UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of English proficiency on length of stay after isolated cardiac bypass surgery Tang, Erin Wei-Man 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2014_spring_tang_erin.pdf [ 3.36MB ]
JSON: 24-1.0167296.json
JSON-LD: 24-1.0167296-ld.json
RDF/XML (Pretty): 24-1.0167296-rdf.xml
RDF/JSON: 24-1.0167296-rdf.json
Turtle: 24-1.0167296-turtle.txt
N-Triples: 24-1.0167296-rdf-ntriples.txt
Original Record: 24-1.0167296-source.json
Full Text

Full Text

   THE EFFECTS OF ENGLISH PROFICIENCY ON LENGTH OF STAY AFTER  ISOLATED CARDIAC BYPASS SURGERY  by  ERIN WEI-MAN TANG B.Sc.N., The University of British Columbia, 2000  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIRMENTS FOR THE DEGREE OF   MASTER OF SCIENCE IN NURSING  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Nursing)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)    April 2014 © Erin Wei-Man Tang, 2014  ii  Abstract Nurses often face the challenge of communicating information to patients who may not fully understand English.  Limited English proficient (LEP) patients are at higher risk of misunderstanding health teaching, leading to impediments in their recovery.  This study examined whether LEP patients have a prolonged length of stay (LOS) following coronary artery bypass graft (CABG) surgery when compared with English proficient (EP) patients.  The study also compared the LOS of the two groups of patients (LEP and EP) with the study hospital’s clinical pathway target. A retrospective chart audit was conducted of all patients undergoing isolated CABG, over a two-year period.  A screening tool was administered to determine each patient’s English proficiency based on documented information.  A data abstraction tool collected pertinent pre-, intra-, and post-operative health indicators.  691 (97.1%) of the 712 charts reviewed were eligible for inclusion.  The “limited or not English proficient” (LEP/NEP) patients had a median post-operative stay that was one day longer than that of the EP patients (7 days versus 6 days, p = .007).  The median LOS stay of all patients, irrespective of English proficiency, exceeded the clinical pathway target of 5 days.  In an unadjusted polynomial regression model, the LEP/NEP patients were found to be 2.2 times more likely to have a LOS of 9+ days (95% [CI 1.5, 7.2]).  The multivariate model revealed that post-operative infection was the strongest predictor of LOS, and language proficiency was no longer statistically significant.  Post-operative infections fully mediated the relationship between language proficiency and LOS.  Relative to having a < 6-day LOS, patients with infections had greater odds of staying longer: ORadj = 3.30 (95% CI [1.51, 7.19]) for 6 days; ORadj = 4.37 (95 % CI [2.09, 9.14]) for 7-8 days; and ORadj = 12.4 (95% iii  CI [6.03, 25.48]) for 9+ days.  Higher infection rates were observed in LEP (29.1%) versus EP patients (16.7%) (OR = 2.05 (95% CI [1.27, 3.30])). The high infection rates suggest that further research is warranted to examine the mechanisms through which post-operative infections are acquired in CABG patients, and to better understand the role of language proficiency.     iv  Preface This thesis is original, unpublished, independent work of the author, Erin Wei-Man Tang. The research reported in Chapters 3 and 4 was approved by the UBC Clinical Ethics Review Board (CREB) (Certificate number H13-00777).  Ethical approval was also obtained from the Vancouver Coastal Health Authority Research Institute (Certificate number V13-00777). The thesis is based on the ExPRESS study, the work of a small collaborative group of cardiac surgical nurses.  I was the lead investigator, responsible for all major areas of concept formation, data collection and analysis, as well as manuscript composition.  None of the text of the thesis is taken directly from previously published or collaborative works. The Limited English Proficiency (LEP) tool and data abstraction forms discussed in Chapter 3 were designed in collaboration with J. Go., A. Kwok, and B. Leung.  All data analyses described in Chapter 4 and the discussion in Chapter 5 are my original work.      v  Table of Contents Abstract ............................................................................................................................... ii Preface................................................................................................................................ iv Table of Contents ................................................................................................................ v List of Tables ................................................................................................................... viii List of Figures .................................................................................................................... ix Acknowledgements ............................................................................................................. x Dedication .......................................................................................................................... xi Chapter One: Introduction .................................................................................................. 1 Limited English Proficiency ........................................................................................... 2 Purpose of the Study ....................................................................................................... 4 Chapter Two: Review of the Literature .............................................................................. 5 Background ..................................................................................................................... 5 Coronary artery disease. .............................................................................................. 5 Clinical pathways and length of stay for coronary artery bypass graft surgery patients. ........................................................................................................... 6 Limited English Proficiency and Patients’ Outcomes .................................................... 9 Effects of limited English proficiency on length of stay. ......................................... 10 Effects of limited English proficiency on adverse events. ........................................ 13 Effects of professional interpretation on patients’ outcomes. ................................... 15 Effects of limited English proficiency on patients’ experiences. .............................. 19 Summary. .................................................................................................................. 23 Theoretical Background ................................................................................................ 24 Health communication. ............................................................................................. 24 Language barriers. ..................................................................................................... 28 Study variables and their relationship to the theoretical framework. ....................... 31 Summary ....................................................................................................................... 32 Chapter Three: Methods ................................................................................................... 34 Research Questions ....................................................................................................... 34 Hypothesis .................................................................................................................... 34 Research Design ........................................................................................................... 35 vi  Inclusion criteria. ...................................................................................................... 36 Exclusion criteria. ..................................................................................................... 36 Research Methods ......................................................................................................... 37 Research protocol. ..................................................................................................... 37 Measurement. ............................................................................................................ 37 Data Quality .................................................................................................................. 42 Limited English proficiency screening tool. ............................................................. 42 Data abstraction tool. ................................................................................................ 44 Data Collection Procedures .......................................................................................... 44 Data Analysis ................................................................................................................ 45 Data entry and cleaning............................................................................................. 45 Determining sample size. .......................................................................................... 46 Statistical analysis. .................................................................................................... 47 Ethical Considerations .................................................................................................. 48 Chapter Four: Results ....................................................................................................... 50 Characteristics of the Sample ....................................................................................... 50 Language proficiency. ............................................................................................... 51 Demographics. .......................................................................................................... 52 Pre-operative characteristics. .................................................................................... 54 Intra-operative characteristics. .................................................................................. 59 Post-operative characteristics. ................................................................................... 60 Length of Stay ............................................................................................................... 62 The association between the patients’ characteristics and their post-operative length of stay. ................................................................................................ 66 Multivariate Analyses: English Proficiency and Length of Stay .................................. 70 Chapter Five: Discussion .................................................................................................. 74 Overview of the Findings ............................................................................................. 74 The Findings in Relation to the Literature .................................................................... 75 Differences in the patients’ characteristics. .............................................................. 75 Factors affecting the relationship between English proficiency and length of stay. ............................................................................................................... 81 English proficiency, post-operative infection, and length of stay............................. 84 Other predictors of length of stay. ............................................................................ 88 Clinical pathway targets and length of stay. ............................................................. 89 Measurement of English proficiency. ....................................................................... 91 Methodological Limitations .......................................................................................... 92 Clinical Implications ..................................................................................................... 95 vii  Nursing practice. ....................................................................................................... 95 Policy.  ....................................................................................................................... 98 Recommendations for Future Research ...................................................................... 100 Conclusion .................................................................................................................. 103 References ....................................................................................................................... 104 Appendix A:  Clinical path – CABG +/- valve surgery: Interdisciplinary discharge planning tool ....................................................................................................... 116 Appendix B:  Clinical path – CABG +/- valve: Interdisciplinary discharge planning tool for 6+ days ................................................................................................... 124 Appendix C:  Limited English Proficiency Screening Tool ........................................... 125 Appendix D:  Data Abstraction Tool .............................................................................. 128 Appendix E:  Canadian Cardiovascular Society grading of angina pectoris and New York Heart Association Functional Classification ............................................. 131    viii  List of Tables  Table 1:  Sample size calculations ................................................................................................ 47 Table 2:  Languages spoken by the patients ................................................................................. 52 Table 3:  Demographic characteristics of the coronary artery bypass graft patients by English proficiency ...................................................................................................................... 54 Table 4:  Pre-operative health characteristics of coronary artery bypass graft patients by English proficiency ......................................................................................................... 57 Table 5:  Intra-operative characteristics of patients by English proficiency ................................ 60 Table 6:  Post-operative characteristics of coronary artery bypass graft patients by English proficiency ...................................................................................................................... 62 Table 7:  Length of stay, in percentiles, of all coronary artery bypass graft patients ................... 65 Table 8:  Length of stay, in percentiles, of all coronary artery bypass graft patients stratified by English proficiency .................................................................................................... 65 Table 9:  Post-operative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patient demographics ............................................ 67 Table 10:  Post-operative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patients’ pre-operative characteristics .................. 68 Table 11:  Postoperative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patients’ post-operative characteristics ................. 70 Table 12:  Multivariate polynomial regression model of post-operative length of stay on language proficiency and other patient characteristics ................................................... 72   ix  List of Figures Figure 1:  Model of Health Communication. .................................................................... 25 Figure 2:  Language proficiency and the relationship with length of stay as influenced by health communication. ................................................................... 32 Figure 3:  Review process flow chart for enrollment of subjects into study. ................... 51 Figure 4:  Post-operative LOS (in days) from CSICU admission to hospital discharge stratified by English proficiency ........................................................................... 63     x  Acknowledgements This thesis would not have been possible without the combined efforts of many Vancouver General Hospital healthcare team members.  In particular, partnerships with individuals in the health records department and cardiac surgical departments allowed this project to come to life.  My fellow research team members, Jeremy Go, Andrea Kwok, and Bonnie Leung, were essential to the success of the ExPRESS Study.  