of clinical trials. A Cochrane Collaborative systematic review involving 10,005 adults and 3,333 pediatric patients and 65 studies with trial durations ranging between 4-52 weeks published before December 2012 compared ICS to LTRA[80]. In this review, the median dose of ICS was 200 \u00C2\u00B5g/day of microfine hydrofluoroalkane- propelled beclomethasone or equivalent, and its use was associated with significant reduction in exacerbations requiring systemic corticosteroids or hospital admission, improvement in lung function testing (e.g., FEV1), asthma control and quality of life, reduction in asthma symptom score, nocturnal awakenings, and use of SABAs. Another recent trial randomly assigned 309 patients 12 to 80 years of age with mild to moderate asthma, reduced asthma- related quality of life and inadequate asthma control to three treatment groups: ICS monotherapy, ICS and LABA combined inhalers, and LTRA[81]. Over two years, this trial compared quality of life, symptom score and exacerbation rates between patient groups and reported no significant differences; however, the validity of this study may have been 19 compromised by the fact that patients were permitted to change or stop taking medications without restriction. By the second year, 50% of patients did not complete asthma symptom diaries, 50 to 67% did not strictly adhere to protocol medications, and 24 to 59% stopped taking prescribed ICS. Similarly to those comparing ICSs to LRTA, studies comparing ICSs to cromolyn sodium showed that ICS treatment groups had significant improvement in lung function, fewer exacerbations, lower asthma symptom scores and reduced use of SABAs[82]. In addition, several studies have examined the efficacies of adjunctive therapy and have reported that addition of LABA to ICS produced more significant improvement in asthma outcomes than addition of LTRA or theophylline or doubling ICS dose[83, 84]. 1.7.1.2 ICS effectiveness in asthma treatment Several studies have demonstrated that the use of ICSs significantly reduces asthma- related hospital admissions and deaths, regardless of severity and age. The earliest cohort study that assessed the risk of hospital admission for asthma in relation to ICS use was in 1994[85]. This study involved 216 children treated with budesonide and 62 children treated with theophylline, inhaled B2-agonists, and/or sodium cromolyn and followed these patients for up to 7 years. On average, only 0.4% of children who received ICS had hospital admissions for asthma each year during follow-up, a value nearly 10-fold lower than the rate for children treated with other asthma medications (3% per year)[85]. This apparent reduction in hospital admission rates among ICS users was supported by subsequent studies[85-88]. Early investigations using asthma mortality as an outcome produced mixed results. A New Zealand study (1989) involved 117 cases of death from asthma and 468 controls 20 matched with respect to age and ethnic group, reported a 34% increase in asthma mortality in patients who received ICS[89]. In stark contrast, a second study conducted from 1988 to 1990 reported a 30% reduction in asthma-related mortality in patients treated with ICS[90]. A major limitation of this study is that the measure of ICS exposure was binary (yes or no) and length of ICS use was not taken into account. In 1992, Ernst and colleagues conducted a study based on 129 cases of fatal or near-fatal asthma and 655 matched controls. After accounting for disease severity and risk of adverse drug events, patients who received 12 or more ICSs (50 \u00C2\u00B5g per actuation; 200 actuations) over a one-year period were 90% less likely to have fatal and near-fatal asthma (RR, 0.1; 95% CI, 0.02 \u00E2\u0080\u0093 0.6) compared to those who did not receive any ICS during the study year. However, patients who received from 1 to 11 ICS inhalers had similar risk of experiencing the outcome events (RR 1.6; 95% CI 0.9-2.7). The authors concluded that there is no sufficient power to detect the differences between ICS treatment groups[91]. Indeed, in 2000, Suissa et al. addressed methodological limitations found in previous studies and conducted a nested case-control study within a cohort of 30,569 patients with asthma identified using the Saskatchewan Health databases[92]. Asthma severity, time of drug use, length of follow-up at the time of death were considered and incorporated into the case and matched control criteria. This analysis seemed to resolve the contrasting results of early studies and showed that each additional ICS canister used in the baseline year resulted in a 20% reduction in the likelihood of asthma-related mortality. 1.7.1.3 Safety of ICS Since ICSs are targeted directly to the site of inflammation directly, the risk of systemic ADRs is much lower compared with other therapeutics. However, ICS use can cause local 21 ADRs including include oral candidiasis and dysphonia [93-95]. Oral candidiasis, an infection of oral mucus membranes by yeast fungi, occurs in 5 to 34 percent of asthma patients[96-98] depending on ICS does[96]. Increased candidiasis occurrence in asthma patients is due to decreased local immunity caused by inhibition of normal host defense functions at the oral mucosal surface and in the esophagus. It has also been suggested that oral candidiasis occurs because of an increase in salivary glucose levels, which stimulates the growth of Candida albicans. To reduce the incidence of oral candidiasis, clinical practice guidelines recommend mouth rinses with water immediately after inhalation and use of a spacer or valved holding chambers (VHCs) with a non-breath-activated metered dose inhaler (MDI). Mouthwashes with antifungal agents are another effective for treatment of oral candidiasis. Dysphonia occurs in 5-50% of patients who use ICS, but its etiology is still not clear. It has been suggested that dysphonia is associated with a fungal infection, vocal stress and dyskinesia of muscles that control vocal cord tension. Use of a spacer device or VHC with a non-breath-activated MDI, or a reduction in the frequency of ICS use can minimize the incidence of dysphonia. 1.7.1.4 Systemic adverse effects Higher doses of ICS are associated with increased risk of systemic ADRs including reduced growth velocity in children, osteoporosis, skin thinning, cataracts and glaucoma and the suppression of the hypothalamic-pituitary-adrenal (HPA) axis[93, 94, 99]. These systemic effects are related to the total amount of corticosteroid that is absorbed and the rate of clearance from the body. For older ICSs (e.g., beclomethasone dipropionate, triamcinolone and flunisolide), systemic presence of corticosteroids is due to deposition of a significant 22 fraction of drug in the mouth[100], this fraction is swallowed, absorbed across the gastrointestinal (GI) lining and undergoes first-pass metabolism in the liver. In addition, older ICSs have low first-pass metabolism (50-70%), which, combined with swallowing and absorption in the GI tract, results in a higher concentration of ICS in systemic circulation after \u00E2\u0080\u009Cinhalation\u00E2\u0080\u009D. Newer ICSs (e.g., fluticasone, budesonide, ciclesonide), oral swallowing is less common due to improved drug delivery. Systemic availability of these corticosteroids results from local absorption through the lung[100]. Overall, evidence has shown that occurrence of systemic ADRs in patients with a low- to median- dose range of ICS is rare and that ADR events may be clinically insignificant and/or reversible. Patients who continuously receive high-dose range of ICS may have higher risk of experiencing systemic ADRs, but this risk can be reduced by use of a spacer device with an MDI, mouth washing, or using dry-powder inhalers. The primary concern of physicians and parents/patients in the regular use of ICSs in children is related to the potential for growth retardation [101]. A putative link between ICS use and growth is suggested by the known association between systemic corticosteroids levels and growth. Systemic corticosteroids have been associated with inhibition of growth hormone secretion, reduction in growth hormone receptor expression and direct suppression of growth[102-104]. However, the magnitude of the effects of ICS on growth are still not completely clear[100]. Evidence has shown that the reduction of linear height is small, nonprogressive, and possibly reversible [105-107]. For example, the Childhood Asthma Management Program conducted a 5-year, 8-center study that involved 1,000 children 5-12 years of age with mild to moderate asthma[105]. This study reported that the mean height increase in children treated with budesonide was about 1 centimeter (cm) less than the mean 23 height increase in children treated with placebo in the first year of the study. However, after the first year and throughout the remaining 4 years of the study, children on budesonide grew at rate identical to the other children. In addition, poorly controlled asthma has also been linked to reduced growth in children[108, 109]. Since growth hormones are primarily secreted at night or during exercise, the limited physical activity and frequent wakening at night caused by poorly controlled asthma result in reduced growth hormone. These observations should be considered when interpreting findings from growth studies[100]. In summary, ICSs are the most effective long-term maintenance therapy for asthma and have an excellent safety profile, especially in low dose ranges that are usually sufficient for asthma treatment. Adverse reactions can occur at higher doses, but can be avoided through using combination therapy or by decreasing ICS dose once asthma control has been achieved. Other controller medications 1.7.2 Leukotriene receptor antagonists (LTRAs) LTRAs (Zafirlukast and montelukast) are an alternative therapy for the treatment of mild persistent asthma, and they function by inhibiting leukotriene-mediated inflammation. As discussed in Section 1.7.1, LTRA efficacy is lower compared with ICS, evinced by less reduction in exacerbations, less improvement in symptom-free days), LTRA is generally reserved for a second-line therapy, except in children who cannot properly apply ICS inhalation. Studies examining the effectiveness of LTRA or comparing the effectiveness of LTRA to ICS are limited and suffer from methodological flaws. Blais et al. conducted a cohort study involving 27,355 children who initiated ICS or LTRA monotherapy between 1998 and 24 2005 using Quebec administrative databases[110]. The primary outcome measured was the rate of exacerbations (defined as ED visit or hospital admissions or dispensed prescription of oral corticosteroids) over the subsequent year. In order to control confounding by severity, the analyses were stratified by the number of exacerbations that patients previously experienced. In patients who did not have previous exacerbations, ICS use was associated with significantly higher rates of exacerbations compared to montelukast use. However, in patients who had one or more previous exacerbations there was no significant association with ICS use. This study has two major limitations: first, the use of stratification as the sole method used to control confounding by severity is insufficient; second, the authors used a linear regression model for analysis, which did not account for the time-dependent nature of medication use. Together, these limitations challenge the validity of the results comparing exacerbation frequency with ICS use. However, this same study did show that adherence rates for montelukast were significantly higher than the rates for ICS. Recently, Ducharme and colleagues used Quebec health service utilization again to compare asthma-related health service utilization outcomes in 227 children 2-17 years of age[111]. There were no significant differences in the use of oral corticosteroids or ED services between montelukast and the ICS use groups. ICS monotherapy group showed significantly higher hospital admission rates and use of \u00CE\u00B22-agonist for asthma than the montelukast group. Patients who received montelukast appeared to have poorer asthma control, but better adherence than those who received ICS during the baseline year. No valid method was used to measure adherence. The authors still failed to adjust for confounding by severity or use time- dependent analysis to account for the time-dependent nature of asthma drug use. In 2008, the US Food and Drug Administration (FDA) issued a safety alert for 25 Montelukast for a possible link to suicide and other adverse neuropsychiatric events (e.g., tremor, depression, and anxiousness). Since then, 800 suicide-related adverse events associated with LTRAs have been reported to the FDA [112]. There is no clear pharmacological mechanism through which LTRAs might cause suicide, but several theories have been proposed. One theory suggests that, since leukotrienes and their receptors are present throught the nervous system and it is known that LTRAs can penetrate the blood- brain barrier, the inhibition of leukotriene receptors in the brain by LRTAs might be responsible for adverse neuropsychiatric effects [113]. Another theory postulates that production of toxic nitric oxide by direct binding of montelukast to the cysteinyl leukotriene receptor 1[114] might be responsible. While it is interesting to speculate, there is currently no solid evidence to support either of these theories. 1.7.3 Mast cells stabilizers Cromolym sodium and nedocromil, two mast cells stabilizers, act by inhibiting the release of histamine and leukotrienes from mast cells. Cromolyn was approved by the FDA in 1973 as maintenance therapeutic for children with asthma, but became less commonly used after the introduction of ICSs in 1990s. Both Cromolyn and nedocromil have been associated with a significant reduction in the use of ED services (50%) and hospital admissions (20%-60%) in population-based cohort studies [87] [115]. However, a Cochrane review pooled results from four available clinical trials that assessed cromolyn efficacy and showed no significant differences in the percentage of symptom-free days between cromolyn and placebo, and only minimal effects of cromolyn treatment on asthma outcomes [116]. However, both drugs are still used in practice, especially in pediatric patients who are ICS- intolerant or are concerned about ICS-related ADRs. 26 1.7.4 Methylxanthines Methylxanthines, such as theophylline, have been used in the treatment of asthma since 1930. Theophylline works as a bronchodilator and is used to prevent asthma symptoms. The bronchodilating action of theophylline is due to its inhibition of phosphodiesterase, which increases cellular cAMP and relaxes airway smooth muscle. Theophylline can also lead to behavioral disturbances and learning difficulties in children and high concentrations of can result in cardiac arrhythmias and seizure; thus, it is important to monitor its plasma levels of theophylline in order to prevent occurrence of serious ADR. Because of high ADR frequency relatively low efficacy, theophylline is generally only used as an adjunct therapeutic and is mainly indicated to patients with severe asthma. 1.7.5 Anti-IgE agents Omalizumab is a monoclonal anti-IgE antibody that is delivered subcutaneously every 2 to 4 weeks and that prevents binding of IgE to high-affinity receptors expressed on the surfaces of basophils and mast cells. Omalizumab was approved by the U.S. Food and Drug Administrative (FDA) in 2003 for the treatment of patients 12 years and older with moderate to severe persistent asthma that is controlled with ICS or ICS combined with a long-acting B2 agonist (discussed below). In clinical trials, omalizumab has demonstrated efficacy compared to placebo[117-121]. The effectiveness and safety of omalizumab in large populations have yet to be analyzed in detail. 1.7.6 Long-acting \u00CE\u00B2 agonists (LABA) Long-acting B2 agonists share similar pharmacological mechanisms with short-acting B2 agonists\u00E2\u0080\u0094they relax airway smooth muscle via \u00CE\u00B22 adrenergic receptor mediated increases 27 in cyclic AMP. Because of their relatively high lipophilicity, LABAs have a more sustained effect on bronchodialtion, a single dose produces an effect that lasts at least 12 hours, permitting twice-daily administration. Randomized controlled trials (RCTs) have shown that adding a LABA to an ICS regimen provides greater improvements in lung function, symptom control and reduced need for SABA compared with increasing ICS dose[122-127]. Safety concerns regarding LABAs arose shortly after their market introduction. Several large clinical trials and meta-analyses reported increases in respiratory and asthma-related deaths in patients who received salmeterol [128-130]. In 2004, the FDA issued black box warnings on all preparations containing LABAs. The mechanism(s) underlying the increased risk of severe respiratory events associated with LABA use are not fully understood but are likely related to genetic, environmental and/or disease factors. A recent study from the UK assessed the role of the B2 receptor Arg16 allele in asthma exacerbations in 1,182 children and adult patients [131] and associated increased risk of exacerbations with extra copies of the Arg16 allele in patients who regularly use inhaled salmeterol. Interestingly, the associated risk was not observed in patients carrying an extra copy of the Arg16 allele who were also taking oral or ICS, suggesting that steroids may protect against LABA-mediated receptor down regulation and desensitization through the ADRB2 glucocorticoid response element (a DNA sequence that regulates transcription). 1.7.7 Systemic corticosteroids Systemic corticosteroids are used in patients with severe asthma exacerbations as adjunct to SABAs since they reduce airway inflammation fast through systemic effects. They are also recommended to be used at the onset of symptoms in upper respiratory tract infection- 28 induced exacerbations[13]. The risk of ADR effects depends on dose and duration, and the likelihood of ADRs occurring in patients with short courses of systemic corticosteroids is small[132]. However, regular use has been associated with a series of serious adverse effects, including growth retardation, osteoporosis, myopathy, adrenal supression and development of cushingoid habitus[3, 133, 134]. In contrast to controller medications designed to manage symptoms over sustained periods of use, reliever medications are designed to rapidly treat acute symptoms and this class of medications include SABAs and anti-cholinergic drugs. 1.7.8 Short-acting bronchodilators SABAs are the most effective medication for relieving acute bronchoconstriction[13] and their efficacy has been established in a number of RCTs trials. SABAs are adrenergic receptor agonists that relax airway smooth muscle via stimulation of \u00CE\u00B22 adrenergic receptors, which leads to bronchodilation. \u00CE\u00B22 receptors are coupled to stimulatory G proteins that activated adenylyl cyclase and increased intracellular cyclic AMP. Increased cyclic AMP leads to brochodilation by a number of mechanisms including activation of protein kinase A and increased calcium permeability. \u00CE\u00B22 agonists also increase the conductance of large calcium- sensitive potassium channels in airway smooth muscle, which promotes membrane hyperpolarization and relaxation[2, 10]. Historically, the first bronchodilator administered for asthma treatment was subcutaneous epinephrine, a nonselective adrenergic agonist that binds to \u00CE\u00B21- and \u00CE\u00B22- adrenergic receptors. By the mid twentieth century, an inhaled formulation of epinephrine became available. Studies in the early 1990s reported a series of epinephrine-induced severe ADRs including cardiac tachycardia and arrhythmias resulting from \u00CE\u00B21 receptor stimulation 29 as well as peripheral blood vessel narrowing due to \u00CE\u00B1-receptor stimulation. Due to these potentially serious ADRs, epinephrine is no longer commonly used to treat asthma. The next bronchodilator developed as an alternative to epinephrine was isoproterenol, a selective \u00CE\u00B2- adrenergic agonist. However, isoproterenol stimulates both \u00CE\u00B21 and \u00CE\u00B22 receptors, like epinephrine it can cause cardiac stimulation in addition to the desired bronchodilation effect; indeed, in the mid-1950s high-dose isoproterenol was found to significantly increase asthma-related deaths in England. This suggested the need for a more selective \u00CE\u00B22 agonist, and in 1980s, the first \u00CE\u00B22-selective SABA, salbutamol, was developed by GlaxoSmithKline, and remains a mainstay in the treatment of acute asthma symptoms. Safety of short-acting bronchodilators In the late 1980s and early 1990s, a series of studies associated chronic SABA use with a significant increase in asthma-related death[89] , [90, 135-137]. This effect is likely due to the down-regulation of \u00CE\u00B2 receptors that occurs with long term use, and could lead to decreased bronchodilation and bronchoprotection. Evidence has shown that regular use of SABAs also reduces drug effectiveness[138, 139], increase airway hyperresponsiveness[140-142], and increases the response of eosinophils and mast cells to allergen challenge[141, 142]. Because SABAs provide immediate effects in reliving symptoms, some patients maintain asthma control solely by SABA use. The negative outcome of regular SABA use highlights the need for physicians and asthma educators to provide patients with information on the improper use of medication, and to regularly monitor patients\u00E2\u0080\u0099 drug use. 1.7.9 Anti-cholinergic medications Anti-cholinergic medications including ipratropium and tiotropium relax airway smooth 30 muscle by inhibiting muscarinic receptors in the lung and submucosal gland cells[143]. Because muscarinic receptors do not play a primary role in the pathophysiology of asthma, anti-cholinergic medications are less potent bronchodilators than SABAs. Ipratropium, a short-acting cholinergic medication is extensively used in EDs in combination with \u00CE\u00B2 agonists to manage acute exacerbations. A recent review of 32 RCTs showed that use of a SABA and anticholinergics in combination produced a significant increase in lung function in pediatric and adult patients, and was associated with a 25% reduction in asthma-related hospital admissions in children and a 32% reduction in adults[144]. Recently, three RCTs showed that addition of tiotropium, a long-acting anticholinergic medication to patients\u00E2\u0080\u0099 regimens of ICS and LABA produced greater efficacy than those patients who added placebo to their regimens[145-148]. Two of these trials have small sample sizes (n=210 and 388) and short study durations (8-16 weeks)[146, 148]. More evidence based on larger study sample and longer study duration is needed to determine the role of tiotropium as adjunct therapy in patients with moderate to severe asthma. 1.8 ASSESSMENT OF ASTHMA CONTROL Well controlled asthma has been associated with significant reductions in ED visits and hospital admissions and with improved quality of life. Success in achieving and maintaining well controlled asthma has been associated with greater adherence to controller medication regimen, regular health professional assessment of asthma control, greater continuity of care, and easy access to care[3]. Asthma control refers to the degree to a patient\u00E2\u0080\u0099s asthma symptoms are under control at a given point of time. The National Health Heart, Lung and Blood Institute (NHLBI) recommend that asthma control be measure asthma based on two criteria: current impairment 31 and future risk[13]. Impairment involves assessment of: 1) frequency and intensity of symptoms; 2) frequency of SABA use; 3) pulmonary function; and 4) reduction in patient activity (e.g., exercise, attendance at work or school). The risk component of asthma control estimates the likelihood of recurrent ED, progressive loss of lung function or reduction of lung growth in children, and ADRs. Risk of future exacerbations can be measured using spirometry. FEV1 values which have been used as markers of the degree of airflow obstruction and used by the NHLBI as a measure of asthma severity, are particularly useful. For example, in a pediatric population with asthma (n=13,842), the risk of having subsequent asthma attacks (i.e., wheezing or shortness of breath) is two-fold greater in patients with an FEV1 < 60% compared to those with and FEV1 > 80%[149]. Assessment of the risk of risk of reduced lung growth in children is measured by prolonged failure to attain predicted lung function values for a given age and by longitudinal assessment of lung function. In contrast, little is known about the risk and prevention of progressive loss of pulmonary function among asthma patients. Since population-based health services utilization data have been widely used nowadays to perform epidemiological studies in asthma, there is growing evidence on the development of indicators of asthma control level based on electronically available data records. Factors related to asthma control level are summarized in Table 1.1. Evidence has shown that patients who have had exacerbations requiring ED visits, hospital admission or intensive care unit (ICU) admission, especially in the past year, have a greater risk of exacerbations in the future[150-152]. In adults, hospital admission in the past 12 months increases the risk of future hospitalization; patients with asthma-related hospital admissions in the previous 12 months were 3 times more likely to have repeated ED visits for exacerbations[150, 151]. In 32 pediatric patients, a hospital admission or ED visit in the past 6 months doubles the risk of future hospital admission or ED visits for exacerbations[152]. Studies conducted by Firoozi et al. and Ungar et al. also reported using number of SABA and ICS inhalers, dispensing records of the second-, third- and fourth-line asthma drug therapy (e.g., LABA, LTRA, theophylline, and oral steroid prescriptions) is valid in measuring patients\u00E2\u0080\u0099 asthma level of control[153, 154]. 33 Table 1.1 Factors affecting asthma control level Characteristic Association with risk of hospital admission or ED visit Rationale Demographic information Age Young children are more likely to visit ED or be admitted to hospitals for asthma than older patients. Young children have smaller airway caliber[155]. Airway constriction and inflammation are more likely to cause airway blocking. Parents tend to take young children to the ED for asthma management mainly due to lack of knowledge or confidence in managing their children\u00E2\u0080\u0099s asthma at home [156]. Gender Boys are more likely to have asthma exacerbations. Risks of using ED or hospital services for asthma are similar in teenager, but significantly higher in female adult patients. Airway hyperresponsiveness is more common and severe among boys in childhood[31]; Atopy (the production of IgE in response to allergens) is more common in males before age 13 years[37]; Fluctuation of estrogen levels due to female menstrual cycles has been reported to activate proteins that produce an inflammatory response[40]. Details see Section 1.3.1.3. 34 Characteristic Association with risk of hospital admission or ED visit Rationale Health service utilization Admission to hospital for asthma in the past 12-month Admission to hospital in previous year increases the risk of hospital admissions or ED visits for asthma in the subsequent year Patients who were hospitalized for asthma may have more severe disease and/or poorer asthma control[157]. ED visit in the past 12- month ED visits for asthma in previous year increases the risk of ED visits or hospital admissions in the subsequent year. Patients who visited ED for asthma may have more severe disease and/or poorer asthma control[157]. Asthma drug dispensing Number of SABA inhalers Patients with frequent SABA use may have higher likelihood of using ED or hospital services for asthma. Frequent use of SABA may imply poor asthma control [157]; higher risk of outcome. Number of theophylline prescriptions Patients with theophylline use for asthma may have higher likelihood of using ED or hospital services for asthma. Use of theophylline may imply poor asthma control. Thophylline is used as a third-line asthma drug therapy and is indicated to patients whose asthma cannot be well controlled by SABA, ICS and LABA[3]. Details see Section 1.7.4. 35 Characteristic Association with risk of hospital admission or ED visit Rationale Number of LABA prescriptions dispensed Patients with LABA use for asthma may have higher likelihood of using ED or hospital services for asthma. Use of LABA may imply poor asthma control. LABA is used as the second-line asthma drug therapy and is indicated to patients whose asthma cannot be well controlled by SABA and low/median-dose ICS[3]. Details see Section 1.7.6. Number of montelukast prescriptions dispensed Patients with montelukast use for asthma may have higher likelihood of using ED or hospital services for asthma. Use of montelukast may imply poor asthma control. Montelukast is used as the third-line asthma drug therapy and is indicated to patients whose asthma cannot be well controlled by SABA, ICS and LABA[3]. Details see Section 1.7.2. Number of Omalizumab prescriptions dispensed Patients with omalizumab use for asthma may have higher likelihood of using ED or hospital services for asthma. Use of omalizumab may imply poor asthma control. Omalizumab is used as the third-line asthma drug therapy and is indicated to patients whose asthma cannot be well controlled by SABA, ICS and LABA[3]. Details see Section 1.7.5. Note: SABA: short-acting broncodilators; ICS: inhaled corticosteroids; LTRA: leukotriene receptor antagonists; LABA: long- acting bronchodilators 36 Another measure of asthma control is the ratio of controller medication to reliever medication, but the utility of this measure is unclear[158-162]. Fuhlbrigge et al. used prescription dispensing data and compared the association between two different asthma prescription measures and subsequent risk of ED visits in children enrolled in three U.S. managed care organizations[162]. The first measure, dispensing of a controller medication was associated with a 70% lower risk of ED visits. However, results acquired using the second measure, the ratio of dispensed controller and dispensed reliever medications were dependent on the number of reliever medication dispensings. A higher ratio of controller-to- reliever dispensing was associated with a lower risk of ED visits in children dispensed <4 relievers over the 1 year study; however, there was no significant relationship between the ratio of controller-to-reliever dispensing and ED visits in children dispensed \u00E2\u0089\u00A54 relievers. Also contradicting are the findings of Griffiths et al., who found a higher prescribing ratio was associated with a lower hospital admission rate[160], and Gottlieb et al., who observed an inverse correlation between the ratio of the ratio of controller-to-reliever dispensing and hospital admissions. Shelley et al. observed no correlation between hospital admission rate and the ratio of controller-to-reliever dispensing. Finally, Schatz et al. showed that patients who use <6 \u00CE\u00B2-agonist canisters per year are less likely than patients who are prescribed \u00E2\u0089\u00A5 6 \u00CE\u00B2- agonist canisters per year to experience subsequent ED visits. This suggests that the number of \u00CE\u00B2-agonist canisters dispensed may be a useful measure that reflects asthma control. 1.9 CLINICAL MANAGEMENT OF ASTHMA 1.9.1 Long-term control Due to the highly variable nature of asthma, a dynamic, closely monitored therapeutic 37 approach is recommended in clinical practice guidelines. Figure 1.2 and Figure 1.3 show the stepwise approach of asthma management recommended by the NHLBI Guidelines for the Diagnosis and Management of Asthma, EPR-3 (2007)[3]. Infrequent symptoms in the presence of expiratory flows can be treated with an inhaled SABA used as needed to relieve symptoms. If the rescue SABA is needed > 3 times a week (excluding 1 dose per day before exercise to prevent exercise-induced asthma), or if lung function is abnormal, an ICS should be added to the regimen at the minimum daily dose (equivalent to < 500 mcg/day of beclomethasone dipropionate in adults, and < 250 mcg/day in children). If symptoms are not adequately controlled by low ICS doses, guidelines recommend two options: 1) add a LABA to the low-dose ICS regimen; or 2) increase ICS dose to the medium-dose range (500 \u00E2\u0080\u0093 1000 mcg/day in adults and 250 \u00E2\u0080\u0093 500 mcg/day in children <12 years). Because of the safety concerns associated of LABA use (see Section 1.7.6), the EPR-3 Panel recommended that equal consideration be given to the options of increasing ICS dose or adding a LABA for patients whose asthma is not sufficiently controlled with a low-dose ICS alone[3]. 38 Figure 1.2 Stepwise approach to therapy for managing asthma in patients 5-11 years of age (copied from the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma, Expert Panel Report-3 with permission)[3]. Note: PRN: as needed; SABA: short-acting broncodilators; ICS: inhaled corticosteroids; LTRA: leukotriene receptor antagonists; LABA: long-acting bronchodilators 39 Figure 1.3 Stepwise approach to therapy for managing asthma in patients 12 years or older (copied from the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma, Expert Panel Report-3 with permission)[3]. Note: PRN: as needed; SABA: short-acting broncodilators; ICS: inhaled corticosteroids; LTRA: leukotriene receptor antagonists; LABA: long-acting bronchodilators For patients whose symptoms cannot be sufficiently controlled by the regimens listed above, the next recommended step is to increase ICS dose to the medium-dose range and add a LABA. If this still does not lead to improvement in asthma control, patients are recommended to take a high dose of ICS (> 1000 mcg/day in adults and > 500 mcg/day in children), or consult an asthma specialist. Upon achieving control, consideration of an ICS dose reduction is recommended. 40 1.9.2 Managing exacerbations of asthma Asthma exacerbations refer to acute episodes of progressively worsening shortness of breath, cough, wheezing and chest tightness. Generally, for patients with mild exacerbations (those with a Peak Expiratory Flow or PEF \u00E2\u0089\u00A570 % of predicted), administration of SABA via metered-dose inhaler or a nebulizer at home should quickly relieve symptoms. Patients with moderate exacerbations should take up to 2 SABA treatments of 2-6 puffs 20 minutes apart at home. Treatments should be followed by a reassessment of PEF and symptoms. Patients who do not achieve a PEF of \u00E2\u0089\u00A580% of predicted value after two SABA treatments should seek physician or emergency medical care[3]. Patients with severe exacerbations usually require ICU admission for more extensive monitoring and treatment. Patients who have previously experienced intubation or ICU admission for asthma, have had \u00E2\u0089\u00A5 2 hospital admissions for asthma in the past year, \u00E2\u0089\u00A5 3 or more ED visits for asthma in the past year, have had a hospital or ED visit for asthma in the past month, or are using >2 canisters of SABA per month may be at high risk for death from asthma. These patients may need more intensive treatment at the first sign of an exacerbation. A critical step in the management of acute asthma exacerbations is to identify the start of asthma exacerbation and initiate drug therapy. This avoids treatments delays, prevents worsening of exacerbations, and helps promote patient control of symptoms. Asthma educational programs that provide comprehensive training in recognizing early signs of exacerbation, using asthma medications and asthma treatment plans will help patients initiate early treatment at home and reduce the need for acute care. The latest clinical practice guidelines include the recommend use of a written asthma action plan that summarizes daily drug treatments and shows how to recognize and react to exacerbations. 41 Despite their potential effectiveness in asthma management, studies assessing the utility of written action plans are inconsistent. Benefits of action plans have been observed in at least two studies. The first, conducted by Cowie et al[163] compared asthma healthcare outcomes in groups of patients using either no written plan, a symptom-based written plan, or a peak flow-based written plan. Over the 6-month follow-up period, patients who utilized a peak flow-based written plan had significantly fewer urgent care visits (5 visits in 46 patients) compared to patients who received a symptom-based plan (45 visits in 48 patients) or no written plan (55 visits in 48 patients). A second study by Gibson et al. systematically reviewed 36 RCTs comparing usual care to care that included self-management education programs coupled with regular medical reviews and written action plans. The pooled analysis showed a 36% reduction in hospital admissions, 18% reduction in ED visits by; 21% reduction of missed school or work days, and a 23% reduction in nocturnal asthma symptoms[164]; however, subgroup analyses were not able to isolate a specific contribution of written action plans to these outcomes. Other studies have shown no effect of written action plans on various asthma outcomes. For example, in 2004, Toelle and Ram conducted a systematic review to determine whether the provision of a written action plan increases adherence and improves asthma outcomes[165]. The authors pooled results from 7 trials and found no consistent evidence for positive effects of written plans, but this may have been due to small sample sizes (75 to 150) and a small number of outcomes. Other RCTs have indicated no effect of written asthma action plans in reducing health service utilization[166, 167]. One 2-year RCT conducted by van der Palen[168] examined effects of education related to self-treatment on self-manageability during exacerbations and found a significant 42 improvement in self-confidence and self-management behavior during a hypothetical asthma exacerbation. These results were confirmed in a follow-up study that measured outcomes in the same groups of patients after 2 years, but showed no significant effect on clinical asthma status (i.e., frequency of exacerbations, mean number of outpatient visits for asthma per patient per year, and percentage of symptom-free days and nights)[169]. The cumulative results of these studies are inconsistent. One possible reason is that some asthma self-management plans are too general; some patients may require more specific information on how to identify and react to the onset of exacerbations. Studies have shown that 80-85 % of asthma patients are triggered by upper respiratory tract infections; these patients may need instructions on when to initiate drug therapy and which medication is needed in cases of infections. Some patients may have a high level of denial with respect to their asthmatic status, or may be unwilling to adhere to drug regimens. Adams et al. conducted a comprehensive asthma education program using monthly telephone contact to emphasize the importance of asthma action plans to participants[170]. This study showed that patients with high levels of disease denial and lower self-confidence had increased numbers of exacerbation-associated ED visits. In summary, asthma exacerbation is a major cause of patients\u00E2\u0080\u0099 poorer quality-of-life and increased burden on healthcare system. An effective written action plan facilitates the early detection and treatment of an exacerbation, further prevents unnecessary health services utilization. Key component of effective written action plan include: when to increase treatment, how to increase treatment, for how long and when to seek medical help[171]. As there is great inter-patient variability in asthma triggers, disease severity and medication use, written action plan has to be individualized in order to produce effectiveness. 43 1.10 ADHERENCE TO ASTHMA DRUG REGIMENS 1.10.1 Regimen adherence In the literature, the term \u00E2\u0080\u009Ccompliance\u00E2\u0080\u009D and \u00E2\u0080\u009Cadherence\u00E2\u0080\u009D are usually used to describe agreement between prescribed medications and patients\u00E2\u0080\u0099 actual use. There has been a decades-long debate about the definition and differentiation between the two terms. Haynes et al. defined compliance as \u00E2\u0080\u009Cthe extent to which a person\u00E2\u0080\u0099s behavior (in terms of medications, diets, or life-style changes) coincides with medical or health advice.\u00E2\u0080\u009D[172] Some researchers suggested that the terms adherence and compliance are interchangeable , but others claim they differ substantial, in particular that \u00E2\u0080\u009Cthe term compliance suggests a restricted medical- centered model of behavior, while the alternative adherence implies that patients have more autonomy in defining and following their medication regimens.