This project would not have been possible without the generous funding support provided through the Vancouver Coastal Health Research Challenge competition and the Vancouver General Hospital & University of British Columbia Hospital Foundation- School of Nursing Alumni Association.  I would also like to recognize the research assistance provided by the Vancouver Coastal Health Research Institute.  I am extremely thankful to these groups for supporting the development of research skills in nurses. Thank you to my thesis committee members, Dr. Sandra Lauck and Dr. Sabrina Wong. Their continuous dedication and guidance throughout this project greatly enriched the quality of my research and their insight encouraged me to step outside the limits of my thinking. I’d like to express my immense appreciation to my thesis chair, Dr. Pam Ratner, who supported me throughout this project.  Her unrelenting dedication, expertise, and assistance over the past year resulted in quality nursing research.  Such attention to detail and commitment to research has given me a new appreciation of what is necessary to produce excellent research.  Dr. Ratner has been instrumental in ensuring my own success in the Master of Nursing program.  I am extremely thankful for the opportunity to have worked with her.        xi  Dedication  To my work colleagues at VGH, your experiences are what have inspired me to pursue my master’s degree in nursing.   To Jeremy, Andrea, and Bonnie, your commitment to the ExPRESS Study allowed my thesis to become a reality.  I am indebted to you all.     To my family, always encouraging me to go further, and knowing before me, that I would.      To Ryan, for “riding shotgun” with me through life.      1  Chapter One: Introduction Canada is a multicultural and multilingual country with a population that speaks over 200 languages.  According to census data, of the 33.1 million people living in Canada in 2011, approximately 3.7 million (11.1%) reported speaking a language other than French or English at home (Statistics Canada, 2013).  The concentration of non-official-language speaking people in British Columbia (BC) was greater than the national rate with 15.4% of the population reporting that their language spoken most often at home was neither English nor French.  The census revealed that 12.9% of people living in the City of Vancouver were not able to conduct a conversation in English.  This multilingual context presents important challenges for the healthcare system.  Adequate and effective patient-provider communication may be compromised because of language barriers that may negatively affect the delivery of safe and effective health care (Divi, Koss, Schmaltz, & Loeb, 2007; Hudelson & Vilpert, 2009). 1   Literature from the US suggests that hospital patients with limited or no English proficiency suffer worse healthcare outcomes, such as more frequent adverse events, prolonged length of stays (LOS), and higher readmission rates when compared with English proficient (EP) patients (Martinez, 2010).  Despite such evidence emerging from the US, limited research has been dedicated to this topic within Canada.  Moreover, few researchers have assessed the financial costs of language barriers and their effects on health care, especially the effects of not addressing or mitigating the barriers (Jacobs, Shepard, Suaya, & Stone, 2004).  Further research is needed to better understand the effects of language proficiency on patient outcomes in Canada.  Understanding the relationships between limited English proficiency (LEP) and patient outcomes is critical to a comprehensive evaluation of current healthcare delivery models.  This research                                                  1 The term “language barriers” refers to the inability of a patient and healthcare provider to verbally communicate effectively due to the lack of a shared or common language.   2  may inform and guide the design of more effective interventions to improve health outcomes and healthcare resource access and utilization for patients that face language barriers.  On a national level, such research may highlight the potential health inequities resulting from language barriers.  This may have implications for healthcare policy to address the needs of Canada’s multi-ethnic population.  Limited English Proficiency Limited language proficiency in a country’s official language(s) has multi-faceted effects on immigrants’ health status (Ng, Pottie, & Spitzer, 2011).  Those individuals (and their communities) that are not proficient in the official language(s) are at risk of inadequate access to, and of receiving poorer quality, health care, which can ultimately affect their health status (Ng et al., 2011).  Epidemiological studies of language barriers and their effects on health outcomes systematically indicate that health and illness are associated with a linguistic gradient for language minority populations (Martinez, 2010).  People with limited English proficiency report poorer health status than their English proficient counterparts (Ng et al., 2011).  In Canada, limited English proficiency has been shown to be strongly associated with poorer self-rated health status, disease prevention activity, and access to healthcare services, including the receipt of mental health care (Ng et al., 2011). Health Canada has recognized four constituencies that may face barriers to health care because of having a non-official first language (Bowen, 2001).  These four groups include: (a) First Nations and Inuit communities, (b) newcomers to Canada (immigrants and refugees), (c) deaf persons, and (d) depending on the location of their residence, speakers of the official languages (i.e., French and English). 3  Despite this knowledge, there is a paucity of research related to the association between language barriers and healthcare access.  Historically, research in Canada regarding health inequities and healthcare access has centred on patterns of utilization by income status, rather than differences related to language (or culture or ethnicity).  This is a result of Canada’s unique history with establishing a universal system of health care, Medicare, and reflects past priorities to address financial barriers to access (Bowen, 2001).   With the expanding multicultural and multi-linguistic makeup of Canada’s population, the influence of cultural and language factors (especially patient-provider cultural and language incongruence) is gaining the interest of major stakeholders.  Specifically, the mosaic of linguistic diversity in BC poses specific challenges for the healthcare system.  Within the hospital setting, patients with LEP have tangible language communication barriers with healthcare providers.  Reduced ability or inability to communicate may lead to poorer health and healthcare outcomes, such as prolonged hospital LOS and higher readmission rates.  These outcomes can put further strain on an already taxed healthcare system.  Thus, it is in the interest of both patients and healthcare decision makers that the effects of LEP on the hospital LOS of patients is understood and alleviated. Coronary artery bypass grafting (CABG) surgery is one of the most commonly performed major surgical procedures in the world and accounts for the most resources expended of any cardiovascular procedure in cardiovascular medicine (Preedy & Watson, 2010).  Patients undergoing CABG surgery are an important population to consider with respect to the effects of LEP because patients’ recovery is dependent upon active participation in therapies and dynamic engagement with their healthcare professionals.  Thus, close inspection of whether LEP patients undergoing CABG surgery recover at the same rate as the mainstream population is of interest. 4  It is acknowledged that the importance of English proficiency cannot be fully understood without studying the influences of other relevant factors such as culture, ethnicity, and the health literacy of patients (Bowen, 2001).  This study is limited to describing the impact of patients’ capacity to communicate verbally with healthcare providers during their hospital stays when they are not English proficient.  It is not within the scope of this study to explore and address the multiple alternative forms of communication that also affect healthcare interactions.  Because the possession of a shared language is a prerequisite for communication between patients and healthcare providers, investigating the effects of patients’ English proficiency (EP) on their length of hospital stay will lend further understanding of this multifaceted phenomenon.  Purpose of the Study The purpose of this study was to investigate the relationship between LEP status and the LOS of post-operative elective CABG patients at a major tertiary care hospital.  The study was designed to determine whether patients with LEP had prolonged length of stays after CABG when compared with patients that were EP.  The study explored whether English language proficiency, after taking into consideration factors such as gender, age, and health status, was associated with differences in the length of the post-operative hospital stay.  The study also explored whether limited English language proficiency was associated with differences in post-operative LOS within the intensive care unit and within the cardiac ward.  The proportions of LEP and EP patients that achieved and did not achieve the outcomes of the expected length of stay were also investigated.      5   Chapter Two: Review of the Literature This review of the published literature addresses patients’ outcomes and experiences within the healthcare setting and their relationship with English proficiency.  The review is limited to scholarly works published between the years 1996 and 2013.  The objective was to explore the association and complexity of patient outcomes in relation to limited English proficiency.  The literature search was conducted with the academic databases, Cumulative Index of Nursing and Allied Health Literature (CINAHL), and the Medical Literature Analysis and Retrieval System Online (MEDLINE).  The terms “limited English proficiency,” “language barriers,” “interpreters,” “length of stay,” “patient care,” “health disparities,” “health status,” “quality of care,” “healthcare access,” “cardiac patients,” and “cardiac surgery” were used in various combinations and resulted in 37 articles being selected for this review. Background  Coronary artery disease. Coronary artery disease (CAD) is the pathologic process of atherosclerosis affecting the major blood vessels of the heart (Woods, 2010).  The accumulation of plaque within the inner lining of the blood vessels results in a narrowing of the vessel lumen.  As a result, the delivery of oxygenated blood to the heart is compromised (Woods, 2010).  The reduced blood flow may cause patients to experience chest pain, otherwise known as angina.  Individuals with CAD are at risk of myocardial infarction when the blockage of blood flow to the heart causes the heart muscle to be injured (Preedy & Watson, 2010).   Treatment options for patients with CAD are dependent on the extent of disease, symptom burden, and the individual patient’s comorbidities and preferences.  Surgical treatment to improve blood flow to the heart includes interventions such as a CABG (Woods, 2010).  This 6  surgery is offered to patients when their disease is not amenable to medical management or less invasive surgical options.  Today, CAD is the leading cause of death in Canada and worldwide (Cardiac Services BC, 2010).  Not surprising then, CABG surgery is one of the most commonly performed procedures in the world (Preedy & Watson, 2010).  The largest cross-sectional survey of cardiovascular risk profiles of Canada’s major ethnic groups found that patients with different ethnic backgrounds have distinct cardiovascular risks profiles (Chiu, Austin, Manuel, & Tu, 2010).  In certain ethnic groups, such as those of South Asian descent, the prevalence of CAD is higher than that of the general population (Chiu et al., 2010).  The combination of the prevalence of CABG surgery, the prevalence of the disease in certain ethnic groups, and the prevalence of large ethnic groups utilizing health care in BC gives impetus to investigating the impact of LEP in the CABG surgery population.   Clinical pathways and length of stay for coronary artery bypass graft surgery patients.  The high costs associated with cardiovascular diseases and procedures have motivated efforts towards standardizing clinical practice as a way of reducing the length of patients’ hospital stays and related healthcare costs while maintaining best practices and quality of care (Cohn, Rosborough, & Fernandez, 1997).  One approach includes the implementation of clinical pathways for CABG surgery.  These clinical pathways are defined as “multidisciplinary guidelines that list patient outcomes and also the major interventions performed by the healthcare team to promote efficient utilization of resources and delivery of cost-effective care within a prescribed length of stay” (Rumble, Jernigan, & Rudisill, 1996, p. 35).  Clinical pathways act as a framework to assist healthcare providers in providing coordinated and efficient care within a 7  predictable timeframe while ensuring quality of care (Cohn, Rosborough, & Fernandez, 1997).  The effectiveness of clinical pathways has been demonstrated through a small number of dedicated studies from the 1990s (Cohn et al., 1997; Rudisill, Phillips, & Payne, 1994; Rumble et al., 1996).  Remarkably though, some research has demonstrated that CABG pathways do not reduce the length of hospital stay in hospital when compared with conventional post-operative care (Pearson, Kleefield, Soukop, Cook, & Lee, 2001; Weingarten, 2001).  Pearson et al. (2001) compared the LOS of postoperative CABG patients in a hospital that had implemented a clinical pathway with that of three neighbouring hospitals that did not employ such a pathway.  Over a two-year period, they found that the comparison hospitals were similar in having achieved a reduction in their patients’ LOS, which suggested that other informal efforts driven by competitive pressures on local healthcare plans and the global attention given to LOS issues contributed to improving efficiencies.  Notwithstanding this contradictory evidence, clinical pathways for CABG surgery have been widely embraced by healthcare organizations and implemented worldwide.   In BC, 1,720 coronary artery bypass graft surgeries were performed in 2012 (Cardiac Services BC, 2013).  Approximately 25% of those 1,720 surgeries were performed at one very large tertiary hospital (Cardiac Services BC, 2013).  Like other healthcare organizations, the hospital’s clinical pathway for CABG surgery was developed and implemented to facilitate timely and efficient recovery from surgery (see Appendix A).  The expected and targeted LOS in this pathway is five days with the day of surgery being considered post-operative day zero.  In clinical practice though, a length of stay of up to eight days for patients of advanced age or with multiple comorbidities is often encountered (Rudisil, Phillips, & Payne, 1994).  Increasing age and multiple comorbidities are well accepted pre-operative predictors of extended length of stay 8  after coronary bypass surgery (Johnston, Goss, Malmgren, & Spertus, 2004).  This is reflected in the clinical practice established at the study hospital; an additional ‘extended’ CABG clinical pathway (see Appendix B) was created for post-operative CABG patients anticipated to require a stay of six to eight days.  This pathway has not been validated, but was implemented because it was recognized that the original clinical pathway did not address issues encountered with older and more complicated patients.  It had been observed that these patients typically had prolonged stays because of ongoing physiological complications that rendered them unready for discharge.  Any patient requiring a length of stay of longer than eight days is no longer considered to be on the CABG clinical pathway.  For the purposes of this study, the expected length of stay was defined as being no more than five days.    To meet the pathway targets and goals, healthcare providers are responsible for important patient teaching, such as teaching patients how to complete deep breathing exercises and sternal wound protection activities.  They also are expected to promote understanding and adherence to prescribed medications and other therapies.  