\u00E2\u0080\u009D[172] This distinction is particularly important when considering chronic diseases, such as asthma, which involve complex drug management. Compliance is a term that refers to \u00E2\u0080\u009Cfollowing doctor\u00E2\u0080\u0099s orders\u00E2\u0080\u009D and indicates that patients are largely responsible for their daily care requirements[172, 173]. Using the term adherence gives patients more freedom to decide whether or not to follow physician recommendations, and indicates an equal role of patients in determining treatment protocol, and will be used in the follow discussions. Patient adherence to medical regimens is a major problem in chronic disease management. Despite numerous clinical practice guidelines that carefully define drug management approaches, many asthma patients still do not adhere to drug regimens; reported rates of non-adherence range from 30 - 70 %[174-176]. Studies have reported that on average, half of asthma medications are taken in adherence with the prescription, while the remaining half are taken at incorrect times or using techniques inconsistent with physician 44 instructions[177]. Non-adherence to controller medications has been reported to result in more frequent exacerbations[178, 179], increased health services utilization[178] and decreased quality of life. In Canada in 2000, the estimated economic burden of hospitalizations attributable to patient non-adherence to controller medications exceeded $1.6 billion[180]. 1.10.2 Measure of drug adherence Adherence to asthma drug regimens has been a concern for decades and a number of studies have attempted to accurately measure regimen adherence. These published measures can be divided into two main categories: direct and indirect methods. Direct methods include measurement of the level of medicine in blood and directly observed therapy. For example, early investigations used blood or urine analysis to directly quantify medication levels[181]. For example, adherence to theophylline\u00E2\u0080\u0094a previously widely used medication \u00E2\u0080\u0093was commonly measured using blood serum theophylline concentration[182]; sufficient use was defined as blood theophylline levels > 10 mg/ml[183]. However, such methods are less useful today, since commonly prescribed asthma medications such as becomethasone and salbutamol cannot be easily detected in bodily fluids due to the rapid and local elimination in respiratory system of these agents[176] and quantifying adherence requires other approaches. Another direct measure includes observing patients\u00E2\u0080\u0099 ability to of using an MDI. Manzella et al. developed a 10-item Inhaler Use Checklist (IUC)[184]. This IUC was completed by a trained administrator who directly observed the patients\u00E2\u0080\u0099 inhaler use, and awarded points for successful completion of each IUC item. There is currently insufficient data to assess the reliability of this approach. Indirect methods of adherence assessment include clinician judgment based on 45 conversations with patients on drug use, patient questionnaires, self-reports, pill counts, rate of prescription refills and patient diaries. Clinician judgment was commonly used in earlier investigations on drug adherence. Several factors have been shown to influence a physicians\u00E2\u0080\u0099 overall judgment including past clinical experience, as well as patient race, socioeconomic status (SES), and personality[176], making this measure subject to low validity and reliability. Previous studies have reported great overestimations of the degree to which patients adhered to medications[185, 186]. Steele et al. examined tape recordings of patients with hypertension communicating with their physicians and reported that physicians failed to identify 47% of patients who were not adherent[187]. Questionnaires polling patients about drug use and symptoms have been widely used to assess adherence and healthcare outcomes. Kandane-Rathnayake et al. found that in a group of Australian adult asthma patients, more than one fourth of them did not use controller medications adequately[188]. Additionally, Marco et al. investigated the adequacy of drug usage in a random sample of Italian patients using the GINA guidelines and found that 48% of patients with persistent asthma did not receive sufficient controller medications[189]. Likewise, Finkelstein et al. reported that 73% of pediatric patients with persistent asthma underused controller medications, and that half of those reported no usage of controller medication[190]. Yoos et al. found that 32% of a small sample of pediatric patients with asthma (n=227) who experienced asthma symptoms during a 3-month period did not report any use of controller medications[191]. Self-reporting can be subject to patients\u00E2\u0080\u0099 recall bias. One early investigation compared self-reported use of inhalers with objective adherence data collected by an electronic monitoring device. They found that though all patients self- reported using the inhaler on more than half of the study days, the objectively measured use 46 showed that only 52.6% of the patients were adherent [192]. Another widely used approach, based on prescription dispensing data, uses prescription refill records to estimate patients\u00E2\u0080\u0099 adherence. Currently, the two most commonly used measures of medication adherence include Medication Possession Ratio (MPR) and the Proportion of Prescribed Days Covered (PPDC). MPR is the ratio of days medication supplied to days in a time frame. For example, Mattke et al. used MPR to measure ICSs adherence and reported a median MPR of 15%[193]. The PPDC measure was calculated by dividing the total days\u00E2\u0080\u0099 supply dispensed by the total days\u00E2\u0080\u0099 supply prescribed. Pando et al. used the PPDC methodology to measure adherence to ICSs in pediatric patients. The authors reported that in patients who required >3 doses of SABA per week, 20% only received one prescription for ICSs with no prescribed renewals. The mean PPDC was 62%, which suggested that patient adherence to these drugs was suboptimal[194]. These two methodologies allow estimating the days that patients are on medications within a time frame. However, these methodologies do not take into account whether patients\u00E2\u0080\u0099 adherence to medications is at an appropriate dose or when patients initiate the drug or change the doses of the medications. In 2000, Lagerlov et al. investigated the appropriateness of SABA and ICS use in adult asthma patients using a traffic light approach. Optimality was determined by patient use of inhaled SABAs and ICSs. Green was used to indicate appropriate therapy, yellow to indicate uncertain therapy, and red to indicate inappropriate therapy[195]. This study reported that 34% of these patients did not use asthma regimens appropriately[195] but unfortunately this development was based on a group of physicians\u00E2\u0080\u0099 opinions and did not link adherence to healthcare outcomes. 47 What is needed is a classification approach that provides health planners, clinicians and patients with outputs that can be pragmatically used to improve asthma care. The traffic light approach developed by Lagerlov et al. is a clinically relevant and easy method to identify patients who were inappropriate regimen users. Our study extended this development through incorporating published asthma clinical practice guidelines and linking health services utilization data in a complete population. Our comparison of obviously optimal users (e.g., Lagerlov\u00E2\u0080\u0099s Green Light patients) to obviously suboptimal users (e.g., Lagerlov\u00E2\u0080\u0099s Red Light patients) would allow all three stakeholder groups to improve asthma care by allowing categorization of patients to drive appropriate intervention strategies. In other words, Green Light patients need different interventions (or perhaps no intervention) while Red Light patients are likely to need different and more complex asthma management strategies. 1.10.2 Reasons for non-adherence A variety of factors contribute to nonadherence in asthma therapy[13, 175, 196], and can generally can be categorized as physician- patient- or medication-related. Other factors include inappropriate expectations, poor supervision/training or follow up, anger about one\u00E2\u0080\u0099s condition or its treatment, underestimation of severity, cultural issues, stigmatization, forgetfulness or complacency, and attitudes toward ill health and religious issues[197]. 1.10.2.1 Physician factors Patient adherence is largely impacted by physician adherence to clinical practice guidelines (e.g., prescribing appropriate medications, assess asthma control and show how to use inhalers). Studies assessing potential barriers to physician adherence to guidelines identified two primary factors: 1) Physician knowledge, including lack of awareness or lack 48 of familiarity with guidelines; and 2) Physician attitude toward guidelines, including lack of agreement, lack of self-efficacy, lack of outcome expectancy, or inertia of previous practice [173, 198]. Wolff et al. reported that only 27% of U.S. FPs knew where to find a clinical practice guideline and that approximately one third of physicians were unfamiliar with the content of specific guidelines[199]. Conducting educational programs at the physician individual level has been shown to be successful, but is labour-intensive and costly. With respect to attitude toward clinical practice guidelines, the majority of physicians tend to agree that guidelines are good educational tools and convenient sources of advice, and that they are developed to improve the quality of health care. This positive attitude, however, was not found in all studies; approximately one quarter of physicians view guidelines negatively, describing them as oversimplified, too rigid to apply to individual patients, and a challenge to physician autonomy. When researchers examined guideline implementation, they discovered that few physicians reported making changes to their clinical practice based on published guidelines[200, 201]. Indeed, guidelines ranked well below other sources of information, including continuing medical education, discussions with colleagues, and review articles, in influencing physician practice[201]. In a 1997 Canadian study, Hayward et al. suggested that a major challenge in implementation is the generation of efficient strategies that facilitate physicians reading, remembering and using new guidelines[200]. The authors suggested the use of an authoritative endorsement in implementing guidelines, or that they are presented to doctors in a format that easy to read and refer to, such as pamphlets or pocket cards. 49 1.10.2.2 Patient factors Several studies highlight patient fear of potential ADRs caused by regular use of medications. In particular, parents of asthmatic children expressed concern that ICS use might cause shorter height and increased irritability, as well ICS addiction or dependency, in their children[172]. Several studies also reported patients and/or caregiver concerns regarding the usefulness of controller medications[202, 203]. Due to the intermittent nature of asthma, many patients and/or caregivers believe that their asthma is not severe enough to require daily medication; rather, they believe it is sufficient to take asthma medications when acute symptoms occur and do not realize the importance of using ICSs regularly. Other evidence indicates that patients misunderstand the importance of taking ICS. Since ICS do not provide immediate effects, some patients believe they offer no real benefits. As such, patients\u00E2\u0080\u0099 adherence to long-term controller medications has been reported to be much lower than that of reliever medications[204, 205]. Patients also reported a number of difficulties related to medication-use, including having to take several different medications or multiple doses per day, inconvenience of carrying medications around, and difficulty using inhalers. Some parents reported difficulty in administering their children\u00E2\u0080\u0099s medications, due in part to the children\u00E2\u0080\u0099s resistance. Different factors interfering with willingness to use daily asthma medications in pediatric patients have been identified in previous studies. Children, particularly adolescents, expressed that taking their medication is often undesirable when it reminds them that they are different from \u00E2\u0080\u009Cnormal\u00E2\u0080\u009D children and when it reveals their illness to other children[202]. This can lead to ongoing struggles between parent and child over the child\u00E2\u0080\u0099s willingness to take 50 medication. As children grow older, it is assumed that the responsibility for remembering to take the medication shifts from the parent to the child. Successful communication between a patients and their physician is a useful tool to enhance adherence. Successful communication facilitates more efficient delivery of information, offers support, and encourages patients to ask questions and express concerns. One study of 101 asthmatic children and their parents, demonstrated that that both children and their parents preferred to obtain more information on the disease and medication than the pediatricians anticipated. Additionally, both physicians and patients may benefit from interventions focused on better understanding the nature of asthma and drug management. For example, clarifying the multiple negative consequences of non-adherence and highlighting the positive outcomes of adherence to patients should increase adherence to controller medications. Negative patient attitudes toward pharmacotherapy should be identified and addressed, and multiple studies suggest interactive methods to deal with this. For example, a physician might ask, \u00E2\u0080\u009Cwhat evidence would convince you that the medicine did help?\u00E2\u0080\u009D and then devise a way to collect that information. Alternatively, physicians might ask a patient if he or she would be willing to take a medication as recommended for a period of time and then examine measures of lung function to see if the drug had helped. Such interventions focused on modifying patient beliefs have been show to improve adherence and positive outcomes[206, 207]. 1.10.2.3 Medication factors The primary medication-related factors include difficulties with inhaler devices, complex regimens and ADRs[172]. Most asthma medications are delivered via inhalation to optimize delivery to the lung 51 and minimize potential ADRs. However, inhalation was reported to be the most disliked aspect of drug management by asthmatic children and their parents. Indeed, multiple studies have indicated a preference for oral over inhaled administration[208]. For a chronic disease like asthma, patients may take multiple medications to control their asthma symptoms. The complex regimens may result in patients\u00E2\u0080\u0099 poor understanding of medication use, and patients\u00E2\u0080\u0099 forgetfulness of taking or refilling medications, Studies have reported that combing two drugs into one inhalation formulation reduces the number of doses required for administration, and thus decreases non-adherence due to forgetfulness. Stoloff et al. reported that this approach doubled adherence (4.06 refills per 12-month period with fixed combinations versus 2.35 refills with each compounds in a single inhaler, respectively). With a fixed combination fluticasone/salmeterol adherence was almost identical to oral treatment with montelukast (4.51 refills per 12-month period). In comparison with the inhalation, oral administration was always associated with better adherence rates. Several other reasons make adherence to asthma treatment regimens problematic: medication regimens can be long in duration and involve multiple medicines; dosing with different medications can occur on both a fixed schedule and an as needed basis; and because patients go through periods of symptom remission[176]. As described in Section 1.7.1, use of ICS may cause ADRs including oral candidiasis and hoarse voice. A Canadian survey of 603 patients with asthma reported the most common fear of taking asthma medication is the potential ADRs[209]. 1.11 PROGNOSIS OF ASTHMA Epidemiological studies have indicated a reduction in asthma symptoms with age. Longitudinal cohort studies have reported that 40-75% of patients whose asthma began in 52 infancy or childhood did not experience any asthma symptoms during their adulthood[210- 213]. Symptoms returned after a period of remission in 12-35%[212-214], while symptoms persisted through adulthood in approximately 30% of patients[214]. Levels of asthma control, as well as asthma severity in childhood and adolescence have been identified as predictors of adult asthma outcomes. Poorer asthma control in childhood and adolescence has been associated with irreversible airway damage in adulthood. Lung growth (i.e., frequent pulmonary alveolar multiplication and lengthening and enlargement of the airways) occurs during childhood and adolescence and is complete by the end of puberty. Evidence has shown that frequent exacerbations during early childhood may result in significantly reduced lung function during adolescence[215, 216]. Furthermore, chronic airway inflammation during childhood and adolescence may lead to more irreversible airway limitation in adulthood[217]. The Melbourne Epidemiological Study of Childhood Asthma is the lengthiest study completed thus far; this study recruited 479 children (374 with asthma and 105 controls) at the age of 7 or 10 years and followed up at 7-year intervals until 42 years of age. At the time of recruitment, patients were divided into groups: those with intermittent symptoms (i.e. symptoms occurring only with upper respiratory tract infections (URTIs)), persistent asthma (symptoms not associated with URTIs), severe persistent asthma (i.e., symptoms begin before the age of 3 years, persistent symptoms at 10 years, or an FEV1/FVC \u00E2\u0089\u00A450%). The disease severity at age 42 has not changed from that at age 35. The proportion of cases who did not have asthma exacerbations in the past 3 years has increased steadily from 20% at age 14 years to 40% (126/317) at age 42[218]. 53 1.12 RESEARCH GOALS, HYPOTHESIS AND OBJECTIVES The overarching goal of this work is to identify patients with suboptimal drug regimens according to published clinical practice guidelines and determine the association between asthma drug regimen optimality and health services utilization. Objectives (1) To develop an optimal asthma case definition and more accurately identify patients with treated asthma in B.C. (2) To characterize the burden of asthma in patients aged 5-55 years of age in B.C. between 1996 and 2009 based on this optimal asthma case definition. (3) To determine the association between asthma drug regimen optimality defined by patients\u00E2\u0080\u0099 dispensings of SABA with or without ICS, and health services utilization (e.g., ED, hospital services) for asthma exacerbations. Hypothesis a) Patients who are on suboptimal regimens are more likely to visit the ED or be admitted to hospitals for asthma compared to those who are being treated optimally. b) Over time switching from suboptimal to optimal regimens is associated with reduced likelihood of asthma-related ED visits and hospital admissions compared to continuing on suboptimal regimens. (4) To determine the association between regular/intermittent use of ICS and asthma- related ED visits or hospital admissions. Hypothesis Patients who are using ICS regularly are less likely to visit ED or be admitted 54 to hospitals for asthma exacerbations compared to those who are using ICS intermittently. 55 CHAPTER 2 : BURDEN OF ASTHMA IN BRITISH COLUMBIA 56 2.1 SYNOPSIS Patients with asthma have been defined according to different criteria in previous studies using health services utilization data. Asthma burden estimates using health services utilization data vary significantly depending on which criterion is used. There are problems in the existing asthma case definitions. Patients who only receive physician-diagnosed asthma once may receive working diagnosis and those who only obtain 1 or 2 asthma prescription drugs may receive these medications for diagnostic purposes (e.g., those with only a single set of prescriptions for the two primary drugs used to treat asthma). Due to a lack of available prescription dispensing data in some jurisdictions, previous studies have mainly used hospital admission and physician visit criteria to identify patients with asthma. This chapter develops an optimal asthma case definition. Using this optimal asthma case definition, this chapter estimates the burden of asthma in patients with treated asthma in the entire province of B.C. from 1996 to 2009. This Chapter defines patients with treated asthma if they meet any of the following criteria during a 12-month period: 1 or more hospital admission with the principal diagnosis as asthma; or 2 or more physician visit claims for asthma; or 3 or more asthma drug dispensings. This case definition is used to characterize the burden of asthma in B.C. from 1996 to 2009, in terms of: 1) epidemiology (e.g., prevalence and incidence); 2) health services utilization (e.g., family physicians (FP), specialist, ED visits and hospital admission); and 3) drug dispensings. Asthma prevalence was stable around 2.3% in patients 5-55 years of age in B.C. from 1996 to 2009. Asthma incidence was stable around 0.7% from 2001 to 2009. On average, each patient had 1.5 FP visits in 1996 and 1.1 FP visits for asthma in 2009. Use of specialist 57 services for asthma changed from 40 to 20 per 100 patients, use of ED services decreased from 23 to 6 per 100 patients, while use of hospital services declined by 67% (3 per 100 patients in 1996 and 1 in 2009). For asthma prescription medications, SABAs were dispensed to an average of 76% of patients in each study year. Percentage of patients receiving ICS decreased from 61% in 1996 to 42% in 2009. LABA and LABA/ICS combined inhalers were dispensed to patients since 1999 and increased to 33% in 2009. Use of FP and specialist services was significantly higher in large cities compared to rural and remote areas. In contrast, use of ED and hospital services was significantly higher in rural and remote areas than the use in large urban cities. In summary, the burden of asthma declined significantly over the 14-year period. The burden estimates were significantly different from findings from previous studies. This discrepancy is likely due to different case definition criteria that have been used in each study. Physicians and researchers need to be cautious when interpreting asthma burden estimates reported in different studies. 2.2 METHODOLOGY 2.2.1 Data Sources The B.C. linked health service utilization and drug prescription dispensing databases contain detailed and comprehensive information on all B.C. residents registered with the Medical Services Plan (MSP). This excludes about 3% of B.C. residents, including Native and Inuit residents (whose medical insurance is covered through Health Canada's First Nations and Inuit Health Branch) as well as federal employees, e.g., members of the Royal Canadian Mounted Police [219]. The linked database integrates physician visits records, 58 hospital admission records, prescription data, comorbidity information, socioeconomic status (SES), and regional information. Residency status is established after being present in the province at least six months. Table 2.1 summarizes the B.C. linked databases used in this study and key variables contained therein. All files used in the analyses were stripped of personal identifiers by the BC Ministry of Health and therefore did not contain any variables that would permit identification of individuals. Table 2.1 Data sources of the present research Data Set Data Records Client Registry Gender, date of birth, Local Health Area (LHA) of patient\u00E2\u0080\u0099s residence, Duration registered with the Medical Services Plan (MSP) MSP Physician Visit Database Practitioner ID, practitioner specialty, date of service, fee item code, patient ID, referring physician code, service location, diagnostic code, amount paid and other codes Discharge Abstracts Database (DAD) of Hospitalization records Patient ID, levels of care, physician services, case mix group, types of diagnosis, procedures performed, resource intensity weight, length of stay, admission and discharge dates, physician IDs, hospital ID, transfer dates, death flag, admission route PharmaNet Prescription Dispensing Database Drug names, drug identification number, strength, days supply, dosage form, prescriber identification number, pharmacist identification number, and date of service provided by pharmacist Comorbidity Database Annual sets of Expanded Diagnosis Cluster (EDC) codes for each patient based on annual application of Johns Hopkins ACG case mix algorithm to MSP and DAD records Census Database Annual sets of socioeconomic status quintiles for each patient There is a substantial overlap between some respiratory diagnoses (e.g., bronchiolitis, 59 reactive airways disease) and asthma, particularly in children, some patients may not be coded as having asthma in health services utilization databases despite having asthma. To ensure that the databases used in the present study captured all asthma patients, seven respiratory diseases (excluding asthma) which are commonly comorbidities in asthma patients were used to form the source population in addition to asthma diagnosis. Specifically, the source population was defined as having at least one of the following respiratory diagnoses between April 1, 1991 and March 31, 2010: Medical Services Plan (MSP) claims database of fee-for-service physician services in which diagnostic codes (based on the International Classification of Diseases, ninth version; ICD-9) 466 (acute bronchitis and bronchiolitis), 472 (chronic pharyngitis and nasopharyngitis), 490 (bronchitis, unspecified), 491 (chronic bronchitis), 492 (emphysema), 493 (asthma), 494 (bronchiectasis), or 496 (COPD) were entered on the record. All ICD-9 and corresponding ICD-10 codes are summarized in Table 2.2. Discharge Abstracts Database (DAD) of hospital admissions in which any of the above ICD-9 (or equivalent ICD-10) codes which occur in any of the 16 diagnosis fields (data 1991-2000) and 25 diagnosis fields (data 2001-2009). 60 Table 2.2 ICD-9 and ICD-10 codes for respiratory conditions tracked in the asthma cohort Diagnosis ICD-9 ICD-10 acute bronchitis or bronchiolitis 466 J20, J21 chronic pharyngitis / nasopharyngitis 472 J31 bronchitis - other 490 J40 chronic bronchitis 491 J41, J42 emphysema 492 J43 asthma 493 J45, J46 bronchiectasis 494 J47 chronic airways obstruction 496 J44 For each patient identified using these criteria, all respiratory-related physician visits, hospitalization records and prescription dispensing records for respiratory disease-related medications records from April 1, 1996 and March 31, 2010 were obtained. Prescription data were obtained from PharmaNet, a province-wide computer network that links pharmacies with a central database and records all prescription drug dispensings regardless of source of payment. British Columbia Medical Services Plan database The MSP database contains fee-for-service (FFS) billing records (e.g. visit dates, locations, principal diagnoses) filed by B.C. physicians for services to B.C. and non-B.C. residents as well as payment information for B.C. residents who received physician services in Quebec, the U.S. and other countries. When physicians claim FFS, clinic staff usually transforms the patient\u00E2\u0080\u0099s medical chart to a computer-based telecommunications (Teleplan) system within 90 days of the service date[220]. Teleplan is available for physicians to submit claims, notes and requests to MSP, and to receive payment statements and reject patient MSP claims[220]. During 1996 and 2009 approximately 70-80% of B.C. physicians work on a 61 FFS basis[221, 222]. In 2002 (the most recent available data), 2,250 physicians were funded by alternative payment arrangements. More specifically, an estimated 24% of psychiatry, 12% of oncology, 10% of paediatric services, and 4% of emergency services were funded through the Alternative Payment Program (APP)[223]. The APP database includes salary, sessional payment, and service agreement data for each physician who is funded through the APP program. Salaried physicians are primarily on staff at hospitals, private corporations, government agencies or universities. For example, many ED physicians working in the three main tertiary hospitals (B.C. Children\u00E2\u0080\u0099s, Vancouver General and St. Paul\u00E2\u0080\u0099s) or academic health centers (physicians employed by the B.C. Cancer Agency, Riverview Hospital) or the Centre for Disease Control, and regional and provincial medical health officers in Greater Vancouver are salaried. Physicians who received sessional payments (based on time, where one sessions equals 3.5 hours) are generally those who work in mental health and palliative care. Based on the most recent data available, the percentage of physicians paid through the APP increased from 21.3% in the fiscal year 1999 (i.e. April 1, 1999 to March 31, 2000) to 28.4% in 2002 and 29.3% in 2005[221, 222, 224]. Despite these alternative forms of payment, which are not included in the MSP database, it has been estimated that MSP data covers visit and payment information for over 90% of FPs [222] and 77% of specialists [225]. Hospital Admission Discharge Abstract Database (DAD) All B.C. hospitals submit information on acute inpatient care and day surgery separation to the Canadian Institute for Health Information (CIHI). When patients are discharged, their medical records are coded and abstracted based on CIHI criteria. The resulting DAD abstract, including coded diagnostic, intervention and patients\u00E2\u0080\u0099 demographic information is submitted to CIHI. After checking record quality, CIHI returns reports to each 62 hospital for further review and corrections, if need. Completed records are then used in the production of CIHI reports and disseminated to the provincial Ministry of Health [226]. British Columbia PharmaNet Database PharmaNet is administered by the Ministry of Health in order to facilitate prescription payment and patient-pharmacist interaction. It also fosters drug utilization and outcome research. It captures comprehensive information from every outpatient prescription dispensed in B.C., regardless of source of payment. As of 2007, B.C. PharmaNet contained over 47 million prescription dispensing records[227]. Upon presentation of a prescription, pharmacists transmit the prescription information on PharmaNet, which includes the patient\u00E2\u0080\u0099s personal health number, identification numbers for the prescriber, pharmacy and pharmacist, and drug information (e.g., drug identification number, quantity, strength, form, and dispensing date). 2.2.2 Quality of data Health service utilization data are frequently used to estimate disease prevalence and healthcare burdens resulting from chronic conditions, as well as to study healthcare outcomes, predict future health service utilization and evaluate prevention and treatment interventions[228]. Disease case definitions are primarily based on ICD diagnostic codes; as such, the accuracy of diagnostic codes documented in population-based health service utilization has been examined. There are several issues to be considered when evaluating the accuracy of diagnostic codes in databases: the accuracy of diagnosis, the accuracy of transcription of ICD codes from medical charts to databases, and the method used for evaluation. As to the quality of MSP database, the quality of the MSP database in B.C. has not 63 been examined. Using similar data from Ontario, To et al. compared consistency between medical charts and physician visit database and reported an overall agreement of 99.4% in the diagnostic field and 98.1% in the use of asthma diagnostic code[71]. Given the consistency of asthma diagnostic information in the physician visit database in Ontario and medical charts, as well as the similar type of information contained in the B.C physician visit database, the use of the MSP data to identify patients with asthma is a reasonable representation of reality. To ensure the accuracy and quality of hospital separation DAD databases, CIHI has conducted a series of cross checks and data quality control measures. These include: implementation of a standardized abstracting procedure through the use of abstracting software; annual tests of abstracting software; offering educational programs about coding and abstracting to abstractors, conducting correction processes and data quality-assessment studies (i.e., re-abstracting patients\u00E2\u0080\u0099 medical records to test the consistency with the original abstracted information)[226]. The most recent (and the largest) validation study in Canada to date involved 14,500 discharges from 18 hospitals in Ontario between April 1 st , 2002 to March 31 st , 2004 and reported an 85% exact match on diagnosis codes over the study period[229]. The accuracy of PharmaNet data for adherence assessment was examined during a six-month prospective study that compared PharmaNet prescription drug dispensing records with the Medication Event Monitoring System for each study patient[230]. This study reported a high level of agreement between the two methods and suggested that using refill records in the PharmaNet data accurately reflect medication adherence for the majority of patients. 64 2.2.3 Data preparation To prepare the data for analysis, a series of examinations of data integrity were conducted. For the MSP data, several issues are associated with the raw physician visit data in the MSP database, including: duplication of records; \u00E2\u0080\u009CNo charge\u00E2\u0080\u009D referral records; unpaid claims; and retroactive adjustments. Since these records do not reflect patients\u00E2\u0080\u0099 physician visits, these records (0.4%) were removed from our analysis. When preparing the hospital admission data for analysis, 0.1% of discharge dates were recorded as occurring after the patients\u00E2\u0080\u0099 date of death. A new variable was created to indicate this error. For B.C. PharmaNet data, SABA and ICS dispensing records were checked. SABAs inhalers contain 50 - 200 inhalations, while ICS inhalers contain 8 - 240 inhalations, depending on brand and packaging. For some prescription drug dispensing records, the supplied quantity was not a multiple of the packaging units, and instead correlated with the number of inhalers provided. As such, these values were adjusted by multiplying the number of inhalations per canister by the recorded supplied quantity. For example, one inhaler provided 200 doses; the value \u00E2\u0080\u009C1\u00E2\u0080\u009D was changed to \u00E2\u0080\u009C200\u00E2\u0080\u009D. In total, 0.26% of dispensation records were modified. Patient birth year and month are included in the demographic, MSP, DAD and PharmaNet databases. To ensure consistency in patient age calculations, birth years and months from the demographic database were used and birth years and months from other databases were not used. All the processes were performed at the level of individual records to maximally ensure claim records accuracy. 65 2.2.4 Cohort definition Patients were classified as having asthma if they satisfied any of the following criteria in each study year: 1 or more hospital admission with asthma as the principal diagnosis based on ICD-9 code 493 or ICD-10 code J45; 2 or more physician visits for asthma as the principal diagnosis based on ICD-9 code 493; or 3 or more asthma drug dispensings. Asthma drugs include SABA, ICS, LABA, LTRA, mast cell stabilizers, xanthenes and anti-IgE agents. Generic names of these drug categories were summarized in Table 2.3. Both two physician visits and three asthma drug dispensings in a given year provide some assurance that these patients have chronic asthma, and were not just seen by a physician a single time or provided drug therapy for trial purposes only. Patients younger than 5 years of age were excluded because of the uncertainties associated with asthma diagnosis. Asthma symptoms - wheezing, coughing and chest tightness can arise from other conditions such as upper respiratory tract infections (URTI), colds or pneumonia, which commonly occur in pediatric patients and may lead to temporary working diagnosis or further misdiagnosis. We also excluded patients aged 55 years or older since older patients are more likely to develop chronic obstructive pulmonary disease (COPD), which is a common comorbidity in asthma patients. Figure 2.1 summarizes the study cohort. 66 Table 2.3 Prescription medications for the treatment of asthma by class Short-acting \u00C3\u009F2-agonists Systematic Corticosteroids Anti-allergic agents \u00EF\u0082\u009F Salbutamol \u00EF\u0082\u009F Hydrocortisone \u00EF\u0082\u009F Ketotifen \u00EF\u0082\u009F Terbutaline \u00EF\u0082\u009F Methylprednisolone \u00EF\u0082\u009F Sodium cromoglycate \u00EF\u0082\u009F Fenoterol \u00EF\u0082\u009F Prednisone \u00EF\u0082\u009F Nedocromil sodium \u00EF\u0082\u009F Dexamethasone Long-acting \u00C3\u009F2-agonists \u00EF\u0082\u009F Betamethasone Leukotriene-receptor antagonists \u00EF\u0082\u009F Salmeterol Anticholinergics \u00EF\u0082\u009F Montelukast \u00EF\u0082\u009F Formoterol \u00EF\u0082\u009F Ipratropium \u00EF\u0082\u009F Zafirlukast \u00EF\u0082\u009F Formoterol/budesonide \u00EF\u0082\u009F Tiotropium \u00EF\u0082\u009F Fluticasone/salmeterol \u00EF\u0082\u009F Xanthines Inhaled corticosteroids \u00EF\u0082\u009F Aminophylline \u00EF\u0082\u009F Beclomethasone \u00EF\u0082\u009F Theophylline \u00EF\u0082\u009F Budesonide \u00EF\u0082\u009F Oxtriphylline \u00EF\u0082\u009F Flunisolide \u00EF\u0082\u009F Fluticasone \u00EF\u0082\u009F Triamcinolone \u00EF\u0082\u009F Ciclesonide 67 Figure 2.1 Form of study cohort Note: MSP indicates Medical Services Plan; ICD indicates International Classification of Diseases. 2.2.5 Validity of asthma case definitions A variety of asthma case definitions on the basis of population-based health services utilization data have been used by researchers (see To et al.[231], Kozyrskyj et al.[69], Wilchesky et al.[68], and Huzel et al. [67] for examples). Asthma cases are typically defined by physician visit, hospital admission and/or prescription dispensing criteria for population- based epidemiological studies. As discussed in the Introduction, there are multiple challenges associated with the clinical diagnosis of asthma, in particular, the accuracy of diagnosis and the accuracy of transcription from medical charts to electronic database can significantly impact the validity of asthma case definition. Using a single physician visit as a criterion is 68 also problematic. Patients may visit a physician for multiple reasons, but MSP database contains only one diagnostic field; thus, that diagnoses other than asthma could be entered. In addition, patients could receive asthma as part of a differential diagnosis, but the specific diagnosis of asthma might be excluded upon further visits. Because of these confounding issues, it is critical to test the validity of asthma case definitions. Three strategies are frequently used to test the validity of case definitions obtained from population-based health service utilization data: testing the agreement between different definitions; using statistical methods; and most commonly, using a \u00E2\u0080\u009Cgold standard\u00E2\u0080\u009D comparison. The traditional approach to testing case definition validity is to examine the agreement between multiple case definitions. This method expresses agreement using a kappa value, but cannot determine which case definition is more accurate. Another approach is to use statistical methods to evaluate case definition validity. One commonly used approach is latent class modeling (LCM), a grouping methodology that calculates scores based on a set of observed categorical variables and assigns patients to different classes according to these scores. The most commonly used validation approach is to use survey data or medical chart review as a \u00E2\u0080\u009Cgold standard\u00E2\u0080\u009D with which to compare results derived from health service utilization data. Wilchesky et al. compared the consistency between physicians\u00E2\u0080\u0099 medical charts in patients 66 years of age and older (the gold standard) with case definitions based on one or more physician visits for asthma identified in the Quebec physician claims data[68], and reported a sensitivity of 43% and specificity of 99%. This indicates that only 43% of asthma cases determined by the case definition were recorded in the medical charts, and that 69 99% of patients who did not have an asthma diagnosis in the database were not diagnosed with asthma in the medical charts. This low sensitivity and high specificity is likely indicative of limitations associated with having only one diagnostic field in the database; this study focused on an older patient group who likely suffer from multiple comorbidities and visit physicians for multiple reasons, one of which could be entered into the database. Prescription drug therapy is the major treatment for asthma, thus, medication dispensing records play an important role in asthma case definition. Kozyrskyj et al. explored the validity of their asthma case definition. Asthma cases in their study were defined by 1 or more physician visit or hospital admission for asthma (ICD-9 code 493) in one year or 1 or more prescriptions for an ICS, cromoglycate, or ketotifen concomitant with an inhaled or oral beta agonists, or 2 or more prescriptions for an inhaled or oral beta agonist in children 5 to 15 years of age[69]. The gold standard was defined as the dispensing of an asthma drug. This study reported a sensitivity of 74% and a specificity of 91%. There were several problems associated with the methodology of the Kozyrskyj et al. study. The authors used 1 or more physician visit as a criterion in the asthma case definition, thus patients who only had 1 physician visit for asthma during their study period were regarded as having asthma. As described previously, these patients may receive a working diagnosis of asthma. Patients were also identified by the criterion of at least one controller medication combined with at least one quick relief medication. Also, patients who were identified as having asthma based on the gold standard may not have asthma with certainty, as those patients who only received 1 controller medication and/or 1 quick relief medication may be prescribed these medications for diagnostic purpose or for virus-induced respiratory infections. Both asthma case definitions and gold standard involve misclassification of patients. 