In turn, patients are expected to play active roles in their recovery.  Their participation is recognized as vital to the achievement of the clinical pathway goals (Cohn et al., 1997).  Accordingly, these clinical pathways presuppose that all patients are able to comprehend and understand their prognosis, the recovery process, and their expected post-operative behaviour.  Such an assumption does not account for the possible contextual influences, such as language proficiency, that can significantly affect patients’ perceptions of their role in recovery from surgery.  At the study hospital, several educational materials, including pre-operative and post-operative teaching videos and discharge instructions had been produced in several languages to mitigate the language barriers in health communication with patients and their families.  Family members were welcome to stay with 9  patients on the ward, staff were often assigned to patients that were not EP and with whom they shared a language, and professional translators could be booked, if necessary, to assist in communication.  The LEP patient and healthcare providers’ inability to communicate with and understand each other can exert powerful influences on interpretation, diagnosis, and care.  Consequently, the increasing cultural and linguistic diversity of BC’s population may affect LEP patients’ progression through the clinical pathway, and ultimately prolong their LOS.  Limited English Proficiency and Patients’ Outcomes  Limited English proficiency is defined “as the limited ability or inability to speak, read, write, or understand the English language at a level that permits the person to interact effectively with healthcare providers or social service agencies” (Divi et al., 2007, p. 60).  Language barriers severely restrict communication between patients and their healthcare providers, potentially diminishing the overall quality of care provided to LEP patients contributing to health disparities between LEP patients and the mainstream patient population (Bartlett, Blais, Tamblyn, Clermont, & MacGibbon, 2008; Divi, 2007; Jacobs et al., 2003; Karliner, 2007).  Research from several disciplines addresses the importance of language and culture to health beliefs, behaviour, and patterns of communication (Bowen, 2001).  This literature review acknowledges the inextricable intertwining of these factors.  Although it is difficult to identify language as the exclusive culprit for communication difficulties between patients and their providers, language is the prerequisite for effective communication and understanding (Bowen, 2001).  The next section of this literature review explores the effects of LEP status on patients’ outcomes, specifically LOS and adverse events.  The effects of professional interpreters on LEP patients’ outcomes are also discussed.  Finally, LEP patients’ experiences are examined. 10  Effects of limited English proficiency on length of stay. In the hospital setting, language barriers may have severe clinical consequences for patients’ safety and care.  Patients with LEP may be unable to adequately communicate their needs, such as pain, which may lead to delayed management and care (Oliva, 2008).  Clinicians’ inability to adequately communicate with their LEP patients can lead to erroneous diagnoses and prescribing of treatments.  As a result, patients with LEP can have poorer adherence to treatment recommendations (Jacobs et al., 2004; Karliner, Jacobs, Chen, & Mutha, 2007; Oliva, 2008).   Length of stay is often utilized as an indirect measurement of patients’ outcomes and the quality of care provided in hospitals (Epps, 2004).  Length of stay is also an indicator used to assess the efficiency of healthcare delivery (Hyman, 2009).  Prolonged LOS is associated with poor health outcomes and complications, and has direct implications for healthcare costs (Epps, 2004).  Curtailing escalating costs in cardiac care has been a major focus of public policy to meet demand and reduce the pressure on the healthcare system.  Driving this focus are policy initiatives that address the number of patients on waiting lists and that maximize the capacity of hospital beds while still investing in research innovation, new technologies, and infrastructure (Cardiac Services BC, 2010).  There is a paucity of literature, however, that specifically examines the critical relationship between LEP and LOS after cardiac surgery.  There is, though, related literature focused on the effects of LEP on LOS in other patient populations.   Among the studies that have described the importance of English proficiency and its relationship with the LOS of inpatients, Baptiste et al. (2004) included 23 medical and surgical conditions in their analysis.  Their Canadian-based retrospective review of approximately 60,000 health records found that LEP patients had relatively longer hospital admissions for the following conditions: acute coronary syndromes and chest pain, CABG, stroke, craniotomy 11  procedures, diabetes mellitus, major intestinal and rectal procedures, and elective hip replacement.  The average LOS for these specific conditions ranged from 0.7 to 4.3 days longer than that of EP patients.  Of interest, this study found that the effect of mainstream English proficiency on LOS was greater for certain conditions than for others.  Specifically within the CABG population, LEP patients’ length of stay averaged 12.7 days as compared with 10.3 days for EP patients (odds ratio, OR = 1.07, 95% confidence interval, CI [1.03, 1.12]).  This finding was statistically significant although modest in magnitude in comparison with the likelihood of a prolonged length of stay for patients with LEP and stroke, acute coronary syndromes, or chest pain (OR = 1.29, 95% CI [1.23, 1.34]).  The researchers did not provide an explanation for the differences but noted that their study was unable to detect processes of care (such as complications or number of procedures) that could have accounted for the observed differences.    Baptiste et al. (2004) compared in-hospital mortality rates of LEP and EP patients.  Again, they found varying adjusted odds ratios for in-hospital cause-specific mortality but no clear pattern emerged.  Overall, the mortality rates were not found to be greater for LEP patients.  Levas, Cowden, and Dowd (2011) examined whether the English proficiency in United States (US) parents affected the LOS of their children admitted to hospital because of infections.  Their retrospective cohort study found that children with LEP parents had prolonged LOS of 1.6 days, on average (95% CI [1.1, 2.3]).  These children stayed 60% longer in hospital compared with the children of EP parents.  Levas et al. (2011) measured the referral rates to home health services for prolonged intravenous antibiotic therapy as a marker of healthcare utilization.  Patients receiving this kind of service typically use fewer resources (Freedman et al., 2006).  The decision of whether a patient will receive a home healthcare referral is dependent upon factors such as the family 12  members’ and providers’ perceptions of the ability to adequately manage home health care, the family members’ comfort level with patient care, insurance coverage, and availability of home healthcare services in the community (Levas et al., 2011).  LEP was significantly associated with a smaller number of home healthcare referrals (OR = 0.20, 95% CI [0.04, 0.80]).  This disparity persisted even after adjustment for relevant covariates such as insurance status, the presence of a primary care provider, and the patients’ comorbidities. A study from BC identified predictors of prolonged LOS for patients receiving hip and knee arthroplasty surgery at two hospitals.  Of those with extended LOS, 26.1% were non-English speaking and of those with short LOS, 10.4% were non-English speaking.  Compared with their English-speaking counterparts, the non-English speaking patients were 4.15 times more likely to have long LOS, after adjusting for age, living alone, having two or more co-morbid conditions, and requiring home support (MacDonald, Ottem, Wasdell, & Spiwak, 2010).  The researchers suggested that this finding may have resulted from the hospital program that solely provided English-language patient teaching and educational materials.  Non-English speakers would have had to access their education through interpreters. Contrary to the studies that have found an association between LEP and LOS, Grubbs et al. (2008) did not find an association between the LOS and the language preferences of patients in the US.  They examined the LOS of patients admitted for acute myocardial infarction by conducting a retrospective study of over 12,000 health records.  The small increase in LOS (3.8%, 95% CI [0.3%, 7.3%]) was associated with non-English preference.  But this finding was not significant once they adjusted for history of cardiac procedures or surgery after admission (2.8%, 95% CI [-0.6%, 6.2%]) or hospital (0.9%, 95% CI [-2.5%, 4.3%]).  The authors suggested that hospital-level factors had greater influence on LOS than patients’ individual characteristics.  13  That is, access to quality services and staff may be a more important determinant of LOS, rather than patients’ English proficiency.  Notable though, this study defined LEP as patient language preference, rather than actual proficiency, which could have accounted for the differences in the findings of other studies. In summary, there are few studies that have demonstrated a relationship between LEP and LOS.  To date, the majority of relevant studies have indicated that LEP leads to relatively longer LOS, although further investigation is needed.  In addition, past studies have utilized various approaches to the measurement of English proficiency, which has contributed to a lack of standardization and validation of the identification of English proficiency.   Effects of limited English proficiency on adverse events. Effective communication is integral to delivering safe and high-quality care (Baker et al., 2004).  Language barriers may affect patient-provider communication and thus, compromise care.  In fact, such communication problems are the most frequent cause of adverse events (Hyman, 2009).  There is a plethora of literature linking language barriers with adverse events or any “unintended harm to the patient by an act of commission or omission rather than by the underlying disease or condition of the patient” (Aspden et al., 2004, p. 30).   Baker et al. (2004) conducted the first Canadian adverse events study.  An estimated 7.5% (95% CI [5.7%, 9.3%]) of patients admitted to acute care hospitals in Canada, in the year 2000, experienced one or more adverse events.  Of those events, 36.9% (95% CI [32.0%, 41.8%]) of the patients were judged to have highly preventable adverse events.  In relation to LOS, patients experiencing adverse events in the study had longer LOS in hospital than those without adverse events.   14  In-hospital patients with LEP have been observed to incur higher rates of error leading to physical harm, which results in longer LOS and higher readmission rates (Karliner, Kim, Meltzer, & Auerbach, 2010).  Bartlett, Blais, Tamblyn, Clermont, and MacGibbon (2008) examined the adverse events of 20 randomly selected acute care hospitals in Quebec.  They found that patients with communication barriers2  were three times more likely to experience a preventable adverse event associated with some level of disability or multiple hospital admissions compared with patients without communication barriers (OR = 3.00, 95% CI [1.43, 6.27]). The majority of preventable adverse events rated severe were either drug related (40%) or caused by poor clinical management (32%).   Divi et al. (2007) explored the relationship between adverse events and LEP patient status.  They examined the different characteristics of adverse events between both LEP and EP patients in six randomly selected accredited hospitals in the US.  Their findings showed that of the total reported adverse events, detectable physical harm occurred in 49.1% of patients with language barriers compared with 29.5% of patients who spoke English fluently.  Of the adverse events that occurred to the LEP patients, 52.4% were also more likely to be the result of communication errors as compared with 35.9% for the English-speaking patients.  These findings clearly demonstrate that there are health disparities for patients with communication barriers, such as language discordance.  The findings also emphasize the need for additional resources to ensure patient safety and to avoid preventable adverse events.   Cohen, Rivara, Marcuse, McPhillips, and Davis (2005) conducted a case-control study to determine whether pediatric patients whose families had language barriers experienced more adverse events than did patients with families that were English proficient.  They did not find an                                                  2The source of communication barriers included both language comprehension and physical problems that interfered with communication.  15  increased risk for the patient group that was not fluent in English or Spanish.  But within this subgroup, compared with LEP patients and families who did not request an interpreter, Spanish-speaking patients and families who requested an interpreter seemed to have a significantly greater risk of serious medical events during pediatric hospitalization.  The Spanish-speaking families that did not request an interpreter had a twofold increased risk of a serious adverse event (OR = 2.26, 95% CI [1.06, 4.81]).  The authors hypothesized that this finding may be a result of healthcare providers erroneously believing that they have a satisfactory command of Spanish and speculated that these providers may not have relied on interpreters for Spanish-speaking patients as much as for patients who spoke neither English nor Spanish.  Like the Grubbs et al. (2008) study presented earlier, patients defined as having language barriers were self-identified.   In summary, language barriers impeding patient-provider communication are a major safety concern.  Several studies have demonstrated that LEP is strongly associated with more frequent adverse events in comparison with the rates of EP patients (Bartlett et al., 2008; Divi et al., 2007; Karliner, Kim, Meltzer, & Auerbach, 2010; Oliva, 2008).  Acknowledged throughout these studies, though, is the fact that there is a lack of comparable data and “gold standard” research demonstrating a conclusive link between language barriers in clinical contexts and effects on patient safety outcomes (Johnstone & Kanitsaki, 2006). Effects of professional interpretation on patients’ outcomes.   Although the study of the relationship between hospital LOS and LEP is limited, there is extensive literature discussing professional interpreters’ impact on patients’ clinical outcomes.  Such studies have examined the effects of language concordance between patients and their healthcare providers.  The majority of these studies have been conducted in outpatient or emergency department settings (Bernstein et al., 2002; Hampers & McNulty, 2002; Karliner et 16  al., 2007).  Very few studies have been conducted within the in-patient setting.  In addition, the majority of published studies have been conducted in the US.   One in-patient study by Lindholm et al. (2012) examined the differences among LEP patients that received professional interpretation services at the time of admission and discharge and those that did not.  The patients that did not receive interpreter services at admission or discharge had a statistically longer LOS (between 0.75-1.47 days, p < .002), than the LOS of LEP patients that received such services.  This latter group of patients was also less likely to be readmitted to the hospital within 30 days of discharge compared with those that did not receive any interpretation (14.9% compared to 24.3% [p < .001] of LEP patients that did not receive professional interpretation services at time of admission and discharge).  Of note, this study found that access to professional interpretation at admission had the greatest effect on LOS.  Intuitively, this finding suggests that during admission, obtaining a thorough and accurate patient history is imperative to constructing a correct diagnosis.   Because the first point of contact with healthcare services for LEP patients is often in the emergency department, several studies have focused on interpreter utilization practices in this fast-paced setting.  Of particular interest in the literature is the impact of interpreters on resource utilization and costs of healthcare delivery (Bernstein et al., 2002; Hampers & McNulty, 2002).  These studies observed that non-English speaking patients that received interpretive services were equal in their resource utilization to that of their English-speaking counterparts.  Bernstein et al.’s (2002) study found that non-English speaking patients that did not receive interpretive services utilized fewer ED resources (i.e., they had the shortest ED stay, underwent the fewest diagnostic tests, and were less likely to be prescribed intravenous infusions and medications) compared with both English-speaking patients and non-English speaking patients with an 17  interpreter.  These findings are in sharp contrast to Hampers and McNulty’s (2002) results where the non-English speaking plus interpreter cohort had the longest LOS in the emergency department (+16.0 minutes, 95% CI [6.2, 26.0) and the language barrier cohort that did not receive interpretive services had a higher rate of, and cost for, diagnostic testing, was most likely to be admitted (OR 2.6, 95% CI [1.4, 4.5]), and to receive intravenous hydration (OR 2.2, 95% [1.2, 4.3]).   Both these studies suggest that language barriers (in the absence of interpreter assistance) put patients at greater risk of undergoing lesser clinical investigation in the emergency department (Bernstein et al., 2002) or an increased chance of undergoing unnecessary procedures (Hampers & McNulty, 2002); both of which can be detrimental to patients.  Only the presence of an interpreter brought the level of care for non-English speaking patients to par with that provided to English-speaking patients.   Ramirez, Engel, and Tang (2008) conducted a systematic literature review of interpreter utilization in the emergency department setting.  They cited several common barriers to interpreter utilization.  In the busy environment of emergency departments, reliance on untrained, ad hoc interpreters, such as patients’ family members and friends and other healthcare staff, was common practice (Ramirez et al., 2008).  But untrained interpreters may be prone to mistakes and are severely limited by their lay interpretation of medical terminology.  For example, Elderkin-Thompson, Cohen Silver, and Waitzkin (2001) found that one hospital’s nursing staff that was regularly called upon to translate made interpretation errors in over 50% of their encounters, which resulted in physicians’ misunderstanding.   The perceived effort associated with obtaining and working with professional interpreters discourages their use.  Perceptions about the difficulty in obtaining an interpreter and the costs 18  associated with implementing interpreter programs prevent interpreter utilization (Ramirez et al., 2008).  Despite studies that have disputed the negative aspects of interpreter use, and that have found that they may be cost-effective in the long-term, providing coordinated and quality interpretive services remains a system-wide challenge (Hyman, 2009).  In Canada, Access Alliance Multicultural Health and Community Services commissioned a literature review to document the impact of not providing professional and coordinated interpretation services within the healthcare system.  Their review considered over 200 peer-reviewed journal articles and an environmental scan of governmental and institutional documents.  The authors found strong evidence of the negative impacts of language barriers on healthcare quality, efficiency, and cost and argued that implementation and coordinated interpretation services in Canadian health care is “…a basis for providing all Canadians their right to an acceptable standard of health care” (Hyman, 2009, p. 1).   Karliner et al.’s (2007) systematic review provided further insight about the impact of professional interpreters on patients’ outcomes.  They concluded that the use of professional interpreters improves the quality of clinical care for LEP patients when compared with the use of ad hoc interpreters.  The review determined that professional interpreters reduced communication errors, improved patients’ comprehension, equalized healthcare utilization, and improved clinical outcomes such that they approached or equaled the outcomes of patients without language barriers.  As a result, LEP patients that were provided professional interpretation services expressed greater satisfaction with the communication and with the clinical services provided.   Despite the strong evidence supporting the use of professional interpreters in clinical practice, such services are underutilized (Karliner et al., 2010; Ramirez et al., 2008).  When 19  interpreter services are employed, their usage varies by clinician contact, with physicians utilizing their services more than do nurses (Schenker, Perez-Stable, Nickleach, & Karliner, 2011).  In particular, Schenker et al. (2011) noted a significantly low rate of interpreter use in encounters with nurses.  When interpreters were needed, ad hoc interpretation provided by family members, friends, or clinical staff was most frequently used (Schenker et al., 2011).  This finding has wide spread implications for nursing practice.  While interactions with nurses may be more routine and shorter than interactions with physicians, critical communication during encounters with nurses may be missed and may negatively affect the quality of care for patients.   In summary, the literature emphasizes the importance of language concordance between healthcare professionals and LEP patients.  Many researchers have recommended the support and implementation of professional translational services in health care and further support the growing body of evidence that LEP affects patient outcomes.  Finally, utilization of professional services is important rather than the common practice of using ad hoc interpreters because professional interpreters are associated with better patient outcomes.  Effects of limited English proficiency on patients’ experiences. A growing body of research reveals that language barriers lead to significant consequences from affecting the quality of care delivered to LEP patients to ultimately altering the trajectory of patients’ recovery from illness (Rivadeneyra, Elderkin-Thompson, Silver, & Waitzkin, 2000).  Another aspect of the problem to consider is the LEP patient’s perspective.   Today’s approach to providing respectful and responsive care to patients is based on a patient-centred care model where patients are considered partners in their care (Bowman, 2001).  Underpinning this model of care is the assumption that equal and open communication between providers and patients exists.  Language barriers prevent the realization of equality in 20  communication between patients and providers, thus presenting a limitation to achieving patient-centred care.  Nurses caring for patients with language barriers experience feelings of stress and helplessness while patients report dissatisfaction (Cioffi, 2003).  Understanding and documenting patients’ experiences provide a substantive account of the complexities involved in patients’ readiness for discharge from hospital and, thus, provide clarity to the multiple processes that affect the situation.   Exploration of the literature reveals that LEP patients’ experiences within in-patient settings are not well understood.  This gap in the literature exemplifies the need for research about the effects of LEP on in-patients and the healthcare system.  A small number of studies have examined LEP patients’ experiences in out-patient settings, as measured by patients’ satisfaction, understanding, and compliance with recommended care.  These studies have consistently reported that LEP patients rate their healthcare experiences more negatively than do English proficient patients (Carrasquillo, Orav, Brennan, & Burstin, 1999; Morales, Cunningham, Brown, Liu, & Hays, 1999; Ngo-Metzger, Phillips, & Sorkin, 2009).   At the time of this review, only one study was found related to LEP patients’ experiences in the in-patient setting.  Karliner et al. (2012) investigated the association between having a language barrier and patients’ understanding of hospital discharge instructions at two urban hospitals in California.  Overall, the LEP and EP patients understood their diagnosis and their follow-up appointment details, but there was disparity between the groups’ understanding of their medications and the type of follow-up appointment required.  In particular, the LEP patients were less likely to know the category of (OR = 0.63, 95% CI [0.42, 0.95]) or both the category and purpose of (OR = 0.59, 95% CI [0.39, 0.89]) their medications compared with the EP patients.   21  Carrasquillo et al. (1999) studied non-English speaking patients in the northeastern US.  They employed a cross-sectional survey and a follow-up interview to examine patients’ satisfaction and willingness to return to an emergency department.  In their multivariate analysis of over 2,300 patients (15% non-English speaking), non-English speaking patients were less satisfied with their care in the emergency department than were English speakers, reported more problems with their overall care, and thus, were less willing to return to the emergency department for care.  Several studies have analyzed the relationship between LEP and patients’ medical comprehension in the primary care setting.  Wilson, Chen, Grumbach, Wang, and Fernandez (2005) conducted a California-wide telephone survey in 11 languages, excluding English.  Their survey of 1,200 respondents found that LEP patients were more likely than were EP patients to report problems understanding their medical situations (OR = 3.7, 95% CI [ 2.9, 4.8]), problems understanding how to use their medications (OR = 1.6, 95% CI [1.3, 2.1]), and medication reactions caused by problems understanding the instructions (OR = 2.2, 95% CI [1.5, 3.3]).  Morales et al. (1999) conducted a similar study regarding patients’ satisfaction with their family physicians’ communication in out-patient clinic settings.  They studied three language groups: Latinos responding in Spanish, Latinos responding in English, and non-Latino whites.  Latino patients responding in Spanish were significantly more dissatisfied with the communication than were the latter two groups even after controlling for confounding factors such as age, gender, education, and insurance status.  This group also had poorer comprehension of their disease, which further reduced their satisfaction with the quality of the medical care they received.   22  Ngo-Metzger et al. (2009) focused on 2,746 LEP Chinese and Vietnamese patients in eight US cities.  Their study compared the LEP patients with language-concordant physicians (providers who spoke their language) with LEP patients with language-discordant physicians (providers who did do not speak their language).  They evaluated the degree of health education and quality of interpersonal care (measured on a scale of nine questions that asked each patient whether each of the nine experiences occurred during their recent visit with a physician) and their associations with satisfaction.  For all of the patients, the quality of interpersonal care was an important predictor of satisfaction with the visit and with the provider.  The researchers reported that the LEP patients with language discordant physicians, and with no access to an interpreter, had poor quality interpersonal care and reported a 4.4-fold greater chance of being dissatisfied with their physician.  The quality and amount of health education given to these patients were also important predictors of satisfaction.  Those patients with language-discordant physicians and no access to an interpreter received limited health education, which in turn affected their self-reported understanding of their disease and treatment plan. Patients’ dissatisfaction with nurses’ care also has been studied.  Cioffi (2003) described nurses’ experiences communicating with LEP patients.  The nurses made multiple efforts to reduce the marginalization of their LEP patients.  They employed a number of strategies such as using charts, a family member, and sign and body language to communicate with their patients.  Still, the nurses experienced frustration with the complexity of the communication process and recognized that the language barrier also contributed to frustration in their patients.   Linguistic disparities and patients’ experiences have not received much attention in Canada.  Liu, So, and Quan (2007) surveyed Chinese and white patients in Calgary to determine their level of satisfaction with their primary physicians.  Even though the Chinese population had 23  lower ratings of satisfaction with their physicians as compared with their white counterparts (OR = 0.70, 95% CI [0.42, 1.18]), this was not statistically significant.  But within the Chinese subgroup population, the participants that were LEP or recent immigrants were less satisfied than were the Chinese Canadians that spoke English (OR = 0.5, 95% CI [0.3, 0.9] and OR = 0.2, 95% CI [0.1, 0.7], respectively).  This finding suggests that communication between physicians and their patients plays a compelling role in patients’ satisfaction and perceptions of their physicians’ care. Summary. The current literature reveals that LEP may contribute to health disparities at multiple points in the continuum of care.  Patients with language barriers experience less satisfaction with their health care, lower utilization rates of health care, and poorer understanding of their health conditions.  These outcomes suggest that LEP may be an important predictor of patients’ health status and healthcare outcomes.  Even when they do have access to care, LEP patients experience disparities because they are less likely to comprehend their diagnosis and treatment, less likely to adhere to treatments and follow-up plans, and more likely to experience medical errors in the hospital than are English proficient patients.  Although the literature supports the value of professional interpretation services for LEP patients, real and perceived perceptions of their impact on patients’ care result in underutilization.   A few studies have demonstrated that LEP status affects patients’ LOS in hospital although some researchers have suggested that hospital system-level differences should be considered and accounted for in the examination of the association.  Amongst the studies presented, the definitions and methods employed in measuring the LEP status of patients were varied and non-standardized.  Very little literature has been dedicated to the study of the effects 24  of LEP on LOS in general and even less so in the Canadian population.  More specifically, there is no dedicated literature that explores the phenomenon of interest within the population of patients having elective CABG surgery. Thus, the lack of available literature and contradictory findings regarding LEP status and LOS warranted further study. Theoretical Background This research study was based upon two theoretical frameworks that provided a relatively holistic representation of the relevant concepts, and described how they were actualized.  The dynamic processes involved in provider-patient communication were considered with Northouse and Northouse’s (1998) model of health communication.  Their theory provides a basis for understanding the multiple factors that influence health communication.  Martinez’s (2010) theory of language barriers further informed this work with its explication of the relationships between language barriers and patients’ health outcomes.   Health communication. Health communication is composed of the transactions between healthcare participants and their health-related issues.  Northouse and Northouse’s (1998) model of health communication explicated the important factors that influence interactions and the health communication process in healthcare settings: relationships, transactions, and contexts.  Considered together, these elements affect interactions, including the concerted efforts that are made to achieve the mutual goal of health for clients.  These elements are presented in Figure 1, and are discussed further below.   25   Figure 1.  Model of Health Communication.  Factors that influence the interactions between client, significant others and health professionals in health care communication.  Adapted from “Health communication: Strategies for health professionals” by L. L. Northouse and P. G. Northouse, 1998, p. 17. Appleton & Lange.    