70 Huzel et al. tested the validity of using 1 or more physician visits for asthma as a case definition in patients 20-44 years of age in Manitoba using a gold standard of acquired survey data and reported a sensitivity of 70% and a specificity of 99.8%[67]. One concern arising from this study involves the self-reported nature of past diagnoses in population- based surveys, which makes case identification susceptible to recall bias. As well, questionnaire-based studies usually define patients as having asthma based on the question \u00E2\u0080\u009Chave you been professionally diagnosed with asthma in the previous 12 months\u00E2\u0080\u009D[232], resulting in the inclusion of one-time and potentially incorrect diagnosis. 2.2.6 ICD-9 and ICD-10 diagnostic codes for asthma The International Classification of Diseases is a diagnostic coding system developed by the World Health Organization and the U.S. National Center for Health Statistics. The ICD-9 was published in 1979, and has gradually transitioned to the ICD-10 since 1999. In the present study, ICD-9 codes were used to identify asthma-related physician visits and hospital admissions from 1991 to 2001 and ICD-10 codes were used to identify hospital admissions since 2001. Most ICD-9 numbers consist of a three-digit code containing one or two decimal places. The first three digits represent the disease classification and the following refer to sub-classifications. For example, ICD-9 code 493.0 indicates extrinsic asthma, 493.2 refers to chronic obstructive asthma, and 493.9 indicates unspecified asthma. On the other hand, ICD- 10 codes are broken down into \u00E2\u0080\u009Cchapters\u00E2\u0080\u009D and \u00E2\u0080\u009Csubchapters\u00E2\u0080\u009D and start with a letter and followed by two numbers. The letter indicates the class of disease, while the numbers report the details of the disease. For example, ICD-10 codes starting with the letter J represent diseases of the respiratory system, ICD-10 codes starting from K mean diseases of the 71 digestive system, and ICD-10 codes starting from J indicate respiratory diseases. Compared with the ICD-9 coding system, ICD-10 contains more detailed categories (8,000 vs. 4,000), three new additional chapters, regrouped conditions and new coding rules. The comparability and accuracy of using ICD-10 and ICD-10 diagnostic codes has been examined in previous studies. The U.S. National Center for Health Statistics (NCHS) began testing the comparability ICD-10 and ICD-9 diagnostic coding system via using population-based mortality data in the mid-1990s. Investigators calculated a comparability ratio by dividing the number of ICD-10-coded deaths by the counts of ICD-9-coded deaths. This ratio has been calculated for 113 selected causes of death. The comparability of asthma- related (ICD-9 code 493 and ICD-10 J45 and J46) mortalities is 0.89[233], thus 89% of code usage is comparable. Juhn et al. tested the accuracy of ICD-9 code 493 use using medical record review as a gold standard and reported an accuracy of 82%[234]. Similarly, Kem et al, reported an accuracy of 86% associated with the use of asthma or respiratory disease-related ICD-9 codes comparing with medical chart review[235]. Juurlink et al. re-abstracted 14,500 hospital DAD records from 18 hospital sites in Ontario and reported an accuracy of 86% in the use of asthma-related ICD-10 codes in the DAD database[236]. Thus, both the comparability and accuracy of ICD-9 and ICD-10 codes are relatively high. 2.2.7 Estimation of asthma burden in B.C. 1996-2009 2.2.7.1 Study design This study is a trend analysis and calculated the annual prevalence and incidence of asthma, as well as asthma-related health services utilization and prescription drug dispensing 72 from 1996 to 2009. 2.2.7.2 Study measures Measure of patients\u00E2\u0080\u0099 characteristics Patient age and gender have an impact on health service utilization. Although multiple databases contain date-of-birth information, we used birth year and month records in the demographic database to calculate age. Patients were grouped into two age categories (5- 11 years and 12-55 years) based on clinical practice guidelines that provided specific age- group recommendations. Because asthma-related patient characteristics differ in young children, adolescents, young adults and older adults, the two age groups were subcategorized into four groups: aged 5-11 years; 12-18 years; 19-34 years; and 35-55 years. Known differing characteristics between these age groups include asthma development, exacerbations, use of health services, and barriers in adherence [3]. 2.2.7.3 Measure of prevalence and incidence of asthma The burden of asthma was estimated at three levels: epidemiological characteristics (prevalence and incidence), use of health services (e.g., use of FP, specialist, ED and hospital services) and prescription dispensing rates. Prevalence and incidence of asthma were calculated in each fiscal year from 1996 to 2009. Annual prevalence was calculated by dividing the number of asthma patients by the provincial population. The population of B.C. residents classified by age and gender from 1996 to 2009 was estimated using B.C. Statistics Population Extrapolation for Organization Planning with Less Error (Version 34 (PEOPLE 34) software, B.C. Ministry of Vital Statistics, Victoria, B.C.)[237]. Patients were defined as having newly diagnosed asthma if they did not fulfill the asthma case definition in the 73 previous all five years. Since the earliest complete datasets that was part of our study was for 1996, rates of newly diagnosed asthma were calculated from 2001 to 2009. Rates of newly diagnosed asthma were calculated by dividing the number of patients who met the criteria for of newly diagnosed asthma in a given year by the number of patients at risk for asthma (the total population minus the number of people who met our case definition criteria in the previous five years). 2.2.7.4 Identify health services utilization and prescription drug dispensing records for asthma Emergency department visits ED visits provide urgent medical care as well as primary care when FPs are unavailable[238] and is an important outcome measure for asthma. Although there is no individual-level ED visit database available in B.C., most ED use can be captured by FFS payments made to physicians through MSP. In 1998, there were 1.6 million ED visits in B.C. and 90% were paid through the FFS system[239]. In 2005, 80% of ED visits were paid through the FFS system [222]. Some hospitals in B.C. in which ED physicians are paid by salary or sessional payments cannot be identified by the MSP FFS system. These hospitals include: Vancouver General, St. Paul\u00E2\u0080\u0099s, and Children\u00E2\u0080\u0099s in Vancouver; Bell Coola General in Bella Coola; R. W. Large Memorial in Waglisa; and Wrinch Memorial in Hazelton. Recent years have seen an increase in salaried or sessional-payment based physicians, and visits to these individuals are also are not captured in the FFS MSP database. Despite these limitations in identifying ED visits, our approach captures 80% of ED use in B.C[239]. The current study used an algorithm developed by the Centre of Health Services and Policy Research (CHSPR)[239] based on service location codes, fee item codes and service 74 codes to identify ED visits from the MSP database Fee items indicate the type of services provided; there are over 4,000 fee items in the FFS system, these are grouped into 40 broad categories according to service code. Key steps in identifying ED visits adapted from CHSPR\u00E2\u0080\u0099s algorithms are listed below: \u00EF\u0082\u00B7 Identify physician services that occurred in a hospital using the \u00E2\u0080\u009Cservice location code\u00E2\u0080\u009D which indicates H for Hospital or E for Emergency; combined with this service location code, the subsequent steps include: \u00EF\u0082\u00B7 Identify out of office visits, evening, night, weekend or Statutory Holiday visits combined with emergency service indicated; \u00EF\u0082\u00B7 Identify fee item codes that show emergency care, emergency visit, on call, on site hospital visit, or worker\u00E2\u0080\u0099s compensation board emergency call out; \u00EF\u0082\u00B7 Identify emergency medicine consultation; \u00EF\u0082\u00B7 If patients were admitted to the hospital through ED, the ED visit dates should match the hospital admission date. Records in which the ED visit date falls between the time of admission and separation were not considered ED visits and excluded. Following the ED visit selection process, patients with records containing diagnostic code ICD-9 493 were selected as having asthma-related ED visits. Physician visits and hospital admissions Physician (FP and specialist) visits containing the diagnostic code ICD-9 493 were selected from the MSP database. Hospital admission for asthma exacerbation was determined if ICD-9 diagnostic code 493 or ICD-10 diagnostic code J45 and J46 are shown in the 75 primary diagnostic field in the DAD database Percentages of patients who had FP, specialist, ED visits or hospital admissions for asthma were calculated. Numbers of visits to FP, specialist, ED or hospital per 100 patients in each study year were calculated to estimate the use of health services for asthma. Prescription drug dispensing Percentages of patients who were dispensed SABAs, ICSs, LABA and LABA/ICS combined inhalers, LTRA and other controller medications were calculated in each year. Other controller medications included methylxanthines, anti-IgE and mast cell stabilizer agents. 2.2.7.5 Regional differences To facilitate health service delivery, B.C. has been divided into 5 health authorities (HA). These HAs are further divided into 16 health service delivery areas (HSDA) and 89 local health areas (LHA). Availability of and access to health services (e.g., travel distance to clinics and hospitals) are different between large cities, small towns and remote areas. LHAs are small geographical health areas that often lack larger health facilities (e.g. hospitals, EDs) and specialists, meaning patients may need to travel to another area to access them. Thus, in order to maintain relatively comparable regional variations, we decided to examine asthma prevalence, incidence, use of health services and prescription drug dispensings by HSDA. 2.2.7.6 Data analysis Venn diagrams were created to demonstrate each component of our asthma case 76 definition. Prevalence and incidence of asthma in B.C. were described using different asthma case definition criteria: (1) 1 or more physician visit with the principal diagnosis as asthma (ICD-9 493) or 1 or more hospital admission for asthma; (2) 2 or more physician visits for asthma or 1 or more hospital admission for asthma; (3) 2 or more physician visits or 1 or more hospital admission for asthma or 1 or more asthma-related prescription drug dispensing; (4) 2 or more physician visits or 1 or more hospital admission for asthma or 2 or more asthma-related prescription drug dispensings. Using our asthma case definition, asthma prevalence and incidence, use of health services utilization for asthma as well as dispensings of asthma medications were calculated by age group, gender and HSDA over time. Since significant regional variations were observed for patients\u00E2\u0080\u0099 use of specialist, ED and hospital services, differences in using these health care services for asthma were mapped using the iMapBC , a web-based mapping tool provided by GeoBC[240], on the basis of the most recent data 2009. Multivariate logistic regression models were produced using the Enter method such that all variables were entered in a single step. Logistic regression models were fitted to test whether asthma prevalence and incidence changed significantly over time, as well as the percentages of using health services or dispensed drugs, adjusting for age group and gender. Fiscal years from 1996 to 2009 were coded as categorical variables from 0 to 13. Since all covariates were categorical variables, the regression coefficient represents the change in the odds of being a prevalent or incident case in a covariate category, compared with a reference. Poisson regression models were used to examine whether visit rates of FP, specialist, ED and 77 hospital services changed significantly over time, and adjusted for age group and gender. Residual deviance and scaled deviance were used as indicators of goodness of fit for Poisson regression models. The ratios between deviance value and degree of freedom were 0.05 for FP, 0.45 for specialist visits, 0.08 when using ED visits for asthma, 0.34 when using hospital admissions as study outcomes. Ratios were smaller than 1.0 suggesting good model fit [241]. 2.3 RESULTS 2.3.1 Identification of patients with treated asthma In our linked database that contains patients with respiratory diseases, 189,627 patients had physician diagnosed asthma only once between 1996 and 2009. 335,799 patients received asthma medications once or twice in a 12-month period but did not receive medications in other years. The number of patients who only visited physicians once for asthma increased from 73,114 in 1996 to 101,325 in 2009 (a 39% increase). The number of patients who only received one asthma prescription drug dispensing increased from 94,098 in 1996 to 128,172 in 2009 (a 36% increase). In total, 336,901 patients met the asthma case definition between 1996 and 2009. 5,794 (1.7%) of these were admitted to hospital at least once for an asthma exacerbation; 230,358 (68.4%) had \u00E2\u0089\u00A52 physician visits for asthma; and 233,567 (69.3%) had \u00E2\u0089\u00A53 asthma prescription dispensings in at least one study year. A Venn diagram has been created to show the number of patients that met each case definition criteria (Figure 2.2). 5,026 patients (1.5%) met both the hospitalization and prescription dispensing criteria; 4,866 patients (1.5%) met the hospitalization and physician visit criteria; while 136,600 patients (41.5%) met the prescription dispensing and physician visit criteria. When the prescription drug criterion was 78 removed from our case definition, 105,917 previously included patients were excluded from the study population in the current analysis. Estimates of asthma prevalence and incidence using different case definitions are demonstrated in Figure 2.3 and Figure 2.4. Based on the case definition of \u00E2\u0089\u00A51 hospital admission, \u00E2\u0089\u00A52 physician visits, or \u00E2\u0089\u00A51 asthma drug dispensing, asthma prevalence increased from 6.2% to 7.8% (30%) from 1996 to 2009, while asthma incidence changed slightly from 2.2% to 2.4% over this period. The above case definition criteria produced the highest asthma prevalence and incidence estimates compared to estimates made using other asthma case definitions. When the criterion of \u00E2\u0089\u00A51 asthma drug dispensing was replaced by \u00E2\u0089\u00A52 asthma drug dispensings, the prevalence decreased 50%; while it\u00E2\u0080\u0099s replaced by \u00E2\u0089\u00A53 asthma drug dispensings, the prevalence decreased another 40%. The lowest asthma prevalence and incidence estimates were produced using the case definition of \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A52 physician visits. The case definition of \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A51 physician visit for asthma produced a stable asthma prevalence of 3% between 1996 and 2009. When \u00E2\u0089\u00A52 physician visits criterion was used, the prevalence of asthma reduced 60% in each year. Similar changes were also observed in asthma incidence estimates when using different asthma case definitions. 79 Figure 2.2 Number of patients that met each case definition criteria in B.C. 5-55 years of age, 1996 - 2009 80 Figure 2.3 Prevalence of asthma based on different asthma case definitions in B.C aged 5-55 years, 1996 \u00E2\u0080\u0093 2009 *This asthma case definition is developed in the current research and is used for all the analyses. Figure 2.4 Incidence of asthma based on different asthma case definitions in B.C. aged 5-55 years, 1996 - 2009 *This asthma case definition is developed in the current research and is used for all the analyses. 0 1 2 3 4 5 6 7 8 9 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 P r e v a le n c e, % Year 1+ phy or 1+ hosp 1+ hosp or 2+ phy 1+ hosp or 2+ phy or 1+ Rx 1+ hosp or 2+ phy or 2+ Rx 1+ hosp or 2+ phy or 3+ Rx* 0 0.5 1 1.5 2 2.5 3 2001 2002 2003 2004 2005 2006 2007 2008 2009 In ci d en ce , % Year 1+ phy or 1+ hosp 1+ hosp or 2+ phy 1+ hosp or 2+ phy or 1+ Rx 1+ hosp or 2+ phy or 2+ Rx 1+ hosp or 2+ phy or 3+ Rx* 81 2.3.2 Overall burden of asthma The number of B.C. residents with treated asthma increased from 62,900 in 1996 to 76,221 in 2009. Taking population growth into account, the age- and gender-adjusted prevalence of asthma was stable at an average rate of 2.6% (OR 1.14; 95% CI 1.13 \u00E2\u0080\u0093 1.15) over the 14 study years (Table 2.4). The rate of newly diagnosed asthma was also stable (average rate of 0.7%; p=0.04) (Table 2.5). On average, 85% of patients used FP services for asthma; this trend was stable over time adjusting for age group, and gender (OR 0.99; 95% CI 0.97 \u00E2\u0080\u0093 1.01). The percentage of patients who visited a specialist for asthma declined from 30% in 1996 to 25% in 2009 (OR 0.71; 95% CI 0.69 \u00E2\u0080\u0093 0.73), and the percentage of patients who visited ED for asthma decreased from 30% to 14% (OR 0.32; 95% CI 0.31 \u00E2\u0080\u0093 0.33). Similarly, the percentage of hospital admissions declined by 60% (4% in 1996 to 1.6% in 2009; OR 0.35; 95% CI 0.32 \u00E2\u0080\u0093 0.39), after adjusting for age group and gender. Trends of patients\u00E2\u0080\u0099 use of health services for asthma are shown in Figure 2.5. On average, there were 150 asthma-related FP visits per 100 patients in 1996 and 113 in 2009. There were 40 specialist visits per 100 patients in 1996; this rate declined 50% by 2009. Use of ED and hospital services declined significantly during the study period. There were 3 hospital admissions and 23 ED visits per 100 patients in 1996, and only 1 hospital admission and 6 ED visits per 100 patients in 2009. 82 Figure 2.5 Trends of using health services for asthma in patients with treated asthma in B.C. aged 5-55 years, 1996 - 2009 Health services utilization rates are shown in patients who met our asthma case definition (i.e., \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A52 physician visits for asthma or \u00E2\u0089\u00A53 asthma drug dispensings) in 12-month intervals between 1996 \u00E2\u0080\u0093 2009. 83 Table 2.4 Prevalence of asthma in British Columbia aged 5-55 years of age, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. of asthma pts Population size Prevalence % No. of asthma pts Population size Prevalence % No. of asthma pts Population size Prevalence % No. of asthma pts Population size Prevalence % Overall 5-11 yrs 10,979 354,284 3.6 9,931 339,008 3.5 9,483 314,042 3.8 -5.2 -11.4 7.0 12-18 yrs 7,363 357,584 2.3 7,012 380,089 2.1 6,802 377,035 2.2 -0.5 5.4 -5.6 19-34 yrs 16,548 938,100 1.9 15,625 892,153 1.9 16,231 974,129 1.9 1.9 3.8 -1.9 35-55 yrs 28,010 1,203,462 2.5 36,459 1,362,645 2.9 43,705 1,412,034 3.4 59.6 17.3 36.0 All ages a,d 62,900 2,853,430 2.4 69,027 2,973,895 2.6 76,221 3,077,240 2.8 25.6 7.8 16.5* ,b Female 5-11 yrs 4,198 172,564 2.8 3,683 164,608 2.7 3,417 151,881 2.9 -8.8 -12.0 3.7 12-18 yrs 3,584 173,614 2.3 3,333 183,480 2.1 3,026 182,569 2.0 -10.2 5.2 -14.6 19-34 yrs 9,663 461,419 2.3 8,954 443,354 2.2 9,191 481,079 2.2 -0.9 4.3 -4.9 35-55 yrs 16,662 600,371 3.0 21,882 687,039 3.5 25,868 713,431 4.0 59.6 18.8 34.3 All ages 34,107 1,407,968 2.7 37,852 1,482,294 2.9 41,502 1,528,960 3.1 26.3 8.6 16.3 e Male 5-11 yrs 6,781 181,720 4.3 6,248 174,400 4.3 6,066 162,161 4.7 -3.0 -10.8 8.7 12-18 yrs 3,779 183,970 2.3 3,679 196,609 2.2 3,776 194,466 2.3 8.9 5.7 3.0 19-34 yrs 6,885 476,681 1.6 6,671 448,799 1.6 7,039 493,050 1.6 5.9 3.4 2.4 35-55 yrs 11,344 603,091 2.0 14,576 675,606 2.3 17,837 698,603 2.8 59.6 15.8 37.8 All ages 28,789 1,445,462 2.2 31,174 1,495,414 2.3 34,718 1,548,280 2.6 24.9 7.1 16.6 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population; P < 0.001 for comparison of rates between gender. 84 Table 2.5 Patients with newly diagnosed asthma in British Columbia aged 5-55 years of age, in 2001, 2005 and 2009 Group 2001 2005 2009 % change from 2001 to 2009 Incident cases Population, No. Incidence, % Incident cases Population, No. Incidence, % Incident cases Population, No. Incidence, % Incident cases Population, No. Incidence, % Overall 5-11 yrs 3,801 342,912 1.11 4,096 318,451 1.29 3,234 305,333 1.06 -14.9 0.03 -4.44 12-18 yrs 2,422 376,510 0.64 2,820 376,163 0.75 2,427 371,308 0.65 0.21 0.03 1.61 19-34 yrs 5,132 883,637 0.58 5,263 881,118 0.60 5,821 961,504 0.61 13.4 0.01 4.23 35-55 yrs 9,901 1,322,272 0.75 10,667 1,345,536 0.79 12,026 1,375,625 0.87 21.5 0.01 16.75 All ages a,d 21,247 2,925,331 0.73 22,560 2,921,268 0.77 23283 3,013,770 0.77 9.6 0.003 6.37* ,b Female 5-11 yrs 1,585 167,943 0.94 1,685 155,924 1.08 1,349 148,876 0.91 -14.9 0.06 -3.98 12-18 yrs 1,309 182,555 0.72 1,488 181,436 0.82 1,219 180,173 0.68 -6.9 0.07 -5.64 19-34 yrs 3,054 437,113 0.70 3,143 437,454 0.72 3,441 473,993 0.73 12.7 0.02 3.90 35-55 yrs 5,986 661,762 0.90 6,486 676,445 0.96 7,192 691,764 1.04 20.1 0.02 14.94 All ages 11,934 1,449,373 0.82 12,802 1,451,259 0.88 13,201 1,494,806 0.88 10.6 0.01 7.25 e Male 5-11 yrs 2,216 174,969 1.27 2,411 162,527 1.48 1,885 156,457 1.20 -14.9 0.06 -4.87 12-18 yrs 1,113 193,955 0.57 1,332 194,727 0.68 1,208 191,135 0.63 8.5 0.07 10.14 19-34 yrs 2,078 446,524 0.47 2,118 443,664 0.48 2,380 487,512 0.49 14.5 0.02 4.90 35-55 yrs 3,914 660,511 0.59 4,181 669,091 0.62 4,834 683,861 0.71 23.5 0.02 19.29 All ages 9,321 1,475,959 0.63 10,042 1,470,009 0.68 10,307 1,518,965 0.68 10.6 0.01 7.45 e *P < 0.001 for comparison of rates over time, from 2001 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 85 SABAs were dispensed to 78% of patients in 1996 and 74% patients in 2009 (adjusted OR 0.79; 95% CI 0.77 \u00E2\u0080\u0093 0.81). LABA dispensings (including LABA alone or in combination with ICS) increased dramatically from 2% in 1996 to 34% in 2009 (OR 23.5; 95% CI 22.19 \u00E2\u0080\u0093 24.88). LABA and ICS combined inhalers were dispensed to patients since 1999. The percentage of patients who received combined inhalers increased from 0.5% to 33% between 1999 and 2009 (OR 98.50; 95% CI 88.29 \u00E2\u0080\u0093 109.90). In contrast, the percentage of patients receiving LABA alone increased from 2% in 1996 to 7% in 2001, and then declined to 1% in 2009. There was a sharp decrease in ICS single inhaler dispensings from 61% in 2000 to 42% in 2009 (adjusted OR 0.46; 95% CI 0.45 \u00E2\u0080\u0093 0.47). However, the percentage of patients who received ICS alone or in combination with LABA increased from 61% in 1996 to 69% in 2009 (OR 1.40; 95% CI 1.37 \u00E2\u0080\u0093 1.43). LTRA were dispensed to patients since 1997. The percentage of patients receiving LTRA increased from 1% in 1997 to 8% in 2009 (OR 9.7; 95% CI 8.9 \u00E2\u0080\u0093 10.5). Other asthma medications (i.e., theophylline, anti-IgE and mast cell stabilizer drug prescriptions) decreased from 12% to 1% during the study period (OR 0.13; 95% CI0.12 - 0.14). Additional tables showing the percentages of patients dispensed each class of asthma medications are listed in Appendix A. 86 Figure 2.6 Percentage of patients dispensed asthma medications in B.C. aged 5-55 years, 1996 \u00E2\u0080\u0093 2009 SABA: short-acting bronchodilators; ICS: inhaled corticosteroids; LABA: long-acting bronchodilators; other asthma medications include theophylline, anti-IgE and mast cell stabilizers. 2.3.3 Burden of asthma by age group and gender Asthma prevalence was highest in children aged 5-11 years, with a stable average rate of 3% during the study period. Prevalence was lowest in patients 12-18 and 19-34 years of age, with rates approximately 50% lower than in children (OR 0.59, 95% CI 0.58 \u00E2\u0080\u0093 0.59 for 12-18-year-olds; OR 0.53, 95% CI 0.52 \u00E2\u0080\u0093 0.53 for 19-24-year-olds). The prevalence of asthma in patients aged 35-55 years was on average 20% less than in children (OR 0.81; 95% 0 10 20 30 40 50 60 70 80 90 P e rc e n ta g e Year SABA ICS LABA or LABA/ICS combined inhalers LTRA Other asthma medications 87 CI 0.81 \u00E2\u0080\u0093 0.82) and stable at a rate of 2.5% (Table 2.4) Consistent with these estimates, rates of newly diagnosed asthma were highest in children aged 5-11 years (1.8%, stable over time), while the rates of newly diagnosed asthma was lowest in patients aged 19-34 years (0.58%, stable over time) (Table 2.5). Table 2.6 shows rates of using FP services for asthma in patients 5-55 years of age in B.C. Use of FP services for asthma was highest in children. Compared to this group, adolescents were 20% less likely (OR 0.82; 95% CI 0.80 \u00E2\u0080\u0093 0.83), young adults were 30% less likely (OR 0.72; 95% CI 0.71 \u00E2\u0080\u0093 0.74), and older adults were 60% less likely to use FP services for asthma (OR 0.43; 95% CI 0.42 \u00E2\u0080\u0093 0.43). As expected, of all age groups, children were most likely to use specialist services (Table 2.7), ED (Table 2.8) and hospital services (Table 2.9) for asthma. The percentage of young children receiving ICS was twice higher than the percentages in adolescent and adult patients. In contrast, the percentages of young children receiving other asthma medications were significantly lower than the percentages in other age groups. In children aged 5 \u00E2\u0080\u0093 11 years, asthma prevalence was 1.8 fold higher in girls than in boys. While there was no significant gender difference in adolescents, asthma prevalence was 1.3 fold higher in female adults aged 19-34 years, and 1.5 fold higher in female adults aged 35-55 years, compared with males of the same age. Gender distributions in asthma incidence were consistent with the prevalence estimates, but in health service use were not significant (Table 2.4 and Table 2.9). 88 Table 2.6 Use of family physician services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of FP visits for asthma. FP visits rates per 100 pts No. with Asthma No. of FP visits for asthma. FP visits rates per 100 pts No. with Asthma No. of FP visits for asthma. FP visits rates per 100 pts No. with Asthma No. of FP visits for asthma. FP visits rates per 100 pts Overall 5-11 12,594 22,104 176 11,910 18,514 155 11,943 16,356 137 -5.2 -26.0 -22.2 12-18 8,194 12,872 155 8,152 11,363 139 8,154 9,758 120 -0.5 -24.2 -22.6 19-34 18,099 28,556 137 17,228 25,826 150 18,446 23,264 126 1.9 -18.5 -8.0 35-55 30,350 40,504 174 39,864 46,910 118 48,435 49,249 102 59.6 21.6 -41.4 All ages a,d 69,237 104,306 151 77,154 102,613 133 86,978 98,627 113 25.6 -5.4 -25.2* ,b Female 5-11 4,772 8,301 174 4,435 6,675 151 4,354 5,787 133 -8.8 -30.3 -23.6 12-18 4,027 6,547 163 3,902 5,614 144 3,615 4,339 120 -10.2 -33.7 -26.4 19-34 10,621 16,672 157 9,966 14,779 148 10,527 13,011 124 -0.9 -22.0 -21.0 35-55 18,081 24,611 136 23,998 28,442 119 28,859 29,679 103 59.6 20.6 -24.3 All ages 37,501 56,131 150 42,301 55,510 131 47,355 52,816 112 26.3 -5.9 -25.3 e Male 5-11 7,822 13,803 176 7,475 11,839 158 7,589 10,569 139 -3.0 -23.4 -21.0 12-18 4,167 6,325 152 4,250 5,749 135 4,539 5,419 119 8.9 -14.3 -21.7 19-34 7,478 11,884 159 7,262 11,047 152 7,918 10,197 129 5.9 -14.2 -18.9 35-55 12,265 15,893 130 15,865 18,468 116 19,576 19,570 100 59.6 23.1 -23.1 All ages 31,732 47,905 151 34,852 47,103 135 39,622 45,755 115 24.9 -4.5 -23.8 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. 89 d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population; P = 0.05 for comparison of rates between gender. 90 Table 2.7 Use of specialist services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of specialist visits for asthma. Specialist visits per 100 pts No. with Asthma No. of specialist visits for asthma Specialist visits per 100 pts No. with Asthma No. of specialist visits for asthma. Specialist visits per 100 pts No. with Asthma No. of specialist visits for asthma. Specialist visits per 100 pts Overall 5-11 12,594 9,427 75 11,910 7,157 60 11,943 5,584 47 -5.2 -40.8 -37.3 12-18 8,194 3,547 43 8,152 2,996 37 8,154 2,454 30 -0.5 -30.8 -30.2 19-34 18,099 5,207 29 17,228 3,394 20 18,446 2,909 16 1.9 -44.1 -44.8 35-55 30,350 9,814 32 39,864 8,196 21 48,435 7,503 15 59.6 -23.5 -53.1 All ages a,d 69,237 27,995 40 77,154 21,743 28 86,978 18,450 21 25.6 -34.1 -47.5* ,b Female 5-11 4,772 3,488 73 4,435 2,646 60 4,354 2,130 49 -8.8 -38.9 -32.9 12-18 4,027 1,793 45 3,902 1,443 37 3,615 1,003 28 -10.2 -44.0 -37.8 19-34 10,621 3,276 31 9,966 2,076 21 10,527 1,853 18 -0.9 -43.4 -41.9 35-55 18,081 6,163 34 23,998 5,093 21 28,859 4,672 16 59.6 -24.2 -52.9 All ages 37,501 14,720 39 42,301 11,258 27 47,355 9,658 20 26.3 -34.4 -48.7 e Male 5-11 7,822 5,939 76 7,475 4,511 60 7,589 3,454 46 -3.0 -41.8 -39.5 12-18 4,167 1,754 42 4,250 1,553 37 4,539 1,451 32 8.9 -17.3 -23.8 19-34 7,478 1,931 26 7,262 1,318 18 7,918 1,056 13 5.9 -45.3 -50.0 35-55 12,265 3,651 30 15,865 3,103 20 19,576 2,831 14 59.6 -22.5 -53.3 All ages 31,732 13,275 42 34,852 10,485 30 39,622 8,792 22 24.9 -33.8 -47.6 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 91 Table 2.8 Use of hospital services for in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of hosp admissions for asthma. Hosp admission rates per 100 pts No. with Asthma No. of hosp admissions for asthma. Hosp admission rates per 100 pts No. with Asthma No. of hosp admission s for asthma. Hosp admission rates per 100 pts No. with Asthma No. of hosp admissions for asthma. Hosp admission rates per 100 pts Overall 5-11 12,594 473 3.8 11,910 272 2.3 11,943 224 1.9 -5.2 -52.6 -50.0 12-18 8,194 236 2.9 8,152 92 1.1 8,154 72 0.9 -0.5 -69.5 -69.0 19-34 18,099 517 2.9 17,228 211 1.2 18,446 145 0.8 1.9 -72.0 -72.4 35-55 30,350 722 2.4 39,864 387 1.0 48,435 290 0.6 59.6 -59.8 -75.0 All ages a,d 69,237 1,948 2.8 77,154 962 1.2 86,978 731 0.8 25.6 -62.5 -71.4* ,b Female 5-11 4,772 198 4.1 4,435 99 2.2 4,354 87 2.0 -8.8 -56.1 -51.2 12-18 4,027 133 3.3 3,902 46 1.2 3,615 23 0.6 -10.2 -82.7 -81.8 19-34 10,621 363 3.4 9,966 153 1.5 10,527 98 0.9 -0.9 -73.0 -73.5 35-55 18,081 527 2.9 23,998 254 1.1 28,859 200 0.7 59.6 -62.0 -75.9 All ages 37,501 1,221 3.3 42,301 552 1.3 47,355 408 0.9 26.3 -66.6 -72.7 e Male 5-11 7,822 275 3.5 7,475 173 2.3 7,589 137 1.8 -3.0 -50.2 -48.6 12-18 4,167 103 2.5 4,250 46 1.1 4,539 49 1.1 8.9 -52.4 -56.0 19-34 7,478 154 2.1 7,262 58 0.8 7,918 47 0.6 5.9 -69.5 -71.4 35-55 12,265 195 1.6 15,865 133 0.8 19,576 90 0.5 59.6 -53.8 -68.8 All ages 31,732 727 2.3 34,852 410 1.2 39,622 929 0.8 24.9 27.8 -65.2 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 92 Table 2.9 Use of ED services for asthma in patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of ED visits for asthma. ED visit rates per 100 pts No. with Asthma No. of ED visits for asthma. ED visit rates per 100 pts No. with Asthma No. of ED visits for asthma ED visit rates per 100 pts No. with Asthma No. of ED visits for asthma. ED visit rates per 100 pts Overall 5-11 12,594 3,640 28.9 11,910 2,683 22.5 11,943 986 8.2 -5.2 -72.9 -71.6 12-18 8,194 2,210 27.0 8,152 1,392 17.1 8,154 590 7.2 -0.5 -73.3 -73.3 19-34 18,099 4,635 25.6 17,228 2,761 16.0 18,446 1,413 7.7 1.9 -69.5 -69.9 35-55 30,350 5,389 17.8 39,864 3,738 9.4 48,435 1,885 3.9 59.6 -65.0 -78.1 All ages a,d 69,237 15,874 22.9 77,154 10,574 13.7 86,978 4,874 5.6 25.6 -69.3 -75.5* ,b Female 5-11 4,772 1,342 28.1 4,435 954 21.5 4,354 343 7.9 -8.8 -74.4 -71.9 12-18 4,027 1,220 30.3 3,902 665 17.0 3,615 285 7.9 -10.2 -76.6 -73.9 19-34 10,621 2,806 26.4 9,966 1,680 16.9 10,527 783 7.4 -0.9 -72.1 -72.0 35-55 18,081 3,504 19.4 23,998 2,395 10.0 28,859 1,188 4.1 59.6 -66.1 -78.9 All ages 37,501 8,872 23.7 42,301 5,694 13.5 47,355 2,599 5.5 26.3 -70.7 -76.8 e Male 5-11 7,822 2,298 29.3 7,475 1,729 23.1 7,589 643 8.5 -3.0 -72.0 -71.0 12-18 4,167 990 23.8 4,250 727 17.1 4,539 305 6.7 8.9 -69.2 -71.8 19-34 7,478 1,829 24.5 7,262 1,081 14.9 7,918 630 8.0 5.9 -65.6 -67.3 35-55 12,265 1,885 15.4 15,865 1,343 8.5 19,576 697 3.6 59.6 -63.0 -76.6 All ages 31,732 7,002 22.1 34,852 4,880 14.0 39,622 2,275 5.7 24.9 -67.5 -74.2 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 93 2.3.4 Burden of asthma by HSDA Asthma prevalence and incidence estimates were not significantly different between HSDA groups. Distributions of FP, specialist, hospital and ED use for asthma in B.C. are mapped in Figure 2.7-Figure 2.10. On average, there were 0.7-0.9 asthma-related FP visits per patient residing in rural areas (e.g., East Kootenay, Kootenay Boundary, Northern B.C.). This rate is 20%-30% less than in large cities (e.g. Richmond, Victoria and Surrey), where there were 1.3 visits per 100 patients (Figure 2.7), when adjusted for age, gender, and study year. Likewise, in rural areas, there were 4-10 asthma-related specialist visits per 100 patients, a rate, 70%-90% less than in large cities (e.g., 36 visits per 100 patients in Richmond, 36 in Surrey, and 33 in Vancouver) (Figure 2.8). However, patients in rural areas were 1.6 to 2.3 times more likely to be admitted to a hospital for asthma (1.3-2 admissions per 100 patients in northern or interior areas compared with 1 admission per 100 patients in large cities) (Figure 2.9). Patients in rural areas were also 2.0 to 2.6 times more likely to visit the ED (20 visits per 100 patients in rural areas compared with 2-3 visits per 100 patients in large cities) (Figure 2.10), when adjusted for age, gender and study year. Additional tables showing the regional variation in patients\u00E2\u0080\u0099 use of health services for asthma are listed in Appendix A. 94 Figure 2.7 Use of family physician services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 Figure 2.8 Use of specialist services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 95 Figure 2.9 Use of hospital services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 Figure 2.10 Use of ED services for asthma by HSDA in asthma patients 5-55 years of age, 1996 - 2009 96 CHAPTER 3 : ASTHMA REGIMEN OPTIMALITY AND HEALTH SERVICES UTILIZATION 97 3.1 SYNOPSIS Despite numerous clinical practice guidelines available for asthma management, patients often receive suboptimal drug therapy. This chapter identifies patients who received suboptimal regimens during a one-year period according to the NHLBI Guidelines for the Diagnosis and Management of Asthma in B.C. and determines the association between patients\u00E2\u0080\u0099 regimens and utilization of healthcare services. To ensure the robustness of findings from this work, two study designs were used in the present study. A cross-sectional study design was initially applied to produce a preliminary assessment. 65,345 asthma patients were identified using provincial health service utilization data (including all respiratory disease-related prescription medication dispensings, physician and hospital visits) for the 2009 fiscal year (the latest year that data are available in this work). Patient-specific regimens of inhaled SABA with or without ICS were categorized as optimal or suboptimal. Logistic regression models were used to determine the association between regimen optimality and health service utilization, adjusted for SES, prior year hospital and ED visits for asthma. Patients with suboptimal regimens had significantly greater risk of using health services than patients with optimal regimens of SABA and/or ICS. In particular, adolescents with suboptimal regimens were the most likely to have hospital admissions (odds ratio (OR) 3.8; 95% CI 1.8 \u00E2\u0080\u0093 7.8), visit the ED (OR 2.2; 95% CI 1.6 \u00E2\u0080\u0093 3.1), and be high users of FP services (OR 5.7; 95% CI 4.0 \u00E2\u0080\u0093 8.1) compared with patients in other age groups. To infer causal relationships between asthma drug regimen optimality and health service utilization, a retrospective cohort study was further conducted. Patients were followed from the first date that they entered the cohort. Regimen optimality was classified during the 98 baseline year (i.e., one-year period after cohort entry). Outcomes were defined as the first ED visit or hospital admission after the baseline year. Outcomes occurred over the subsequent year after baseline year were analyzed. Propensity score matching methodology was applied to ensure the comparability between patients with suboptimal and optimal regimens. Cox proportional regression models were used for statistical analysis. Optimal and suboptimal regimens were able to categorize in 268,090 patients. Patients with suboptimal regimens were approximately 4 times more likely to visit ED and 5 times more likely to be admitted to hospitals for asthma exacerbations compared to those with optimal regimens. Suboptimal regimens were associated with significantly higher usage of health services. Identifying patients with suboptimal regimens and improving their medication management in accordance with asthma clinical guidelines is likely to result in lower health service utilization. 3.2 METHODOLOGY 3.2.1 Cross-sectional study design B.C. population-based health services utilization data and prescription drug dispensing data were used for this study. Information on sources, quality and preparation of data has been described in detail in CHAPTER 2 Sections 2.2.1-2.2.3. 3.2.1.1 Study design and study patients A cross-sectional study was used to examine the relationship between asthma drug regimen optimality and health services utilization in 2009 fiscal year (i.e., April 1, 2009 to March 31, 2010). Patients were classified as having asthma if they satisfied any of the following criteria during 2009: at least one hospital admission with asthma as the principal 99 diagnosis based on the International Classification of Diseases, 10 th version, code J45 (ICD- 10 J45); or at least two physician visits for asthma as the principal diagnosis based on the ICD 9 th version, code 493 (ICD-9 493); or at least three asthma drug dispensings. Detailed information in relation to the asthma case definition, validity of asthma case definition, validity and accuracy of ICD diagnostic codes has been described previously in CHAPTER 2 Sections 2.2.5 and 2.2.6. 3.2.1.2 Categorization of optimal, suboptimal and unclassified asthma regimens SABA and ICS are the two main classes of medication used to treat asthma, with 80% of patients in the study cohort received at least one SABA and 70% received at least one ICS prescription in each study year (Chapter 2 Section 2.3.2). As SABA and ICS are the gold standard for asthma management according to asthma clinical practice guidelines[3], SABA and ICS dispensings were focused in the categorization of optimal and suboptimal asthma drug regimens. To define optimal and suboptimal regimens, each SABA dispensing was converted to salbutamol (albuterol) 100\u00C2\u00B5g metered dose inhaler (MDI) equivalents and all prescriptions for ICS, alone or in a combination inhaler with LABA were converted to beclomethasone MDI 50\u00C2\u00B5g MDI equivalents. Salbutamol and fenoterol are considered equipotent, while terbutaline is 0.4 times the potency of salbutamol[242, 243]. Beclomethasone is one half the potency of fluticasone and ciclesonide, 2.5 times as the potency of flunisolide and 0.8 times the potency of budesonide[244, 245]. These data were obtained from the Compendium of Pharmaceuticals and Specialties in years concurrent with the study period of the present analyses[246]. Nebulised \u00CE\u00B22-agonists were not included in the analysis because they are infrequently used in patients 5 years of age and older. Due to the seasonal nature of asthma and the inter- and intrapatient variability in 100 medication use, we chose a 12-month interval to define regimen optimality. The corresponding SABA/ICS dispensings were plotted into a matrix (Figure 3.1 and Figure 3.2) based on the NHLBI Guidelines for the Diagnosis and Management of Asthma EPR-3[244]. The calculation of regimen optimality cutoff points for adults is summarized in Table 3.1. Some regimen optimality patterns could be regarded as optimal or suboptimal depending on other clinical factors such as symptoms, lung function, and the patients\u00E2\u0080\u0099 quality-of-life, as well as patients\u00E2\u0080\u0099 dispensing behaviors (e.g., some patients may need multiple inhalers concurrently at home, in the car or at work). Without these records, characterization of these annualized regimens as optimal/suboptimal may be appropriate. These regimens were therefore deemed unclassifiable by two asthma experts (respirologist, clinical pharmacologist) and were excluded from the analysis. This study focused on health outcomes of patients with clearly optimal and suboptimal asthma medication dispensings. For children 5-11 years of age .we defined the annual drug quantities to be one half of those for adults and older children, based on manufacturer drug label recommendations (Figure 3.2). 101 Figure 3.1 Asthma regimen optimality classification (patient ages 12-55 years) Green color indicates optimal regimens. Red color indicates suboptimal regimens. Yellow color indicates regimens which were deemed unclassifiable. 