Definitions of concepts and their relationship to the study. The factors that influence healthcare interactions and communication processes, as specified in Northouse and Northouse’s (1998) model, are defined as follows. Relationships.  Four major types of relationships exist within healthcare settings: professional-professional, professional-client, professional-significant other, and client-significant other.  Health communication occurs within each of these relationships.  Individuals bring their unique characteristics, personal experiences, and beliefs to the healthcare setting.  These characteristics influence the interactions that compose health communication.  These relationships are interlinked and, thus, influence one another either directly or indirectly, as shown in Figure 1.   26  This study focused on the dynamic relationship between the professional and the client, which is a crucial component of effective healthcare delivery and of clients’ health outcomes.  The professional-client relationship is essential to a patient’s ability to follow prescribed therapies and to achieve compliance with prescribed healthcare regimens.  Positive professional-client relationships are important to overall patient satisfaction with the healthcare system and to the ability to manage and cope with illness (Northouse & Northouse, 1998).  The concepts of unique individual characteristics were operationalized in this study with selected patients’ characteristics, including their age and LEP status.  Such variables affect the perception and interpretation of information, influencing health communication -- the transactions between professional and client.  These variables were analyzed for their potential influence upon LOS, an indirect measure of patients’ health status.   Transactions. Transactions, in this context, are the health-related interactions between the participants of the health communication process.  These transactions include verbal and nonverbal communication behaviour.  Successful health transactions depend upon effective and compatible components of verbal and nonverbal communication.  In the health communication model, represented in Figure 1, transactions are represented by a central circular element with an unending spiral, which is meant to depict how transactions are part of the continual feedback that occurs between participants.   Operationalizing these concepts in this study, the health communication model suggested that compatible language within the professional-client relationship is a necessary component of effective verbal communication.  Incompatible language is a barrier to effective transactions, and negatively influences health communication and thus patients’ health status.  27  Other transactions in health communication include the dissemination of health education through pre-operative teaching and media materials, including pamphlets, videos, and posters.  This study was limited to the investigation of the transactions dependent upon verbal communication during the hospital stay and did not explore the influences of other forms of transaction.   Contexts.  Contexts pertain to the setting in which health communication takes place and the systemic properties of the setting.  The context is dependent upon the specific physical setting in which communication takes place, such as a hospital nursing unit or a physician’s office.  These different settings affect the dynamics and patterns of communication.  On another level, the context may refer to the number of participants within a particular healthcare setting.  For example, one-to-one situations affect interactions differently than do small group discussions.   In this research study, the context was limited to two in-hospital settings: a cardiac surgical intensive care unit (CSICU) and a cardiac nursing unit.  These environments were fast paced work environments for staff and subject to unpredictable medical emergencies, which may have affected the patterns of communication, especially for LEP patients.  Challenges to communication were further compounded in the early stages of post-operative recovery, when the patients were intubated, often sedated, in pain, and at risk of experiencing delirium and hemodynamic instability.    Northouse and Northouse’s (1998) model of health communication emphasized how relationships, transactions, and contexts affect health communication.  Their theory postulated that language compatibility or concordance between patient and provider mitigates potential barriers to effective health communication.  English proficiency of a patient influences such 28  concepts, thus, affecting health communication.  Complementing these concepts, Martinez (2010) provided a theoretical framework that directly linked language barriers and patients’ health outcomes.   Language barriers. Martinez (2010) provided a theory of language barriers that served as an underpinning of this study.  He sought to account for the multi-faceted effects of linguistic inequality on health outcomes.  Two core assumptions of his theory are: 1.  Language barriers occur across a continuum of care – equally in the clinic as in the community and 2.  When language discordance exists, language barriers perpetuate a negative effect that blocks communication, regardless of the patient’s and their healthcare providers’ ability to converse in the dominant language.  Language barriers are performances in which the process of meaning making is intentionally and unintentionally blocked by the speakers.  For example, the absence of meaning in a conversation is overshadowed with preconceived rationalizations of the speakers.  Rationalizations such as, “Why can’t they just learn English?” unintentionally contribute to the blocking of communication.      Martinez (2010) purported that health and illness follow a “linguistic gradient” for language minority populations.  The extent or degree to which patients experience negative health outcomes is related to both their proficiency with English and the presence of a language other than English in their lives.  Research has demonstrated this phenomenon; LEP patients have been shown to have worse health outcomes compared with EP patients, and bilingual 29  patients also have poorer outcomes compared with those who speak only English (Cheng, Chen, & Cunningham, 2007; Kandula, Lauderdale, & Baker, 2007). The systematic pattern of negative health outcomes associated with LEP raises important questions about the nature of language barriers.  Martinez (2010) proposed that there are multiple dimensions on which language differences between patients and healthcare providers operate in creating barriers to communication.  Dimensions of language barriers. Interaction dimension. The interaction dimension is the most readily recognizable dimension of language barriers.  This dimension refers to the lack of a shared code between those involved in an interaction.  Limited or no English proficiency creates a barrier to communication because a lack of mutual understanding between patient and provider results.  Communication exchange may be shortened, as a result, potentially compromising patient care and threatening to alienate patients from their providers (and vice versa).   Informational dimension.  The informational dimension refers to the disparity in the distribution of health information that is readily available in a society.  Language barriers generate gaps in effective health information delivery, which limit health knowledge acquisition and the potential for health action by patients.    In this study’s setting, health information including discharge teaching videos, educational pamphlets, and physiotherapy classes were widely available in English.  Some of these materials had been translated into other languages, such as Chinese and Punjabi, but despite the predominance of culturally diverse patients having CABG surgeries, the majority of the health information was available in English only.  According to Martinez’s theory (2010), this informational dimension influences a LEP patient’s ability to acquire health knowledge.  The 30  inability to acquire health knowledge leads to an inability to act on such knowledge, and restricts engagement in the behaviour and activity that are required for successful surgical recovery and hospital discharge.  For example, if physiotherapy classes are conducted only in English, then a LEP patient may miss vital information pertinent to understanding why such therapy is important to recovery.  The patient may then, in turn, not recognize the importance of such classes and may not actively participate in therapy, affecting the likelihood of successful recovery. Acceptance dimension.  Language barriers between participants block the intersubjectivity of the linguistic exchange.  Intersubjectivity refers to the activities involved in establishing shared cognition and consensus, which are essential to shaping a relationship.  These activities include mutual recognition of sociocultural differences between participants, and the authentication of those differences.  This process facilitates trust in the healthcare provider and patient relationship and legitimizes the cultural stance of the individuals involved.  Thus, this theory provides a possible explanation of how language barriers between LEP patients and their healthcare providers may negatively affect the patient-provider relationship and influence patients’ health outcomes. Performance dimension.  The fourth dimension of Martinez’s (2010) theory is the performance dimension.  Martinez described this dimension as the obstruction of patients’ ability to take over the management of their own disease processes.  Patients who do not speak the predominant language may lack the confidence, knowledge, and tools to adequately care for themselves compared with patients who understand their therapy and can communicate at ease within the healthcare system.  Martinez’s theory guided this study because it acknowledged that LEP patients face obstacles in communicating with their healthcare providers, may be less inclined toward participating in their own care, and may result in being more heavily reliant and 31  dependent on the healthcare system.  This study proposed that this dependence could lead to a prolonged length of stay after surgery.   In summary, language barriers prevent LEP patients from interactions, access to health information, and intersubjective communication.  Northouse and Northouse’s (1998) model of communication provided the foundational concepts that health communication is dependent upon.  Martinez’s (2010) theory of language barriers identified specific dimensions of language barriers that negatively influence patients’ health outcomes.  This study operationalized the key concepts in both theories and measured health outcomes operationalized as CABG surgery patients’ post-operative LOS. Study variables and their relationship to the theoretical framework. In this study, the outcome variable, LOS, served as an indirect measure of the possible poor health outcomes that occur following isolated CABG surgery (see Figure 2).  Poor outcomes including atelectasis, infection, delirium, stroke, and so forth, are post-operative complications that may be influenced by patients’ abilities to communicate with their providers and result in prolonged lengths of stay.  For example, pulmonary hygiene (i.e., deep breathing exercises) is important to prevent patients’ airways from collapsing and the potential complications of pneumonia.  Language discordance prevents a LEP patient from understanding the concept of pulmonary hygiene and consequently the patient may not actively participate or self-initiate regular deep breathing exercises.  This patient then would be prone to the development of atelectasis, which could lead to infection (Woods et al., 2010).  Another poor outcome, post-operative delirium could be exacerbated by language barriers because the nurse is unable to orient and provide effective psychological support to a patient who does not share the nurse’s language(s).  Recovery from a post-operative stroke could be complicated by language 32  barriers.  Similarly, the inability to communicate effectively could affect patients recovering from both surgery and stroke.  Lack of English proficiency is strongly associated with a lack of stroke knowledge in both the general population and in patients with prior history of stroke (Cruz-Flores et al., 2011).  Figure 2 displays how several possible confounding variables such as demographics and comorbidities could affect patients’ outcomes and thus, their LOS.                        POSSIBLE CONFOUNDING VARIABLES Gender Age Employment status Marital status Comorbidities   Figure 2.  Language proficiency and the relationship with length of stay as influenced by health communication.  Summary This literature review has highlighted the clinical consequences of limited English proficiency on patients’ outcomes.  Utilizing Northouse and Northouse’s (1998) health Language Proficiency• English speaking• Limited English• No EnglishHealth Communication• Dependent upon                               relationships and transactions contexts • Language barriersAdverse Events• Infection• Delerium• Stroke• Other LOS: in CSICU, in nursing unit, and entire post-operative stay•Expected•Prolonged Predictors Outcomes 33  communication model and Martinez’s (2010) language barriers model to provide a theoretical framework for this study, it is hypothesized that language barriers have a negative effect on patients’ health outcomes and on healthcare efficiency as measured by LOS.   In the current literature, the varying methods to measure LEP status, in conjunction with contradictory evidence of the effects of LEP on LOS, paint an unclear picture of the relationship between English proficiency and LOS.  In particular, there is a dearth of evidence dedicated to studying this phenomenon in the cardiac surgical population.  Furthermore, in the multilingual and multicultural population of Canada, dedicated Canadian research to this topic is sparse at best.  Examining patients’ English proficiency and its relationship with length of stay in the post elective CABG population in a Canadian hospital will assist in determining whether changes to practice and policy are warranted.  Determining if a relationship between LEP and LOS exists will assist nurses and other healthcare professionals to acknowledge and address the role they play in mitigating the potential adverse health outcomes of cardiac surgical patients.  34  Chapter Three: Methods This chapter provides a review of the research questions that guided the examination of the phenomenon of interest: the effects of patients’ English proficiency on length of stay following isolated cardiac bypass surgery.  Details of the study’s research design and methods are described, including a discussion of the sample inclusion and exclusion criteria and the two measurement tools utilized: the data abstraction tool and the LEP screening tool.  Next, the determination of the sample size required for this study is presented.  To provide the reader with further insight about how the results were derived, this chapter explains the statistical analysis used to obtain the results from the data.  The ethical considerations of the study are also discussed.   Research Questions The main research question for this study was: Do LEP patients undergoing isolated cardiac artery bypass graft (CABG) surgery have prolonged LOS when compared with EP patients?  A secondary research question was: Do LEP patients undergoing isolated CABG surgery consistently meet the targeted LOS established by the CABG surgery clinical pathway as compared with their EP counterparts?   Hypothesis  The study had two directional hypotheses:  1. On average, patients with LEP have a longer LOS after isolated CABG surgery when compared with EP patients. 35  2. Relatively fewer LEP patients undergoing isolated CABG surgery meet the established targeted LOS of the CABG surgery clinical pathway compared with EP patients. Research Design The design of this study was a retrospective chart review, which set out to determine whether a relationship existed between the post-operative LOS and the English proficiency status of isolated CABG surgery patients.  The study took place at a single tertiary care hospital.  During the period of interest, the hospital was one of four provincial centres that offered major cardiac surgery; it performed approximately 25% of the total share of open heart surgeries conducted in the province (Cardiac Services BC, 2010).  Patients were scheduled for isolated CABG surgery Mondays through Fridays.  Up to four potential patients were scheduled for cardiac surgery each day.  Emergency cardiac surgeries were performed after hours or on the weekends, as needed, and as operating room staff were available.  The total number of open heart cardiac surgeries performed in 2012, in BC, was 3,348 (Cardiac Services BC, 2013).  Of these surgeries, 1,720 patients underwent isolated CABG procedures (Cardiac Services BC).  At the study hospital, 409 patients underwent isolated CABG surgery (Cardiac Services BC).  All of the charts of patients receiving isolated CABG surgery between January 1, 2011 and December 31, 2012 were screened according to the inclusion and exclusion criteria.  All of the charts of patients meeting the inclusion criteria were reviewed in detail.     Before the retrospective chart review was conducted, it was recognized that a standardized tool to assess the LEP status of patients, as described in their medical records, was not available.  Thus, a LEP screening tool was developed and evaluated. 36  Inclusion criteria. The target population consisted of both male and female adults who at the time of surgery: (a) were 19 years of age or older and (b) had undergone isolated coronary artery bypass (CABG) surgery.  These criteria encompass patients whose surgeries were elective and in-patient cases of all priority levels.3 Exclusion criteria. The study excluded any patients who had additional procedures at the time of their isolated CABG surgery.  Patients who had concomitant procedures, such as a valve repair or replacement, aneurysm repair, or implantation of a pacemaker or implantable cardioverter defibrillator were excluded from the study.  Patients who were not placed on a bypass machine for their surgery or who received robotic-assisted CABG surgery were also excluded.  These patients had a different expected length of stay as compared with isolated CABG surgery (Rudisil et al., 1994).  Patients that did not have their sternum closed post-operatively were also excluded from the study.  Typically, after the surgery is completed, the sternum is sutured closed, although surgeons may decide to not close the chest wall cavity, postoperatively (known as “open chest”).  Indications for open chest after CABG surgery include postoperative low cardiac output, hemodynamic breakdown upon closure, bleeding, and arrhythmias (Woods et al., 2010).  These patients remain intubated and sedated postoperatively until they are hemodynamically stable to return to the operating room for sternum closure and thus, have an extended length of stay.  Patients were excluded if their language status could not be determined from their health records.                                                       3 All patients awaiting CABG surgeries in Canada are designated a priority level (based on Urgency Rating Scores) mandated by the federal and provincial health ministries (Cardiac Services BC, 2008).  This study did not control for the level of urgency for the surgery.  37  Research Methods Research protocol. After their open heart surgery, all of the patients were admitted to the critical care unit of the CSICU.  A master list of the admitted patients was kept on this unit; it included a record of each patient’s name, operating surgeon, type of surgery, date of hospital admission, date of surgery, date of CSICU admission, and date of discharge to the nursing unit.  From this list, isolated CABG surgery patients were identified and their medical charts were obtained from the health records department.  Thirty to fifty records were requested on a weekly basis and were reviewed on site.    Consistent with the requirements of the hospital ethics board, patients’ consent was not required to obtain the medical records.  Each record was assessed for eligibility for the study.  If the patient’s records met the inclusion criteria, a LEP screening tool was administered to assess the patient’s English proficiency status, as indicated in the record, and a standardized data abstraction tool was utilized to collect all the information relevant to the variables of interest.  Only the investigators, co-investigators, and research personnel had access to the anonymized data.  Measurement. The English language proficiency screening tool. As previously mentioned, a review of the literature revealed that previous retrospective chart audits have relied upon a variety of unvalidated measures or definitions of LEP.  Many US studies have determined LEP status by noting a patient’s language preference recorded in the electronic health record.  Language preference and English proficiency are two concepts commonly interchanged (Gee, Walsemann, & Takeuchi, 2010).  The use of patients’ language 38  preferences as an indicator of an individual’s English proficiency is an inaccurate indicator.  For example, patients may prefer to speak in their native tongues, but would be found to be English proficient if interviewed and assessed by a validated tool, such as the LEP Census tool (Karliner, Napoles-Springer, Schillinger, Bibbins-Domingo, & Perez-Stable, 2008). To overcome the complications associated with language preference as an indicator, other researchers have determined a patient’s LEP status by noting the use of a professional interpreter recorded in the medical records.  As recognized in the literature review, this measurement is also a problematic indicator of LEP status because it is common practice to rely on non-professional or ad hoc sources of interpretation for patients, such as other healthcare providers or family members (Flores, 2006; Karliner et al., 2007; Jacobs et al., 2004; Lindholm et al., 2012).  This practice may not result in a notation in the record.  This practice was also common at the study hospital.   To measure English proficiency status, the primary investigator and research team developed a LEP assessment tool.  This tool relied upon documentation and confirmation of the patient’s LEP status from four sections of the medical record (see Appendix C).  In each of these four areas, the data collector specifically searched for evidence that the patient identified a language other than English as her or his primary language, was not fluent in English, or required professional or non-professional translation services, and searched for evidence of language status recorded in the record.   In the first of the four areas of the record searched, the data collector reviewed the patient’s registration form.  This form was completed at the time of admission to the hospital.  The admission clerk would have recorded the patient’s preferred language of medical communication.  The second area reviewed was the nursing admission or assessment form.  This 39  form would have been completed when the patient was first admitted to a nursing unit, and provided a detailed assessment of the patient’s medical and social history.  The first section of the form required the interviewer, most likely a nurse, to indicate whether the patient was “fluent in English and/or fluent in … (specify).”  A second question asked whether there was a family translator and if so, the contact information of that person.  Directly following this section of the form, there was a comment section that allowed the interviewer to record additional information (e.g., a history could not be taken because of a language barrier).  The third area of the record reviewed by the study data collector was the preoperative checklist.  This checklist would have been completed before the patient was transferred to the operating room and consisted of important information for the operating room staff.  On the form, a notation recorded the “patient[’s] language … (if not English).”  Lastly, the nurses’ and consultants’ (e.g., physicians, social workers, occupational/physiotherapists) notes were appraised to further clarify the patient’s English proficiency.  For each of the four areas of assessment, the data collector indicated whether the patient was noted to be EP, LEP, or NEP (not English proficient).  The data collectors could also indicate whether the relevant information was not recorded, not available (the record was missing from the chart), or ambiguous (not clearly noted to be EP, LEP, or NEP).  These four areas of inspection were deemed sufficient to facilitate the identification of the patient as EP, LEP, or NEP.  If the data collector could not determine a patient’s English proficiency status from these four sections of the health record, the patient was categorized as having indeterminate language status.  Data abstraction tool. For each patient’s medical record, after determining the LEP/EP status, a data abstraction tool was used to efficiently collect all relevant information including socio-demographic 40  information, pre- and post-procedural conditions, and other health status variables that could have influenced the outcome variables of the study (see Appendix D). Several risk factors have been shown to predict length of stay following elective CABG surgery (Johnston et al., 2004).  Such risk factors include patients’ socio-demographics and clinical variables.  The following variables were included in the data abstraction tool. Socio-demographic variables.  The patient’s age, gender, marital status, and employment status were recorded in the data abstraction tool.  This information was collected from the admission face sheet or the nursing admission history notes. Pre-operative conditions.  The data abstraction tool recorded the following characteristics of the patient’s health status, disease severity, and comorbidities: height, weight, body mass index (BMI), history of hypertension, history of dyslipidemia, history of diabetes, history of respiratory disease, number of diseased vessels, history of left main artery disease, previous cardiac procedures or surgery, previous myocardial infarction (MI), history of valvular disease (mitral or aortic), pre-operative left ventricular ejection fraction (LVEF), extracardiac arterial disease, history of stroke or transient ischemic attack (TIA), pre-operative use of intravenous inotropic agents or intra-aortic balloon pump (IABP).4   In addition, each patient’s current disease symptomology was captured with two scales and recorded in the record: the Canadian Cardiovascular Society functional classification (CCS) for chest pain and the New York Heart Association functional classification (NYHA).  The CCS class is a widely accepted grading system based on four levels from I to IV that determines the severity of a patient’s chest pain burden on her or his quality of life (Woods & Motzer, 2010).  The NYHA classification measures if a patient experiences concomitant heart failure symptoms.                                                   4 IABP is a mechanical device inserted into the aorta and assists in reducing the workload of the heart.  41  Like the CCS class, classification for NYHA is based on a grading system from I to IV (See Appendix E for Class definitions of CCS and NYHA).   These variables were important to consider because some of these risk factors may have been confounding variables in the relationship between LEP and LOS.  Accordingly, all information related to these risk factors was recorded.  All information related to pre-operative conditions was obtained from the physician’s consultation notes. Peri- and post-operative conditions.  Data were collected related to: (a) the number of surgical grafts, (b) the length of time on a cardiopulmonary bypass pump, (c) any return to the operating room for additional surgery, (d) prolonged intubation (> 24 hours) or reintubation,5 (e) post-operative stroke or transient ischemic attack (TIA), as confirmed through a neurologist’s clinical diagnosis, (f) post-operative delirium, as defined by a psychiatric clinical diagnosis, (g) infection, and (h) whether the patient was awaiting transitional care or long-term care placement.  These specific conditions were recorded because they were key descriptors of the sample and were potentially confounding variables.  In addition, the data abstraction tool captured significant surgical or postoperative complications (i.e., the use of an intra-aortic balloon pump, the need for dialysis, or post-operative MI) in the “notes/additional comments” section.  This information was obtained from the operating room summary report, and the physicians’ and nurses’ notes. Length of stay.  For all of the study patients, the dates of when they were admitted and discharged from the nursing units were collected to calculate their length of stay.  Dates recorded included: the hospital admission and discharge dates, the CSICU admission and discharge dates, and the nursing unit discharge date.  This information was obtained from the operating room report and discharge summary report.                                                  5 Reintubation refers to the situation when patients that were sufficiently stable to have their endotracheal tube (breathing tube) removed post-operatively experienced declining health and required a second intubation. 42  Data Quality Limited English proficiency screening tool.   Important considerations were made to reduce any systematic or random error related to determining the English proficiency status of each patient.  A small pilot study was conducted to verify the reliability and validity of the LEP screening tool developed specifically for this study (see Appendix C).  All research personnel (the principal investigator and the three co-investigators who served as data collectors) were trained to use the LEP screening tool.  They concurrently assessed 40 currently admitted cardiac surgery patients’ records to determine the patients’ EP status (i.e., EP, LEP, NEP, or indeterminate).  The four raters’ assessments were evaluated for inter-rater reliability with Fleiss’s (1971) kappa and were interpreted according to the guidelines provided by Landis and Koch (1977):  .81 - 1.00 was considered almost perfect agreement, .61 - .80 was substantial agreement, and .41 - .60 was moderate agreement.  Anything below .41 was treated as unacceptable; thus, agreement of .80 or better was sought for this study.  Review of the definitions and indicators of English proficiency were undertaken until this level of agreement was achieved. The assessment of 40 records was determined to be sufficient to estimate the inter-rater reliability of the raters based on the following assumptions: (a) that the probability of chance agreement among the four raters was equal to zero and (b) the desired  agreement between the raters was .80, allowing for 20% relative error (Gwet, 2010).   Inter-rater reliability and validity results. The principal investigator and the three data collectors assessed 40 admitted cardiac patients’ records during June 2013.  Using Fleiss’s (1971) kappa calculation of k = (P – Pe)/(1 –Pe), the four coders obtained agreement of k = .83.  According to Landis and Koch (1977), the 43  result indicated that the inter-rater reliability was strong, almost near perfect.  In addition to assessing the inter-rater reliability of the chart reviews, the data collectors’ classification of patient English proficiency status was compared with the assessment of the nurse caring for each of the 40 patients.  Following a structured number of questions (see Appendix C) for each patient, the assigned nurse was asked whether she or he believed the patient to be English proficient.  We calculated the positive, negative, and overall agreement of the research team’s determination of the English proficiency of the 40 patients with that of the nurses’ responses.  This served as a method to confirm the validity of the LEP screening tool. Of the 40 patients, the majority of the research team’s classification agreed with the nurse’s; 37 of the 40 patients were agreed to be either LEP/NEP or EP.  This overall percentage of agreement between the majority of the coders and the nurse was 92.5%.  The majority of coders determined that 31 of the 40 patients were EP, whereas the nurses stated that 32 of the 40 patients were EP.  Thus, the percent agreement between the majority of the coders and the nurses’ assessment that their patients were EP was 96.9%.  Of the 8 patients that were deemed to be LEP or NEP by the nurses, the majority of coders agreed for 6 of these patients, indicating an agreement percentage of 75%.  