102 Figure 3.2 Asthma regimen optimality classification (patient ages 5-11 years) Green color represents optimal regimens. Red color represents suboptimal regimens. Yellow color represents regimens which were deemed unclassifiable. 103 Table 3.1 Regimen optimality classification for asthma patients aged 12 years or older Optimality of Drug Regimens Calculation and Definition Optimal regimens \u00EF\u0082\u00B7 A maximum of 3 doses of SABA per week is recommended by the NHLBI guidelines, resulting in a total of 6 inhalations per week (up to 2 inhalations per dose) x 52 weeks = 312 inhalations per year or 1.5 inhalers per year (1 inhaler = 200 inhalations); \u00EF\u0082\u00B7 An additional dose of SABA to be used prior to exercise should be added to the above total inhalations (2 inhalations per day x 7 days x 52 weeks = 728 inhalations per year, or approximately 3 inhalers per year; \u00EF\u0082\u00B7 Daily dose of ICS is recommended to be no more than 500\u00C2\u00B5g/day. Suboptimal regimens \u00EF\u0082\u00B7 More than 8 inhalers of SABA/year combined with no more than 1000 \u00C2\u00B5g/day of ICS; \u00EF\u0082\u00B7 4 or fewer inhalers of SABA/year combined with more than 1000\u00C2\u00B5g/day of ICS. Unclassified optimality \u00EF\u0082\u00B7 5-8 inhalers of SABA/year with or without ICS; \u00EF\u0082\u00B7 No more than 4 inhalers of SABA/year combined with ICS dispensings ranging between 500 \u00E2\u0080\u0093 1000 \u00C2\u00B5g/day; \u00EF\u0082\u00B7 More than 8 inhalers of SABA combined with more than 1000 \u00C2\u00B5g/day of ICS. Definition of asthma drug optimality was based on the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma Expert Panel Report-3. 3.2.1.3 Measure of study outcomes The primary outcomes of interest were hospitalization and ED visits for asthma. Use of physician services was a secondary outcome. Databases used to select hospital admission and physician visits and identify ED visits were described in CHAPTER 2 Section 2.2.7.4. Asthma-related ED visits were selected based on the health service location codes (e.g., indicating that service location was hospital ED), and health service fee item codes (e.g., indicating urgent care) from the provincial MSP database described earlier. We examined 104 physician services for both FPs and specialists. High users of FP services were defined as those patients who used FP most frequently, namely, in the top 5% of the total number of FP for asthma in the study year. 3.2.1.4 Covariates As patients who received suboptimal regimens may have poorer asthma control than those who received optimal regimens, a series of variables were adjusted to control confounding by disease severity: age, gender, prior-year hospital admission or ED visits, comorbid mental disorders (i.e., depression or anxiety), SES, whether a patient received LABA and ICS as separate or combined inhalers, theophylline prescriptions since these factors have been shown to be significantly associated with patients\u00E2\u0080\u0099 use of asthma medications and the use of health services for asthma in previous studies[232, 247-251]. A patient\u00E2\u0080\u0099s SES was measured using the neighborhood income quintile provided by the Ministry of Health in BC[252, 253]. 3.2.1.5 Statistical analysis Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) (version 19 th , Chicago IL). A series of logistic regression models were used to examine the relationship between health service utilization outcomes and indicators of optimal and suboptimal drug therapy, and calculate ORs. Statistical analyses were adjusted for confounding factors. Because a significant age by regimen optimality interaction effect was detected, analyses were stratified by age group. Patients were grouped into young children (5-11 years), adolescents (12-18 years), adults19-34 years and 35-55 years of age. As LABA and ICS combined inhalers are increasingly used in patients with asthma, a 105 secondary analysis was conducted to describe LABA dispensings in each regimen optimality classification and its impact on health services utilization. 3.2.2 Retrospective cohort study design 3.2.2.1 Study design To infer the causal relationship between asthma drug regimen optimality, a retrospective cohort study was subsequently conducted to address the same research questions. Patients were followed from the first date that they met the asthma case definition described previously. To be in the analysis, patients need to have at least one year\u00E2\u0080\u0099s complete drug dispensing and health services utilization records after the cohort entry date. Figure 3.3 shows how the study population was formed for this analysis. The one year time period after cohort entry was regarded as the baseline year. Time 0 is 365 days after the cohort entry date. The initial cohort contains 336,901 patients with valid age and gender records. During the baseline year, 5,703 patients turned 56 years of age, death occurred in 2,500 patients and the latest data available date March 31, 2010 occurred in 19,574 patients. These patients were excluded from the analysis, resulting in 308,319 patients included in the analysis. 106 Figure 3.3 Study population in the comparison of health services utilization between patients with suboptimal and optimal regimens Initial cohort included patients who met the asthma case definition: \u00E2\u0089\u00A51 hospital admission for asthma or \u00E2\u0089\u00A52 physician visits for asthma or \u00E2\u0089\u00A53 asthma prescription drug dispensings during a 12-month period. The baseline year was defined as the one-year period after cohort entry date. 3.2.2.2 Measure of study exposures Optimal and suboptimal regimens were categorized based on patients\u00E2\u0080\u0099 dispensing records of SABAs and ICSs during the baseline year. Detailed methods used for regimen optimality classification were described in Section 3.2.1.2. 107 3.2.2.3 Measure of study outcomes The study outcomes included the first ED visit and hospital admission for asthma (i.e., the primary diagnostic code ICD-9 943 or ICD-10 J45, J46) over the subsequent year after baseline year. How ED visits were identified in the MSP database and the data sources for determining hospital admissions have been described in CHAPTER 2 Section 2.2.7.2. Patients were followed until they had the outcome events, or reach the end of one-year follow-up period, or died, or reach their 56 th birthday or March 31 st , 2010 which is the latest date available in our data. 3.2.2.4 Covariates As described in Section 3.2.1, patients who received suboptimal regimens may have poorer asthma control compared to those who received optimal regimens. A series of variables were used to measure patients\u00E2\u0080\u0099 asthma control level during the baseline year, including demographic factors, use of health services during the baseline year, whether patients had comorbidities which are significantly related to asthma drug management and health services utilization (i.e., allergy, allergic rhinitis, URTI, acute lower respiratory tract infections, COPD, chronic bronchitis or emphysema); dispensing of the second/third line asthma drug prescriptions (i.e., LABA, theophylline, leukotriene receptor antagonists). The associations between these variables and patients\u00E2\u0080\u0099 asthma control level have been established in previous studies[72, 232]. 3.2.2.5 Data analysis Propensity score matching Propensity score matching was used to ensure the two patient groups had similar 108 distributions of confounding variables described above. The propensity score method was developed by Rosenbaum and Rubin[254] and is defined as the probability of assignment of a medication conditional on a series of observed covariates. Multiple approaches can be used to calculate propensity score, such as neural networks, discriminant function analysis, and classification trees[255]. In the analysis of this chapter, propensity score was calculated using a logistic regression model, with the outcome being the probability of using asthma regimens suboptimally or optimally. To determine which confounding factor should be included in the propensity score calculation and avoid over-adjustification, univariate logistic regression models were conducted to examine the relationship between each potential confounding variable and the likelihood of using asthma drug regimens suboptimally. The ORs were listed in Table 3.2. All potential confounding variables, which had statistically significant (p<0.05 for 2-tailed Wald test) unadjusted association with using asthma drug regimens suboptimally were included in the propensity score model. To maximize the comparability of the two patient groups, patients\u00E2\u0080\u0099 ages, ED visits, hospital admissions and dispensings of the second and third asthma medications were matched by the frequency. Patients\u00E2\u0080\u0099 comorbidity status was matched as binary variables. The propensity score model yielded a c-statistic of 0.78, which indicated a strong ability to differentiate between the two ICS use groups. To maximize the accuracy of matching, we used an \u00E2\u0080\u0098exact matching\u00E2\u0080\u0099 methodology, which means the matched pair had the same propensity score. The standardized differences of each confounding variable before and after propensity score matching were calculated and shown in Figure 3.4. A Quantile-quantile plot (QQ plot) was used to evaluate whether there is a good balance between patients who had suboptimal and optimal asthma drug regimens after the propensity score matching (Figure 3.5). The QQ plot has been regarded as the best 109 approach to compare distributions of confounding variables between comparison groups[256]. Patients with suboptimal asthma drug regimens were matched 1:1 with those with optimal regimens during the baseline year. Table 3.2 Propensity score variables and association with suboptimal use of asthma regimens Univariate OR Confidence Interval Variables Demographic information Age (years) 1.012 1.011 \u00E2\u0080\u0093 1.013 Gender (men vs. women) 1.44 1.40 \u00E2\u0080\u0093 1.47 Number of the 2 nd - or 3 rd -line asthma medications dispensed during the baseline year LABA 1.33 1.32 \u00E2\u0080\u0093 1.35 Theophylline 1.44 1.41 \u00E2\u0080\u0093 1.47 LTRA 1.01 0.99 \u00E2\u0080\u0093 1.02 Omalizumab 1.09 0.76 \u00E2\u0080\u0093 1.57 Use of health services for asthma ED visits 1.52 1.50 \u00E2\u0080\u0093 1.55 Hospital admissions 2.25 2.13 \u00E2\u0080\u0093 2.38 Comorbidity Allergy (y/n) 0.67 0.60 \u00E2\u0080\u0093 0.75 Allergic rhinitis (y/n) 0.69 065 \u00E2\u0080\u0093 0.72 URTI (y/n) 0.73 0.71 \u00E2\u0080\u0093 0.75 Depression or anxiety (y/n) 1.06 1.03 \u00E2\u0080\u0093 1.09 LRTI (y/n) 1.37 1.33 \u00E2\u0080\u0093 1.42 COPD, emphasema or chronic bronchitis (y/n) 3.22 3.05 \u00E2\u0080\u0093 3.41 Abbreviations: LABA \u00E2\u0080\u0093 long-acting bronchodilators; LTRA \u00E2\u0080\u0093 leukotriene receptor antagonists; URTI \u00E2\u0080\u0093 upper respiratory tract infections; LRTI \u00E2\u0080\u0093 lower respiratory tract infections; COPD \u00E2\u0080\u0093 chronic obstructive pulmonary diseases. 110 Figure 3.4 Standardized difference in the distribution of each confounding variable before and after propensity score matching 111 Figure 3.5 QQ plots \u00E2\u0080\u0093 assessment of propensity score matching 112 The QQ plots compare the distribution of each confounding variable before and after propensity score matching. The black points represent the distribution of each variable before propensity score matching; while the red points represent the distribution after propensity score matching. The linearity of the red points suggests how well the two comparison groups (i.e., patients with suboptimal and optimal regimens) are matched by each confounding variable. 113 Cox regression models Cox proportional hazards regression models were used to account for the time- dependent nature of drug use and calculate the hazard of ED visits or hospital admissions among patients who had suboptimal or optimal regimens, at a given time, t. The hazard of outcome event occurrence was defined as the conditional probability of the event occurring at time t, given survival to time t or later. The Cox proportional hazard model assumes a baseline hazard function, and the hazard of study outcomes (i.e., ED and hospital admissions) is a constant multiple of this hazard function. Log minus log (LML) plot of survival is a well- accepted method of testing if the two comparison groups meet the Cox model assumption [257] and was used in the present analysis. This LML plot produced parallel lines (Figure 3.6), suggesting the proportional hazards holds for the two groups (i.e., patients who received suboptimal asthma drug regimens and those who received optimal asthma drug regimens). For continuous variables, the HR represents the reduction in risk associated with a one-unit covariate increase at each time point. For categorical variables, HRs were calculated for each category compared to the reference category. The significance of the individual regression estimates were tested using the Wald statistic. The SPSS version 20 was used to clean and prepare the data, and conduct descriptive analysis and multivariate logistic regression models. Propensity score matching and time- dependent analysis were performed using R version 3.0.0 (\u00C2\u00A9 The R Foundation for Statistical Computing). 114 Figure 3.6 Log minus log plot to assess Cox proportional hazard model assumption 3.2.2.6 Sensitivity analysis A standard proportional hazards regression was used to repeat the analysis, adjusting for all potential confounding factors, rather than matching on the propensity score. 3.3 RESULTS 3.3.1 Results based on the cross-sectional study design 3.3.1.1 Description of the patient groups In 2009, 83,157 patients qualified for inclusion in the asthma cohort. Based on the dispensed quantities of SABA and ICS and established clinical practice guidelines for the use of these drugs in asthma management, drug regimens in 65,345 patients were able to be categorized into optimal and suboptimal group. Characteristics of patients with optimal or suboptimal regimens are displayed in Table 3.3. Overall, 67% of these patients (43,870) 115 received optimal asthma drug regimens while 15% (9,461) were on suboptimal regimens. 12,014 patients\u00E2\u0080\u0099 drug regimens were categorized into the unclassified group and were therefore not included in any further analysis. Table 3.3 Characteristics of study patients Patients with asthma (n = 65, 345) Patient groups with different regimens P value Optimal regimens Suboptimal regimens Gender, no. (and %) female 25,039 (57.1) 4,430 (46.8) P < 0.001 Age, mean yrs (and SD) 32.5 (15.6) 36.7 (13.9) P < 0.001 Service use for asthma in the year prior to the study period No. (and %) of patients admitted to hospitals 229 (0.5) 154 (1.6) P < 0.001 No. (and %) of patients visited ED 1,923 (4.4) 700 (7.4) P < 0.001 Neighborhood income quintile (%) 1 (low) 21.3 27.3 2 20.5 21.3 3 20.0 18.8 4 19.2 16.8 5 (high) 17.4 13.4 P < 0.001 Dispensing of other asthma medications (%) leukotriene receptor antagonists 7.3 6.8 0.115 Theophylline 0.3 1.8 P < 0.001 Oral steroids 21.2 24.8 P < 0.001 In total, 9,326 patients (15%) were dispensed SABA frequently (i.e., >8 inhalers in patients aged 12 years or older, or > 4 inhalers in patients younger than 12 years). Of these patients with frequent SABA dispensings, one fourth did not receive any ICS. An additional 55% of patients with frequent SABA dispensings did not receive sufficient amount of ICS. 3.3.1.2 Concurrent health services utilization in patients with suboptimal asthma regimens In patients with suboptimal asthma regimens, 1.6% were admitted to hospital at least 116 once for asthma, 7.4% visited the ED, 8.1% used specialist services, and 67.9% used physician services for asthma. Overall, patients with suboptimal regimens were approximately three times more likely to be admitted to hospital (adjusted OR 2.7, 95% CI 2.1-3.3), and 1.5 times more likely to visit the ED (adjusted OR 1.5, 95% CI 1.4-1.7) compared to those with optimal regimens. In contrast, the risk of using physician services in patients with suboptimal regimens only increased by 20% (adjusted OR 1.2, 95% CI 1.17-1.30). When utilization of physician services was explored further, we found that patients on suboptimal regimens were 3.4 times more likely to be high users of FP services (i.e., patients who used FP services in the top 5% of the total number of physician visits for asthma in the study year) compared to patients on optimal regimens (adjusted OR 3.4, 95% CI 3.1-3.7). The ORs and corresponding 95% CIs are summarized in Table 3.4. The LABA-adjusted ORs did not differ significantly from the unadjusted ORs (Table 3.5). In the age-group stratified analyses (Table 3.6), adolescents (12-18 years of age) who were on suboptimal regimens had the greatest risk of using health services compared to other age groups. Patients in this age group with suboptimal regimens were 4 times more likely to be admitted to the hospital (OR 3.8, 95% CI 1.8 to 7.8), and twice more likely to use ED services (OR 2.2, 95% CI 1.6 to 3.1) compared to those with optimal regimens. Adolescent group who received suboptimal regimens were 6 times more likely to be high users of FP services (OR 5.7, 95% CI 4.0 to 8.1). 117 Table 3.4 Comparison of health service utilization between patients with suboptimal and optimal regimens Use of health services for asthma Unadjusted ORs (and 95% CI) Adjusted ORs (and 95% CI)** Hospital admissions 3.2 (2.6 \u00E2\u0080\u0093 3.9) 2.7 (2.1 - 3.3) ED visits 1.7 (1.6 - 1.9) 1.5 (1.4 - 1.7) Physician visits 1.2 (1.1 \u00E2\u0080\u0093 1.2) 1.2 (1.17 \u00E2\u0080\u0093 1.30) FP visits 1.3 (1.2 \u00E2\u0080\u0093 1.3) 1.3 (1.26 \u00E2\u0080\u0093 1.39) Specialist visits 0.8 (0.7 \u00E2\u0080\u0093 0.8) 0.8 (0.75 \u00E2\u0080\u0093 0.89) High users of physician services* 2.3 (2.1 \u00E2\u0080\u0093 2.5) 2.2 (2.0 \u00E2\u0080\u0093 2.4) High users of FP services 3.8 (3.5 \u00E2\u0080\u0093 4.1) 3.4 (3.1 \u00E2\u0080\u0093 3.7) High users of specialist services 0.6 (0.5 \u00E2\u0080\u0093 0.7) 0.6 (0.5 \u00E2\u0080\u0093 0.7) *High users of physician services were defined as those patients who used physician services the top 5% of the total number of physician visits for asthma in 2009. **Adjusted for age group, gender, patients\u00E2\u0080\u0099 comorbid depression and anxiety, receiving of long-acting beta agonists or theophylline prescriptions, prior year hospital admission and/or ED visit for asthma and socioeconomic status. 118 Table 3.5 Unadjusted and LABA use-adjusted ORs in the comparison of health service utilization between patients with suboptimal and optimal regimens Use of health services for asthma Unadjusted ORs (and 95% CI) LABA use - Adjusted ORs (and 95% CI)** Hospital admissions 3.2 (2.6 \u00E2\u0080\u0093 3.9) 3.0 (2.5 - 3.7) ED visits 1.7 (1.6 - 1.9) 1.7 (1.6 - 1.9) Physician visits 1.2 (1.1 \u00E2\u0080\u0093 1.2) 1.1 (1.09 \u00E2\u0080\u0093 1.20) FP visits 1.3 (1.2 \u00E2\u0080\u0093 1.3) 1.3 (1.19 \u00E2\u0080\u0093 1.31) Specialist visits 0.8 (0.7 \u00E2\u0080\u0093 0.8) 0.8 (0.70 \u00E2\u0080\u0093 0.82) High users of physician services* 2.3 (2.1 \u00E2\u0080\u0093 2.5) 2.2 (2.0 \u00E2\u0080\u0093 2.4) High users of FP services 3.8 (3.5 \u00E2\u0080\u0093 4.1) 3.6 (3.3 \u00E2\u0080\u0093 4.0) High users of specialist services 0.6 (0.5 \u00E2\u0080\u0093 0.7) 0.6 (0.5 \u00E2\u0080\u0093 0.7) *High users of physician services were defined as those patients who used physician services the top 5% of the total number of physician visits for asthma in 2009. Table 3.6 Comparison of health service utilization between patients with suboptimal and optimal regimens in each age group *High users of FP services were defined as patients who used FP services in the top 5% of the total number of FP visits for asthma in 2009. **Adjusted for gender, patients\u00E2\u0080\u0099 comorbid depression and anxiety, receiving of long-acting beta agonists or theophylline prescriptions, prior year hospital admission and/or ED visit for asthma and socioeconomic status. Use of Health Services Age Groups 5 \u00E2\u0080\u0093 11 years 12 \u00E2\u0080\u0093 18 years 19 \u00E2\u0080\u0093 34 years 35 \u00E2\u0080\u0093 55 years ORs** (95% CIs) ORs** (95% CIs) ORs** (95% CIs) ORs** (95% CIs) High use of FP services Suboptimal regimens 2.3 (1.7 \u00E2\u0080\u0093 3.1) 5.7 (4.0 \u00E2\u0080\u0093 8.1) 5.2 (4.4 \u00E2\u0080\u0093 6.3) 2.9 (2.6 \u00E2\u0080\u0093 3.3) Hospital admissions Suboptimal regimens 2.2 (1.4 \u00E2\u0080\u0093 3.4) 3.8 (1.8 \u00E2\u0080\u0093 7.8) 3.1 (1.9 \u00E2\u0080\u0093 5.0) 2.9 (2.1 \u00E2\u0080\u0093 4.0) ED visits Suboptimal regimens 1.4 (1.1 \u00E2\u0080\u0093 1.8) 2.2 (1.6 \u00E2\u0080\u0093 3.1) 1.4 (1.2 \u00E2\u0080\u0093 1.6) 1.6 (1.3 \u00E2\u0080\u0093 1.8) 119 3.3.2 Results based on the cohort study design 308,319 patients qualified for inclusion in the asthma cohort. Based on the dispensed quantities of SABA and ICS and established clinical practice guidelines for the use of these drugs in asthma management, optimal and suboptimal regimens were able to categorize in 268,090 patients. Overall, 78.9% of these patients (243,252) received optimal asthma drug regimens while 8.1% (24,845) were on suboptimal regimens. 12.9% (39,639) patients\u00E2\u0080\u0099 drug regimens were categorized into the unclassified group and were therefore not included in any further analysis. Characteristics of patients with optimal or suboptimal regimens are displayed in Table 3.7. A significantly higher percentage of patients who received suboptimal regimens were older and males compared to those who received optimal regimens. Patients with suboptimal regimens received higher numbers of LABA, theophylline and LTRA prescriptions than patients who received optimal regimens (p<0.001). Patients with suboptimal regimens had more ED visits and hospital admissions for asthma exacerbations during the baseline year compared to those with optimal regimen years (P<0.001). A higher percentage of patients with suboptimal regimens experienced acute lower respiratory tract infections, emphysema, chronic bronchitis or COPD than those with optimal regimens (p<0.001). In contrast, a lower percentage of patients with suboptimal regimens had comorbidities like allergy, allergic rhinitis, or URTI than those with optimal regimens (p<0.001). The two patient groups had similar percentages having the comorbidities of depression or anxiety (22%). The propensity score matching ensured the balance of characteristics between the two study groups (c statistic = 0.72) based on matched variables. The matching produced 19,060 pairs. The standardized differences between suboptimal and optimal regimen users were zero 120 for all other matched variables (Table 3.7). 121 Table 3.7 Characteristics of patients before and after propensity score matching Before propensity score matching After propensity score matching Suboptimal regimen users Optimal regimen users Standard difference, % P Value Suboptimal regimen users Optimal regimen users Standard difference, % P Value # of patients 24,843 243,247 19,060 19,060 Demographic information Mean age (SD), y 32.05 (14.77) 28.99 (16.06) 19.8 < 0.001 31.8 (14.5) 31.8 (14.5) 0.0 1.000 Female, % 46.9 56.0 18.1 <0.001 54.3 54.3 0.0 1.000 Asthma drug dispensings Mean # of LABA inhalers dispensed 0.48 (1.77) 0.14 (0.64) 25.7 < 0.001 0.11 (0.66) 0.11 (0.66) 0.0 1.000 Mean # of theophylline prescriptions dispensed 0.23 (1.26) 0.03 (0.57) 20.8 <0.001 0.02 (0.32) 0.02 (0.32) 0.0 1.000 Mean # of LTRA prescriptions dispensed 0.11 (0.86) 0.11 (0.78) 0.5 0.45 0.02 (0.32) 0.02 (0.32) 0.0 1.000 Health services utilization for asthma Mean # of ED visits for asthma 0.35 (1.19) 0.11 (0.51) 27.0 0.001 0.10 (0.37) 0.10 (0.37) 0.0 1.000 Mean # of hospital admissions for asthma 0.06 (0.30) 0.02 (0.15) 15.8 <0.001 0.01 (0.08) 0.01 (0.08) 0.0 1.000 Cormobidity (%) Allergy 1.5 2.3 5.3 <0.001 0.8 0.8 0.0 1.000 Allergic rhinitis 8.1 11.3 11.1 <0.001 6.2 6.2 0.0 1.000 URTI 25.9 32.4 14.4 <0.001 23.6 23.6 0.0 1.000 Depression or anxiety 22.8 21.9 2.3 <0.001 21.6 21.6 0.0 1.000 122 Before propensity score matching After propensity score matching Suboptimal regimen users Optimal regimen users Standard difference, % P Value Suboptimal regimen users Optimal regimen users Standard difference, % P Value Acute lower respiratory tract infection 21.9 17.0 12.5 <0.001 18.3 18.3 0.0 1.000 Emphysema, chronic bronchitis, COPD 7.0 2.3 22.5 <0.001 4.0 4.0 0.0 1.000 * LABA \u00E2\u0080\u0093 long-acting bronchodilators; LTRA \u00E2\u0080\u0093 leukotriene receptor antagonists; COPD \u00E2\u0080\u0093 chronic obstructive pulmonary diseases; URTI \u00E2\u0080\u0093 upper respiratory tract infection 123 3.3.2.1 The first ED as the study outcome In total, 2,526 ED visits as the study outcome occurred over the subsequent year after time 0. In total, 9.2% of patients who received suboptimal regimens and 3.0% of patients who were on optimal regimens had at least one ED visit for asthma during the follow-up year. Patients with suboptimal asthma regimens were 3.9 times more likely to have ED visits (HR 3.93; 95% CI 3.75 \u00E2\u0080\u0093 4.13) compared to patients who received optimal regimens over the subsequent year of time 0 (Table 3.8). Figure 3.7 shows the significant differences in the cumulative incidence of ED visits for asthma during the entire follow up for the matched cohort. 3.3.2.2 The first hospital admission as the outcome During the one-year follow-up, 378 hospital admissions as the outcome occurred over the subsequent year of after time 0. 1.5% of patients who received suboptimal regimens and 0.3% of patients who were on optimal regimens had at least one hospital admission for asthma during the follow-up. Patients who received suboptimal regimens were 5.4 times more likely to be admitted to hospitals (HR 5.36; 95% CI 4.77 \u00E2\u0080\u0093 5.23) compared to patients who were on optimal regimens over the subsequent year of time 0 (Table 3.8). Figure 3.8 shows the significant differences in the cumulative incidence of ED visits for asthma during the entire follow up for the matched cohort. 124 Table 3.8 Adjusted hazard ratios for ED and hospital admissions associated with suboptimal and optimal use of asthma regimens Suboptimal use of regimens Optimal use of regimens Propensity Score- Matched Regression Standard Covariate- Adjusted Regression Had outcomes, % Median time to outcome, days Had outcomes, % Median time to outcome, days Adjusted HR (95% CI) Adjusted HR (95% CI) ED visits One year follow-up 9.2 141 3.0 166 3.93 (3.75 \u00E2\u0080\u0093 4.13) 4.46 (4.24 \u00E2\u0080\u0093 4.70) Hospital admissions One year follow-up 1.5 172 0.3 192 5.36 (4.77 \u00E2\u0080\u0093 6.03) 6.06 (5.38 \u00E2\u0080\u0093 6.83) 125 Figure 3.7 Cumulative incidence for asthma-related ED visits for propensity score matched cohort Figure 3.8 Cumulative incidence for asthma-related hospital admissions for propensity score matched cohort 126 3.3.2.3 Age group stratified analysis HRs for each age group were summarized in Table 3.9. Consistent with the cross- sectional analysis described previously, adolescent patients who received suboptimal regimens had the greatest risk of using health services than those who received optimal regimens in this age group. Adolescent patients who received suboptimal regimens were 6 times more likely to visit ED (HR 5.98; 95% CI 5.23 \u00E2\u0080\u0093 6.85), and 13 times more likely to be admitted to hospitals for asthma (HR12.81; 95% CI 9.07 \u00E2\u0080\u0093 18.08) over the subsequent year after baseline year compared to patients who were on optimal regimens in this age group. 3.3.2.4 Sensitivity analysis The standard proportional hazards regression models which adjusted for all the potential confounding factors produced similar results as the propensity score models (Table 3.8). 127 Table 3.9 Adjusted hazard ratios for ED and hospital admissions associated with suboptimal and optimal use of asthma regimens by age group Suboptimal use of regimens Optimal use of regimens Propensity Score- Matched Regression Had outcomes, % Median time to outcome, days Had outcomes, % Median time to outcome , days Adjusted HR (95% CI) Children 5-11 years of age ED visits One year follow-up 11.5 134 6.7 152 2.55 (2.29 \u00E2\u0080\u0093 2.83) Hospital admissions One year follow-up 1.9 172 0.8 122 2.78 (2.15 \u00E2\u0080\u0093 3.59) Adolescents 12-18 years of age ED visits One year follow-up 16.1 127 3.4 138 5.98 (5.23 \u00E2\u0080\u0093 6.85) Hospital admissions One year follow-up 2.8 153 0.2 170 12.81 (9.07 \u00E2\u0080\u0093 18.08) Adults 19-34 years of age ED visits One year follow-up 11.0 140 3.1 186 4.84 (4.44 \u00E2\u0080\u0093 5.28) Hospital admissions One year follow-up 1.6 173 0.3 269 9.46 (7.35 \u00E2\u0080\u0093 12.18) Adults 35-55 years of age ED visits One year follow-up 6.3 152 1.7 178 4.47 (4.10 \u00E2\u0080\u0093 4.87) Hospital admissions One year follow-up 1.1 165 0.2 179 5.46 (4.46 \u00E2\u0080\u0093 6.67) 128 CHAPTER 4 : CHANGING FROM SUBOPTIMAL TO OPTIMAL ASTHMA DRUG REGIMENS: SIGNIFICANT IMPROVEMENTS IN HEALTH SERVICES UTILIZATION OUTCOMES 129 4.1 SYNOPSIS Chapter 3 demonstrates that suboptimal asthma drug regimens are associated with increased risk of using health services for asthma exacerbations over a one-year period. However, since patients with asthma frequently change their drug regimens to accommodate symptomatic changes, using a one-year study time frame limits assessment of optimal and suboptimal asthma medication use over time. To address this limitation of the previous research work, the study presented in this chapter aims to identify patients who changed from suboptimal to optimal drug regimens over time, and to determine the association between regimen optimality changes and use of ED or hospital services for asthma exacerbations between 1996 and 2009. In total, 336,901 asthma patients were identified using provincial health service utilization data (including dispensings of respiratory disease-related medications, as well as physician visits and hospital admissions). Dispensings of inhaled SABA either with or without ICS were categorized as optimal or suboptimal according to asthma practice guidelines at 12-month intervals after cohort entry. To have a clear picture of patients\u00E2\u0080\u0099 health services utilization as a result of regimen optimality changes, this study defined a regimen optimality change as a change from a suboptimal regimen in one year to an optimal regimen in the following years. Regimen change was modeled as a time- dependent variable. The study outcomes were repeated ED visits or hospital admissions for asthma exacerbations during the entire follow-up period up to 14 years. Cox regression models were used to take into account the time-dependent nature of asthma 130 medication use and outcome occurrence. As asthma disease severity and patients\u00E2\u0080\u0099 comorbidity status may change over time, measures of these two factors were also modeled as time-dependent variables. In total, 4,059 patients changed their regimens from suboptimal to optimal between 1996 and 2009. These patients contributed to 33,766 person-years, with a mean follow-up period of 8.3 years. Switching to an optimal drug regimen was associated with a reduction of 30% in the use of hospital services for asthma (HR 0.71, 95% CI 0.54 \u00E2\u0080\u0093 0.93), as well as a reduction of 50% in the use of ED services for asthma management (HR 0.49, 95% CI 0.33 \u00E2\u0080\u0093 0.73). With one additional optimal regimen year, the use of hospital services decreased by 10% (HR 0.88, 95% CI 0.81 \u00E2\u0080\u0093 0.97); and the use of ED services for asthma decreased by 30% (HR 0.70, 95% CI 0.58 \u00E2\u0080\u0093 0.85), after adjusting for patients\u00E2\u0080\u0099 age, gender, dispensing of asthma add-on therapy, disease severity and comorbidity status over time. Changing from suboptimal to optimal therapy is associated with significant reductions in health services utilization for asthma. Patients with frequent prescribing and dispensing of SABA but without sufficient dispensing of ICS (suboptimal regimen) are likely to benefit from an increase in ICS use to reduce their need for healthcare services. Patients\u00E2\u0080\u0099 close self-monitoring of asthma symptoms, regular assessments by physicians and better communications between patients and physicians are suggested in order to improve asthma drug use and patients\u00E2\u0080\u0099 asthma-related health services utilization outcomes. 131 4.2 METHODOLOGY This study used B.C. linked population-based health services utilization data and prescription drug dispensing data. Information on data sources quality and preparation are described in detail in CHAPTER 2 Sections 2.2.1 \u00E2\u0080\u0093 2.2.3. 4.2.1 Study design and study patients This is a retrospective cohort study. Similar to the study outlined in Chapter 3, patients were classified as having asthma if they satisfied any of the following criteria during a 12-month period between April 1, 1996 and March 31, 2009 using moving windows: \u00E2\u0089\u00A51 hospital admission with asthma as the principal diagnosis based on ICD-10 code J45 or J46; \u00E2\u0089\u00A52 physician visits with asthma as the principal diagnosis based on the ICD-9 code 493; or \u00E2\u0089\u00A53 asthma drug dispensings. Patients were followed from the first date that they met the asthma case definition. Patients <5 years of age were excluded from this study because of the uncertainties of asthma diagnosis in this age group. Those aged \u00E2\u0089\u00A555 years were also excluded in order to minimize the number of patients who concurrently had COPD. 4.2.2 Exposure assessment-measure of changes in regimen optimality In total, 336,901 asthma patients were identified and entered the cohort on the first date that they met the case definition described above. Dispensings of inhaled SABAs either with or without ICSs over a 12-month interval were used to categorize patient drug regimens as either optimal or suboptimal according to practice guidelines. 12-month intervals were used to account for the seasonal nature of asthma and for inter- and intra- 132 patient variability in medication use. Some regimens were not obviously classifiable as optimal or suboptimal based on drug dispensing data alone. Factors that would have assisted classification of these borderline regimens, including frequency of symptoms, lung function, and the patients\u00E2\u0080\u0099 quality-of-life (details see Chapter 3 Section 3.2.1.2), were not available for this study. Borderline regimens were assessed and deemed unclassifiable by two asthma experts (a respirologist and a clinical pharmacologist). In total, 283,848 patients had at least 2 years in which drug regimens could be classified, and these patients exhibited 35,000 different regimen optimality patterns during the 14-year study period. 198,413 (70%) patients continued receiving optimal asthma drug regimens during their follow-up, while 5,161 patients continued receiving suboptimal regimens (2%). It is difficult to interpret the clinical meaning of changes from or to unclassifiable regimens or to explain the impact of unclassifiable regimens on subsequent drug dispensings and health services utilization. In order to obtain meaningful clinical implications, a regimen optimality change was defined as a change from a suboptimal regimen in one year to an optimal regimen in the following years. Follow-up periods started from the first suboptimal regimen year that was eventually followed by 1 or more optimal years, with no unclassifiable regimens in between. Time 0 was defined as the start of the first suboptimal regimen year during the follow-up period. Patients having only consecutive suboptimal and optimal years were followed until death, age 56 or March 31, 2010 (the last available health utilization data). Patients whose regimens switched from suboptimal to unclassifiable were excluded from the analysis. 133 Unclassifiable regimen years that occurred before a valid optimality change were not included in the follow-up period. Patients who switched to an unclassifiable regimen year after a valid optimality change were censored on the last date of the defined drug regimen. Patients frequently change drug regimens to accommodate changes in symptoms in clinical practice, and thus patients were allowed to switch back from optimal to suboptimal regimens in this analysis. Figure 4.1diagrams the scenarios described above. Figure 4.1 Examples of study scenarios Red indicates suboptimal regimens; green indicates optimal regimens and yellow indicates unclassifiable regimens. Example 1 represents a scenario where a patient had consecutive suboptimal and optimal regimen years from the cohort entry date and were followed until death, age 56 years or March 31, 2010. Example 2 represents a scenario where an unclassifiable regimen optimality year occurred before a defined drug regimen optimality change (i.e., a 134 change from a suboptimal regimen in one year to an optimal regimen in the following year). Unclassifiable regimen years were not included in the follow-up period. Example 3 shows a scenario where a patient\u00E2\u0080\u0099s regimen switched unclassifiable after a defined regimen optimality change; such patients were censored on the last date of the defined drug regimen. 4.2.3 Research outcomes assessment Asthma is characterized by recurrent symptoms[3], and patients may have multiple exacerbations and seek medical assistance repeatedly from ED or hospitals. Thus, the study outcomes of this analysis were defined as multiple ED visits or hospital admissions with ICD-9 code 493 or ICD-10 codes J45 and J46 between 1996 and 2009. Details regarding identification of asthma-related ED visits (from the MSP physician visit database), hospital admissions (from the DAD), as well as the accuracy of ICD-9 and ICD-10 diagnostic codes are presented in Chapter 2 Section 2.2.7.4. 4.2.4 Statistical analysis To account for both intermittent and permanent changes in asthma drug regimen optimality, changes were evaluated for up to 14 years and regimen optimality was analyzed as an annually-updated, time-dependent variable. Suboptimal regimens were followed by up to 13 optimal regimen years; more optimal regimen years may lead to more significant reductions in asthma-related health service use. To account for this, the number of optimal years after regimen optimality change was modeled as a continuous time-dependent variable. 135 Statistical analyses were adjusted for demographic information (i.e., age and gender), dispensing of 2 nd , 3 rd , or 4 th line drug therapy, and common comorbidities (i.e., URTI, allergy, mental illness, chronic bronchitis, emphysema and COPD), since these factors have been shown to be associated with asthma-related health services use[157, 232, 249, 250]. Patients\u00E2\u0080\u0099 ED and hospital admissions in previous years are associated with asthma-related health service use in subsequent years[3]. When ED visits were examined as the study outcome in the current studies, the number of hospital admissions during follow up was adjusted as a measure of disease severity. Likewise, when hospital admissions were used as the outcome, the number of ED visits during follow up was adjusted. In British Columbia, patients are admitted to hospitals mainly through ED admissions, thus in our analysis, those asthma-related ED visits that resulted in discharge but were followed by hospital admissions for asthma within 14 days were treated as the same asthma exacerbation event. These ED visits were not used in the adjustment for the association between asthma drug regimen optimality and repeated hospital admissions for asthma. Instead, these ED visit/hospitalization combinations were used as hospitalization events only. Since asthma severity can change over time and since patients were followed for up to 13 years, severity was measured at 12-month intervals using a time- varying approach. In this analysis, age at time 0 was modeled as a time-fixed variable to prevent a correlation of age and follow-up time. Since patients were followed for up to 13 years, patient characteristics including regimen optimality, disease severity, comorbidity status as well as use of ED and hospital 136 services all varied from year to year during the follow-up period. A Cox proportional hazards model with time-varying exposure, outcomes and measures of disease severity and comorbidities was applied to account for the variability over time. 4.2.5 Sensitivity analyses In our primary analysis, ED visits and hospital admissions during the follow up period were modeled using a time-varying approach. As asthma-related ED and hospital visits in previous year may impact subsequent visits and future drug use, and these effects cannot be taken into account in a multi-event model. Therefore, the primary analysis was repeated using the first ED visit or hospital admission as the study outcome. The aim of this sensitivity analysis was to confirm that changing from suboptimal to optimal regimens was associated with reduced health service use when noise due to associations between previous health services use, subsequent drug use and study outcomes was removed. The second sensitivity analysis was designed to address the effects of unclassifiable drug regimens on health service use. Since unclassifiable regimens were present in one-third of the drug regimen optimality changing patterns, the primary analysis was repeated in order to better understand health service use in patients having an unclassifiable regimen that fell in between suboptimal years or between a suboptimal and an optimal year. The unclassified components of such regimen optimality patterns likely represent transition times during which patients adjust their drug use based on symptoms. We therefore conducted a two-part sensitivity analysis in which the 137 unclassifiable regimen was first regarded as the same as the regimen in the previous year and then regarded as the same as the regimen in the following year. 4.3 RESULTS 4.3.1 Baseline characteristics The study cohort included 4,059 patients who changed from suboptimal to optimal regimens over a 14-year period and together contributed to 33,766 person-years of follow-up between 1996 and 2010. The average follow-up period was 8.3 years, mean patient age at time 0 was 29.3 \u00C2\u00B1 12.6, and women comprised 45.2% of the cohort. The study cohort included 4,059 patients who changed from suboptimal to optimal regimens over a 14-year period and together contributed to 33,766 person-years of follow-up between 1996 and 2010. The average follow-up period was 8.3 years, mean patient age at time 0 was 29.3 \u00C2\u00B1 12.6, and women comprised 45.2% of the cohort. Switching from suboptimal to optimal asthma drug regimens was more likely to occur in male patients, 19-34 and 35-55 years of age. In contrast, 2,013 patients changed from optimal to suboptimal regimens over a 14-year period. This switching was more likely to occur in female patients 12-18 years of age. Patient characteristics are summarized in Table 4.1. 138 Table 4.1 Characteristics of study patients (n=4,059) Patient characteristics* Demographic information Age (y), mean \u00C2\u00B1 SD 29.3 \u00C2\u00B1 12.6 Age groups Young children 5-11 years of age, no (%) 478 (11.8%) Adolescence 12-18 years of age, no (%) 461 (11.4%) Adults 19-34 years of age, no (%) 1,600 (39.4%) Adults 35-55 years of age, no (%) 1,520 (37.4%) Female sex, no. (%) 1,836 (45.2%) Dispensings of adds-on asthma drug therapy Dispensed oral steroids, no. (%) 896 (22.1%) Dispensed nebulized bronchodilators, no. (%) 218 (5.4%) Dispensed theophylline, no. (%) 182 (4.5%) Dispensed omalizumab, no. (%) 0 Dispensed LABA or LABA/ICS combo, no. (%) 376 (9.3%) Dispensed LTRA, no. (%) 128 (3.2%) Comorbidities Allergy, no. (%) 50 (1.2%) Allergic rhinitis, no. (%) 262 (6.5%) Acute URTI**, no. (%) 920 (18.5%) Acute LRTI**, no. (%) 750 (18.5%) Chronic bronchitis, emphysema and COPD**, no. (%) 167 (4.1%) Depression and anxiety, no. (%) 861 (21.2%) *Patient characteristics evaluated at time 0. **Abbreviations: LABA: long-acting bronchodilators; LTRA: leukotriene receptor antagonists; URTI: upper respiratory tract infections; LRTI: lower respiratory tract infections; COPD: chronic obstructive pulmonary diseases. 4.3.2 Switching from suboptimal to optimal regimens Prior to switching from suboptimal to optimal regimens, the median number of years that patients had suboptimal asthma drug regimens was 4, with a range of 1 to 13. After switching, the average number of optimal asthma drug regimen years was 3.4 years, with a range of 1 to 12. Patient characteristics at the one year prior to the first regimen optimality change are summarized in Table 4.2. Percentages of patients who received oral 139 steroids, nebulized bronchodilators or theophylline prescriptions were 8%, 2.1% and 1.2%, respectively, which reduced more than 50% compared to patients\u00E2\u0080\u0099 dispensing at time 0 (Table 4.1). 18.3% of patients received LABA or LABA/ICS combo prescriptions, which was doubled the percentage at time 0 (Table 4.1). On average, each patient received 13 SABA inhalers during each suboptimal regimen year; the averaged ICS daily dose was stable at 50 mcg beclomethasone equivalent. After patients switched to optimal asthma drug regimens, on average, each patient received 2-3 SABA inhalers during each optimal regimen year; the averaged ICS daily dose each patient received changed from 90 mcg to 30 mcg beclomethasone over time. 140 Table 4.2 Characteristics of patients over the one year prior to their first change from suboptimal to optimal asthma drug regimens during follow-up Patient characteristics Demographic information Age (y), mean \u00C2\u00B1 SD 32.8 \u00C2\u00B1 12.7 Age groups Young children 5-11 years of age, no (%) 226 (5.6%) Adolescence 12-18 years of age, no (%) 446 (11.0%) Adults 19-34 years of age, no (%) 1,466 (36.1%) Adults 35-55 years of age, no (%) 1,921 (47.3%) Female sex, no. (%) 1,836 (45.2%) Dispensings of adds-on asthma drug therapy Dispensed oral steroids, no. (%) 325 (8.0%) Dispensed nebulized bronchodilators, no. (%) 87 (2.1%) Dispensed theophylline, no. (%) 50 (1.2%) Dispensed omalizumab, no. (%) 4 (0.1%) Dispensed LABA or LABA/ICS combo, no. (%) 744 (18.3%) Dispensed LTRA, no. (%) 152 (3.7%) Comorbidities Allergy, no. (%) 42 (1.0%) Allergic rhinitis, no. (%) 165 (4.1%) Acute URTI**, no. (%) 605 (14.9%) Acute LRTI**, no. (%) 407 (10.0%) Chronic bronchitis, emphysema and COPD**, no. (%) 99 (2.4%) Depression and anxiety, no. (%) 620 (15.3%) **Abbreviations: LABA: long-acting bronchodilators; LTRA: leukotriene receptor antagonists; URTI: upper respiratory tract infections; LRTI: lower respiratory tract infections; COPD: chronic obstructive pulmonary diseases. 4.3.3 Health services use associated with switching from suboptimal to optimal regimens 5,515 asthma-related ED visits occurred during the follow-up period. The majority (4,800) occurred during suboptimal regimen years. Among the 1,467 patients with at least one asthma-related ED visits, the median time between time 0 and visits was 92 days (minimum, 1 day; maximum, 1,325 days). 141 616 hospital admissions occurred during the follow-up period, with the majority (541) occurred during suboptimal regimen years. Among 391 patients who had at least one hospital admissions, the median time between time 0 and admissions was 175 days (minimum, 32 days; maximum, 2,103 days). On average, the hospital admission rate for asthma was 4 per 100 patients per year during suboptimal regimen years. After switching to optimal asthma drug regimens, the hospital admission rate was 1 per 100 patients per year. Similarly, on average, the ED visit rate for asthma was 31 per 100 patients per year during suboptimal regimen years. After switching to optimal asthma drug regimens, the ED visit rate was 6 per 100 patients per year. When the association between regimen optimality changes and use of health services was examined, suboptimal to optimal switches were associated with a 50% reduction in the likelihood of an ED visit (HR 0.49; 95% CI 0.33 \u00E2\u0080\u0093 0.73) (Table 4.3) and a 30% reduction in the likelihood of a hospital admission (HR 0.71; 95% CI 0.54 \u00E2\u0080\u0093 0.93) after adjusting for age, gender, asthma disease severity, and comorbidity status (Table 4.4). In the relationship between asthma drug regimen changing from suboptimal to optimal and use of ED services for exacerbations, previous hospital admissions for asthma was associated with a twice increase in the likelihood of subsequent ED visits (HR 1.98; 95% CI 1.76 \u00E2\u0080\u0093 2.23). Dispensing of add-on therapy (i.e., theophylline, nebulized bronchodilators, or LABA) was not associated with asthma-related ED visits significantly except the dispensing of LTRA prescriptions, which decreased the likelihood 142 of ED visits by 10% (HR 0.90; 95% CI 0.86 \u00E2\u0080\u0093 0.94). In the relationship between asthma drug regimen changing from suboptimal to optimal and hospital admissions for exacerbations, as expected, previous ED visits for asthma were associated with a 30% increase in the likelihood of subsequent hospital admissions (HR 1.31; 95% CI 1.29 \u00E2\u0080\u0093 1.33). Dispensing of nebulized bronchodilators and LTRA increased the likelihood of hospital admission for asthma significantly. 143 Table 4.3 Unadjusted and adjusted hazard ratios and 95% confidence intervals for risk of ED visits for asthma exacerbations Unadjusted Hazard Ratio (95% Confidence Interval) P-value Adjusted Hazard Ratio (95% Confidence Interval) p-value Primary Exposure (Time- Dependent) Changing from suboptimal to optimal regimens 0.18 (0.17 \u00E2\u0080\u0093 0.20) <0.001 0.49 (0.33 \u00E2\u0080\u0093 0.73) <0.001 Time-dependent covariates No. of hospital admissions for asthma 3.72 (3.64 \u00E2\u0080\u0093 3.81) <0.001 1.98 (1.76 \u00E2\u0080\u0093 2.23) <0.001 Dispensings of theophylline 2.15 (1.93 \u00E2\u0080\u0093 2.39) <0.001 1.04 (0.75 \u00E2\u0080\u0093 1.43) 0.81 Dispensings of nebulizers 3.99 (3.68 \u00E2\u0080\u0093 4.32) <0.001 0.77 (0.56 \u00E2\u0080\u0093 1.04) 0.09 Dispensings of LABA * 2.38 (2.22 \u00E2\u0080\u0093 2.55) <0.001 1.17 (0.92 \u00E2\u0080\u0093 1.49) 0.19 Dispensings of LTRA 0.90 (0.86 \u00E2\u0080\u0093 0.93) <0.001 0.90 (0.86 \u00E2\u0080\u0093 0.94) <0.001 Comorbid allergy 1.40 (1.11 \u00E2\u0080\u0093 1.78) 0.005 1.38 (0.63 \u00E2\u0080\u0093 3.00) 0.42 Comorbid allergic rhinitis 1.20 (1.07 \u00E2\u0080\u0093 1.35) 0.003 0.57 (0.38 \u00E2\u0080\u0093 0.84) 0.005 Comorbid URTI* 1.63 (1.53 \u00E2\u0080\u0093 1.74) <0.001 1.14 (0.90 \u00E2\u0080\u0093 1.45) 0.28 Comorbid LRTI* 2.04 (1.91 \u00E2\u0080\u0093 2.18) <0.001 1.31(1.03 \u00E2\u0080\u0093 1.67) 0.03 Comorbid COPD* 2.60 (2.34 \u00E2\u0080\u0093 2.88) <0.001 1.23 (0.77 \u00E2\u0080\u0093 1.97) 0.38 Comorbid depression or anxiety 1.13 (1.05 \u00E2\u0080\u0093 1.21) <0.001 1.17 (0.91 \u00E2\u0080\u0093 1.50) 0.21 Time-fixed covariates Age 0.97 (0.97 \u00E2\u0080\u0093 0.98) <0.001 0.97 (0.97 \u00E2\u0080\u0093 0.98) <0.001 Gender (male vs. female) 1.10 (1.04 \u00E2\u0080\u0093 1.16) <0.001 1.04 (0.84 \u00E2\u0080\u0093 1.28) 0.71 *Abbreviations: LABA: long-acting bronchodilators; LTRA: leukotriene receptor antagonists; URTI: upper respiratory tract infections; LRTI: lower respiratory tract infections; COPD: chronic obstructive pulmonary diseases; 144 Table 4.4 Unadjusted and adjusted hazard ratios and 95% confidence intervals for risk of hospital admissions for asthma exacerbations Unadjusted Hazard Ratio (95% Confidence Interval) P-value Adjusted Hazard Ratio (95% Confidence Interval) p-value Primary Exposure (Time- Dependent) Changing from suboptimal to optimal regimens 0.19 (0.14 \u00E2\u0080\u0093 0.24) <0.001 0.71 (0.54 \u00E2\u0080\u0093 0.93) 0.01 Time-dependent covariates No. of ED visits for asthma 1.42 (1.40 \u00E2\u0080\u0093 1.44) <0.001 1.31 (1.29 \u00E2\u0080\u0093 1.33) <0.001 Dispensings of theophylline 4.58 (3.51 \u00E2\u0080\u0093 5.99) <0.001 1.26 (0.95 \u00E2\u0080\u0093 1.67) 0.11 Dispensing of nebulizers 9.65 (7.97 \u00E2\u0080\u0093 11.69) <0.001 2.23 (1.82 \u00E2\u0080\u0093 2.73) <0.001 Dispensings of LABA* 5.08 (4.20 \u00E2\u0080\u0093 6.15) <0.001 1.19 (0.96 \u00E2\u0080\u0093 1.48) 0.11 Dispensings of LTRA 1.14 (1.09 \u00E2\u0080\u0093 1.20) <0.001 1.06 (1.00 \u00E2\u0080\u0093 1.12) 0.04 Comorbid allergy 1.12 (0.46 \u00E2\u0080\u0093 2.70) 0.8 0.57 (0.24 \u00E2\u0080\u0093 1.39) 0.22 Comorbid allergic rhinitis 2.21 (1.63 \u00E2\u0080\u0093 3.00) <0.001 0.84 (0.60 \u00E2\u0080\u0093 1.17) 0.30 Comorbid URTI* 2.45 (2.04 \u00E2\u0080\u0093 2.95) <0.001 0.91 (0.75 \u00E2\u0080\u0093 1.10) 0.34 Comorbid LRTI* 3.76 (3.15 \u00E2\u0080\u0093 4.49) <0.001 1.07 (0.89 \u00E2\u0080\u0093 1.29) 0.46 Comorbid COPD* 4.87 (3.73 \u00E2\u0080\u0093 6.36) <0.001 0.77 (0.56 \u00E2\u0080\u0093 1.06) 0.12 Comorbid depression or anxiety 2.42 (2.00 \u00E2\u0080\u0093 2.93) <0.001 1.22 (1.00 \u00E2\u0080\u0093 1.49) 0.05 Time-fixed covariates Age 0.98 (0.97 \u00E2\u0080\u0093 0.99) <0.001 0.98 (0.97 \u00E2\u0080\u0093 0.99) <0.001 Gender (male vs. female) 1.10 (1.04 \u00E2\u0080\u0093 1.16) 0.02 0.77 (0.65 \u00E2\u0080\u0093 0.91) 0.003 *Abbreviations: LABA: long-acting bronchodilators; LTRA: leukotriene receptor antagonists; URTI: upper respiratory tract infections; LRTI: lower respiratory tract infections; COPD: chronic obstructive pulmonary diseases. 145 When regimen change was modeled as a continuous time-dependent variable that accounted for additional optimal years after switching, the likelihood of ED visit and hospital admission decreased by 30% (HR 0.70, 95% CI 0.58 \u00E2\u0080\u0093 0.85) and 10% (HR 0.88, 95% CI 0.81 \u00E2\u0080\u0093 0.97), respectively. Patients who switched from suboptimal to optimal regimens for 1-2 years were 15% less likely to be admitted to hospitals for asthma (HR 0.84; 95% CI 0.56 \u00E2\u0080\u0093 1.25), while patients who switched to optimal regimens years for 5-6 years were 90% less likely to use hospital services (HR 0.10; 95% CI 0.02 \u00E2\u0080\u0093 0.45) (Figure 4.3). Similarly, switching to an optimal regimen for 1-2 years was associated with a 40% decrease in the use of ED services (HR 0.59; 95% CI 0.51 \u00E2\u0080\u0093 0.68) and switching to optimal regimens for 5-6 years was associated with a 70% reduction (HR 0.29; 95% CI 0.21 \u00E2\u0080\u0093 0.41) (Figure 4.3). 146 Figure 4.2 Association between the number of optimal regimen years and the likelihood of hospital admission for asthma Figure 4.3 Association between the number of optimal regimen years and the likelihood of ED visits for asthma 147 4.3.4 Results from sensitivity analyses Sensitivity analysis 1: first ED or hospital admission for asthma during the follow up as the study outcome When using the first ED or hospital admission for asthma as the research outcome, 1,323 ED visits occurred during suboptimal years and 160 ED visits occurred during optimal years. Similarly, 356 hospital admissions occurred during suboptimal years and 60 occurred during optimal years. Patients who changed from suboptimal to optimal regimens were 80% less likely to visit ED (HR 0.23; 95% CI 0.19 \u00E2\u0080\u0093 0.28) and 85% less likely to be admitted to hospitals for asthma (HR 0.16; 95% CI 0.10 \u00E2\u0080\u0093 0.23). This association is consistent with findings from the primary analysis. Sensitivity analysis 2: unclassified regimen optimality regarded as the same as the optimality class in the previous or subsequent year In total, 26,467 patients had one unclassifiable regimen year between two suboptimal regimen years, or two optimal regimen years, or one suboptimal and optimal regimen year. When this unclassifiable regimen was regarded as the same as the optimality class for the previous or subsequent year, 11,476 patients made a regimen optimality change, namely a change from a suboptimal regimen in one year to an optimal regimen in the following year between 1996 and 2009. When this unclassifiable regimen was regarded as the same as the optimality class for the previous year, patients who changed from suboptimal to optimal regimens 148 experienced 60% fewer in ED visits (HR 0.42; 95% CI 0.40 \u00E2\u0080\u0093 0.44) and were 70% less likely to be admitted to hospitals for asthma (HR 0.32; 95% CI 0.29 \u00E2\u0080\u0093 0.36). When unclassifiable regimens were regarded as the same as the optimality class in the subsequent year, switching to optimal regimens was associated with a 60% reduction in ED visits for asthma (HR 0.42; 95% CI 0.41 \u00E2\u0080\u0093 0.43) and a 60% reduction in hospital admissions (HR 0.41; 95% CI 0.37 \u00E2\u0080\u0093 0.45) after adjusting for age at time 0, gender, asthma severity, and comorbidity status. 149 CHAPTER 5 : REGULAR VS. INTERMITTENT USE OF INHALED CORTICOSTEROIDS 150 5.1 SYNOPSIS Inhaled corticosteroids are the most effective and safest controller medications for asthma management and critical in reducing airway inflammation. NHLBI guidelines recommend daily ICS use for patients who require >3 doses (per week) of short-acting bronchodilators. However, due to concerns of potential risk of ADR associated with ICS use (e.g., growth retardation in children; osteoporosis, cataracts and glaucoma in adults) some clinicians recommend that patients only use them during flu/allergy season or if symptoms flare up. A few randomized clinical trials have examined the efficacy of intermittent ICS use, but these show conflicting results. The aim of this chapter is to compare asthma-related ED and hospital service use in patients who received ICS regularly and those who received ICS intermittently via a retrospective cohort study. Patients entered the cohort on the first date they met our asthma case definition, and ICS use status in the first year after entry (the index year) was categorized as either regular or intermittent. To categorize ICS use status, all ICSs were converted to beclomethasone 50\u00C2\u00B5g metered dose inhaler (MDI) equivalents. In accordance with clinical practice guidelines, regular ICS use was defined by \u00E2\u0089\u00A54 beclomethasone inhaler dispensings in the index year, and \u00E2\u0089\u00A51 inhaler every 100 days. Intermittent ICS use was defined as 1-2 inhaler dispensings during the index year. Asthma severity was measured using dispensings of SABA and add-on asthma drug therapies, visits to the ED or hospital for asthma as well as comorbidity status in the index year. Study outcomes included the first asthma-related ED visit or hospital admission in the 365 days following the index 151 year. Propensity score matching was used to ensure the comparability of disease severity between regular and intermittent ICS users, and Cox proportional regression models were used to take into account the time dependent nature of asthma drug use over time. A total of 16,652 patients received ICS regularly and 13,255 patients received ICS intermittently during the index year. Compared to patients who received ICS intermittently, those who received ICS regularly were 20% less likely to visit ED (HR 0.81; 95% CI 0.74-0.90) and 30% less likely to be admitted to hospitals (HR 0.72 (0.56 \u00E2\u0080\u0093 0.92) for asthma over the subsequent year. These findings emphasize that regular ICS use is critical for asthma management and indicate that improved ICS adherence may lead to a reduction in asthma-related health service use and an increase in patient quality of life. 5.2 METHODOLOGY 5.2.1 Study design and study patients B.C. population-based health services utilization data and prescription drug dispensing data were used for this study. Information on sources, quality and preparation of data has been described in detail in CHAPTER 2 Sections 2.2.1 \u00E2\u0080\u0093 2.2.3. A retrospective cohort study was conducted. Patients were classified as having asthma if they satisfied any of the following criteria during 12-month moving windows: at least one hospital admission with asthma as the principal diagnosis based on the International Classification of Diseases, 10 th version, code J45 (ICD-10 J45); or at least two physician visits for asthma as the principal diagnosis based on the ICD 9 th version, 152 code 493 (ICD-9 493); or at least three asthma drug dispensings. Patients younger than 5 years of age were excluded because of the uncertainties of the diagnosis of asthma in this age group. We also excluded those aged 55 years or older to minimize patients who concurrently had chronic obstructive pulmonary disease. Detailed information in relation to the asthma case definition, validity of asthma case definition, validity and accuracy of ICD diagnostic codes has been described previously in CHAPTER 2 Section 2.2.5 and 2.2.6. The initial asthma cohort contained 336,901 patients. The cohort entry date was defined as the first date that patients met above asthma case definition. The one-year period after cohort entry date was used as the index year. In order to be categorized into regular and intermittent ICS use groups, patients needed to have complete records during the index year. More specifically, 5,703 patients turned 56 years of age, 2,500 died and March 31, 2010 (i.e., the last date our data is available) occurred during the index year in 19,574 patients. These patients were excluded and the remaining 308,319 patients were included in subsequent analyses. Figure 5.1shows how the study population for this analysis was formed. Patients were followed for up to 365 days after the index year. The start date of the follow-up period was defined as time 0. 153 Figure 5.1 Form of study patients in the comparison of health services utilization between patients with regular and intermittent use of ICS Note: March 31 st , 2010 which is final date of our study data. Patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the index year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. 5.2.2 Study measures 5.2.2.1 Measure of study exposure All ICS prescriptions, either alone or in a combination inhaler, were converted to 154 beclomethasone metered dose inhaler (MDI) 50 mcg, 200 doses equivalent. Beclomethasone is 0.5 the potency of fluticasone, 2.5 times as the potency of flunisolide and 0.8 times the potency of budesonide[2]. Based on ICS refill records during the baseline year, patients were classified into two groups: those with regular use of ICS and those with intermittent use of ICS. Based on the NHLBI Asthma Management Guidelines, the recommended minimum maintenance dose of beclomethasone MDI 50 mcg, 200 doses for patients 5-11 years of age is 1 puff, twice daily. For patients \u00E2\u0089\u00A512 years, the maintenance dose is 1-2 puffs, twice daily. According to the minimum maintenance dose, a 50 mcg, 200 dose beclomethasone MDI could last up to 100 days. As B.C. has a 100 Day Maximum Supply policy (i.e., for repeated prescriptions of maintenance drugs, the maximum dispensed is 100 days supply), patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the baseline year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. In patients who used ICS intermittently, 70% of ICS dispensings occurred during winter and spring seasons (Figure 5.2). 155 Figure 5.2 Distribution of ICS dispensings in patients who used ICS intermittently during the index year The cohort entry date was defined as the first date that patients met above asthma case definition. The one-year period after cohort entry date was used as the index year. 5.2.2.2 Measure of study outcomes Outcomes included the first ED visit and hospital admission for asthma during the follow-up year after time 0. The databases used to identify ED visits and hospital admissions for asthma were described in Chapter 2 Section 2.2.7.4. Patients were followed until they manifested the outcome events, died, turned 56 years of age, or until March 31 st , 2010 which is final date of our study data. 5.2.2.3 Measure of disease severity and comorbidity status Several variables are associated with patients\u00E2\u0080\u0099 use of asthma medications as well 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec N u m b er o f IC S d is p en si n g s Month 156 as use of health services. These are regarded as confounding variables (i.e., covariates) in the analysis, including demographic factors (i.e., age, gender), level of asthma control, as well as comorbidities during the baseline year. Certain comorbidities have been reported to significantly affect patients\u00E2\u0080\u0099 use of asthma drugs and health services, including allergy, allergic rhinitis, URTI, lower respiratory tract infections (LRTI), COPD, emphysema, chronic bronchitis, depression and anxiety[72, 232]. As such, comobidities in patients with asthma were identified during the baseline year using the Expanded Diagnosis Clusters (EDC) \u00C2\u00AE [258] data provided by the B.C. Ministry of Health. The EDC \u00C2\u00AE system, developed by investigators at Johns Hopkins University, is a useful tool for identifying patients with specific comorbidities. Each of the 264 EDCs correlates with an ICD diagnostic code reported in physician visit, hospital admission or encounter databases. Another important potential confounding factor is asthma control status, which has been associated with a higher likelihood of ED visit or hospital admission, and with a greater dispensing of asthma medications[250, 259, 260]. In the present study, level of asthma control is indicated by use of health services (i.e., number of ED visits or hospital admissions) and number of the second/third line asthma drug prescriptions; (i.e., LABA, theophylline, LTRA). The association between these factors and asthma control level has been well established in previous studies[153, 154, 250] (details in Chapter I Section 1.8). 5.2.3 Data analysis 5.2.3.1 Descriptive analysis Characteristics, numbers and percentages of patients who received ICSs regularly 157 and intermittently were described. 5.2.3.2 Propensity score matching Propensity score matching was used to ensure the two patient groups had comparable age, gender, level of asthma control and comorbidity status. The propensity score method was developed by Rosenbaum and Rubin[254] and is defined as the probability of assignment of a medication conditional on a series of observed covariates. Multiple approaches can be used to calculate propensity score, such as neural networks, discriminant function analysis, and classification trees[255]. In the present analysis, propensity score was calculated using a logistic regression model, with the outcome being the probability of receiving ICS regularly or intermittently. As described earlier, a series factors were used to measure patients\u00E2\u0080\u0099 disease severity and comorbidity status, these factors could be distributed unevenly between patients with regular and intermittent use of ICS. To determine which factor should be used in the propensity score calculation, univariate logistic regression models were conducted to examine the relationship between each of those potential confounding variable and the likelihood of using ICS regularly. The ORs were listed in Table 5.1. All potential confounding variables, which had statistically significant (p<0.05 for 2-tailed Wald test) unadjusted association with using ICS regularly were included in our propensity score model. The propensity score model yielded a c-statistic of 0.78, which indicated a strong ability to differentiate between the two ICS use groups. To maximize the accuracy of matching, we used an \u00E2\u0080\u0098exact matching\u00E2\u0080\u0099 methodology, which means the matched pair had the same propensity score. The 158 standardized differences of each confounding variable before and after propensity score matching were calculated and shown in Figure 5.3. A Quantile-quantile plot (QQ plot) was used to evaluate whether there is a good balance between patients who received ICS regularly and those who received ICS intermittently after the propensity score matching (Figure 5.4). The QQ plot has been regarded as the best approach to compare distributions of confounding variables between comparison groups[256]. Patients with regular ICS use were matched 1:1 with those with intermittent ICS use during the baseline year. 159 Table 5.1 Propensity score variables and association with using ICS regularly Univariate OR* p-value Variables Demographic information Age (years) 1.042 1.041 \u00E2\u0080\u0093 1.044 Gender (men vs. women) 0.91 0.87 \u00E2\u0080\u0093 0.95 Number of asthma medications dispensed during the index year SABA inhalers 1.078 1.073 \u00E2\u0080\u0093 1.083 LABA prescriptions 2.40 2.27 \u00E2\u0080\u0093 2.53 Theophylline prescriptions 1.44 1.41 \u00E2\u0080\u0093 1.47 LTRA prescriptions 1.15 1.11 \u00E2\u0080\u0093 1.18 Use of health services for asthma ED visits 1.05 1.02 \u00E2\u0080\u0093 1.08 Hospital admissions 1.34 1.22 \u00E2\u0080\u0093 1.47 Comorbidity Allergy (y/n) 0.98 0.83 \u00E2\u0080\u0093 1.15 Allergic rhinitis (y/n) 1.09 1.01 \u00E2\u0080\u0093 1.17 URTI (y/n) 0.65 0.62 \u00E2\u0080\u0093 0.68 Depression or anxiety (y/n) 1.51 1.43 \u00E2\u0080\u0093 1.60 LRTI (y/n) 1.10 1.05 \u00E2\u0080\u0093 1.17 COPD, emphasema or chronic bronchitis (y/n) 3.67 3.25 \u00E2\u0080\u0093 4.14 LABA \u00E2\u0080\u0093 long-acting bronchodilators; LTRA \u00E2\u0080\u0093 leukotriene receptor antagonists; URTI \u00E2\u0080\u0093 upper respiratory tract infections; LRTI \u00E2\u0080\u0093 lower respiratory tract infections; COPD \u00E2\u0080\u0093 chronic obstructive pulmonary diseases *Odds ratios reflect the likelihood of using ICS regularly. 160 Figure 5.3 Standardized difference in the distribution of each confounding variable before and after propensity score matching URTI \u00E2\u0080\u0093 upper respiratory tract infection; LRTI \u00E2\u0080\u0093 lower respiratory tract infection; LTRA \u00E2\u0080\u0093 leukotriene receptor antagonists; COPD \u00E2\u0080\u0093 chronic obstructive pulmonary disease; LABA \u00E2\u0080\u0093 long-acting bronchodilator; ED \u00E2\u0080\u0093 emergency department 161 Figure 5.4 Quantile-quantile plots: evaluation of propensity score matching 162 Black dots represent distributions of each variable before propensity score matching; and red dots represent distributions of each variable after propensity score matching. 163 5.2.3.3 Cox regression models Cox proportional hazards regression models were used to account for the time- dependent nature of drug use and calculate the hazard of ED visits or hospital admissions among patients who received ICS regularly or intermittently, at a given time, t. The hazard of outcome event occurrence was defined as the conditional probability of the event occurring at time t, given survival to time t or later. The Cox proportional hazard model assumes a baseline hazard function, and the hazard of study outcomes (i.e., ED and hospital admissions) is a constant multiple of this hazard function. Log minus log (LML) plot of survival is a well-accepted method of testing if the two comparison groups meet the Cox model assumption [257] and was used in the present analysis. This LML plot produced parallel lines (Figure 5.5), suggesting the proportional hazards holds for the two groups (i.e., patients who received ICS regularly and those who received ICS intermittently). For continuous variables, the HR represents the reduction in risk associated with a one-unit covariate increase at each time point. For categorical variables, HRs were calculated for each category compared to the reference category. The significance of the individual regression estimates were tested using the Wald statistic. We cleaned and prepared the data, and conducted descriptive analysis and multivariate logistic regression models using the Statistical Package for the Social Sciences (SPSS) version 20. Propensity score matching and time-dependent analysis were performed using R version 3.0.0 (\u00C2\u00A9 The R Foundation for Statistical Computing). 164 Figure 5.5 Log minus log plot to asses Cox model hazard function assumption 5.2.4 Sensitivity analysis In order to ensure the robustness of the association between ICS use group and the health services utilization outcome, we conducted several sensitivity analyses using different definitions of ICS use as well as different study cohort. 5.2.4.1 Cohort definition Our primary analysis focused on a cohort of patients with newly diagnosed asthma. However, in our sensitivity analysis, we followed patients beginning the third year since the first date that they met our case definition. This analysis was designed to test the relationship between exposure and outcomes in a cohort with stable asthma. 165 5.2.4.2 Exposure The primary analysis of this study defined regular use of ICS as having being dispensed \u00E2\u0089\u00A54 ICS inhalers during the baseline year with at least one dispensed in each quarter. Since patients may take doses different from those prescribed, or miss doses (due to forgetfulness or delays in re-filling prescriptions), we used an alternate definition of regular ICS to test whether we would obtain similar findings using another definition. Regular ICS use for this sensitivity analysis was defined as \u00E2\u0089\u00A53 ICS inhalers during the baseline year, with at least one dispensed inhaler every 120 days. 5.2.4.3 Analytical approach A standard proportional hazards regression was used to repeat the analysis, adjusting for all potential confounding factors, rather than matching on the propensity score. 5.3 RESULTS 5.3.1 Baseline characteristics A total of 16,652 patients received ICS regularly and 13,255 patients received ICS intermittently during the baseline year. The characteristics of these two groups are summarized in Table 5.2. Prior to matching, patients who were on ICS regularly were older, more likely to be female, more likely to receive SABA and the second-/third-line asthma drug therapy, more likely to visit ED or be admitted to hospitals for asthma, and more likely to have allergic rhinitis, depression/anxiety, lower respiratory tract infections or COPD compared to patients who received ICS intermittently. Matching on propensity score resulted in 4,512 pairs. Propensity score matching ensured the balance of characteristics between the matched cohorts, with the standardized differences 166 of 0 for all matching variables between patients with regular and intermittent use of ICS (Table 5.2). 167 Table 5.2 Characteristics of patient groups before and after propensity score matching (based on the 90-day definition) Before propensity score matching After propensity score matching Regular ICS use group Intermittent ICS use group Standard difference, % P Value Regular ICS use group Intermittent ICS use group Standard difference, % P Value # of patients 16,652 13,255 4,512 4,512 Demographic information Mean age (SD), y 34.7 (15.9) 22.4 (15.4) 78.7 < 0.001 27.8 (16.2) 28.2 (15.7) 2.4 0.247 Female, % 53.6 51.3 4.7 <0.001 46.6 46.6 0.0 1.000 Dispensing of asthma medications Mean # of SABA inhalers dispensed 6.5 (7.7) 3.6 (5.2) 43.6 < 0.001 3.8 (4.4) 3.8 (4.4) 0.0 1.000 Mean # of LABA inhalers dispensed 1.7 (3.1) 0.1 (0.3) 72.5 < 0.001 0.01 (0.16) 0.01 (0.16) 0.0 1.000 Mean # of theophylline prescriptions dispensed 0.2 (1.4) 0.04 (0.49) 15.9 <0.001 0.0007 (0.33) 0.0007 (0.33) 0.0 1.000 Mean # of LTRA prescriptions dispensed 0.3 (1.3) 0.1 (0.8) 13.9 <0.001 0.02 (0.34) 0.02 (0.34) 0.0 1.000 Use of health services for asthma Mean # of ED visits for asthma 0.3 (1.0) 0.2 (0.7) 3.8 0.001 0.05 (0.24) 0.05 (0.24) 0.0 1.000 Mean # of hospital admissions for asthma 0.04 (0.26) 0.03 (0.20) 5.1 <0.001 0.0 (0.05) 0.0 (0.05) 0.0 1.000 Comorbidity Allergy 2.0 2.0 0.3 0.77 0.4 0.4 0.0 1.000 Allergic rhinitis 12.0 11.1 2.7 0.19 6.6 6.6 0.0 1.000 URTI 27.6 37.1 20.4 P<0.001 25.6 25.6 0.0 1.000 Depression/anxiety 24.7 17.8 16.9 P<0.001 16.1 16.1 0.0 1.000 168 Before propensity score matching After propensity score matching Regular ICS use group Intermittent ICS use group Standard difference, % P Value Regular ICS use group Intermittent ICS use group Standard difference, % P Value Acute lower respiratory tract infection 22.7 21.0 4.1 P<0.001 13.1 13.1 0.0 1.000 Emphysema, chronic bronchitis, COPD 8.6 2.5 26.9 P<0.001 0.6 0.6 0.0 1.000 LABA \u00E2\u0080\u0093 long-acting bronchodilators; LTRA \u00E2\u0080\u0093 leukotriene receptor antagonists; URTI \u00E2\u0080\u0093 upper respiratory tract infections; LRTI \u00E2\u0080\u0093 lower respiratory tract infections; COPD \u00E2\u0080\u0093 chronic obstructive pulmonary disease 169 5.3.2 Comparison between regular and intermittent use of ICS In total, 470 ED visits and 52 hospital admissions occurred over the subsequent year after time 0. The incidence of ED visit was 4.6% in patients who used ICS regularly and 5.9% in patients who used ICS intermittently, while the incidence of hospital admission was 0.6% for both patient groups over the subsequent year after time 0. Patients who received ICS regularly were 20% less likely to visit ED (HR 0.81; 95% CI 0.74-0.90) and 30% less likely to be admitted to hospitals (HR 0.72 (0.56 \u00E2\u0080\u0093 0.92) for asthma over the subsequent year after time 0 compared to those who received ICS intermittently (Table 5.3). The cumulative incidence curves of hospital admission for unmatched and matched cohort, were shown in Figure 5.6 and Figure 5.7. In unmatched cohort, the cumulative incidence of hospital admission was higher in patients who used ICS regularly than the incidence in those who used ICS intermittently. However, after adjusting for propensity score, the cumulative incidence was significantly higher in those with ICS intermittently. In unmatched cohort, the cumulative incidence of ED visits for asthma was lower in patients who used ICS regularly compared to those with ICS intermittently (Figure 5.8). However, after adjusting for propensity score, the differences in cumulative incidence were more significant between these two study groups (Figure 5.9). 170 Table 5.3 Adjusted hazard ratios for ED and hospital admissions associated with regular (90 days definition) and intermittent use of ICS Regular use of ICS Intermittent use of ICS Propensity Score- Matched Regression Standard Covariate- Adjusted Regression Had outcomes, % Median time to outcome, days Had outcomes, % Median time to outcome , days Adjusted HR (95% CI)* Adjusted HR (95% CI) ED visits One year follow-up 4.6 156 5.9 166 0.81 (0.75 \u00E2\u0080\u0093 0.88) 0.70 (0.64 - 0.77) Hospital admissions One year follow-up 0.6 236 0.6 156 0.72 (0.56 \u00E2\u0080\u0093 0.92) 0.69 (0.55 \u00E2\u0080\u0093 0.87) *Reflects the risk of health services utilization for asthma in the regular ICS use group compared to the intermittent ICS use group. Hazard ratios are adjusted for the propensity score, calculated using age, gender, number of SABA, LABA, LTRA, theophylline prescriptions, number of ED and hospital admissions, patients\u00E2\u0080\u0099 comorbidity status during the index year. 171 Figure 5.6 Cumulative incidence curves of hospital admissions for unmatched cohort Patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the index year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. 172 Figure 5.7 Cumulative incidence curves of hospital admissions for matched cohort Patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the index year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. 173 Figure 5.8 Cumulative incidence curves of emergency department visits for unmatched cohort Patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the index year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. 174 Figure 5.9 Cumulative incidence curves of emergency department visits for matched cohort Patients with regular ICS use were defined as receiving \u00E2\u0089\u00A54 ICS inhalers during the index year, and were required to receive \u00E2\u0089\u00A51 inhaler in each quarter. Intermittent ICS use was defined as receiving 1-2 inhalers per year, which equates to approximately 1 puff daily for 6 months, or 2 puffs twice daily for up to 3 months. 5.3.3 Sensitivity analyses Definition of ICS regular use based on 120-day intervals In total, 27,344 patients received ICS regularly (based on the120-day interval definition) 175 and 13,255 patients received ICS intermittently during the index year. Patients with this alternative definition of regular ICS use were older, more likely to be female, to receive asthma drug therapy, to use ED or hospital services for asthma, had comorbid depression, anxiety, acute lower respiratory tract infection or COPD, but less likely to have a URTI during the baseline year compared to patients who were on ICS intermittently (Table 5.4). After propensity score matching, 6,441 pairs were matched. The standardized difference for age between the two comparison groups was 1.4% (p = 0.42) and 0 for all other matching variables (Table 5.4). Results based on this alternative definition of regular ICS use (i.e., patients received at least one ICS in each 120-day during the baseline year) were consistent with our main analysis. Using first ED visit as outcome, patients who were on ICS regularly were 20% less likely to visit ED (HR 0.82, 95% CI 0.76-0.89) compared with those who used ICS intermittently over the subsequent year after time 0 (Table 5.5). When examining the first hospital admission as the outcome, compared to patients with intermittent ICS use, patients with regular ICS use were 20% less likely to be hospitalized for asthma (HR 0.78, 95% CI 0.63 \u00E2\u0080\u0093 0.96) during the first year of follow-up (Table 5.5). Standard proportional hazards regression model The standard proportional hazards regression models which adjusted for all the potential confounding factors produced similar results as the propensity score models (Table 5.3). 