Agreement at 75% between the majority of the coders and the nurses provided assurance that the LEP assessment tool was fairly effective in correctly classifying the patients as LEP or NEP, assuming that the nurses knew the status of their patients.   It is recognized that treating the nurses’ assessments as the “gold standard” in determining the patients’ English proficiency was an imperfect and weak standard.  For example, in one case of disagreement, the nurse divulged that she was unsure of her patient’s English proficiency status.  Two other disagreements between the nurse and the coders resulted because all documentation in the record indicated that the patients were EP, but upon interviewing the nurse, 44  the nurse believed that the patient was LEP.  These findings demonstrate how the nurses may not have been fully informed or objective in their assessment.  The nurses’ classifications may have been subject to biases.  Although, it was assumed that the nurse caring for a patient would be objective and accurate in determining a patient’s English proficiency, and standardized questions were developed to facilitate the nurses’ ability to indicate their patients’ English proficiency status, the nurses’ classification was ultimately subjective in nature.  Directly interviewing the patient and using a linguistic assessment tool would be most ideal, but time constraints and design prevented this method from being employed. Data abstraction tool.   Once the reliability and validity of the English proficiency assessment was established, confidence in the developed LEP assessment tool allowed the study to commence.  All research personnel (the principal investigator and three data collectors) were trained together and reviewed the first seven patient charts together to ensure consistency of data collected. Data Collection Procedures A record of all patients admitted for open heart surgery was kept in the CSICU ward.  From this record, a list was compiled of all patients who had undergone isolated CABG surgery during the period of interest.  The hospital completed 813 isolated CABG surgeries during the years 2011 and 2012 (Cardiac Services, 2013).  The researchers consecutively sampled and requested 712 of these patients’ health records for review.  Each patient record was reviewed once by a researcher.  The health record was first assessed to determine whether the patient met the inclusion criteria.  Next, the health record was read to determine the patient’s language status (i.e., EP, LEP, or NEP) and then information was extracted with a data abstraction tool.  If the researcher could not determine the language status of the patient from the chart (e.g., insufficient 45  documentation or illegible handwriting), the patient was considered not eligible for inclusion.  Any required data that were missing from the chart were documented as ‘missing’ in the data abstraction tool.     Data Analysis  Descriptive analysis of the study participants’ information was conducted.  Descriptive statistics, including the frequency distributions, measures of central tendency, and dispersion were obtained to describe the attributes of the study participants.  Inferential statistics were used to answer the research questions.  Categorical variables were analyzed with the chi-square statistic and polynomial regression.  Data entry and cleaning. All data obtained from the health records were initially recorded on study-specific paper-based case report forms.  The data were then anonymized and entered into the IBM® SPSS® version 21software package.  Once the data were entered electronically, they were thoroughly reviewed to ensure that all the variables had valid and usable values (Plichta Kellar & Kelvin, 2013).  All invalid and out-of-range values were defined as “missing” and were not included in the final analysis.   Data entry is prone to error thus data cleaning ensures further accuracy and verification (Polit & Beck, 2012).  The data set was searched for missing or erroneous data.  The data were randomly checked and compared with the original paper case report forms.  Efforts were made to complete the data sets; missing data were addressed when possible through re-examination of the patients’ records.  Next, the distributional properties of the data were inspected and the validity of the outliers was assessed.   46  Determining sample size. In comparing the LEP and EP patients, we required sufficient statistical power to detect a statistically significant difference between the two groups’ LOS, if a difference was truly present.  The hospital completed 404 and 409 isolated CABG surgeries during the years of 2011 and 2012, respectively (Cardiac Services, 2013).  This study’s goal was to review patients’ charts that met the screening criteria between January 2011 to December 2012.  Thus, in total, 813 patient charts were available for consideration.   There was insufficient a priori information about the rate of English proficiency, in the population, to determine an adequate sample size with confidence.  We estimated the requisite sample size a priori based on several assumptions.  Kelsey et al.’s (1996) method of calculating sample sizes for cross-sectional cohort studies was used.   In one scenario (see Table 1), specifying an alpha level of .05 and power of .80, assuming the ratio of EP patients to LEP patients was 6.8:1 (i.e., 13% of the sample would be LEP as derived from the 2011 Census for Greater Vancouver), and assuming that the number of EP status patients with extended LOS was 5% (i.e., exceeded the care pathway target) and the number of LEP status patients with an extended LOS was 13%, a sample size of 624 patients would have been required.  Of those 624 patients, 80 patients would have been LEP.  Two other scenarios with some modification of these assumptions are also presented in Table 1.  Given these scenarios and the proposed sample sizes of patients, we anticipated that there would be sufficient statistical power to detect a difference in the LOS of these two populations of patients, if there was a difference.  Thus, a minimum sample size of 624 patients (scenario #1) was selected for this study.    47  Table 1   Sample size calculations Assumptions Scenario Alpha Power Ratio of EP to LEP patients* % of EP patients with extended LOS % of LEP patients with extended LOS Estimated detectable OR Required sample size (EP) Required sample size (LEP) Required Sample size (Total) #1 .05 .80 6.8:1  5  13 2.8 544 80 624  #2 .05 .80 6.8:1 10 20 2.3 610 91 710  #3 .05 .80 6.8:1 15 20 1.4 3267 476 3713 *Assumed that 13% of all patients were LEP.  Statistical analysis. The primary and secondary endpoints of this study were: (a) postoperative LOS (entire hospital stay from admission to the cardiac surgical intensive care unit [CSICU] to discharge) stratified by English proficiency status and adjusted for possible confounding variables, including demographics, co-morbidities, number of bypass grafts, length of pump time, and postoperative complications and (b) LOS in the cardiac surgical intensive care unit (CSICU) and LOS in the nursing unit, specifically, stratified by English proficiency status with adjustment for the possible confounding variables mentioned above.  Figure 2 identifies the variables of the study and how they were operationalized within the guiding theoretical frameworks of Northouse and Northouse (1998) and Martinez (2010).  The study sample was described using standard descriptive statistics including means, frequencies, percentages, and standard deviations.  Associations between categorical variables were assessed with Chi square statistics for independence and Fischer’s exact tests, when cell counts did not meet the minimum expectation of five cases.  The group-specific means of 48  continuous variables were analyzed using independent sample t-tests.  Inferential statistics were applied to answer the study’s research questions: do LEP patients undergoing isolated cardiac artery bypass graft surgery have longer LOS when compared with EP patients?  Do LEP patients undergoing isolated CABG surgery consistently meet the targeted LOS established by the CABG surgery clinical pathway as compared with their EP counterparts?  Specifically to address the first question, multivariate analysis was undertaken with polynomial regression analysis of LOS on language proficiency with bivariate statistically significant variables of both English proficiency and LOS also entered into the model to adjust for potential confounders and intermediate variables.  LOS was categorized into quartiles (i.e., 1 to 5 days, 6 days, 7 to 8 days, and 9 or more days).  Next, the odds of a particular duration of post-operative LOS given English proficiency status and other patient characteristics were computed.  The ratio of LEP patients’ odds relative to EP patients’ odds for each quartile was computed with 95% confidence intervals.  The level of significance (α) was set at .05.  To address the second question, the distribution of LOS was categorized by percentiles (weighted average method6) and the Mann Whitney U test was applied to compare the LOS between EP and LEP patients (the percentages of patients with LOS less than six days were compared).   Ethical Considerations This study was reviewed and granted ethical approval by the University of British Columbia, Clinical Research Ethics Board.  Because this study was a chart audit, the data were considered to be a secondary source of information and no informed consent, from patients, was required.                                                  6Weighted average method equation:  (n + 1) p = i + f, where Percentile Value = (1 - f)xi + fxi=1.  Note that i = integer part of (n + 1)p; f = fractional part of (n + 1)p; n is the number of observations; p is the percentile value divided by 100. 49  Confidentiality of all patient information and collected data was safeguarded throughout the study.  To protect the privacy of the patients, each enrolled participant was assigned a unique subject identifier unrelated to information about the individual and which could not be used to identify the individual.  Any direct identifiers were removed from the study data records and replaced with the unique subject identifier.  It was possible to re-identify specific participants because the principal investigator retained a master list that linked the participants’ unique subject identifiers with their names.  The purpose of the master list was to allow for data to be re-linked only if necessary (i.e., in the event that a record needed to be reviewed at a later date to complete missing or ambiguous data entries).  This master list was kept in a locked cabinet separated from the coded study data.  All related hard copies with the patients’ information were kept in a locked cabinet in the Cardiac Surgical Intensive Care office.  Electronic versions of the study data were anonymized and password protected.  Only the principal investigator and related study personnel were able to access the study data and patients’ information.     50  Chapter Four: Results Characteristics of the Sample The study population included patients undergoing isolated cardiac bypass graft surgery with cardiopulmonary bypass between January 1, 2011 and December 31, 2012.  Within the study duration, 813 patients underwent isolated cardiac bypass surgery (Cardiac Services, 2013); 712 patient charts were screened for inclusion in the study (time constraints prevented review of the remaining 101 patients’ records).  Of the 712 reviewed, 691 (97.1%) patients’ charts were eligible for inclusion.  Eleven (1.5%) of the charts were excluded because they were found to be ineligible (i.e., they were not placed on a bypass machine for their surgery, received robotic assisted coronary artery bypass surgery, or were ‘open-chest’ patients).  Ten (1.4%) of the patients’ charts were excluded because of inadequate documentation.  Of those 10 charts, 8 were missing from the Health Records Department and 2 were incomplete (i.e., they were missing volumes).  Figure 3 depicts the enrollment of the patients into the study.    51   Figure 3.  Flow chart of patient enrollment into the study.  Note.  EP = English proficient; LEP = Limited English proficiency; and NEP = Not English proficient.   Language proficiency.  Of the study’s total sample (N = 691), 85.1% of the patients (n = 588) were determined to be English proficient (EP).  Eight percent (n = 56) were determined to have limited English proficiency (LEP) and 6.8% (n = 47) were not English proficient (NEP).  Because of the small numbers of LEP and NEP patients in the sample, the analyses were conducted by comparing EP patients with the LEP and NEP patients combined (n = 103; 14.9%).  Among the 103 LEP and NEP patients, the most common language spoken was Punjabi at 41.7% (n = 43) (see Table 2).  About one quarter of the LEP/NEP patients spoke Cantonese (n = 29; 28.2%), 8.7% spoke Isolated CABG patient populationN = 813Assessed for eligibilityN = 712 (87.6%)EligibleN = 701 (98.5%)Positive for inclusion screening criteriaN = 691 (98.6%)EPN = 588 (85.1%)LEP/NEPN = 103 (14.9%)Excluded (inadequate documentation)N = 10 (1.4%)Ineligible N = 11 (1.5%)52  Mandarin (n = 9), and 4.9% spoke Vietnamese (n = 5).  Table 2 also shows that all other languages combined, made up 16.5% (n = 17) of the languages spoken by the LEP/NEP group (including Albanian, Amharic, Arabic, Bulgarian, Farsi, French, Hindi, Hungarian, Italian, Japanese, Russian, and Tagalog).  Table 2   Languages spoken by the patients (N = 691) Language N (%) English 588 (83.9) Punjabi 43 (6.1) Cantonese 29 (4.1) Mandarin 9 (1.3) Vietnamese 5 (0.7) Other* 17 (2.4) *Including Albanian, Amharic, Arabic, Bulgarian, Farsi, French, Hindi, Hungarian, Italian, Japanese, Russian, and Tagalog.  Demographics. A summary of the patients’ demographic characteristics is presented in Table 3.  The average age of the patients was 65.7 years (SD = 9.1).  The EP patients’ average age was 65.3 years (SD = 9.1) and the LEP/NEP patients’ average age was 67.7 years (SD = 8.4), which was statistically significantly older (t (689) = -2.51, p =.012 (2-tailed)).    Of the sample, 585 (84.5%) were male and 107 patients (15.5%) were female (see Table 3).  The men were more likely to be EP than were the women (87.2% vs. 73.8%) (x2 (1, 691) = 11.63, p =.001; odds ratio (OR) = 2.41; 95% CI: [1.47, 3.94]).  The majority of the patients were married.  In all, 510 (78.8%) of the patients were married or in a common-law relationship.  A greater percentage of LEP/NEP patients (81.3%, n 53  = 74) were married compared with the EP patients (78.4%, n = 436).  A greater percentage of LEP/NEP patients were divorced or widowed (16.5%, n = 15) compared with the EP patients (11.7%, n = 65).  Only 2.2% (n = 2) of the LEP/NEP patients were single compared with 9.9% (n = 55) of the EP patients, x2 (2, 647) = 6.79, p = .034.   At the time of surgery, over one half of the patients were retired (53.4%, n = 334), 37.9% (n = 237) were employed, and 8.8 % (n = 55) were on disability insurance or unemployed.  There was a statistically significant association between employment status and language proficiency.  Relatively more EP patients were employed and fewer were unemployed or on disability compared with the LEP/NEP patients.  Specifically, 41.3% (n = 222) of the EP patients were employed compared with 17.0% (n = 15) of the LEP/ NEP patients.  Eight percent (n = 43) of the EP patients were on disability or unemployed compared with 13.6 % (n = 12) of the LEP/NEP patients, x2 (2, 626) = 19.35, p < .001.  In summary, there were statistically significant differences in the demographic profiles of the EP and LEP/NEP groups.  There were more women in the LEP/NEP group.  The LEP/NEP group tended to be older and was more likely to be retired, unemployed, or on disability than were the EP patients (see Table 3).  In the EP group, there were a larger proportion of single marital status patients and relatively more of them were employed.   54  Table 3   Demographic characteristics of the coronary artery bypass graft patients by English proficiency  All  EP   LEP or NEP  Characteristic N = 691 (100%) N = 588 (85.1%) N = 103 (14.9%) Statistic Age in years (x̅, SD) 65.7 (9.1) 65.3 (9.1) 67.7 (8.4) t (689) = -2.51*       Sex (N (%))     Male 584 (84.5) 509 (86.6) 75 (72.8) x2 (1) = 11.63* Female 107 (15.5) 79 (13.4) 28 (27.2)       Marital Status (N (%))    x2 (2) = 6.79* Single 57 (8.8) 55 (9.9) 2 (2.2)  Married/common-law 510 (78.8) 436 (78.4) 74 (81.3)   Divorced/widowed 80 (12.4) 65 (11.7) 15 (16.5)       Employment status (N (%))    x2 (2) = 19.35* Employed 237 (37.9) 222 (41.3) 15 (17.0)  Disability/unemployed 55 (8.8) 43 (8.0)  12 (13.6)  Retired  334 (53.4) 273 (50.7) 61 (69.3)  Note.  