176 Table 5.4 Characteristics of patient groups before and after propensity score matching (based on the 120-day definition) Before propensity score matching After propensity score matching Regular ICS use group Intermittent ICS use group Standard difference, % P Value Regular ICS use group Intermittent ICS use group Standard difference, % P Value # of patients 27,344 13,255 6,442 6,442 Demographic information Mean age (SD), y 33.5 (16.2) 22.4 (15.4) 70.4 < 0.001 25.6 (16.1) 25.8 (15.8) 1.4 0.420 Female, % 53.7 51.3 4.9 <0.001 47.4 47.4 0.0 1.000 Dispensing of asthma medications Mean # of SABA inhalers dispensed 6.1 (7.4) 3.6 (5.2) 39.4 < 0.001 3.5 (4.2) 3.5 (4.2) 0.0 1.000 Mean # of LABA inhalers dispensed 1.3 (3.1) 0.1 (0.3) 65.2 < 0.001 0.01 (0.16) 0.01 (0.16) 0.0 1.000 Mean # of theophylline prescriptions dispensed 0.2 (1.3) 0.04 (0.49) 14.5 <0.001 0.0014 (0.06) 0.0014 (0.06) 0.0 1.000 Mean # of LTRA prescriptions dispensed 0.2 (1.3) 0.1 (0.8) 12.0 <0.001 0.02 (0.32) 0.02 (0.32) 0.0 1.000 Use of health services for asthma Mean # of ED visits for asthma 0.3 (1.0) 0.2 (0.7) 4.4 0.001 0.06 (0.29) 0.06 (0.29) 0.0 1.000 Mean # of hospital admissions for asthma 0.05 (0.28) 0.03 (0.20) 6.8 <0.001 0.01 (0.08) 0.01 (0.08) 0.0 1.000 Comorbidity Allergy 2.0 2.0 0.2 0.91 0.6 0.6 0.0 1.000 177 Before propensity score matching After propensity score matching Regular ICS use group Intermittent ICS use group Standard difference, % P Value Regular ICS use group Intermittent ICS use group Standard difference, % P Value Allergic rhinitis 12.1 11.1 3.2 0.02 7.1 7.1 0.0 1.000 URTI 28.5 37.1 18.4 P<0.001 28.9 28.9 0.0 1.000 Depression or anxiety 23.8 17.8 14.8 P<0.001 15.8 15.8 0.0 1.000 Acute lower respiratory tract infection 22.8 21.0 4.3 P<0.001 14.9 14.9 0.0 1.000 Emphysema, chronic bronchitis, COPD 7.5 2.5 22.9 P<0.001 0.7 0.7 0.0 1.000 178 Table 5.5 Adjusted hazard ratios for ED and hospital admissions associated with regular (120 days definition) and intermittent use of ICS Regular use of ICS Intermittent use of ICS Propensity Score- Matched Regression Standard Covariate- Adjusted Regression Had outcomes, % Median time to outcome, days Had outcomes, % Median time to outcome, days Adjusted HR (95% CI)* Adjusted HR (95% CI) ED visits One year follow-up 4.9 158 6.5 157 0.82 (0.76 \u00E2\u0080\u0093 0.89) 0.73 (0.63 - 0.85) Hospital admissions One year follow-up 0.6 152 0.7 143 0.78 (0.63 \u00E2\u0080\u0093 0.96) 0.88 (0.56 \u00E2\u0080\u0093 1.34) *Reflects the risk of health services utilization for asthma in the regular ICS use group compared to the intermittent ICS use group. Hazard ratios are adjusted for the propensity score, calculated using age, gender, number of SABA, LABA, LTRA, theophylline prescriptions, number of ED and hospital admissions, patients\u00E2\u0080\u0099 comorbidity status during the index year. 179 CHAPTER 6 : CONCLUDING CHAPTER 180 This thesis identified asthma patients who received suboptimal drug regimens according to clinical practice guidelines and examined the link between regimen optimality and health service use for asthma exacerbations. This work represents an epidemiologic evaluation of the use of asthma medication and health services utilization. This chapter will summarize key results, discuss the strengths and limitations of this research work, suggest directions for future research, and consider potential policy implications. 6.1 KEY FINDINGS 6.1.1 Key findings in the asthma burden estimate project 6.1.1.1 Asthma case definition Chapter 2 summarized case definitions used in previous studies and applied them to the same population (B.C. residents) during the same study period (1996 \u00E2\u0080\u0093 2009). Results showed that asthma prevalence and incidence estimates varied significantly (from 1 - 7%) depending on the case definition used. Evaluations based on the criteria of \u00E2\u0089\u00A51 physician visit, \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A51 asthma drug dispensing resulted in the highest prevalence and incidence estimates (8% and 2%, respectively); while the criteria of \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A52 physician visits produced the lowest estimates (1% and 0.5%, respectively). Population- based burden estimates provide comprehensive information about asthma epidemiology, health service use and asthma drug dispensings, this work will thus be useful for the improvement of health and health resources relocations. Because of the usefulness of disease burden estimates in population health, stakeholders (such as policy-makers, physicians and researchers) must be cautious when interpreting estimates presented in the literature, and should use a case definition that accurately reflects the burden of disease in a large 181 population. This study used the criteria of 2 or more asthma-related physician visits and 3 or more asthma prescription drug dispensings to identify patients. These two criteria were used to account for patients that received a working (rather than final) diagnosis or received medication once or twice for diagnostic purposes. Hallmark asthma symptoms such as wheezing, coughing or chest tightness can arise from other conditions, such as bronchitis, colds, pneumonia or COPD, which may result in repeated clinic visits in order to reach an accurate diagnosis. Asthma could be misdiagnosed as other diseases and vice versa during this process. Results presented in Chapter 2 (page 74) showed that 189,627 patients had a single occurrence of physician-diagnosed asthma between 1996 and 2009; these patients likely do not have asthma and were excluded from this analysis by our case definition. Similarly, 335,799 patients received asthma medication once or twice in a 12-month period but did not receive medications in other years. Since medication is the primary treatment for most asthma patients, those who only received medication once or twice during the 14 study years likely received them for diagnostic purposes or for some acute respiratory diseases (e.g., viral pneumonia) when wheezing and shortness of breath are presented as symptoms. Patients identified by our case definition continuously received treatment and required asthma management. Studies that use single drug dispensings as a criterion[67-69, 261] may overestimate the burden of asthma. Another important consideration when determining a case definition is that databases available in some jurisdictions may only capture information for limited age groups or drug insurance programs. In Ontario, for example, comprehensive drug dispensing data are only available for patients over 65 years[262]; likewise, the National Prescription Drug Utilization 182 Information System Database only captures patients aged 65 years or older within public drug insurance programs in Alberta, Saskachewan, Manitoba, New Brunswick and Nova Scotia[263]. Results presented in Chapter 2 (page 74) show that drug dispensing criteria is important for identifying asthma cases: 105,917 patients were identified based on this criterion alone. Inclusion of drug dispensing data is important because a patient may be visiting the physician for multiple purposes, and, since the Medical Services Plan database only has one diagnostic field, asthma may not be entered as the primary reason for the visit. Thus, it\u00E2\u0080\u0099s critical to include drug dispensings as a criterion in asthma case definitions. Some studies also used asthma-related ED visits as a criterion[70, 72, 264]. Macy et al.[70] and Schatz et al.[72] defined patients as having asthma if they were admitted to hospitals at least once, or received \u00E2\u0089\u00A52 medication dispensations, or \u00E2\u0089\u00A51 ED or outpatient clinic visit. Because in B.C, the majority of ED visits (83%) are contained in the Medical Services Plan physician visit database[239] which has been also used as a platform for the physician visit criterion of our asthma case definition. Thus it\u00E2\u0080\u0099s not necessary to use ED visits as a separate case definition criterion in the present study. In summary, the case definition utilized for this study is superior to previous asthma case definitions in capturing an entire population with treated asthma in B.C. between 1996 and 2009, specifically, those patients who continuously required medications and physician services to manage asthma. Importantly, this case definition excludes patients who received a working diagnosis or medication for diagnostic purposes (e.g., those with only a single set of prescriptions for the two primary drugs used to treat asthma), as well as some patients with very mild asthma (i.e., those who only require SABA 3 doses or less per week, and do not need daily ICS use to maintain their asthma control according to asthma clinical practice 183 guidelines). This group of patients is not the target of the present work since they have few asthma-related health services. 6.1.1.2 Estimates of asthma burden in B.C. 1996 - 2009 Asthma prevalence and incidence estimates were stable in B.C. over the 14 study years. This trend is consistent with other\u00E2\u0080\u0099s findings, but the prevalence and incidence estimates (2% and 0.7% respectively) were significantly lower compared those reported in similar studies based on population-based databases over comparable time periods[262, 265]. This may be due to the case definitions used in this analysis. Using \u00E2\u0089\u00A52 asthma prescription drug dispensings as a criterion resulted in 50% lower asthma prevalence and incidence estimates, and using \u00E2\u0089\u00A53 dispensings resulted in an additional 30% reduction when applied to the cohort who had \u00E2\u0089\u00A51 physician visit for asthma or \u00E2\u0089\u00A51 asthma prescription drug dispensings in each 12-month period between 1996 and 2009. A key finding of the asthma burden estimate study is that asthma-related health service use (FP, specialists, ED and hospitals) all declined significantly from 1996 to 2009 (Chapter 2, Page 78); this may reflect an overall improvement in asthma management driven by a series of Canada- and provincial-wide education programmes during 1990s and 2000s. Between 1997 and 2000 (the latest evidence available), 74 asthma education programs were identified in Canada (outside Quebec), 8 of these occurred in B.C[266]. Such programs have shown success in increasing public awareness, improving understanding of asthma and helping patients avoid triggers. Patients involved in education programmes had significantly reduced unscheduled physician visits, ED visits or hospital admissions[267-269]. For example, Hopman and colleagues evaluated the effects of a Canada-wide multicenter program for adult patients on health service utilization, absenteeism from school or work and 184 quality of life[269]. In this intervention, educators provided information on trigger avoidance, environmental control, medication role, action plans and self-monitoring skills. Patients were followed for up to six months. At 6 months, unscheduled physician visits were reduced by 80%, hospital admissions and ED visits were reduced by 70%. In a more recent study, McGhan and colleagues developed a \u00E2\u0080\u009CRoaring Adventures of Puff (RAP)\u00E2\u0080\u009D, a school-based asthma education program for children that incorporated multiple factors including physical, mental, personal and social development[267]. In total, 34 schools were randomly assigned to receive either RAP intervention or usual care. The intervention group showed significant improvement in the frequency of ICS use, the avoidance of triggers, limitation of activity, understanding of asthma, and overall quality of life compared to usual care. Other reasons contributing to the decline in health service use for asthma may be related to improved asthma drug formulation and delivery techniques, changes in the Canadian healthcare system and availability of other health resources. During the study period (1996 \u00E2\u0080\u0093 2009), hydrofluoroalkane (HFA) formulations replaced chlorofluorocarbon (CFC) formulations. CFC formulated drugs contain large particles (for example, the average particle size of CFC-beclomethasone (BDP) was 3.5-4.0 microns), which cannot be effectively delivered to the lung. HFA formulation produces particles of about 1.1 microns (for HFA-beclomethasone)[270, 271], because of this smaller size and resulting improved delivery, lower doses of HFA formulated inhalers were required to achieve efficacy equivalent to that of CFC formulated inhalers[3] Another improvement involved the use of aerosol holding chambers to deliver drugs. Before the 1990s, inhaled asthma drugs were delivered by the pressurized metered dose inhaler (pMDI), which required patient skills and coordination. Poor technique may result in 185 low (~20%) lung deposition [272], and large drug amounts are delivered to the oropharynx, which increases risk of oral candidiasis, dysphonia and potential systemic adverse effects caused by gastrointestinal absorption. Spacers or valved holding chambers attached to the pMDI mouthpiece have been recommended in the international asthma practice guidelines since mid-1990s[3]; these make inhalers simpler to operate, and as a result show higher lung deposition (i.e., > 50% ), and decreased oropharyngeal deposition (from 28% to 4%), compared to a pMDI[271, 272]. New formulations and improved delivery techniques have demonstrated lower risk of local ADRs caused by ICS and more effective drug delivery to patients of all ages. ICS-related ADR is a barrier to adherence (details see Chapter I Section 1.7.1); thus, reducing the risk of ICS-related ADR may increase ICS use and improve asthma control. The decreasing trend in hospital admissions may also reflect changes in the Canadian health care system; specifically, care has been relocated from expensive inpatient care to home and outpatient care[273]. As a result, the number of hospital beds in Canada decreased from 140,000 in 1996 to 120,000 in 2002 (the most recent evidence available)[273]. This reduction was balanced by an expansion of outpatient and home care services. The number of FPs in outpatient clinics increased by 13% (4,186 in 1997 and 4,731 in 2006), while the number of specialists increased by 14% (3,431 in 1997 and 3,904 in 2006)[274]. The availability of after-hour (e.g., 5pm \u00E2\u0080\u0093 11pm), weekend, and holiday walk-in clinics[275, 276] may potentially use of ED services for asthma. In Canada, Jones et al. evaluated the association between implementation of an outpatient after-hour clinic and number of ED visits over a 3-year period. They reported a reduction of 37 urgent and 49 semi-urgent ED visits per month during the after-hour clinic operating hours. A U.S study 186 shows that 1.2 million people in Colorado had \u00E2\u0089\u00A51 ED visit during a 12-month period. 75% of these patients reported the needs for after-hour care, while 57% reported inability to get an appointment with their FPs[277]. The availability of after-hour clinics appears to contribute to reduced ED service use for asthma, as they provide places for patients/parents to seek emergency medical aid and access information. Notably, while the decreasing trend of ED visits and hospital admissions is consistent with those reported in comparable studies[262, 265], the magnitude of the decrease (60-70%) is larger in this study. For example, Gershon et al. reported a decrease of 20% in ED service use and 25% in hospital service use for asthma in Ontario between 1996 and 2006[262]. This is likely due to the use of different case definitions. In contrast to our study, Gershon et al. defined asthma patients using the criteria of \u00E2\u0089\u00A51 hospital admissions or \u00E2\u0089\u00A52 physician visits for asthma and calculated patients cumulatively in each year. Since the number of asthma patients increased each year from 1996 to 2006, the magnitude of the decline in ED and hospital service use reported by Gershon et al. was smaller than that we report here. Compared with other age groups, asthma patients 5-11 years of age had the highest prevalence and incidence, use of physician, ED and hospital services, indicating that young children placed the greatest burden on the healthcare system. This may be due to biological characteristics: for example, since children have smaller airways[278], microvascular leakage and mucus plugs can cause blockages that and lead to exacerbations. Another reason for highest health service use in young children could be that parents are often eager to seek medical help when managing a child\u00E2\u0080\u0099s exacerbations[156], either because they do not have enough knowledge or lack the confidence to make decisions about symptom management. Home management plans have been shown to be effective in reducing unnecessary clinic 187 visits and ED service use[164, 168]. Gibson et al. reviewed 36 trials and examined the effects of asthma self-management programmes on healthcare outcomes in adults[164]. They reported that self-management plans, coupled with regular health practitioner review, reduced hospitalizations by 36%, ED visits by 18% and unscheduled doctor visits by 32%, compared to regular care. Self-management also decreased days off work or school due to asthma by 30%, reduced nocturnal symptoms by 33%, and significantly improved quality of life score. A recent Cochrane review by Bhogal et al. evaluated the effects of self-management in 355 children and adolescents using results from four trials[279]. They reported a 30% reduction in the use of ED services in patients with symptom-based self-management plans. Plans that are easy to read and clearly describe the onset of exacerbations, early interventions, self- monitoring strategies, and define a threshold for presentation are likely to improve asthma management and reduce the need for urgent care[280]. Trend of asthma prevalence was stable during the study period for all age groups except patients 35-55 years, in which, asthma prevalence increased from 2.0% in 1996 to 3.4% in 2009, representing an increase of 40%. A number of reasons could contribute to this increasing trend. This trend could be affected by the method used in the present study to identify the treated asthma cohort. Drug dispensing criterion is the most important component of the asthma case definition used in the current studies, particularly in patients 35-55 years of age. More specifically, a significantly higher percentages of patients aged 35- 55 years entered the cohort based on the asthma drug dispensing criterion (83% in 1996 and 87% in 2009) compared to the percentages in children 5-11 years of age (57%), patients aged 12-18 years (67%), and patients aged 19-34 years (78%). The increasing trend of patients entering the cohort on the basis of regular asthma drug use may result in the increase in 188 asthma prevalence in patients 35-55 years of age observed in the present study. Progressive decline in lung function may occur more significantly in this age group with asthma over the 13-year study period, which has been associated with frequent exacerbations over time, or synergistic effect of other factors, such as smoking[281] [215, 282]. Bai and colleagues in British Columbia, Canada examined the effects of frequency of asthma exacerbations on annual lung function decline in a cohort of 93 patients (initially aged 25 years) who never smoked[282]. Patients were followed for a median time period of 11 years and were divided into two groups based on the median exacerbation rate (0.1/year). Patients with an exacerbation rate above the median were defined as having frequent exacerbations, while patients with an exacerbation rate below the median were defined as having infrequent exacerbations. The Bai et al. study reported that patients with more frequent exacerbations had an annual decline of 32ml/year in FEV1, which was significantly higher compared to the annual decline in patients with infrequent exacerbations (15 ml/year). Another prospective longitudinal study conducted by Apostol et al. involved 4,000 patients initially aged 18-30 years and a follow-up period over 10 years[215]. The authors found an 8.5% decline in FEV1 in individuals who neither had asthma nor smoked, a 10% decline in nonsmokers with asthma, and 18% decline in those with both asthma and current smoking of 15 or more cigarettes per day. Patients with more significant declining in lung function may require more frequent use of bronchodilators, higher dose of ICS and more dispensings of the third- or fourth-line asthma drug therapy. This increasing trend of asthma prevalence in adults 35-55 years of age is also consistent with the hypothesis in asthma etiology research that more people at the working age are exposed to industrial environment such as chemicals, metal and smoke 189 exposures[283]. These exposures have been associated with increased respiratory sensitization and further development of asthma[3]. Additionally, during the past decade, there has been an increased recognition of adult-onset asthma, which may also affect the trend of asthma prevalence in patients 35-55 years of age. Chapter 2 (page 90) also shows that asthma prevalence and incidence did not vary significantly by health service delivery areas (HSDA); however, significant variations were observed in the use of specialist, ED and hospital services for asthma. For example, patients in large cities were 3 times more likely to use specialist services compared to those in remote areas; in contrast, city patients were 20-30% less likely to use ED or hospital services than those in remote areas. These differences confirm previous studies showing regional variations in asthma-related health service use[284], which likely result from differences in availability and access[284, 285]. Specialists are primarily located in large cities; their absence in remote areas is likely the main reason for low use. The distribution of ED and hospital service use (lower in large cities and higher in rural areas) is likely to be due to the abundance of FPs, walk-in clinics and large ambulatory clinics in cities, which reduced the need for ED or hospital service use. In contrast, due to a limited numbers of FPs and specialists in remote areas, patients may need to seek medical assistance in the ED or hospital [285], thus may be more likely to use hospital or ED services. 6.1.2 Key findings in regimen optimality and health services utilization project Since drug therapy is the primary intervention for asthma management, it is critical to identify patients with suboptimal regimens and high health service use in order to improve patient outcomes and decrease the burden on the healthcare system. The data presented in 190 Chapter 3 demonstrates how suboptimal and optimal drug regimen users interact with the healthcare system. 80% of patients who were dispensed large quantities of SABA (i.e., annual quantities of >8 inhalers in patients aged \u00E2\u0089\u00A512 years, or >4 inhalers in patients <12 years) did not receive sufficient ICS according to internationally accepted clinical practice guidelines. Compared to patients with optimal regimens, those with high SABA use but insufficient ICS dispensings were three times more likely to be admitted to hospital for exacerbations, and four times more likely to be high users of FP services. Since, insufficient ICS is likely a factor for these poor outcomes, these patients would likely benefit from increased ICS use. This change alone \u00E2\u0080\u0093 made in accordance with clinical practice guidelines - would likely result in fewer hospital admissions and FP visits for asthma. Patients with suboptimal drug regimens were 1.5 times more likely to use ED services for asthma compared to those with optimal regimens. The absolute difference in the proportion of patients who made asthma-related ED visits between regimen optimality groups was 3%, indicating that a number needed to be treated optimally to prevent an asthma-related ED visit in a one-year period is 33. In 2009, 9,461 B.C. patients received suboptimal regimens; based on the study described in Chapter 3, 300 asthma-related ED visits could have been prevented. A previous study reported a unit cost of $324 for each ED visit[286]; thus, $92,889 in asthma-related healthcare costs could have been saved. 6.1.2.1 Impact of suboptimal regimens by age group This study is the first to report the impact of suboptimal drug regimens on healthcare service use in adolescent asthma patients. We found that adolescent patients with suboptimal regimens were more likely to use health services compared with other age groups. This 191 finding is contrary to previous studies that suggested reduced asthma symptoms or remission (rates ranged between 20%-60%) in adolescents compared with children [287-290]. Several factors likely contribute to high service use in adolescents with suboptimal regimens. For example, adolescence is a period when individuals seek independence; as such, adolescents may be less likely to share their health experiences with parents or physicians, which may lead to suboptimal therapy. The variable nature of asthma may contribute to denial of symptoms or the need for controller medications[291]. Accordingly, studies have found that < 50% of adolescent patients were adherent to asthma drug therapies [292, 293]. Unfortunately, current asthma educational programs are focused more on young children and parents, rather than adolescents[294, 295]; indeed, up to 80% of adolescents are not receiving sufficient information on asthma management[296]. Frequent asthma exacerbations and chronic inflammation are associated with airway remodeling, narrowed airways, and decreased lung function [297, 298]. Adolescence is the final stage of lung growth; thus, poorly controlled asthma during this period can lead to reduced lung capacity. Reduced capacity has been associated with greater declines in lung function[215, 216] and more frequent irreversible airway obstructions in adulthood[217]. In order for health planners to apply target intervention strategies, it is critical to identify the factors associated with suboptimal regimen use and use of health services in adolescent patients. The high impact of suboptimal regimens on healthcare systems continued into the 19- 34 age group; patients with suboptimal regimens were 3 times more likely to be hospitalized and 5 times more likely to be high users of FP services compared to those with optimal regimens. 192 This study is the first to measure asthma regimen optimality according to the latest international clinical practice guidelines and to link this factor with health service use. These results have increased generalizability compared with other studies because they are derived from a provincial population of 4.5 million. Anis et al. described physician service use and hospital admissions in patients who were dispensed excessive SABA (i.e., \u00E2\u0089\u00A59 canisters of salbutamol equivalents per year) without sufficient ICS (i.e., < 100 \u00C2\u00B5g/puff beclomethasone equivalents) in the same region 17 years ago[299]. In that study, 3,069 patients filled SABA prescriptions excessively, and of this group, 55% did not use sufficient ICS. This proportion is almost double that of the present study, in which 9,326 patients were dispensed SABA frequently (i.e., >8 inhalers in patients aged \u00E2\u0089\u00A512 years or >4 inhalers in patients <12 years per year), and 31% did not receive sufficient ICS. This discrepancy is likely due to differences in how the data were used; for example, the study population used by Anis et al. only included patients who received social assistance or had annual drug expenditures of more than $600 in 1995[299]. In addition, it was unknown if patients had asthma, or rather suffered from COPD or other respiratory conditions warranting the use of these drugs. Patients who received suboptimal asthma drug regimens might have different characteristics. Patients who received suboptimal asthma drug regimens could be those who experienced more severe asthma that required frequent SABA use to treat their symptoms, but did not use sufficient ICS. This notion is supported by previous studies that reported up to 50% of patients with inadequate asthma control did not receive sufficient amounts of controller medications[300, 301]. A small number of patients (n=500) with suboptimal regimens did not require frequent SABA use but did receive high average daily doses of ICS (i.e., >1000 mcg/day for patients 12 years or older and >500 mcg/day for patients 5-11 years 193 of age). These patients may not have frequent asthma exacerbations that required SABA use, but received high doses of ICS. Due to their high ICS use, these patients have higher risk of experiencing ICS-related ADRs and may be receiving more ICS than is needed to control asthma symptoms based on their SABA use. Among these suboptimal use patients, only 35% were dispensed asthma add-on (third line) therapies (e.g., LABA, LTRA), which are guideline-supported options for patients instead of high-dose ICS. As described previously, majority of patients who received suboptimal regimens (85%) were those who received SABA but did not use sufficient ICS. Some of these patients could also be those with infrequent exacerbations but administer SABA more frequently than necessary due to poor inhaler technique. These patients may have poor inhaler technique for both SABA and ICS, however, SABA onset of action is within 5 minutes for the majority of patients. Those who have poor inhaler technique may require additional doses of medication to achieve the desired effect. On the contrary, ICS reduce airway inflammation within 10-15 days, which may not allow patients to realize the underdosing that poor technique results from. Thus, reflected in our data, these patients were frequently dispensed SABA but did not receive enough ICS. Poor inhaler technique has been associated with significantly reduced deposition of medications in the lung, loss of bronchodilator effect and increased SABA use [302, 303]. Several errors have been reported in the use of inhalers, for example, patients may have difficulty in coordinating inhaler activation with inhalation, may fail to hold their breath long enough time, or may apply multiple actuations without waiting or shaking the unit between doses[302, 304]. Checklists of assessing inhaler technique have been developed and summarized in the NHLBI asthma clinical practice guidelines[3]. A number of studies have demonstrated significantly improved asthma control and health services utilization 194 outcomes after assessment of inhaler technique by physicians, nurses and pharmacists[305]. Findings from the present study emphasized the importance of regularly assessment of patients\u00E2\u0080\u0099 asthma control and inhaler technique in accordance to the asthma clinical practice guidelines. Impact of LABA use on regimen optimality and health services utilization LABA is recommended for asthma patients whose symptoms cannot be controlled by low or moderate dose ICS according to international clinical practice guidelines 18 . However, in our analysis, LABA-adjusted ORs were not significantly different from the unadjusted ORs (Chapter 3, page 114). This may be due to the fact that LABA therapy is often prescribed and dispensed to patients with ICS either as two separate products or a combined inhaler. The use of combination inhalers as first-line therapy for asthma is known to occur, despite practice guideline recommendations 43 . This use may blind the effect of LABA on asthma-related use of ED, physician and hospital services as examined in this study. 6.1.2.2 Changes in regimen optimality over time The efficacy of ICSs in reducing asthma exacerbations is well established, and although significant non-adherence rates have been reported, the effects of regimen optimality changes on health service use have not been described in large populations over extended periods. The results presented in Chapter 3 examine asthma drug regimen optimality over one year; however, since patients frequently modify their drug regimens to accommodate symptomatic changes, monitoring drug use over a longer period of time would improve the assessment of optimal versus suboptimal medication use. To address this, a subsequent study was conducted to examine the association between optimality changes and health service use in 130,057 asthma patients between 1996 and 2009 (Chapter IV). The key 195 finding is that changing from suboptimal to optimal regimens led to 50% reduction in the use of ED and 30% reduction in the use of hospital services for asthma exacerbations. Additionally, we found that one additional optimal regimen year further reduced ED service use (by 30%) and hospital service use (by 10%). These results highlight the importance of optimal regimens in asthma management and suggest that increasing ICS adherence is likely to reduce health service utilization outcomes in patients with asthma. These results are consistent with a recent study by Williams et al., which showed that ICS adherence is associated with reduced asthma exacerbations[306]; however, it focused on 298 patients during a 2-year follow-up, that study has limited generalizability. During the primary analysis, repeated ED and hospital visits were used to reflect the recurrence of asthma symptoms over time. Since health service use in previous year can impact drug and health service use in the subsequent year [157], a sensitivity analysis was performed to avoid interference from repeated service use in previous years. Results of this analysis, which used the first ED visit and hospital admission during follow up as study outcomes, showed that switching from suboptimal to optimal regimens was associated with 70% reduction in ED visits and 50% reduction in hospital admissions (Chapter IV, page 124). The protective effects of switching asthma drug regimen optimality appear to be larger in the sensitivity analysis than in the primary analysis (HR 50% for ED visits and 30% for hospital admissions) (Chapter IV, page 124). The main reason could be over a long time period (up to 14 years), a number of patient characteristics (e.g., ability to using inhalers effectively, medication use and health service use) changed. Changes of these characteristics may have affected patients\u00E2\u0080\u0099 use of medications and health services utilization over time, although these changes were not reflected in our primary analysis due to a lack of patient level information 196 in the data. Nonetheless, findings from the primary and sensitivity analysis clearly demonstrate that switching from suboptimal to optimal regimens lead to a significant reduction in asthma-related ED and hospital service use. 6.1.3 Key findings in the comparison between regular and intermittent use of ICS Results presented in Chapter V demonstrate that regular (compared with intermittent) ICS use is associated with a 20% reduction in the use of ED services and a 30% reduction in the use of hospital services. The latter result is consistent with findings of Suissa et al.[251], who examined the effects of regular versus irregular ICS use on asthma-related hospital service use in patients who received \u00E2\u0089\u00A53 asthma medications, and found that regular use was associated with a 31% decrease in the rate of hospital admissions. We expanded the study population by using 2 additional case definition criteria (i.e., \u00E2\u0089\u00A51 hospital admission or \u00E2\u0089\u00A52 physician visits), which accounted for 40% of the patients in our cohort. This large population increases the generalizability of our findings to patients outside the region of study. ICS is widely viewed as the most important asthma drug since it is by far the most effective and safest controller medication. Daily use of ICS is recommended to patients who have persistent asthma (i.e., require >3 doses of SABA each week) according clinical practice guidelines[3]. However, some clinicians recommend patients take ICS only intermittently during flu season or in response to symptoms[307], due to concern about potential ADRs associated with regular ICS use, such as osteoporosis, cataracts and glaucoma. A few trials have examined the efficacy of intermittent ICS use, but show conflicting results, and there are ongoing debates about whether patients with mild persistent asthma should use ICS 197 regularly[308-310]. This analysis involved patients with persistent asthma and variable disease severity emphasized the benefits of using ICS regularly in asthma control. In this study, we found that 60% of patients aged 5-18 years received ICS intermittently during the index year compared with 30% of adult patients. This large difference could because physicians may not prescribe ICS to children regularly or parents do not prefer children to take ICS regularly due to potential risk of ADRs associated with ICS. 6.2 STUDY STRENGTHS This research was performed on the basis of B.C. health services utilization and prescription drug dispensing databases. These databases provided a large sample size (n = 325, 073) of patients with asthma, which minimized selection bias and increased the external validity of our findings. Other advantages of using population-based data for health research are discussed elsewhere[311]. Briefly, since data are collected in a systematic manner, studies can be performed faster and at a lower cost compared with those based on censuses or survey data. Further, in this study, comprehensive information including health service utilization and prescription dispensing records is available over a 14-year period, which allows us to track patients\u00E2\u0080\u0099 histories over time and permits longitudinal studies (time-to-event analyses that avoid immortal time biases)[311]. The electronic PharmaNet database captures drug dispensings from pharmacies and is a more accurate data source than self-reported drug use, particularly for patients who suffer from multiple conditions, since it prevents recall bias. Database records, linked by unique patient identifiers, provide information on patient demographics, location, hospital admissions, outpatient physician visits, socio-economic status and comorbidities. Another strength of this study is the development of a unique case definition that 198 maximally captures asthma patients. This definition excludes patients with mild asthma and those with working (rather than absolute) asthma diagnoses, and therefore allowed us to focus on a study cohort with persistent asthma. The use of \u00E2\u0089\u00A53 drug dispensing records as a criterion allowed us to capture 100,000 patients that may have been missed in previous studies. Since we followed patients for up to 14 years, the use of an extended Cox model with time-varying covariates representing changes in disease severity and comorbidities is also strength of this study. Although the latest guidelines emphasize that asthma severity changes frequently, studies often assume it is constant. For example, Kozyrskyj et al. developed an asthma severity measure in children using health services utilization data. Children were categorized as having mild to moderate, moderate-severe, and severe asthma based on frequency of exacerbations (indicated by hospitalizations) and medication use over 15 months; however, the study was limited by the assumption that disease severity did not change over this time. Indeed, the dynamic nature of asthma symptomatology is rarely taken into account, even though it can confound results. Schatz and colleagues examined the relationship between controller/total medication ratios and the likelihood of ED visits for asthma in the subsequent year[72]. They measured asthma severity by assessing ED visit and hospital admission, as well as use of SABA and oral corticosteroids over a one-year period, and then measured the study outcome in the subsequent year. Since disease severity may change over time, this approach may still confound their findings. To date, the severity and variability of asthma have not been adequately addressed and incorporated into epidemiological studies. In order to take into the dynamic nature of asthma symptoms over a 14-year period, we used an annually-updated disease severity measure (Chapter IV) that 199 varies with dispensings of add-on drug therapies and comorbid depression, anxiety, or COPD. 6.3 STUDY LIMITATIONS 6.3.1 Selection biases Since the B.C. health services databases capture the entire provincial population, this study is not subject to selection bias associated with patient recruitment. In this study, patients are censored when records show a 56 th birth date, March 31 st , 2010 (the last date available in our data), or death - whichever comes first. For the censorship of death, bias could be incurred if censorship occurs differentially between two comparison groups (i.e., suboptimal vs. optimal regimen users or regular vs. intermittent ICS users). However, since the mortality rate of asthma is low (~21 per 100,000 patients), the small magnitude of bias resulting from censorship due to death will not have a significant effect on the study\u00E2\u0080\u0099s estimates or conclusions. 6.3.2 Information biases 6.3.2.1 Diagnosis misclassification Cohort definition and study outcomes in this study were based on asthma diagnosis, which is coded in the databases used. As discussed above, misdiagnosis and miscoding can occur; however, these errors are most often random or systematic, and not likely lead to significant bias. Nonetheless, we have accounted for this possibility by requiring at least two physician visits with a primary diagnostic code of asthma. The ideal approach to testing the validity of our case definition would be to conduct surveys in a random patient population and check the consistency of patients identified using the case definition in the population-based data and those who reported having an asthma 200 diagnosis in the surveys. The time consuming nature of this approach prevented it from being done during the time allotted for this study; however, the case definition validity was tested in an earlier study using latent class modeling[231]. Latent class modeling is a clustering methodology for classifying patients into groups using scores calculated on the basis of a set of variables and well suited for diseases that have diagnostic challenges, such as asthma, which may need repeated visits, repeated spirometry tests and physical examinations to establish diagnosis. In a study by Prosser et al.[231], patients identified using a latent class model was treated as the gold standard to assess the sensitivity and specificity of the case definition used here. Seven different variables were used to calculate the scores and probabilities of clustering patient groups, including number of FP visits with ICD-9 code 493, number of specialist visits with coded 493, number of hospital admissions with ICD-9 as the primary diagnostic code, number of short-acting bronchodilator prescription fills, number of ICS fills, and number of other asthma drug fills in the study year. The Prosser et al. study produced a sensitivity of 70% and a specificity of 99%. A sensitivity of 70% suggests that the probability of a patient being identified as having asthma by our case definition is 70%; in other words, according to the gold standard, our case definition has a 30% chance of missing a patient. However, since it was formed on the basis of pure statistical calculation without clinical information, it is possible the gold standard includes patients with mild asthma or a working asthma diagnosis; such patients do not often use health services for asthma or experience significantly reduced quality of life. As previously discussed, we considered patients with a single diagnosis or 1-2 drug prescriptions as unlikely to have asthma and excluded them from this study. Likewise, patients with mild asthma (i.e., adequately controlled by 1 or 2 short-acting bronchodilators) are not the focus of this research, as they 201 are not likely to experience significantly reduced quality of life or place a substantial burden on healthcare systems. Since this work is focused on patients with moderate to severe asthma that are likely to be captured by our case definition, the extent of asthma case misclassification is minimized. 6.3.2.2 Misclassification of drug use Another potential limitation of this study is using drug dispensing records as a surrogate for medication use. Actual drug use is difficult to infer from electric drug dispensing data, since consumption is widely variable. For example, patients may not take medications at all, take them intermittently, or take doses different from those prescribed. Actual consumption may not always be accurately reflected by drug dispensing data, which can result in patient misclassification (e.g., intermittent ICS users being misclassified as regular ICS users). In this study, however, drug dispensing is tracked at the patient level and we monitor use via drug refill patterns (a drug that is continuously refilled is likely utilized). Coupled with our very large sample size, this suggests that limitations arising from differences between drug dispensing and actual use are not likely to be significant. Nonetheless, we used sensitivity analysis to account for the possibility that patients change their usage behavior over time. In the primary analysis, regular ICS use were defined by a 90-day interval between refills (since B.C. has a 100-day reimbursement policy, patients were expected to refill medications every 90 days). Patients with regular ICS use were 20% less likely to visit ED and 30% less likely to be admitted to the hospitals for asthma. When we performed a similar analysis using a definition based on 120-day intervals, findings were consistent with those of the primary analysis, but the effects on hospital admissions was smaller, (regular ICS use was associated with a 20% reduction in hospital admissions for 202 asthma). Regular use of ICS was defined as having \u00E2\u0089\u00A53 or more ICS inhalers in the index year and at least one ICS inhaler within each 120-day period (Table 5.2). Patients who missed doses may have been misclassified as regular users when using the 120-day interval; this misclassification would have blinded the effects of regular ICS use on health services utilization outcomes. Nonetheless, the results of these studies emphasize the importance of regular use of a controller medication in patients with persistent asthma. In addition, the databases may contain errors in ICS and SABA quantities and doses, and these may influence drug exposure definitions. We have taken measures to correct the majority of these errors based on quantities listed on the drug package (details see Chapter 2 Section 2.2.4); any bias induced by remaining errors is minimal. 6.3.2.3 Underestimate of ED visits ED visits in this analysis were identified from the MSP fee-for-service physician visit database. As physicians at some EDs in Vancouver city (e.g., B.C. Children\u00E2\u0080\u0099s Hospital, Vancouver General Hospital or St. Paul Hospital) are on salary-based or sessional payment, visits to these physicians at the ED are not captured in the MSP database. However, according to the validation study conducted by the Centre of Health and Services Policy Research, the MSP database captures approximately 80% of ED visits in the entire province[239]. As the majority of ED visits was captured in our analysis, the effects of underestimation were likely to be small. 6.3.3 Confounding The data used for this study does not include measures of ability to effectively use inhaler devices, perspective on using medications or health services, or parent preference in 203 using medications and/or health services to manage a child\u00E2\u0080\u0099s disease, all of which could be associated with suboptimal regimen use and the frequency of health service use. However, patients who use inhalers ineffectively, or do not use the appropriate amount of ICS are probably those who received SABA frequently without sufficient ICS, and were captured in our suboptimal drug regimen classification. Thus, the unavailability of these factors is not likely to have a significant effect on our findings. Another important confounder in this study is asthma severity. In previous studies, SABA dispensings, add-on drug therapies, and use of hospital and ED services for exacerbations have been used to control for confounding by severity. In the present study, SABA use with or without ICS was used to define drug regimens as optimal or suboptimal; neither could be used for confounding adjustment. Other severity indicators derived from clinical information (e.g. frequency of nocturnal awakenings due to symptoms, or spirometry test results) were not available in the databases used. There is potential residual confounding by disease severity in this study. We used novel methods including propensity score matching and time-dependent measurements of severity indicators to maximize our ability account for the possibility of confounding by disease severity and to ensure our two comparison groups had the same comparability in terms of defined severity indicators. 6.4 FURTHER RESEARCH The results from the present study indicated that 15% of patients with treated asthma were still using suboptimal drug regimens, despite wide dissemination of a number of asthma clinical practice guidelines. As discussed above, patients using suboptimal regimens may be those with more severe asthma who required frequent administration of SABA but did not receive sufficient ICS either because this was not prescribed, was not prescribed in sufficient 204 doses, the patient was not adherent to the regimen or did not get sufficient drug effect due to improper inhaler technique. It is also possible that suboptimal regimen users did not receive add-on (third line) therapy (e.g., LABA, LTRA). The amount of clinical information available for this study was limited to physician visit, hospitalization, drug dispensing and mortality data. This may have prevented identification of the specific reasons underlying suboptimal asthma drug regimens in individual cases. Future longitudinal cohort studies linking patient clinical information, health services and prescription drug dispensing data will be helpful in pinpointing the specific reasons for suboptimal drug use and in the development of individualized interventions. Additional clinical data that would be helpful includes assessment of inhaler technique, spirometry test results, frequency of awakenings at night due to asthma and other asthma symptoms. With available health administrative data linked to each patient\u00E2\u0080\u0099s clinical data, long-term asthma drug use and health services could be tracked. Identification of the specific reasons underlying suboptimal regimen use will permit individualized educational programs aimed at optimizing patients\u00E2\u0080\u0099 asthma drug use, improving patients\u00E2\u0080\u0099 asthma control and quality of life. Before-and-after comparisons designed to assess the efficacy of such programs could be developed using asthma-related health services utilization as the study outcomes LABA/ICS combined inhalers were dispensed to one third of patients with treated asthma in B.C (Chapter 2). When examining the association between asthma drug regimen optimality and health services utilization, the LABA-adjusted ORs were not significantly different from the unadjusted ORs (Chapter 3). As discussed previously, it\u00E2\u0080\u0099s likely the combined inhalers were dispensed to some patients as the first-line therapy for asthma, despite that LABA is recommended for patients whose symptoms cannot be well controlled 205 by low dose ICS according to international asthma clinical practice guidelines 18 . This use may blind the effect of LABA on health services utilization outcomes. The available study that examined the extent of using LABA as the first-line therapy for asthma focused on a sample of 1,283 patients in the US[312]. Although health services use records were linked to their data, the time-dependent nature of using asthma medications and health services for exacerbations was not taken into account in that study. Further study is needed to examine the appropriateness and cost-effectiveness of using the combined inhalers as the first-line asthma therapy via overcoming methodological limitations in previous study. We found that switching from suboptimal to optimal regimens was associated with a significant reduction in ED use and hospital admissions over time. In our analysis, add-on drug dispensings and comorbidities were measured using an annually-updated time-variable approach to account for changes in asthma severity and patients\u00E2\u0080\u0099 health status over time. However, our database does not contain information on, for example, patients\u00E2\u0080\u0099 perspective on medication use, ability to effectively using inhalers, and level of asthma knowledge; these factors may have a significant impact on patients\u00E2\u0080\u0099 use of medication and health services. Further work is needed to understand how these factors contribute to asthma management over time, and whether they are associated with a switch from suboptimal to optimal drug regimens. The ICS use comparison study was limited to a one-year study period. Since patients often switch from regular to intermittent ICS use once they become familiar with their symptoms, one year may not be sufficient to associate changes in ICS use with health services use. Longitudinal studies would be helpful in addressing the impact of switching from regular to intermittent ICS use on asthma-related health services use. 206 6.5 POLICY IMPLICATIONS This thesis work describes the burden of asthma in patients aged 5-55 years in B.C., provides an evidence-based classification of optimal and suboptimal medication use, and examines the association between suboptimal drug use and healthcare service use over time. Importantly, this study provides an ideal asthma case definition developed using health services utilization and prescription drug dispensing databases. Our drug regimen optimality classification is important to stakeholders involved in guideline development and healthcare service planning. It is a clinically relevant and visual method to help physicians identify patients who are using their asthma medications suboptimally. The association between asthma drug regimen and health services utilization should be reported to both physicians and patients. Two specific examples are the poor health services outcomes associated with suboptimal drug use and the poor outcomes associated with intermittent ICS use. A better understanding of the healthcare utilization patterns of patients who follow and do not follow clinical practice guidelines for asthma will allow for more pragmatic feedback to physicians and patients about the consequences of suboptimal drug use. These key findings could be delivered to physicians through educational sessions (e.g. workshops) and material dissemination (e.g., printed pamphlets, CD-ROM and internet). Lougheed and colleagues developed a medical education program specifically on asthma management to 2,783 primary care providers (including 1,313 outpatient physicians and 1,470 allied health professionals)[313]. A total of 137 workshops were provided between September 2002 and March 2005. Each workshop contained two 30-min presentations, and 207 each was followed by 1-hour of interactive small group case discussions led by trained facilitators. The case discussion involved a variety of information including asthma diagnosis, examples of spirometry testing, differential diagnoses, triggers, pharmacological management, and development of a written action plan. At the end, participants received a package including a copy of the latest Canadian Asthma Guideline, a flow chart to reflect the guideline key points, action plans available on the guideline website, and a pocket card showing the asthma control measures and ICS dose equivalency table. An evaluation was completed by 335 physicians three-month after the workshop. 67% of these physicians reported that they didn\u00E2\u0080\u0099t provide a written action plan before the workshop but had incorporated one into their practice. 44% reported that their patients had noticed changes in their practices, including more time spent on education, prescription changes, increased use of objective measures and increased patient asthma control. Forsetlund et al. conducted a randomized controlled trial to evaluate whether an educational program increased the use of health research evidence in outpatient physicians\u00E2\u0080\u0099 practice[314]. All outpatient physicians working in municipalities of Norway with more than 3,000 residents were invited to participate in the study. A total of 148 physicians consented, 73 were randomized to the intervention group and 75 to the control group. The intervention included 11 courses on critical review of research evidence and where to locate evidence; goal setting on what they would change in their practice; access to web-based information services including a question and answer service; discussion list and ongoing support services; and 3 newsletters over a 1.5-year period. In contrast, the control group only received free access to library services for one year. As in Norway there are no free but comprehensive library services for practitioners to seek literature and research evidence, this 208 service was regarded an important way in encouraging physicians to improve knowledge. Assessments were conducted before and immediately after the interventions. This before- and-after comparison showed significant difference in knowledge scores about both evidence-based medicine-resources and importance for critical review was statistically different between the two groups after interventions. The negative consequences of suboptimal drug use (e.g., a need for more urgent medical care) should also be delivered to patients and in particular by their prescribing physicians. Physician-patient dialogue regarding the benefits of self-monitoring, drug therapy adjustments and regular physician assessment is important for effective asthma management. This work indicates that adolescents are particularly vulnerable to suboptimal therapy and highlights the need for better communication between patients, parents and physicians. Current asthma clinical practice guidelines suggest that physicians use a variety of communication techniques to emphasize the importance of appropriate asthma drug use during patient visits[3]. Physicians\u00E2\u0080\u0099 attitude to patients plays an important role in patient education: a willingness to address questions and to listen actively can significantly improve patient asthma medication use[3]. Studies also suggested that by simply asking open-ended questions, such as \u00E2\u0080\u009Cwhat worries you most about your asthma?\u00E2\u0080\u009D physicians can encourage their patients and parents to discuss their concerns[3]. The large sample size resulting from the use of population-based databases provides guideline developers and health service planners with a comprehensive picture of drug use patterns and long-term asthma-related services use. 6.6 CONCLUSIONS Using a case definition containing the criteria of 2 or more physician visits and 3 or 209 more prescription drug dispensings improves the ability to identify patients with treated asthma. The prevalence (~2%) and incidence (0.7%) of asthma was stable in patients aged 5- 55 years in B.C. from 1996 to 2009. Asthma-related ED visits and hospital admissions declined by over 50% during this period. Much of the healthcare burden associated with asthma could be prevented by optimizing drug therapy; in particular, by improved ICS adherence. Identifying patients with suboptimal management practices is thus a critical step in reducing the burden of asthma on the healthcare system, decreasing asthma-related morbidity and mortality and improving quality of life. Categorizing asthma patients based on drug regimen optimality according to clinical practice guidelines is a useful way to identify those who are receiving suboptimal care. These findings indicated that optimizing drug regimens in asthma patients will result in lower health service utilization and a reduced burden on the healthcare system. 210 BIBLIOGRAPHY 1. Zhang T, Smith A, Camp P and Carleton B. Asthma drug regimen optimality and health service utilization: a population-based study in British Columbia. Pharmacoepidemiology & Drug Safety. 2013 Apr 5. doi: 10.1002/pds.3444. 2. Barnes, P., \"Chapter 36. Pulmonary Pharmacology\" (Chapter). Brunton LL, Chabner BA, Knollmann BC: Goodman & Gilman's The Pharmacological Basis of Therapeutics, 12e: http://www.accessmedicine.com/content.aspx?aID=16671685. 3. Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma (EPR-3 2007). NIH Publication No.08-4051. Bethesda, MD: Source: National Heart, Lung, and Blood Institute; National Institutes of Health; U.S. Department of Health and Human Services. 4. Guilbert, T. and M. Krawiec, Natural history of asthma. Pediatr Clin North Am, 2003. 50(3): p. 523-38. 5. Lemiere, C., et al., Adult Asthma Consensus Guidelines update 2003. Can Respir J, 2004. 11 Suppl A: p. 9A-18A. 6. Rennard, S.I. and S.G. Farmer, Exacerbations and progression of disease in asthma and chronic obstructive pulmonary disease. Proc Am Thorac Soc, 2004. 1(2): p. 88- 92. 7. Cairns, C.B., Acute asthma exacerbations: phenotypes and management. Clin Chest Med, 2006. 27(1): p. 99-108, vi-vii. 8. ten Brinke, A., et al., Risk factors of frequent exacerbations in difficult-to-treat asthma. Eur Respir J, 2005. 26(5): p. 812-8. 9. Bahadori, K., et al., Economic burden of asthma: a systematic review. BMC Pulm Med, 2009. 9: p. 24. 10. Biaggioni, I. and D. Robertson, Basic & Clinical Pharmacology, 12edition. Chapter 9. Adrenoceptor Agonists & Sympathomimetic Drugs. 11. Gern, J.E., Viral respiratory infection and the link to asthma. Pediatr Infect Dis J, 2004. 23(1 Suppl): p. S78-86. 12. Fireman, P., Understanding asthma pathophysiology. Allergy Asthma Proc, 2003. 24(2): p. 79-83. 13. Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma (EPR-3 2007). NIH Publication No.08-4051. Bethesda, MD: U.S. Department of Health and Human Services; National Institutes of Health; National Heart, Lung, and Blood Institute; National Asthma Education and Prevention Program,. 14. Cohn, L., J.A. Elias, and G.L. Chupp, Asthma: mechanisms of disease persistence and progression. Annu Rev Immunol, 2004. 22: p. 789-815. 15. Homer, R.J. and J.A. Elias, Consequences of long-term inflammation. Airway remodeling. Clin Chest Med, 2000. 21(2): p. 331-43, ix. 16. Eder, W., M.J. Ege, and E. von Mutius, The asthma epidemic. N Engl J Med, 2006. 355(21): p. 2226-35. 17. Gern, J.E., R.F. Lemanske, Jr., and W.W. Busse, Early life origins of asthma. J Clin Invest, 1999. 104(7): p. 837-43. 18. Asthma and Genetics. Available at: http://acc6.its.brooklyn.cuny.edu/~scintech/asthma/Genetics2.htm. 211 19. Sarafino, E.P. and J. Goldfedder, Genetic factors in the presence, severity, and triggers of asthma. Arch Dis Child, 1995. 73(2): p. 112-6. 20. Van Eerdewegh, P., et al., Association of the ADAM33 gene with asthma and bronchial hyperresponsiveness. Nature, 2002. 418(6896): p. 426-30. 21. Zhang, Y., et al., Positional cloning of a quantitative trait locus on chromosome 13q14 that influences immunoglobulin E levels and asthma. Nat Genet, 2003. 34(2): p. 181-6. 22. Allen, M., et al., Positional cloning of a novel gene influencing asthma from chromosome 2q14. Nat Genet, 2003. 35(3): p. 258-63. 23. Laitinen, T., et al., Characterization of a common susceptibility locus for asthma- related traits. Science, 2004. 304(5668): p. 300-4. 24. Walley, A.J., et al., Gene polymorphism in Netherton and common atopic disease. Nat Genet, 2001. 29(2): p. 175-8. 25. Cookson, W., The immunogenetics of asthma and eczema: a new focus on the epithelium. Nat Rev Immunol, 2004. 4(12): p. 978-88. 26. de Marco, R., et al., Differences in incidence of reported asthma related to age in men and women. A retrospective analysis of the data of the European Respiratory Health Survey. Am J Respir Crit Care Med, 2000. 162(1): p. 68-74. 27. Bjornson, C.L. and I. Mitchell, Gender differences in asthma in childhood and adolescence. J Gend Specif Med, 2000. 3(8): p. 57-61. 28. Williams, H. and K.N. McNicol, Prevalence, natural history, and relationship of wheezy bronchitis and asthma in children. An epidemiological study. Br Med J, 1969. 4(5679): p. 321-5. 29. Gerritsen, J., et al., Prognosis of asthma from childhood to adulthood. Am Rev Respir Dis, 1989. 140(5): p. 1325-30. 30. Sherrill, D., et al., The effects of airway hyperresponsiveness, wheezing, and atopy on longitudinal pulmonary function in children: a 6-year follow-up study. Pediatr Pulmonol, 1992. 13(2): p. 78-85. 31. Le Souef, P.N., M.R. Sears, and D. Sherrill, The effect of size and age of subject on airway responsiveness in children. Am J Respir Crit Care Med, 1995. 152(2): p. 576- 9. 32. Ernst, P., H. Ghezzo, and M.R. Becklake, Risk factors for bronchial hyperresponsiveness in late childhood and early adolescence. Eur Respir J, 2002. 20(3): p. 635-9. 33. Gustafsson, P.M. and B. Kjellman, Asthma from childhood to adulthood: course and outcome of lung function. Respir Med, 2000. 94(5): p. 466-74. 34. Roorda, R.J., et al., Follow-up of asthma from childhood to adulthood: influence of potential childhood risk factors on the outcome of pulmonary function and bronchial responsiveness in adulthood. J Allergy Clin Immunol, 1994. 93(3): p. 575-84. 35. Godden, D.J., et al., Outcome of wheeze in childhood. Symptoms and pulmonary function 25 years later. Am J Respir Crit Care Med, 1994. 149(1): p. 106-12. 36. Leynaert, B., et al., Is bronchial hyperresponsiveness more frequent in women than in men? A population-based study. Am J Respir Crit Care Med, 1997. 156(5): p. 1413- 20. 37. Sears, M.R., et al., Atopy in childhood. I. Gender and allergen related risks for 212 development of hay fever and asthma. Clin Exp Allergy, 1993. 23(11): p. 941-8. 38. Barbee, R.A., et al., Longitudinal changes in allergen skin test reactivity in a community population sample. J Allergy Clin Immunol, 1987. 79(1): p. 16-24. 39. Cline, M.G. and B. Burrows, Distribution of allergy in a population sample residing in Tucson, Arizona. Thorax, 1989. 44(5): p. 425-31. 40. Haggerty, C.L., et al., The impact of estrogen and progesterone on asthma. Ann Allergy Asthma Immunol, 2003. 90(3): p. 284-91; quiz 291-3, 347. 41. Dimitropoulou, C., et al., Estrogen reduces carbachol-induced constriction of asthmatic airways by stimulating large-conductance voltage and calcium-dependent potassium channels. Am J Respir Cell Mol Biol, 2005. 32(3): p. 239-47. 42. Nafstad, P., P. Magnus, and J.J. Jaakkola, Early respiratory infections and childhood asthma. Pediatrics, 2000. 106(3): p. E38. 43. Diette, G.B., et al., Home indoor pollutant exposures among inner-city children with and without asthma. Environ Health Perspect, 2007. 115(11): p. 1665-9. 44. Phipatanakul, W., Environmental factors and childhood asthma. Pediatr Ann, 2006. 35(9): p. 646-56. 45. Huss, K., et al., House dust mite and cockroach exposure are strong risk factors for positive allergy skin test responses in the Childhood Asthma Management Program. J Allergy Clin Immunol, 2001. 107(1): p. 48-54. 46. Eggleston, P.A., et al., Relationship of indoor allergen exposure to skin test sensitivity in inner-city children with asthma. J Allergy Clin Immunol, 1998. 102(4 Pt 1): p. 563- 70. 47. Sporik, R., et al., Exposure to house-dust mite allergen (Der p I) and the development of asthma in childhood. A prospective study. N Engl J Med, 1990. 323(8): p. 502-7. 48. Wahn, U., et al., Indoor allergen exposure is a risk factor for sensitization during the first three years of life. J Allergy Clin Immunol, 1997. 99(6 Pt 1): p. 763-9. 49. Stein, R.T., et al., Respiratory syncytial virus in early life and risk of wheeze and allergy by age 13 years. Lancet, 1999. 354(9178): p. 541-5. 50. Sigurs, N., et al., Respiratory syncytial virus bronchiolitis in infancy is an important risk factor for asthma and allergy at age 7. Am J Respir Crit Care Med, 2000. 161(5): p. 1501-7. 51. Gern, J.E., Viral respiratory infection and the link to asthma. Pediatr Infect Dis J, 2008. 27(10 Suppl): p. S97-103. 52. Gern, J.E., et al., Effects of viral respiratory infections on lung development and childhood asthma. J Allergy Clin Immunol, 2005. 115(4): p. 668-74; quiz 675. 53. Dezateux, C., et al., Impaired airway function and wheezing in infancy: the influence of maternal smoking and a genetic predisposition to asthma. Am J Respir Crit Care Med, 1999. 159(2): p. 403-10. 54. Mannino, D.M., D.M. Homa, and S.C. Redd, Involuntary smoking and asthma severity in children: data from the Third National Health and Nutrition Examination Survey. Chest, 2002. 122(2): p. 409-15. 55. Horta, B.L., et al., Low birthweight, preterm births and intrauterine growth retardation in relation to maternal smoking. Paediatr Perinat Epidemiol, 1997. 11(2): p. 140-51. 56. Elliot, J.G., et al., Airway alveolar attachment points and exposure to cigarette smoke 213 in utero. Am J Respir Crit Care Med, 2003. 167(1): p. 45-9. 57. Gilliland, F.D., et al., Effects of early onset asthma and in utero exposure to maternal smoking on childhood lung function. Am J Respir Crit Care Med, 2003. 167(6): p. 917-24. 58. Manitoba Child Health Altas Update, November 2008. . 59. It\u00E2\u0080\u0099s Your Health: Asthma. Accessed September 19th, 2011, at http://www.hc- sc.gc.ca/hl-vs/alt_formats/pdf/iyh-vsv/diseases-maladies/asthm-eng.pdf. 60. Masoli, M., et al., The global burden of asthma: executive summary of the GINA Dissemination Committee report. Allergy, 2004. 59(5): p. 469-78. 61. The burden of asthma in Ontario. ICES Investigative Report, September 2006. Available at: http://www.ices.on.ca/file/Burden_of_Asthma_Aug-07.pdf. 62. McLeod, C., et al., Asthma in British Columbia, 2007. Available at: http://pwhr.sites.olt.ubc.ca/files/2012/03/Asthma-Report-2007.pdf. 63. Sadatsafavi, M., et al., Direct health care costs associated with asthma in British Columbia. Can Respir J. 17(2): p. 74-80. 64. American Lung Association. Trends in asthma morbidity and mortality. 2008. 65. Van Wonderen, K.E., et al., Different definitions in childhood asthma: how dependable is the dependent variable? Eur Respir J. 36(1): p. 48-56. 66. Senthilselvan, A., et al., Stabilization of an increasing trend in physician-diagnosed asthma prevalence in Saskatchewan, 1991 to 1998. Chest, 2003. 124(2): p. 438-48. 67. Huzel, L., et al., Diagnosing asthma: the fit between survey and administrative database. Can Respir J, 2002. 9(6): p. 407-12. 68. Wilchesky, M., R.M. Tamblyn, and A. Huang, Validation of diagnostic codes within medical services claims. J Clin Epidemiol, 2004. 57(2): p. 131-41. 69. Kozyrskyj, A.L., C.A. Mustard, and A.B. Becker, Identifying children with persistent asthma from health care administrative records. Can Respir J, 2004. 11(2): p. 141-5. 70. Macy, E., et al., The prevalence of reversible airflow obstruction and/or methacholine hyperreactivity in random adult asthma patients identified by administrative data. J Asthma, 2005. 42(3): p. 213-20. 71. To, T., et al., Case verification of children with asthma in Ontario. Pediatr Allergy Immunol, 2006. 17(1): p. 69-76. 72. Schatz, M., et al., Asthma quality-of-care markers using administrative data. Chest, 2005. 128(4): p. 1968-73. 73. Aaron, S.D., et al., Overdiagnosis of asthma in obese and nonobese adults. CMAJ, 2008. 179(11): p. 1121-31. 74. Stout, J.W., et al., Classification of asthma severity in children: the contribution of pulmonary function testing. Arch Pediatr Adolesc Med, 2006. 160(8): p. 844-50. 75. Bacharier, L.B., et al., Classifying asthma severity in children: mismatch between symptoms, medication use, and lung function. Am J Respir Crit Care Med, 2004. 170(4): p. 426-32. 76. Brown, H.M., G. Storey, and W.H. George, Beclomethasone dipropionate: a new steroid aerosol for the treatment of allergic asthma. Br Med J, 1972. 1(5800): p. 585- 90. 77. Brown, H.M. and G. Storey, Beclomethasone dipropionate steriod aerosol in treatment of perennial allergic asthma in children. Br Med J, 1973. 3(5872): p. 161-4. 214 78. Eriksson, N.E., S. Lindgren, and N. Lindholm, Long term treatment of steroid dependent asthmatic patients with beclomethasone dipropionate. Allergol Immunopathol (Madr), 1975. 3(2): p. 77-86. 79. Brown, H.M., G. Storey, and F.A. Jackson, Beclomethasone dipropionate aerosol in long-term treatment of perennial and seasonal asthma in children and adults: a report of five-and-half years' experience in 600 asthmatic patients. Br J Clin Pharmacol, 1977. 4 Suppl 3: p. 259S-267S. 80. Chauhan, B.F. and F.M. Ducharme, Anti-leukotriene agents compared to inhaled corticosteroids in the management of recurrent and/or chronic asthma in adults and children. Cochrane Database Syst Rev. 5: p. CD002314. 81. Price, D., et al., Leukotriene antagonists as first-line or add-on asthma-controller therapy. N Engl J Med. 364(18): p. 1695-707. 82. Guevara, J.P., et al., Inhaled corticosteroids versus sodium cromoglycate in children and adults with asthma. Cochrane Database Syst Rev, 2006(2): p. CD003558. 83. Wang, Y., et al., Comparison of inhaled corticosteroid combined with theophylline and double-dose inhaled corticosteroid in moderate to severe asthma. Respirology, 2005. 10(2): p. 189-95. 84. Greenstone, I.R., et al., Combination of inhaled long-acting beta2-agonists and inhaled steroids versus higher dose of inhaled steroids in children and adults with persistent asthma. Cochrane Database Syst Rev, 2005(4): p. CD005533. 85. Agertoft, L. and S. Pedersen, Effects of long-term treatment with an inhaled corticosteroid on growth and pulmonary function in asthmatic children. Respir Med, 1994. 88(5): p. 373-81. 86. Konig, P. and J. Shaffer, The effect of drug therapy on long-term outcome of childhood asthma: a possible preview of the international guidelines. J Allergy Clin Immunol, 1996. 98(6 Pt 1): p. 1103-11. 87. Donahue, J.G., et al., Inhaled steroids and the risk of hospitalization for asthma. JAMA, 1997. 277(11): p. 887-91. 88. Blais, L., et al., First treatment with inhaled corticosteroids and the prevention of admissions to hospital for asthma. Thorax, 1998. 53(12): p. 1025-9. 89. Crane, J., et al., Prescribed fenoterol and death from asthma in New Zealand, 1981- 83: case-control study. Lancet, 1989. 1(8644): p. 917-22. 90. Pearce, N., et al., Case-control study of prescribed fenoterol and death from asthma in New Zealand, 1977-81. Thorax, 1990. 45(3): p. 170-5. 91. Ernst, P., et al., Risk of fatal and near-fatal asthma in relation to inhaled corticosteroid use. JAMA, 1992. 268(24): p. 3462-4. 92. Suissa, S., et al., Low-dose inhaled corticosteroids and the prevention of death from asthma. N Engl J Med, 2000. 343(5): p. 332-6. 93. Kelly, H.W. and H.S. Nelson, Potential adverse effects of the inhaled corticosteroids. J Allergy Clin Immunol, 2003. 112(3): p. 469-78; quiz 479. 94. Hanania, N.A., K.R. Chapman, and S. Kesten, Adverse effects of inhaled corticosteroids. Am J Med, 1995. 98(2): p. 196-208. 95. Roland, N.J., R.K. Bhalla, and J. Earis, The local side effects of inhaled corticosteroids: current understanding and review of the literature. Chest, 2004. 126(1): p. 213-9. 215 96. Rinehart, J.J., et al., Effects of corticosteroid therapy on human monocyte function. N Engl J Med, 1975. 292(5): p. 236-41. 97. Shaw, N.J. and A.T. Edmunds, Inhaled beclomethasone and oral candidiasis. Arch Dis Child, 1986. 61(8): p. 788-90. 98. Toogood, J.H., et al., Candidiasis and dysphonia complicating beclomethasone treatment of asthma. J Allergy Clin Immunol, 1980. 65(2): p. 145-53. 99. Barnes, P.J., S. Pedersen, and W.W. Busse, Efficacy and safety of inhaled corticosteroids. New developments. Am J Respir Crit Care Med, 1998. 157(3 Pt 2): p. S1-53. 100. Pedersen, S., Do inhaled corticosteroids inhibit growth in children? Am J Respir Crit Care Med, 2001. 164(4): p. 521-35. 101. Leone, F.T., et al., Systematic review of the evidence regarding potential complications of inhaled corticosteroid use in asthma: collaboration of American College of Chest Physicians, American Academy of Allergy, Asthma, and Immunology, and American College of Allergy, Asthma, and Immunology. Chest, 2003. 124(6): p. 2329-40. 102. Allen, D.B., Growth suppression by glucocorticoid therapy. Endocrinol Metab Clin North Am, 1996. 25(3): p. 699-717. 103. Allen, D.B., Influence of inhaled corticosteroids on growth: a pediatric endocrinologist's perspective. Acta Paediatr, 1998. 87(2): p. 123-9. 104. Allen, D.B., Effects of inhaled steroids on growth, bone metabolism, and adrenal function. Adv Pediatr, 2006. 53: p. 101-10. 105. Long-term effects of budesonide or nedocromil in children with asthma. The Childhood Asthma Management Program Research Group. N Engl J Med, 2000. 343(15): p. 1054-63. 106. Agertoft, L. and S. Pedersen, Effect of long-term treatment with inhaled budesonide on adult height in children with asthma. N Engl J Med, 2000. 343(15): p. 1064-9. 107. Guilbert, T.W., et al., Long-term inhaled corticosteroids in preschool children at high risk for asthma. N Engl J Med, 2006. 354(19): p. 1985-97. 108. Van Bever, H.P., et al., Does treatment of asthmatic children with inhaled corticosteroids affect their adult height? Pediatr Pulmonol, 1999. 27(6): p. 369-75. 109. McCowan, C., et al., Effect of asthma and its treatment on growth: four year follow up of cohort of children from general practices in Tayside, Scotland. BMJ, 1998. 316(7132): p. 668-72. 110. Blais, L., et al., Inhaled corticosteroids vs. leukotriene-receptor antagonists and asthma exacerbations in children. Respir Med. 105(6): p. 846-55. 111. Ducharme, F.M., et al., Clinical effectiveness of inhaled corticosteroids versus montelukast in children with asthma: prescription patterns and patient adherence as key factors. Curr Med Res Opin. 28(1): p. 111-9. 112. Schumock, G.T., et al., Association between leukotriene-modifying agents and suicide: what is the evidence? Drug Saf. 34(7): p. 533-44. 113. Brenner, Update on singulair (montelukast), 2008. Available at: http://www.beforeyoutakethatpill.com/2008/4/update-on-singulair. 114. Singulair mast cells, mast cell, dust mites, asthma sufferers, immune cells 2008. Available at: http://www.medications.com/effect/view/35943. . 216 115. Adams, R.J., et al., Impact of inhaled antiinflammatory therapy on hospitalization and emergency department visits for children with asthma. Pediatrics, 2001. 107(4): p. 706-11. 116. van der Wouden, J.C., et al., Inhaled sodium cromoglycate for asthma in children. Cochrane Database Syst Rev, 2008(4): p. CD002173. 117. Ayres, J.G., et al., Efficacy and tolerability of anti-immunoglobulin E therapy with omalizumab in patients with poorly controlled (moderate-to-severe) allergic asthma. Allergy, 2004. 59(7): p. 701-8. 118. Deniz, Y.M. and N. Gupta, Safety and tolerability of omalizumab (Xolair), a recombinant humanized monoclonal anti-IgE antibody. Clin Rev Allergy Immunol, 2005. 29(1): p. 31-48. 119. Niebauer, K., et al., Impact of omalizumab on quality-of-life outcomes in patients with moderate-to-severe allergic asthma. Ann Allergy Asthma Immunol, 2006. 96(2): p. 316-26. 120. Humbert, M., et al., Benefits of omalizumab as add-on therapy in patients with severe persistent asthma who are inadequately controlled despite best available therapy (GINA 2002 step 4 treatment): INNOVATE. Allergy, 2005. 60(3): p. 309-16. 121. Busse, W.W., et al., Randomized trial of omalizumab (anti-IgE) for asthma in inner- city children. N Engl J Med. 364(11): p. 1005-15. 122. Greening, A.P., et al., Added salmeterol versus higher-dose corticosteroid in asthma patients with symptoms on existing inhaled corticosteroid. Allen & Hanburys Limited UK Study Group. Lancet, 1994. 344(8917): p. 219-24. 123. van Noord, J.A., et al., Addition of salmeterol versus doubling the dose of fluticasone propionate in patients with mild to moderate asthma. Thorax, 1999. 54(3): p. 207-12. 124. Condemi, J.J., et al., The addition of salmeterol to fluticasone propionate versus increasing the dose of fluticasone propionate in patients with persistent asthma. Salmeterol Study Group. Ann Allergy Asthma Immunol, 1999. 82(4): p. 383-9. 125. Pauwels, R.A., et al., Effect of inhaled formoterol and budesonide on exacerbations of asthma. Formoterol and Corticosteroids Establishing Therapy (FACET) International Study Group. N Engl J Med, 1997. 337(20): p. 1405-11. 126. Lalloo, U.G., et al., Budesonide and formoterol in a single inhaler improves asthma control compared with increasing the dose of corticosteroid in adults with mild-to- moderate asthma. Chest, 2003. 123(5): p. 1480-7. 127. Bateman, E.D., et al., Combination therapy with single inhaler budesonide/formoterol compared with high dose of fluticasone propionate alone in patients with moderate persistent asthma. Am J Respir Med, 2003. 2(3): p. 275-81. 128. Castle, W., et al., Serevent nationwide surveillance study: comparison of salmeterol with salbutamol in asthmatic patients who require regular bronchodilator treatment. BMJ, 1993. 306(6884): p. 1034-7. 129. Nelson, H.S., et al., The Salmeterol Multicenter Asthma Research Trial: a comparison of usual pharmacotherapy for asthma or usual pharmacotherapy plus salmeterol. Chest, 2006. 129(1): p. 15-26. 130. Levenson, M., Long-Acting Beta-Agonists and Adverse Asthma Events Meta-Analysis. Available at: http://www.fda.gov/ohrms/dockets/ac/08/briefing/2008-4398b1-01- FDA.pdf. Joint Meeting of the Pulmonary-Allergy Drugs Advisory Committee, Drug 217 Safety and Risk Management Advisory Committee and Pediatric Advisory Committee 2008. 131. Basu, K., et al., Adrenergic beta(2)-receptor genotype predisposes to exacerbations in steroid-treated asthmatic patients taking frequent albuterol or salmeterol. J Allergy Clin Immunol, 2009. 124(6): p. 1188-94 e3. 132. Ducharme, F.M., et al., Safety profile of frequent short courses of oral glucocorticoids in acute pediatric asthma: impact on bone metabolism, bone density, and adrenal function. Pediatrics, 2003. 111(2): p. 376-83. 133. Linde, K., K. Jobst, and J. Panton, Acupuncture for chronic asthma. Cochrane Database Syst Rev, 2000(2): p. CD000008. 134. McCarney, R.W., et al., Acupuncture for chronic asthma. Cochrane Database Syst Rev, 2004(1): p. CD000008. 135. Grainger, J., et al., Prescribed fenoterol and death from asthma in New Zealand, 1981-7: a further case-control study. Thorax, 1991. 46(2): p. 105-11. 136. Spitzer, W.O., et al., The use of beta-agonists and the risk of death and near death from asthma. N Engl J Med, 1992. 326(8): p. 501-6. 137. Sears, M.R., et al., Regular inhaled beta-agonist treatment in bronchial asthma. Lancet, 1990. 336(8728): p. 1391-6. 138. Cheung, D., et al., Long-term effects of a long-acting beta 2-adrenoceptor agonist, salmeterol, on airway hyperresponsiveness in patients with mild asthma. N Engl J Med, 1992. 327(17): p. 1198-203. 139. O'Connor, B.J., S.L. Aikman, and P.J. Barnes, Tolerance to the nonbronchodilator effects of inhaled beta 2-agonists in asthma. N Engl J Med, 1992. 327(17): p. 1204-8. 140. Cockcroft, D.W., et al., Regular inhaled salbutamol and airway responsiveness to allergen. Lancet, 1993. 342(8875): p. 833-7. 141. Gauvreau, G.M., R.M. Watson, and P.M. O'Byrne, Kinetics of allergen-induced airway eosinophilic cytokine production and airway inflammation. Am J Respir Crit Care Med, 1999. 160(2): p. 640-7. 142. Swystun, V.A., et al., Mast cell tryptase release and asthmatic responses to allergen increase with regular use of salbutamol. J Allergy Clin Immunol, 2000. 106(1 Pt 1): p. 57-64. 143. Runton, L., et al., Goodman & Gilman's The Pharmacological Basis of Therapeutics. 12th ed. New York: McGraw-Hill, 2011. ISBN 13:978-0-07-1624428. Page 171. 144. Rodrigo, G.J. and J.A. Castro-Rodriguez, Anticholinergics in the treatment of children and adults with acute asthma: a systematic review with meta-analysis. Thorax, 2005. 60(9): p. 740-6. 145. Kerstjens, H.A., et al., Tiotropium in asthma poorly controlled with standard combination therapy. N Engl J Med. 367(13): p. 1198-207. 146. Peters, S.P., et al., Tiotropium bromide step-up therapy for adults with uncontrolled asthma. N Engl J Med. 363(18): p. 1715-26. 147. Kerstjens, H.A., et al., Tiotropium improves lung function in patients with severe uncontrolled asthma: a randomized controlled trial. J Allergy Clin Immunol. 128(2): p. 308-14. 148. Bateman, E.D., et al., Tiotropium is noninferior to salmeterol in maintaining improved lung function in B16-Arg/Arg patients with asthma. J Allergy Clin Immunol. 218 128(2): p. 315-22. 149. Blaiss, M.S. and B. Hill, Outcomes in pediatric asthma. Curr Allergy Asthma Rep, 2005. 5(6): p. 431-6. 150. Adams, R.J., B.J. Smith, and R.E. Ruffin, Factors associated with hospital admissions and repeat emergency department visits for adults with asthma. Thorax, 2000. 55(7): p. 566-73. 151. Eisner, M.D., et al., Risk factors for hospitalization among adults with asthma: the influence of sociodemographic factors and asthma severity. Respir Res, 2001. 2(1): p. 53-60. 152. Lieu, T.A., et al., Computer-based models to identify high-risk children with asthma. Am J Respir Crit Care Med, 1998. 157(4 Pt 1): p. 1173-80. 153. Firoozi, F., et al., Development and validation of database indexes of asthma severity and control. Thorax, 2007. 62(7): p. 581-7. 154. Ungar, W.J., K.R. Chapman, and M.T. Santos, Assessment of a medication-based asthma index for population research. Am J Respir Crit Care Med, 2002. 165(2): p. 190-4. 155. Santillanes, G. and M. Gausche-Hill, Pediatric airway management. Emerg Med Clin North Am, 2008. 26(4): p. 961-75, ix. 156. Yoffe, S.J., et al., A reduction in emergency department use by children from a parent educational intervention. Fam Med. 43(2): p. 106-11. 157. Schatz, M., et al., Relationship of asthma control to asthma exacerbations using surrogate markers within a managed care database. Am J Manag Care. 16(5): p. 327- 33. 158. Schatz, M., et al., The controller-to-total asthma medication ratio is associated with patient-centered as well as utilization outcomes. Chest, 2006. 130(1): p. 43-50. 159. Aveyard, P., Assessing the performance of general practices caring for patients with asthma. Br J Gen Pract, 1997. 47(420): p. 423-6. 160. Griffiths, C., et al., Prescribing and hospital admissions for asthma in east London. Bmj, 1996. 312(7029): p. 481-2. 161. Shelley, M., et al., Is the ratio of inhaled corticosteroid to bronchodilator a good indicator of the quality of asthma prescribing? Cross sectional study linking prescribing data to data on admissions. Bmj, 1996. 313(7065): p. 1124-6. 162. Fuhlbrigge, A., et al., Evaluation of asthma prescription measures and health system performance based on emergency department utilization. Med Care, 2004. 42(5): p. 465-71. 163. Cowie, R.L., et al., The effect of a peak flow-based action plan in the prevention of exacerbations of asthma. Chest, 1997. 112(6): p. 1534-8. 164. Gibson, P.G., et al., Self-management education and regular practitioner review for adults with asthma. Cochrane Database Syst Rev, 2003(1): p. CD001117. 165. Toelle, B.G. and F.S. Ram, Written individualised management plans for asthma in children and adults. Cochrane Database Syst Rev, 2004(2): p. CD002171. 166. Baldwin, D., et al., Outcome of asthmatics attending asthma clinics utilizing self- management plans in general practice. Asthma in General Practice, 1997. 5(2): p. 31- 33. 167. Jones, K.P., et al., Peak flow based asthma self-management: a randomised controlled 219 study in general practice. British Thoracic Society Research Committee. Thorax, 1995. 50(8): p. 851-7. 168. van der Palen, J., et al., Behavioural effect of self-treatment guidelines in a self- management program for adults with asthma. Patient Educ Couns, 2001. 43(2): p. 161-9. 169. Klein, J.J., et al., Benefit from the inclusion of self-treatment guidelines to a self- management programme for adults with asthma. Eur Respir J, 2001. 17(3): p. 386-94. 170. Adams, R.J., et al., A randomized trial of peak-flow and symptom-based action plans in adults with moderate-to-severe asthma. Respirology, 2001. 6(4): p. 297-304. 171. Gibson, P.G. and H. Powell, Written action plans for asthma: an evidence-based review of the key components. Thorax, 2004. 59(2): p. 94-9. 172. Gillissen, A., K. Busch, and U. Juergens, [Adherence to therapy in bronchial asthma]. Dtsch Med Wochenschr, 2007. 132(23): p. 1281-6. 173. Burgess, S., P. Sly, and S. Devadason, Adherence with preventive medication in childhood asthma. Pulm Med. 2011: p. 973849. 174. Krishnan, J.A., et al., Corticosteroid use after hospital discharge among high-risk adults with asthma. Am J Respir Crit Care Med, 2004. 170(12): p. 1281-5. 175. Bender, B.G. and S.E. Bender, Patient-identified barriers to asthma treatment adherence: responses to interviews, focus groups, and questionnaires. Immunol Allergy Clin North Am, 2005. 25(1): p. 107-30. 176. Rand, C.S. and R.A. Wise, Measuring adherence to asthma medication regimens. Am J Respir Crit Care Med, 1994. 149(2 Pt 2): p. S69-76; discussion S77-8. 177. Bender, B. and T. Creer, Promoting adherence and effective self-management in patients with asthma. In: Leung D, Sampson H, Geha R, Szefler S, editors. Pediatric asthma: Principles and practice. St. Louis: Mosby; 2003. P. 457 - 64. 2003. 178. Bauman, L.J., et al., Relationship of adherence to pediatric asthma morbidity among inner-city children. Pediatrics, 2002. 110(1 Pt 1): p. e6. 179. Bender, B. and L. Zhang, Negative affect, medication adherence, and asthma control in children. J Allergy Clin Immunol, 2008. 122(3): p. 490-5. 180. Iskedjiian, M., A. Addis, and T. Einarson, estimating the economic burden of hosptialization due to patient nonadherence in Canada. Available: http://www.pharmideas.com/pharmideas_images/pdf/COA2%20Poster%20Final.pdf Assessed on Feb 10th, 2012. 2005. 181. Sackett, D. and R. Haynes, Compliance with therapeutic regimens. Johns Hopkins University Press, 1976. 182. Cluss, P.A., et al., Effect of compliance for chronic asthmatic children. J Consult Clin Psychol, 1984. 52(5): p. 909-10. 183. Radius, S.M., et al., Factors influencing mothers' compliance with a medication regimen for asthmatic children. J Asthma Res, 1978. 15(3): p. 133-49. 184. Manzella, B.A., et al., Assessing the use of metered dose inhalers by adults with asthma. J Asthma, 1989. 26(4): p. 223-30. 185. Davis, M.S., Variations in patients' compliance with doctors' orders: analysis of congruence between survey responses and results of empirical investigations. J Med Educ, 1966. 41(11): p. 1037-48. 186. Kasl, S.V., Issues in patient adherence to health care regimens. J Human Stress, 1975. 220 1(3): p. 5-17. 187. Steele, D.J., T.C. Jackson, and M.C. Gutmann, Have you been taking your pills? The adherence-monitoring sequence in the medical interview. J Fam Pract, 1990. 30(3): p. 294-9. 188. Kandane-Rathnayake, R.K., et al., Adherence to asthma management guidelines by middle-aged adults with current asthma. Thorax, 2009. 64(12): p. 1025-1031. 189. de Marco, R., et al., Are the asthma guideline goals achieved in daily practice? A population-based study on treatment adequacy and the control of asthma. Int Arch Allergy Immunol, 2005. 138(3): p. 225-34. 190. Finkelstein, J.A., et al., Underuse of controller medications among Medicaid-insured children with asthma. Arch Pediatr Adolesc Med, 2002. 156(6): p. 562-7. 191. Yoos, H.L., et al., Treatment regimens and health care utilization in children with persistent asthma symptoms. J Asthma, 2006. 43(5): p. 385-91. 192. Spector, S.L., et al., Compliance of patients with asthma with an experimental aerosolized medication: implications for controlled clinical trials. J Allergy Clin Immunol, 1986. 77(1 Pt 1): p. 65-70. 193. Mattke, S., et al., Anti-inflammatory medication adherence and cost and utilization of asthma care in a commercially insured population. J Asthma. 47(3): p. 323-9. 194. Pando, S., et al., Suboptimal use of inhaled corticosteroids in children with persistent asthma: inadequate prescription, poor drug adherence, or both? Pharmacotherapy. 30(11): p. 1109-16. 195. Lagerlov, P., et al., Structuring prescribing data into traffic-light categories; a tool for evaluating treatment quality in primary care. Fam Pract, 2001. 18(5): p. 528-33. 196. Williams, L.K., et al., Patients with asthma who do not fill their inhaled corticosteroids: a study of primary nonadherence. J Allergy Clin Immunol, 2007. 120(5): p. 1153-9. 197. Wells, K., et al., Race-ethnic differences in factors associated with inhaled steroid adherence among adults with asthma. Am J Respir Crit Care Med, 2008. 178(12): p. 1194-201. 198. Corsico, A.G., et al., Factors affecting adherence to asthma treatment in an international cohort of young and middle-aged adults. Respir Med, 2007. 101(6): p. 1363-7. 199. Wolff, M., et al., US family physicians' experiences with practice guidelines. Fam Med, 1998. 30(2): p. 117-21. 200. Hayward, R.S., et al., Canadian physicians' attitudes about and preferences regarding clinical practice guidelines. CMAJ, 1997. 156(12): p. 1715-23. 201. Cabana, M.D., et al., Why don't physicians follow clinical practice guidelines? A framework for improvement. JAMA, 1999. 282(15): p. 1458-65. 202. Wamboldt, F.S., B.G. Bender, and A.E. Rankin, Adolescent decision-making about use of inhaled asthma controller medication: results from focus groups with participants from a prior longitudinal study. J Asthma. 48(7): p. 741-50. 203. Stempel, D.A., et al., Adherence to asthma controller medication regimens. Respir Med, 2005. 99(10): p. 1263-7. 204. Hand, C.H. and C. Bradley, Health beliefs of adults with asthma: toward an understanding of the difference between symptomatic and preventive use of inhaler 221 treatment. J Asthma, 1996. 33(5): p. 331-8. 205. Kelloway, J.S., et al., Effect of salmeterol on patients' adherence to their prescribed refills for inhaled corticosteroids. Ann Allergy Asthma Immunol, 2000. 84(3): p. 324- 8. 206. Roman, B., N. Borges, and A.K. Morrison, Teaching motivational interviewing skills to third-year psychiatry clerkship students. Acad Psychiatry. 35(1): p. 51-3. 207. Tousman, S., et al., A pilot study on a cognitive-behavioral asthma self-management program for adults. Chron Respir Dis, 2006. 3(2): p. 73-82. 208. Britten, N., Patients' ideas about medicines: a qualitative study in a general practice population. Br J Gen Pract, 1994. 44(387): p. 465-8. 209. Boulet, L.P., Perception of the role and potential side effects of inhaled corticosteroids among asthmatic patients. Chest, 1998. 113(3): p. 587-92. 210. Zeiger, R.S., C. Dawson, and S. Weiss, Relationships between duration of asthma and asthma severity among children in the Childhood Asthma Management Program (CAMP). J Allergy Clin Immunol, 1999. 103(3 Pt 1): p. 376-87. 211. de Marco, R., et al., Prognostic factors of asthma severity: a 9-year international prospective cohort study. J Allergy Clin Immunol, 2006. 117(6): p. 1249-56. 212. Taylor, D.R., et al., Asthma in remission: can relapse in early adulthood be predicted at 18 years of age? Chest, 2005. 127(3): p. 845-50. 213. Vonk, J.M., et al., Childhood factors associated with asthma remission after 30 year follow up. Thorax, 2004. 59(11): p. 925-9. 214. Sears, M.R., et al., A longitudinal, population-based, cohort study of childhood asthma followed to adulthood. N Engl J Med, 2003. 349(15): p. 1414-22. 215. Apostol, G.G., et al., Early life factors contribute to the decrease in lung function between ages 18 and 40: the Coronary Artery Risk Development in Young Adults study. Am J Respir Crit Care Med, 2002. 166(2): p. 166-72. 216. Lange, P., et al., A 15-year follow-up study of ventilatory function in adults with asthma. N Engl J Med, 1998. 339(17): p. 1194-200. 217. Strachan, D.P., et al., Ventilatory function in British adults after asthma or wheezing illness at ages 0-35. Am J Respir Crit Care Med, 1996. 154(6 Pt 1): p. 1629-35. 218. Horak, E., et al., Longitudinal study of childhood wheezy bronchitis and asthma: outcome at age 42. BMJ, 2003. 326(7386): p. 422-3. 219. British Columbia Statistical Profile of Aboriginal Peoples 2006. Aboriginal Peoples Compared to the non-Aboriginal Population with Emphasis on Labour Market and Post Secondary Issues. Available at: http://www.bcstats.gov.bc.ca/data/cen01/abor/aborbc.pdf. 220. B.C. Medical Services Plan. Teleplan: Claims Submission and Processing. Available at: http://www.health.gov.bc.ca/msp/infoprac/teleplan.html. 221. Alternative Payments and the National Physician Database (NPDB). The status of alternative payment programs for physicians in Canada, 1999/2000. https://secure.cihi.ca/free_products/AlternativePaymentsNPDB.pdf. 222. The status of alternative payment programs for physicians in Canada, 2005/2006. https://secure.cihi.ca/free_products/AltPay2005_2006_e.pdf. 223. Alternative Payments to Physicians: A Program in Need of Change. http://www.bcauditor.com/pubs/2003/report4/alternative-payments-physician- 222 program-compensation. 2003. 224. The status of alternative payment programs for physicians in Canada, 2002-2003 and preliminary information for 2003-2004. 225. Watson, D., et al., Planning for renewal: mapping primary health care in British Columbia. 2005. 226. Discharge Abstract Database, 2010\u00E2\u0080\u00932011, Canadian Institute for Health Information. 227. Price, M., et al., Assessing accuracy of an electronic provincial medication repository. BMC Med Inform Decis Mak. 12: p. 42. 228. Peat, J.K., et al., Continuing the debate about measuring asthma in population studies. Thorax, 2001. 56(5): p. 406-11. 229. Juurlink, D., et al., Canadian Institute for Health Information Discharge Abstract Database: A Validation Study. June 2006. 230. Dahri, K., et al., Accuracy of a provincial prescription database for assessing medication adherence in heart failure patients. Ann Pharmacother, 2008. 42(3): p. 361-7. 231. Prosser, R.J., B.C. Carleton, and M.A. Smith, Identifying persons with treated asthma using administrative data via latent class modelling. Health Serv Res, 2008. 43(2): p. 733-54. 232. Zhang, T., et al., The added burden of comorbidity in patients with asthma. J Asthma, 2009. 46(10): p. 1021-6. 233. Anderson, R., et al., Comparability of cause of death between ICD-9 and ICD-10: preliminary estimates. 2001. 234. Juhn, Y., et al., Characterisation of children's asthma status by ICD-9 code and criteria-based medical record review. Prim Care Respir J. 20(1): p. 79-83. 235. Krueger, K.P., E.P. Armstrong, and P.C. Langley, The Accuracy of Asthma and Respiratory Disease Diagnostic Codes in a Managed Care Medical Claims Database. Disease Management, 2001. 4(4). 236. Juurlink, D., C. Preyra, and R. Croxford, Canadian Institute for Health Information Discharge Abstract Database: A Validation Study. Available at: http://www.ices.on.ca/file/cihi_dad_reabstractors_study.pdf: Institute for Clinical Evaluative Sciences; 2006. 237. BC STATs P.E.O.P.L.E software available at: http://www.bcstats.gov.bc.ca/data/pop/pop/dynamic/PopulationStatistics/Query.asp?c ategory=Health&type=HAbyHY&topic=Estimates&agegrouptype=Custom. 2010. 238. Mustard, C.A., et al., Emergency department use as a component of total ambulatory care: a population perspective. CMAJ, 1998. 158(1): p. 49-55. 239. McKendry, R., et al., Emergency rooms in British Columbia: a pilot project to validate current data and describe users. 2002. 240. iMap BC http://webmaps.gov.bc.ca/imfx/imf.jsp?site=imapbc. 241. Kirkwood, B., Essentials of medical statistics. 1988. 242. Hughes, D.A., A. Woodcock, and T. Walley, Review of therapeutically equivalent alternatives to short acting beta(2) adrenoceptor agonists delivered via chlorofluorocarbon-containing inhalers. Thorax, 1999. 54(12): p. 1087-92. 243. Brocklebank, D., et al., Comparison of the effectiveness of inhaler devices in asthma and chronic obstructive airways disease: a systematic review of the literature. Health 223 Technol Assess, 2001. 5(26): p. 1-149. 244. Garner, R. and D. Kohen, Changes in the prevalence of asthma among Canadian children. Health Rep, 2008. 19(2): p. 45-50. 245. Barnes, P.J., \"Chapter 36. Pulmonary Pharmacology\" (Chapter). Brunton LL, Chabner BA, Knollmann BC: Goodman & Gilman's The Pharmacological Basis of Therapeutics, 12e: http://www.accessmedicine.com/content.aspx?aID=16671685. 246. Compendium of Pharmaceuticals and Specialties. The Canadian Drug Reference for Health Professionals. Ottawa: Canadian Pharmacists Association. 247. Boudreaux, E.D., et al., Acute asthma among adults presenting to the emergency department: the role of race/ethnicity and socioeconomic status. Chest, 2003. 124(3): p. 803-12. 248. Cowie, R.L., et al., Predicting emergency department utilization in adults with asthma: a cohort study. J Asthma, 2001. 38(2): p. 179-84. 249. Fernandes, A.K., et al., Characteristics of acute asthma patients attended frequently compared with those attended only occasionally in an emergency department. J Asthma, 2003. 40(6): p. 683-90. 250. Griswold, S.K., et al., Asthma exacerbations in North American adults: who are the \"frequent fliers\" in the emergency department? Chest, 2005. 127(5): p. 1579-86. 251. Suissa, S., P. Ernst, and A. Kezouh, Regular use of inhaled corticosteroids and the long term prevention of hospitalisation for asthma. Thorax, 2002. 57(10): p. 880-4. 252. Hofer, T.P., et al., Use of community versus individual socioeconomic data in predicting variation in hospital use. Health Serv Res, 1998. 33(2 Pt 1): p. 243-59. 253. Wilkins, R., Use of postal codes and addresses in the analysis of health data. Health Rep, 1993. 5(2): p. 157-77. 254. Rosenbaum, P.R. and D.B. Rubin, The central role of the propensity score in observational studies for causal effects. Biometrika, 1983. 70(1): p. 41-55. 255. Glynn, R.J., S. Schneeweiss, and T. Sturmer, Indications for propensity scores and review of their use in pharmacoepidemiology. Basic Clin Pharmacol Toxicol, 2006. 98(3): p. 253-9. 256. Ho, D.E., Matching as nonparametric preprocessing for reducing model dependence in parametric causal inference. Political Analysis, 2007. 15: p. 199-236. 257. Kleinbaum, D. and M. Klein, Survival analysis: a self-learning text. 2005. 258. The Johns Hopkins ACG System. Technical Refrence Guide Version 9.0. 2009. 259. Vollmer, W.M., et al., Association of asthma control with health care utilization: a prospective evaluation. Am J Respir Crit Care Med, 2002. 165(2): p. 195-9. 260. Chipps, B.E., et al., Assessment of asthma control and asthma exacerbations in the epidemiology and natural history of asthma: outcomes and treatment regimens (TENOR) observational cohort. Curr Respir Care Rep. 1(4): p. 259-269. 261. Erzen, D., et al., Changes in asthma severity in Manitoba. Chest, 1995. 108(1): p. 16- 23. 262. Gershon, A.S., et al., Trends in asthma prevalence and incidence in Ontario, Canada, 1996-2005: a population study. Am J Epidemiol. 172(6): p. 728-36. 263. The Canadian Institute for Health Information (CIHI). Drug use among seniors on public drug programs in Canada, 2002 to 2008. Available at: https://secure.cihi.ca/free_products/drug_use_in_seniors_2002-2008_e.pdf. 2010. 224 264. Yiannakoulias, N., D.P. Schopflocher, and L.W. Svenson, Using administrative data to understand the geography of case ascertainment. Chronic Dis Can, 2009. 30(1): p. 20-8. 265. Akinbami, L.J., et al., Status of childhood asthma in the United States, 1980-2007. Pediatrics, 2009. 123 Suppl 3: p. S131-45. 266. Cowie, R.L., L. Cicutto, and L.P. Boulet, Asthma education and management programs in Canada. Can Respir J, 2001. 8(6): p. 416-20. 267. McGhan, S.L., et al., A children's asthma education program: Roaring Adventures of Puff (RAP), improves quality of life. Can Respir J. 17(2): p. 67-73. 268. Poureslami, I., et al., Effectiveness of educational interventions on asthma self- management in Punjabi and Chinese asthma patients: a randomized controlled trial. J Asthma. 49(5): p. 542-51. 269. Hopman, W.M., et al., Outcomes of asthma education: results of a multisite evaluation. Can Respir J, 2004. 11(4): p. 291-7. 270. Leach, C.L., Improved delivery of inhaled steroids to the large and small airways. Respir Med, 1998. 92 Suppl A: p. 3-8. 271. Leach, C.L., et al., Lung deposition of hydrofluoroalkane-134a beclomethasone is greater than that of chlorofluorocarbon fluticasone and chlorofluorocarbon beclomethasone : a cross-over study in healthy volunteers. Chest, 2002. 122(2): p. 510-6. 272. Leach, C.L. and G.L. Colice, A pilot study to assess lung deposition of HFA- beclomethasone and CFC-beclomethasone from a pressurized metered dose inhaler with and without add-on spacers and using varying breathhold times. J Aerosol Med Pulm Drug Deliv. 23(6): p. 355-61. 273. Canadian Institute for Health Information. Hospital Trends in Canada - results of a project to create a historical series of statistical and financial data for Canadian hospitals over twenty-seven years. https://secure.cihi.ca/free_products/Hospital_Trends_in_Canada_e.pdf. 2005. 274. Canadian Institute for Health Information. Canada's Health Care Providers, 1997 to 2006, A Reference Guide. Available at: https://secure.cihi.ca/free_products/HealthCareProv_RefGuideEN_Final_Sep18.pdf. 2008. 275. Wong, W.B., et al., Can after-hours family medicine clinics represent an alternative to emergency departments? Survey of ambulatory patients seeking after-hours care. Can Fam Physician, 2009. 55(11): p. 1106-1107 e4. 276. Jones, D., L. Carroll, and L. Frank, After-Hours Care in Suburban Canada: Influencing Emergency Department Utilization. Journal of Primary Care & Community Health, 2011. 277. An examination of emergency department use in Colorado. Available at: http://www.coloradotrust.org/attachments/0001/9017/FINALEmergency_Dept_IB_10 -15-12-PDF.pdf. 2011. 278. Masters, I.B., et al., Airway sizes and proportions in children quantified by a video- bronchoscopic technique. BMC Pulm Med, 2006. 6: p. 5. 279. Bhogal, S., R. Zemek, and F.M. Ducharme, Written action plans for asthma in children. Cochrane Database Syst Rev, 2006(3): p. CD005306. 225 280. Yin, H.S., et al., Readability, suitability, and characteristics of asthma action plans: examination of factors that may impair understanding. Pediatrics. 131(1): p. e116-26. 281. Vonk, J.M., et al., Risk factors associated with the presence of irreversible airflow limitation and reduced transfer coefficient in patients with asthma after 26 years of follow up. Thorax, 2003. 58(4): p. 322-7. 282. Bai, T.R., et al., Severe exacerbations predict excess lung function decline in asthma. Eur Respir J, 2007. 30(3): p. 452-6. 283. Koehoorn, M., et al., Population-based surveillance of asthma among workers in British Columbia, Canada. Chronic Dis Inj Can. 33(2): p. 88-94. 284. Wong, S.T. and S. Regan, Patient perspectives on primary health care in rural communities: effects of geography on access, continuity and efficiency. Rural Remote Health, 2009. 9(1): p. 1142. 285. Miller, J., et al., Developing strategies to enhance health services research capacity in a predominantly rural Canadian health authority. Rural Remote Health, 2009. 9(4): p. 1266. 286. Awadh Behbehani, N., A. Grunfeld, and J.M. FitzGerald, Health care costs associated with acute asthma: a prospective economic analysis. Can Respir J, 1999. 6(6): p. 521-5. 287. Martin, A.J., et al., The natural history of childhood asthma to adult life. Br Med J, 1980. 280(6229): p. 1397-1400. 288. De Marco, R., et al., Incidence and remission of asthma: a retrospective study on the natural history of asthma in Italy. J Allergy Clin Immunol, 2002. 110(2): p. 228-35. 289. Phelan, P.D., C.F. Robertson, and A. Olinsky, The Melbourne Asthma Study: 1964- 1999. J Allergy Clin Immunol, 2002. 109(2): p. 189-94. 290. Burgess, J.A., et al., Factors influencing asthma remission: a longitudinal study from childhood to middle age. Thorax. 66(6): p. 508-13. 291. de Benedictis, D. and A. Bush, The challenge of asthma in adolescence. Pediatric Pulmonology, 2007. 42(8): p. 683-692. 292. McQuaid, E.L., et al., Medication adherence in pediatric asthma: reasoning, responsibility, and behavior. J Pediatr Psychol, 2003. 28(5): p. 323-33. 293. DiMatteo, M.R., et al., Patient adherence and medical treatment outcomes: a meta- analysis. Med Care, 2002. 40(9): p. 794-811. 294. Roorda, R.J., Prognostic factors for the outcome of childhood asthma in adolescence. Thorax, 1996. 51 Suppl 1: p. S7-12. 295. Perez-Yarza, E.G., Adolescent asthma. Introduction. Thorax, 1996. 51 Suppl 1: p. S1. 296. Price, J.F., Issues in adolescent asthma: what are the needs? Thorax, 1996. 51 Suppl 1: p. S13-7. 297. Shifren, A., et al., Mechanisms of remodeling in asthmatic airways. J Allergy (Cairo). 2012: p. 316049. 298. Rasmussen, F., et al., Risk factors for airway remodeling in asthma manifested by a low postbronchodilator FEV1/vital capacity ratio: a longitudinal population study from childhood to adulthood. Am J Respir Crit Care Med, 2002. 165(11): p. 1480-8. 299. Anis, A.H., et al., Double trouble: impact of inappropriate use of asthma medication on the use of health care resources. CMAJ: Canadian Medical Association Journal = Journal De L'association Medicale Canadienne, 2001. 164(5): p. 625-631. 226 300. Smith, L.A., et al., Modifiable risk factors for suboptimal control and controller medication underuse among children with asthma. Pediatrics, 2008. 122(4): p. 760-9. 301. Gustafsson, P.M., et al., Poor asthma control in children: evidence from epidemiological surveys and implications for clinical practice. Int J Clin Pract, 2006. 60(3): p. 321-34. 302. Inhaler technique in adults with asthma or COPD. Available at: http://www.nationalasthma.org.au/uploads/content/237- Inhaler_technique_in_adults_with_asthma_or_COPD.pdf, National Asthma Council Australia. 303. Al-Jahdali, H., et al., Improper inhaler technique is associated with poor asthma control and frequent emergency department visits. Allergy Asthma Clin Immunol. 9(1): p. 8. 304. Papi, A., et al., Inhaler devices for asthma: a call for action in a neglected field. Eur Respir J. 37(5): p. 982-5. 305. Giraud, V., F.A. Allaert, and N. Roche, Inhaler technique and asthma: feasability and acceptability of training by pharmacists. Respir Med. 105(12): p. 1815-22. 306. Williams, L.K., et al., Quantifying the proportion of severe asthma exacerbations attributable to inhaled corticosteroid nonadherence. J Allergy Clin Immunol. 128(6): p. 1185-1191 e2. 307. Hyland, M.E., et al., Guidelines versus practice: UK asthma nurses often recommend intermittent, symptom-driven use of inhaled corticosteroids. Prim Care Respir J, 2009. 18(2): p. 114-7. 308. Boulet, L.P., Once-daily inhaled corticosteroids for the treatment of asthma. Curr Opin Pulm Med, 2004. 10(1): p. 15-21. 309. Boushey, H.A., et al., Daily versus as-needed corticosteroids for mild persistent asthma. N Engl J Med, 2005. 352(15): p. 1519-28. 310. O'Connor, G.T. and J. Reibman, Inhaled corticosteroid dose adjustment in mild persistent asthma. Jama. 308(10): p. 1036-7. 311. Suissa, S. and E. Garbe, Primer: administrative health databases in observational studies of drug effects--advantages and disadvantages. Nat Clin Pract Rheumatol, 2007. 3(12): p. 725-32. 312. Colice, G.L., et al., Costs and resource use of mild persistent asthma patients initiated on controller therapy. J Asthma, 2008. 45(4): p. 293-9. 313. Lougheed, M.D., et al., Impacts of a provincial asthma guidelines continuing medical education project: The Ontario Asthma Plan of Action's Provider Education in Asthma Care Project. Can Respir J, 2007. 14(2): p. 111-7. 314. Forsetlund, L., et al., Randomised controlled trial of a theoretically grounded tailored intervention to diffuse evidence-based public health practice [ISRCTN23257060]. BMC Med Educ, 2003. 3: p. 2. 227 228 APPENDIX 229 Appendix A Chapter 2 Additional Tables Table A.1 Percentage of patients who were dispensed short-acting bronchodilators in asthma patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of pts with SABA % No. with Asthma No. of pts with SABA % No. with Asthma No. of pts with SABA % No. with Asthma No. of pts with SABA % Overall 5-11 yrs 12,594 9,504 75 11,910 8,756 74 11,943 9,740 82 -5.2 2.48 8.07 12-18 yrs 8,194 6,972 85 8,152 6,478 79 8,154 6,620 81 -0.5 -5.05 -4.58 19-34 yrs 18,099 15,241 84 17,228 13,858 80 18,446 14,802 80 1.9 -2.88 -4.71 35-55 yrs 30,350 23,467 77 39,864 28,731 72 48,435 35,752 74 59.6 52.35 -4.54 All ages a,d 69,237 55,184 80 77,154 57,823 75 86,978 66,914 77 25.6 21.26 -3.48* ,b Female 5-11 yrs 4,772 3,505 73 4,435 3188 72 4,354 3516 81 -8.8 0.31 9.94 12-18 yrs 4,027 3,448 86 3,902 3112 80 3,615 2963 82 -10.2 -14.07 -4.27 19-34 yrs 10,621 8,861 83 9,966 7920 79 10,527 8424 80 -0.9 -4.93 -4.08 35-55 yrs 18,081 13,700 76 23,998 17030 71 28,859 21263 74 59.6 55.20 -2.76 All ages 37,501 29,514 79 42,301 31,250 74 47,355 36166 76 26.3 22.54 -2.96 e Male 5-11 yrs 7,822 5,999 77 7,475 5568 74 7,589 6224 82 -3 3.75 6.94 12-18 yrs 4,167 3,524 85 4,250 3366 79 4,539 3657 81 8.9 3.77 -4.73 19-34 yrs 7,478 6,380 85 7,262 5938 82 7,918 6378 81 5.9 -0.03 -5.59 35-55 yrs 12,265 9,767 80 15,865 11701 74 19,576 14489 74 59.6 48.35 -7.06 All ages 31,732 25,670 81 34,852 26,573 76 39,622 30,748 78 24.9 19.78 -4.07 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. 230 b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population; P < 0.001 for comparison of rates between gender. 231 Table A.2 Percentage of patients who were dispensed ICS in asthma patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of pts with ICS % No. with Asthma No. of pts with ICS % No. with Asthma No. of pts with ICS % No. with Asthma No. of pts with ICS % Overall 5-11 yrs 12,594 8,304 66 11,910 8,136 68 11,943 8,851 74 -5.2 6.59 12.40 12-18 yrs 8,194 5,072 62 8,152 4,484 55 8,154 4,331 53 -0.5 -14.61 -14.19 19-34 yrs 18,099 11,011 61 17,228 8,134 47 18,446 7,376 40 1.9 -33.01 -34.27 35-55 yrs 30,350 19,628 65 39,864 19,800 50 48,435 18,958 39 59.6 -3.41 -39.48 All ages a,d 69,237 44,017 64 77,154 40,554 53 86,978 39,516 45 25.6 -10.23 -28.54* ,b Female 5-11 yrs 4,772 3,108 65 4,435 2,970 67 4,354 3,172 73 -8.8 2.06 11.86 12-18 yrs 4,027 2,492 62 3,902 2,134 55 3,615 1,907 53 -10.2 -23.48 -14.75 19-34 yrs 10,621 6,742 63 9,966 4,964 50 10,527 4,399 42 -0.9 -34.75 -34.17 35-55 yrs 18,081 11,932 66 23,998 12,328 51 28,859 11,757 41 59.6 -1.47 -38.27 All ages 37,501 24,274 65 42,301 22,396 53 47,355 21,235 45 26.3 -12.52 -30.72 e Male 5-11 yrs 7,822 5,196 66 7,475 5,166 69 7,589 5,679 75 -3.0 9.30 12.65 12-18 yrs 4,167 2,580 62 4,250 2,350 55 4,539 2,424 53 8.9 -6.05 -13.75 19-34 yrs 7,478 4,269 57 7,262 3,170 44 7,918 2,977 38 5.9 -30.26 -34.14 35-55 yrs 12,265 7,696 63 15,865 7,472 47 19,576 7,201 37 59.6 -6.43 -41.38 All ages 31,732 19,741 62 34,852 18,158 52 39,622 18,281 46 24.9 -7.40 -25.84 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. 232 d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 233 Table A.3 Percentage of patients who were dispensed LABA or LABA/ICS combined inhalers in asthma patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 No. with Asthma No. of pts with LABA % No. with Asthma No. of pts with LABA % No. with Asthma No. of pts with LABA % Overall 5-11 yrs 12,594 39 0.31 11,910 1,238 10.39 11,943 1,198 10.03 12-18 yrs 8,194 109 1.33 8,152 1,489 18.27 8,154 2,159 26.48 19-34 yrs 18,099 302 1.67 17,228 3,396 19.71 18,446 5,970 32.36 35-55 yrs 30,350 857 2.82 39,864 10,321 25.89 48,435 18,997 39.22 All ages a,d 69,237 1307 1.89 77,154 16,444 21.31 86,978 28,324 32.56* ,b Female 5-11 yrs 4,772 16 0.34 4,435 446 10.06 4,354 408 9.37 12-18 yrs 4,027 71 1.76 3,902 727 18.63 3,615 935 25.86 19-34 yrs 10,621 199 1.87 9,966 1,998 20.05 10,527 3,433 32.61 35-55 yrs 18,081 492 2.72 23,998 6,241 26.01 28,859 11,261 39.02 All ages 37,501 778 2.07 42,301 9,412 22.25 47,355 16,037 33.87 Male 5-11 yrs 7,822 23 0.29 7,475 792 10.60 7,589 790 10.41 12-18 yrs 4,167 38 0.91 4,250 762 17.93 4,539 1,224 26.97 19-34 yrs 7,478 103 1.38 7,262 1,398 19.25 7,918 2,537 32.04 35-55 yrs 12,265 365 2.98 15,865 4,080 25.72 19,576 7,736 39.52 All ages 31,732 529 1.67 34,852 7,032 20.18 39,622 12,287 31.01 *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. 234 b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population. 235 Table A.4 Percentage of patients who were dispensed leukotriene receptor antagonists in asthma patients aged 5-55 years in British Columbia, in 1997, 2003 and 2009 Group 1997 2003 2009 No. with Asthma No. of pts with LTRA % No. with Asthma No. of pts with LTRA % No. with Asthma No. of pts with LTRA % Overall 5-11 yrs 12,594 8 0.06 11,910 1319 11.07 11,943 1,805 15.11 12-18 yrs 8,194 47 0.57 8,152 688 8.44 8,154 925 11.34 19-34 yrs 18,099 146 0.81 17,228 707 4.10 18,446 1,007 5.46 35-55 yrs 30,350 441 1.45 39,864 2318 5.81 48,435 3,183 6.57 All ages a,d 69,237 642 0.93 77,154 5,032 6.52 86,978 6,920 7.96* ,b Female 5-11 yrs 4,772 3 0 4,435 444 10.01 4,354 626 14.38 12-18 yrs 4,027 26 0 3,902 336 8.61 3,615 356 9.85 19-34 yrs 10,621 67 0 9,966 437 4.38 10,527 622 5.91 35-55 yrs 18,081 245 0 23,998 1502 6.26 28,859 2,104 7.29 All ages 37,501 332 0 42,301 2,719 6.43 47,355 16,037 33.87 Male 5-11 yrs 7,822 5 0 7,475 875 11.71 7,589 1,179 15.54 12-18 yrs 4,167 22 0 4,250 352 8.28 4,539 569 12.54 19-34 yrs 7,478 81 0 7,262 270 3.72 7,918 385 4.86 35-55 yrs 12,265 205 0 15,865 816 5.14 19,576 1,079 5.51 All ages 31,732 310 0 34,852 3,708 10.64 39,622 3,212 8.11 *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. b Rates between age groups are standardized by gender to the 2009 B.C. population. 236 d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population; P < 0.001 for comparison of rates between gender. 237 Table A.5 Percentage of patients who were dispensed other asthma medications in asthma patients aged 5-55 years in British Columbia, in 1996, 2003 and 2009 Group 1996 2003 2009 % change from 1996 to 2009 No. with Asthma No. of pts with other Rx % No. with Asthma No. of pts with other Rx % No. with Asthma No. of pts with other Rx % No. with Asthma No. of pts with other Rx % Overall 5-11 yrs 12,594 2531 20.10 11,910 579 4.86 11,943 255 2.14 -5.2 -89.92 -89.38 12-18 yrs 8,194 886 10.81 8,152 235 2.88 8,154 119 1.46 -0.5 -86.57 -86.50 19-34 yrs 18,099 1169 6.46 17,228 259 1.50 18,446 126 0.68 1.9 -89.22 -89.42 35-55 yrs 30,350 3418 11.26 39,864 1279 3.21 48,435 712 1.47 59.6 -79.17 -86.95 All ages a,d 69,237 8004 11.56 77,154 2,352 3.05 86,978 1,212 1.39 25.6 -84.86 -87.95* ,b Female 5-11 yrs 4,772 985 20.64 4,435 189 4.26 4,354 84 1.93 -8.8 -91.47 -90.65 12-18 yrs 4,027 439 10.90 3,902 109 2.79 3,615 49 1.36 -10.2 -88.84 -87.57 19-34 yrs 10,621 715 6.73 9,966 162 1.63 10,527 72 0.68 -0.9 -89.93 -89.84 35-55 yrs 18,081 2108 11.66 23,998 757 3.15 28,859 395 1.37 59.6 -81.26 -88.26 All ages 37,501 4,247 11.33 42,301 1,217 2.88 47,355 600 1.27 26.3 -85.87 -88.81 e Male 5-11 yrs 7,822 1546 19.76 7,475 390 5.22 7,589 171 2.25 -3.0 -88.94 -88.60 12-18 yrs 4,167 447 10.73 4,250 126 2.96 4,539 70 1.54 8.9 -84.34 -85.62 19-34 yrs 7,478 454 6.07 7,262 97 1.34 7,918 54 0.68 5.9 -88.11 -88.77 35-55 yrs 12,265 1310 10.68 15,865 522 3.29 19,576 317 1.62 59.6 -75.80 -84.84 All ages 31,732 3,757 11.84 34,852 1,135 3.26 39,622 3,212 8.11 24.9 -14.51 -31.5 e *P < 0.001 for comparison of rates over time, from 1996 to 2009. a Rates are standardized to the 2009 British Columbia population. 238 b Rates between age groups are standardized by gender to the 2009 B.C. population. d P < 0.001 for comparison of rates between age groups. e Rates between gender are standardized by age to the 2009 B.C. population; P =0.14 for comparison of rates between gender. Other asthma medications include anti-IgE agents, methyxanthines and mast cell stabilizers. 239 Table A.6 Family physician visits for asthma per patient 5-55 years of age by health service delivery areas, 1996 - 2009 Health Services Delivery Areas Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 East Kootenay 0.99 1.07 1.16 1.06 1.17 0.98 1.02 1.04 0.99 1.02 0.91 0.89 0.96 0.95 Kootenay Boundary 1.17 1.26 1.19 1.12 1.00 0.90 0.94 0.90 0.88 0.89 0.79 0.83 0.79 0.80 Okanagan 1.39 1.38 1.31 1.28 1.25 1.28 1.34 1.30 1.23 1.30 1.26 1.24 1.18 1.12 Thompson cariboo Shuswap 1.44 1.42 1.38 1.32 1.28 1.25 1.28 1.28 1.17 1.11 1.12 1.03 0.98 0.92 Fraser East 1.49 1.50 1.57 1.50 1.44 1.46 1.42 1.42 1.48 1.47 1.37 1.36 1.30 1.24 Fraser North 1.47 1.40 1.39 1.40 1.37 1.40 1.41 1.42 1.41 1.44 1.34 1.32 1.28 1.30 Fraser South 1.62 1.63 1.55 1.50 1.49 1.46 1.49 1.48 1.45 1.47 1.31 1.32 1.28 1.26 Richmond 1.27 1.35 1.28 1.36 1.29 1.30 1.28 1.28 1.24 1.50 1.29 1.35 1.27 1.32 Vancouver 1.38 1.35 1.34 1.32 1.31 1.29 1.41 1.41 1.40 1.38 1.27 1.27 1.23 1.29 North Shore Coast Garibaldi 1.01 1.06 0.98 0.96 0.96 0.98 1.01 1.04 1.05 1.12 1.11 1.09 1.12 1.17 South Vancouver Island 1.31 1.39 1.29 1.26 1.33 1.32 1.35 1.33 1.32 1.36 1.28 1.33 1.33 1.28 Central Vancouver Island 1.48 1.49 1.43 1.43 1.43 1.39 1.28 1.30 1.27 1.31 1.24 1.23 1.15 1.12 North Vancouver Island 1.50 1.49 1.42 1.49 1.44 1.34 1.25 1.25 1.24 1.29 1.23 1.23 1.13 1.19 Northwest 1.33 1.33 1.25 1.15 1.10 1.01 0.89 0.89 0.97 0.99 0.88 0.83 0.77 0.69 Northern Interior 1.66 1.58 1.50 1.45 1.42 1.31 1.29 1.25 1.21 1.21 1.15 1.18 1.11 1.02 Northeast 1.27 1.16 1.14 1.21 1.21 1.14 1.13 1.02 1.13 1.07 1.00 0.96 0.95 0.91 240 Table A.7 Specialist visits for asthma per 100 patients 5-55 years of age by health service delivery areas, 1996 - 2009 Health Services Delivery Areas Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 East Kootenay 11.87 13.58 15.82 10.74 4.639 4.10 4.498 4.459 1.99 2.156 2.68 4.137 3.67 4.34 Kootenay Boundary 27.53 30.39 30.7 24.92 24.75 20.03 13.14 14.37 14.77 16.96 13.10 12.53 10.56 10.43 Okanagan 40.71 37.31 40.22 36.29 34.62 30.56 27.41 22.59 20.00 21.11 24.88 23.84 24.98 19.65 Thompson cariboo Shuswap 17.19 17.58 19.36 16.05 16.08 17.85 14.80 11.42 9.78 10.56 15.12 12.31 12.47 11.57 Fraser East 23.39 22.30 21.23 21.18 19.15 16.93 15.33 14.57 13.46 14.98 16.41 15.92 14.35 11.58 Fraser North 44.33 38.59 35.47 36.86 35.10 33.20 39.95 33.03 30.66 32.87 35.07 33.44 31.04 27.28 Fraser South 36.28 36.39 37.39 41.27 39.81 35.51 38.21 35.50 35.98 40.85 47.10 46.64 43.83 35.96 Richmond 47.46 48.31 48.58 42.52 33.64 28.46 34.29 36.68 34.43 35.81 43.08 38.18 41.05 36.33 Vancouver 49.53 49.52 50.03 52.78 48.29 42.66 44.29 37.80 38.50 43.73 45.36 39.88 37.90 33.22 North Shore Coast Garibaldi 35.62 36.19 36.06 30.81 29.68 25.72 25.08 22.25 21.31 22.51 26.13 24.93 21.47 21.71 South Vancouver Island 25.30 22.67 21.24 23.47 21.91 16.78 16.16 18.57 20.97 23.31 23.33 22.72 22.99 24.20 Central Vancouver Island 24.71 23.25 18.42 17.22 18.48 21.47 20.42 18.90 17.73 14.34 11.76 10.76 9.39 9.132 North Vancouver Island 18.88 19.34 19.27 16.00 12.29 8.40 13.84 20.21 18.05 18.40 17.09 15.88 13.99 10.34 Northwest 18.66 16.75 11.28 8.17 7.31 8.37 6.89 6.20 7.45 5.42 7.43 7.34 8.23 8.80 Northern Interior 21.82 22.65 19.42 17.79 16.81 15.49 16.81 13.97 14.59 15.03 18.84 15.54 15.81 14.50 Northeast 7.50 7.24 4.18 3.61 6.83 3.93 3.17 3.61 1.40 1.55 2.051 1.60 2.261823 2.746 241 Table A.8 Hospital admissions for asthma per 100 patients 5-55 years of age by health service delivery areas, 1996 - 2009 Health Services Delivery Areas Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 East Kootenay 3.43 3.11 4.09 2.77 2.41 2.05 2.42 2.39 3.21 3.36 3.55 1.58 1.43 2.01 Kootenay Boundary 4.42 4.65 2.389 2.30 3.18 2.25 1.60 1.78 2.16 2.22 0.81 1.45 1.29 1.29 Okanagan 3.45 2.78 2.98 3.09 2.46 1.96 1.80 1.81 1.92 1.87 1.48 1.05 0.96 0.98 Thompson cariboo Shuswap 4.59 4.09 2.90 2.74 2.50 2.38 1.86 1.62 1.97 1.95 1.44 0.92 0.88 1.00 Fraser East 2.85 2.80 2.81 3.12 2.99 1.84 1.83 1.69 1.92 1.68 1.59 1.37 1.22 1.13 Fraser North 2.50 2.41 2.32 2.21 1.55 1.39 1.29 1.176 1.32 1.09 0.90 0.77 0.82 0.64 Fraser South 2.42 2.53 2.05 2.14 1.66 1.24 1.02 1.21 1.26 1.05 1.08 0.96 0.98 1.08 Richmond 2.52 2.97 2.09 2.14 1.86 0.99 1.27 1.37 1.53 1.70 2.11 1.14 0.94 0.95 Vancouver 2.63 2.75 2.64 2.13 2.34 1.73 1.25 1.18 1.31 1.16 1.18 0.97 0.80 0.74 North Shore Coast Garibaldi 2.61 3.16 2.72 2.45 1.67 1.33 1.23 1.06 0.94 1.181 1.37 0.96 0.77 0.91 South Vancouver Island 3.14 3.69 3.67 2.65 2.31 1.63 1.31 1.20 1.25 1.54 1.36 0.96 0.82 0.92 Central Vancouver Island 3.03 3.14 2.10 2.18 2.64 1.48 1.44 1.06 1.39 1.43 1.15 1.00 1.04 0.99 North Vancouver Island 3.06 3.00 2.61 2.41 2.08 1.43 0.90 1.03 1.67 0.97 0.99 0.73 0.79 0.33 Northwest 3.82 3.78 3.55 2.86 2.35 2.83 1.61 2.20 1.73 2.14 1.54 1.83 1.20 1.32 Northern Interior 3.80 3.63 2.50 2.48 2.99 1.89 1.48 1.52 1.40 1.47 1.29 1.26 0.88 0.74 Northeast 3.7 3.62 2.36 4.52 2.85 1.28 1.76 1.53 1.63 1.18 1.10 1.04 0.75 0.55 242 Table A.9 ED visits for asthma per 100 patients 5-55 years of age by health service delivery areas, 1996 - 2009 Health Services Delivery Areas Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 East Kootenay 36.68 30.52 36.62 30.74 30.41 24.33 31.31 29.31 27.73 24.83 27.19 20.51 17.82 19.28 Kootenay Boundary 35.177 40.31 33.09 23.78 28.26 18.86 20.78 19.70 21.08 23.61 19.85 22.01 18.39 19.48 Okanagan 31.77 30.65 31.00 28.31 26.40 22.20 19.13 22.08 19.49 18.65 12.55 11.25 10.84 9.136 Thompson cariboo Shuswap 38.57 37.42 34.01 30.97 30.08 27.48 26.85 24.56 23.79 20.13 18.57 13.55 11.59 9.13 Fraser East 30.80 31.71 29.69 27.08 26.79 21.97 19.24 18.96 19.79 20.06 15.76 13.78 11.15 11.38 Fraser North 22.91 19.60 19.37 16.53 13.18 11.93 8.51 9.11 9.66 8.93 4.58 3.19 3.26 3.31 Fraser South 24.17 24.30 20.9434 19.56 17.70 14.45 12.12 12.89 11.39 10.53 8.03 6.90 6.41 6.18 Richmond 25.34 25.18 19.66 21.01 18.05 12.93 11.32 13.70 11.43 13.25 5.50 1.01 1.26 1.57 Vancouver 14.52 15.25 14.76 12.54 13.31 11.40 8.91 10.72 8.21 7.25 3.28 2.03 2.24 1.73 North Shore Coast Garibaldi 28.41 30.33 31.28 28.57 23.82 22.25 19.25 19.11 17.51 19.49 13.273 8.52 8.41 9.19 South Vancouver Island 31.64 31.51 31.20 25.09 22.41 14.20 12.67 10.73 6.17 6.74 4.01 2.00 1.85 2.02 Central Vancouver Island 26.50 28.73 25.00 24.29 23.76 17.99 20.31 22.03 16.69 10.96 8.34 7.44 6.59 6.64 North Vancouver Island 30.536 31.581 26.73 25.73 25.43 22.913 21.98 19.17 18.72 16.83 10.69 10.54 8.63 6.96 Northwest 34.37 33.63 32.20 27.77 28.11 27.93 24.88 26.92 31.87 28.42 21.82 24.03 25.94 21.34 Northern Interior 35.13 32.76 27.83 26.27 26.55 26.77 26.60 22.53 16.36 17.21 12.52 10.39 6.96 10.04 Northeast 38.8 41.45 38.33 42.28 35.00 36.46 37.06 35.34 39.14 36.50 30.29 31.94 30.50 25.07 "@en . "Thesis/Dissertation"@en . "2013-11"@en . "10.14288/1.0074010"@en . "eng"@en . "Pharmaceutical Sciences"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "Attribution-NonCommercial 3.0 Unported"@en . "http://creativecommons.org/licenses/by-nc/3.0/"@en . "Graduate"@en . "Asthma drug regimen optimality and health services utilization : a population-based analysis in British Columbia"@en . "Text"@en . "http://hdl.handle.net/2429/44726"@en .