EP = English proficient; LEP = Limited English proficiency; and NEP = Not English proficient.  * p < .05.  Pre-operative characteristics. The two language groups were comparable in several of their pre-operative medical characteristics (see Table 4).  Consistent with established risk factors for coronary artery disease, the majority of the patients had hypertension and hyperlipidemia.  A larger percentage of the LEP/NEP patients had hypertension (97.1 %, n = 100) compared with the EP patients (89.3%, n = 525) (OR = 4.00 (95% CI [1.23, 12.99])). Overall, the average body mass index (BMI) of the patients was 28.0 (SD = 4.8).  The LEP/NEP group had a statistically significantly lower mean BMI of 26.3 (SD = 3.6) than the EP group (M = 28.3, SD = 4.9) (t (174.8) = 4.74, p < .001 (2-tailed)).  55  Almost one half of the patients (43.8%, n = 302) had diabetes.  A statistically significant association was found between language proficiency and diabetes status; 58.3% of the LEP/NEP patients (n = 60) had diabetes versus 41.2% (n = 242) of the EP group (OR = 1.99; 95% CI [1.30, 3.04]).  Among those with diabetes, the majority had Type 2 diabetes (94.7%; n = 287).   Almost all of the patients (86.5%) had three or more diseased vessels (n = 597), one third (31.0%, n = 214) had left main equivalent disease, and one third (31.7%, n = 219) had pre-existing valvular disease.  There were no differences found in the rates of these pre-operative characteristics between the language proficiency groups (see Table 4).  The patients’ average left ventricular ejection fraction (LVEF) was 55.3% (SD = 10.9).  There was no difference in the average LVEF of the EP (M = 55.3%, SD = 10.8) and the LEP/NEP groups (M = 55.8%, SD = 11.7) (t [688] = -0.487, p =.626 [2 -tailed]).   The patients’ coronary artery disease (CAD) burden, measured by Canadian Cardiovascular Society (CCS) class, revealed that the majority of the patients (75.9%, n = 521) experienced a significant level of angina (and other clinical symptoms) related to their CAD with a CCS class of III or IV.  In addition, the majority of the patients (78.8%, n = 530) had a New York Heart Association (NYHA) class of I indicating that they did not have heart failure symptoms.  Neither CCS class nor NYHA class was associated with English proficiency. As displayed in Table 4, almost one half of the patients had had a myocardial infarction (MI) (48.4%, n = 332).  There was no association found between MI history and English proficiency.  Among those who had had a MI, 39.6% (n = 129) of the patients experienced the event more than six weeks before their surgery, 27.9% (n = 91) experienced their event within one to six weeks of their surgery, and 32.5% (n = 106) of the patients experienced the event within the week before their surgery.  There was no difference in the number of cardiac 56  procedures completed prior to surgery for the EP and LEP/NEP groups.  Overall, 17.2% (n = 118) of the patients had undergone a previous cardiac procedure (i.e., catheterization, angioplasty, coronary stent placement, pacemaker). Requiring inotropic support or IABP pre-operatively is indicative of the severity of a patient’s disease and hemodynamic stability at the time of surgery.  Table 4 shows that only 3.8% (n = 26) of the patients required inotropic support and 1.7% (n =12) required IABP support; there were no statistically significant differences noted between the language proficiency groups.  There was a small subset of the sample that had had a stroke or transient TIA before surgery (8.0%, n = 55); this was not associated with English proficiency.  Similarly, extracardiac disease (i.e., carotid stenosis, peripheral vascular disease, arterial disease) was not associated with English proficiency.  Approximately one sixth (13.4%, n = 92) of the sample had a history of extracardiac disease.   The patients’ smoking status was statistically significantly associated with their language proficiency (see Table 4).  There were more former and current smokers among the EP group than among the LEP/NEP group.  Of the EP group, 48.8% (n = 286) were former smokers and 9.7% (n = 57) were current smokers compared with 28.2% (n = 29) and 5.8% (n = 6), respectively, of the LEP/NEP group.  Two thirds (66.0%, n = 68) of the LEP/NEP group had never smoked compared with 41.5% (n = 243) of the EP group.  There were no differences in the groups’ respiratory disease prevalence rates: 22.3% (n = 131) of the EP group had respiratory disease (i.e., chronic obstructive pulmonary disease, obstructive sleep apnea, pulmonary hypertension, asthma, emphysema) compared with 17.6% (n = 18) of the LEP/NEP. A small percentage of the sample (8.1%, n = 56) had a documented history of substance abuse and there was no statistically significant difference noted between the groups (EP group, 57  8.7%, n = 51) versus LEP/NEP group (4.9% [n = 5]).  One in eight patients (12.5%, n = 86) had a psychiatric disorder, including Alzheimer’s disease, dementia, depression, anxiety, bipolar disorder, or schizophrenia.  There was no statistically significant difference between the EP and LEP/NEP group-specific rates.     In summary, the two language proficiency groups were similar in their pre-operative characteristics with the exception of their rates of hypertension, diabetes, and smoking, and average BMIs.  The LEP/NEP group had a greater proportion of patients with hypertension and diabetes.  The EP group had a greater proportion of current and former smokers and a greater average BMI.   Table 4  Pre-operative health characteristics of coronary artery bypass graft patients by English proficiency   All EP LEP or NEP  Characteristic N = 691 (100%) N = 588 (85.1%) N = 103 (14.9%) Statistic Hypertension (N (%)) 625 (90.4) 525 (89.3) 100 (97.1) x2 (1) = 5.31a*      Hyperlipidemia (N (%))  635 (92.0) 538 (84.7) 97 (94.2) x2 (1) = 0.46a      BMI in kg/m2 (x̅, SD) 28.0 (4.8) 28.3 (5.0) 26.3 (3.6) t (174.8) = 4.74b*      Diabetes (N (%)) 302 (43.8) 242 (41.2) 60 (58.3) x2 (1) = 9.64a*      Diabetes Type (N (%))    x2 (1) = 0.21a,c Type 1 16 (5.3) 14 (5.8) 2 (3.3)  Type 2 287 (94.7) 228 (94.2) 59 (96.7)       Diseased vessels     x2 (1) = 0.56a 1-2 93 (13.5) 82 (14.0) 11 (10.7)  3 or more 597 (86.5) 505 (86.0) 92 (89.3)  58   All EP LEP or NEP  Characteristic N = 691 (100%) N = 588 (85.1%) N = 103 (14.9%) Statistic      Left main disease 214 (31.0) 186 (31.7) 28 (27.2) x2 (1) = 0.63a      Mitral or aortic valve disease (N (%)) 219 (31.7) 183 (31.1) 36 (35.3) x2 (1) = 0.52a      LVEF in % (x̅, SD) 55.3 (10.9) 55.3 (10.8) 55.8 (11.9) t (688) = 0.63      CCS class (N (%))    x2 (2) = 0.66 I – II 165 (24.1) 143 (24.4) 22 (21.8)  III 295 (43.0) 248 (42.4) 47 (46.5)  IV 226 (32.9) 194 (33.2) 32 (31.7)       NYHA class (N (%))    x2 (2) = 0.24 I 530 (78.8) 458 (71.3) 72 (71.3)  II 68 (10.0) 55 (9.5) 13 (12.9)  III-IV 84 (12.3) 68 (11.7) 16 (15.8)       Previous MI (N (%)) 332 (48.4) 287 (49.1) 45 (44.1) x2 (1) = 0.69a      MI occurrence (N (%))    x2 (2) = 2.21 < 7 days ago 106 (32.5) 90 (31.9) 16 (36.4)  1 - 6 weeks ago 91 (27.9) 79 (28.0) 12 (27.3)  > 6 weeks ago 129 (39.6) 113 (40.1) 16 (36.4)       Previous cardiac procedure (N (%)) 118 (17.2) 103 (17.6) 15 (14.6) x2 (1) = 0.38a      Inotropic support (N (%)) 26 (3.8) 24 (4.1) 2 (1.9) x2 (1) = 0.60a,c      IABP support (N (%))  12 (1.7) 11 (1.9) 1 (1.0) x2 (1) = 0.06a,c      Stroke or TIA (N (%)) 55 (8.0) 43 (7.3) 12 (11.8) x2 (1) = 1.78a      Extracardiac disease (N (%)) 92 (13.4) 82 (14.0) 10 (10.9) x2 (1) = 0.38a      Smoking history (N (%))    x2 (2) = 21.39* never 311 (45.1) 243 (41.5) 68 (66.0)   current 63 (9.1) 57 (9.7) 6 (5.8)  former 315 (45.7) 286 (48.8) 29 (28.2)  59   All EP LEP or NEP  Characteristic N = 691 (100%) N = 588 (85.1%) N = 103 (14.9%) Statistic      Respiratory disease 149 (21.6) 131 (22.3)  18 (17.6) x2 (1) = 0.86a      Substance abuse (N (%)) 56 (8.1) 51 (8.7) 5 (4.9) x2 (1) = 1.25a      Psychiatric illness (N (%)) 86 (12.5) 75 (12.8) 11 (10.8) x2 (1) = 0.17a Note.  EP = English proficient; LEP = Limited English proficiency; and NEP = Not English proficient; BMI = Body Mass Index; LVEF = Left ventricular ejection fraction; CCS = Canadian Cardiovascular Society; NYHA = New York Heart Association; MI = Myocardial infarction; IABP = Intra-aortic balloon pump; and TIA = Transient ischemic attack. * p < .05. a Yates continuity correction applied. b Equal variances not assumed. c Fisher’s exact test applied (expected cell count < 5).  Intra-operative characteristics.  The average time a patient was on a bypass pump was 116.4 minutes (SD = 39.0) (see Table 5).  There was no statistically significant difference noted when the patients were stratified by their English proficiency.  The EP patients had an average bypass pump time of 116.7 minutes (SD = 39.1) and the LEP/NEP patients had an average time of 114.6 minutes (SD = 38.8).  The majority of the patients (61.9%, n = 428) had four or more bypasses completed and this rate was observed in both the EP and LEP/NEP patients.     60  Table 5    Intra-operative characteristics of patients by English proficiency Characteristic All  EP   LEP or NEP Statistic Bypass pump time in minutes (?̅?, SD)  116.4 (39.0)  116.7 (39.1)  114.6 (38.8)  t (683) = 0.51      Number of bypass grafts    (N (%))     x2 (1) = 0.00a 1 - 3 263 (38.1) 224 (38.1) 39 (37.9)  4 - 7 428 (61.9) 364 (61.9) 64 (62.1)  Note.  EP = English proficient; LEP = Limited English proficiency; and NEP = Not English proficient. a Yates continuity correction applied.  Post-operative characteristics.  Very few of the patients (2.3%, n = 16) returned to the operating room because of surgical complications following their initial CABG surgery (see Table 6) and the rate did not differ by language proficiency status.  There also was no statistically significant difference in the two groups’ rates of post-operative stroke (the LEP/NEP group = 3.9% [n = 4] versus the EP group = 2.4% [n =14]).  One in five patients (20.5%, n = 141) experienced post-operative delirium (see Table 6), and the group-specific rates were comparable (EP = 20.3% [n = 119] and LEP/NEP = 21.1% [n = 22]). There were some statistically significant associations between specific adverse post-operative events and English proficiency status (see Table 6).  Reintubation was required in 5.8% (n = 6) of the LEP/NEP patients and in 1.2% (n = 7) of the EP patients (OR = 5.13; 95% CI [1.69, 15.57]).  In addition, prolonged intubation (≥ 24 hours) was required by 8.7% (n = 9) of the LEP/NEP patients compared with 3.9% (n = 23) of the EP patients (OR = 2.35; 95% CI [1.06, 5.24]).      61  The LEP/NEP had greater odds of developing a post-operative infection (see Table 6).  Overall, the incidence of post-operative infection was 18.6% (n = 128).  Among the LEP/NEP group, 29.1% (n = 30) had a documented post-operative infection compared with 16.7% (n = 98) of the EP group (OR = 2.05; 95% CI [1.27, 3.30]).  Of the 128 patients with infections, the most prevalent types were: oral candidiasis (40.6%, n = 52), pneumonia (26.6%, n = 34), urinary tract infection (24.2%, n = 31), sternal wound infection (10.9%, n = 14), and others (15.6%, n = 20).7   Some patients required a transfer to a long-term care facility following their surgery (see Table 6).  The LEP/NEP patients were more likely to experience a delay in their transfer to another care facility (i.e., the patient was deemed ready for hospital discharge but accommodation was not available at another, more suitable, facility (8.7% [n = 9]) of the LEP/NEP patients experienced a delay in transfer compared with 2.7% (n = 16) of the EP patients (OR = 3.41; 95% CI [1.46, 7.93]).                                                       7 The total exceeds 100% because some patients had more than one type of infection. 62  Table 6   Post-operative characteristics of coronary artery bypass graft patients by English proficiency  All  EP   LEP or NEP  Characteristic N = 691 (100%) N = 588 (85.1%) N = 103 (14.9%) Statistic Returned to operating room (N (%))  16 (2.3) 14 (2.4) 2 (1.9) x2 (1) = 0.00a,b      Re-intubated (N (%))  13 (1.9) 7 (1.2) 6 (5.8) x2 (1) = 7.82 a,b*      Prolonged intubation (N (%))  32 (4.6) 23 (3.9) 9 (8.7) x2 (1) = 3.59 a,b*      Post-operative stroke (N (%))  18 (2.6) 14 (2.4) 4 (3.9) x2 (1) = 0.30 a,b       Delirium (N (%))  141 (20.5) 119 (20.3) 22 (21.4) x2 (1) = 0.01 a      Infection (N (%)) 128 (18.6) 98 (16.7) 30 (29.1) x2 (1) = 8.11 a*      Long-term care transfer delay (N (%))  25 (3.6) 16 (2.7) 9 (8.7) x2 (1) = 7.38 a,b* Note.  EP = English proficient; LEP = Limited English proficiency; and NEP = Not English proficient.  a Yates continuity correction applied. b Fisher’s exact test (2-sided) applied. * p < .05.  Length of Stay The outcome variable of interest, LOS after isolated cardiac bypass surgery, was assessed in three intervals: (a) the total LOS spent in the cardiovascular surgical intensive care unit (CSICU), (b) the total LOS spent in the post-operative nursing ward, and (c) the overall LOS from CSICU admission to hospital discharge (i.e., the sum of “a” and “b”).  The distribution of post-operative LOS from CSICU admission to hospital discharge stratified by English proficiency is presented in Figure 4. 63   Figure 4.  Post-operative LOS (in days) from CSICU admission to hospital discharge stratified by English proficiency. Note. Two patients were excluded from the figure because they were extreme outliers with LOS greater than 60 days.  The median LOS in the CSICU (see Table 7) was 1 day with a range of 1 to 31 days; 95% of the patients stayed five or fewer days.  When stratified by their English proficiency, Table 8 shows that the CSICU LOS was similar for the two groups (i.e., a median CSICU LOS of 1 day with the 95th percentile being 5 days).  The EP patients CSICU LOS ranged from 1 to 31 days and the LEP/NEP patients CSICU LOS ranged from 1 to 9 days. 64  The median post-operative nursing unit LOS was 5 days and ranged from 0 to 94 days (see Table 7).8  The median post-operative nursing unit LOS for the EP and LEP/NEP groups was also 5 days (see Table 8).  However, their LOS diverged at the 75th percentile; 25% of the EP patients stayed 6 to 64 days, whereas 25% of the LEP/NEP patients stayed 7 to 93 days (see Table 8).  This difference in LOS was greater at the 90th and 95th percentiles.  The post-operative nursing unit LOS for EP patients was 8 days and 12 days at the 90th and 95th percentile, respectively, and for LEP/NEP patients they were 13 and 20 days, respectively.   Table 7 shows that the patients’ median post-operative LOS, from CSICU admission to hospital discharge, was 6 days (ranging from 4 to 98 days).  Table 8 shows that when stratified by English proficiency, the EP patients stayed an average of 6 days, post-operatively, whereas the LEP/NEP patients stayed 7 days, on average, which was a statistically significant difference.  The groups differed in their post-operative LOS starting at the 25th percentile, which was 5 days for EP patients and 6 days for LEP/NEP patients.  The LEP/NEP group stayed 1 day longer than their EP cohorts at the 25th, 50th, and 75th percentiles; deviating further at the 90th percentile; 10% of the EP patients stayed 11 to 70 days and 10% of the LEP/NEP patients stayed 15 to 93 days.                                                     8 Two patients were discharged directly home from the CSICU because there was a bed shortage on the postoperative nursing unit (reflected by the LOS of 0 days).  65  Table 7   Length of stay, in percentiles, of all coronary artery bypass graft patients  Percentilea Outcome 5 10 25 50 75 90 95 Total LOS in CSICU in days 1 1 1 1 2 4 5         Total LOS in post-operative nursing ward in days 3 3 4 5 6 9 13         Total LOS from CSICU admission to hospital discharge in days 5 5 5 6 8 12 15 Note.  LOS = Length of stay and CSICU = Cardiovascular surgical intensive care unit.  a Weighted average method.  Table 8   Length of stay, in percentiles, of all coronary artery bypass graft patients stratified by English proficiency   Percentilea Mann Whitney U Outcome  5 10 25 50 75 90 95 p-value Total LOS in CSICU in days EP 1.0 1.0 1.0 1.0 2.0 4.0 5.0 .527 LEP/NEP 1.0 1.0 1.0 1.0 2.0 4.0 5.0            Total LOS in post-operative nursing ward in days EP 3.0 3.0 4.0 5.0 6.0 8.0 11.6 .002* LEP/NEP 3.0 3.4 4.0 5.0 7.0 13.0 19.8            Total LOS from CSICU admission to hospital discharge in days EP 5.0 5.0 5.0 6.0 8.0 11.0 15.0 .007* LEP/NEP 5.0 5.0 6.0 7.0 9.0 15.0 24.4  Note.  EP = English proficient; LEP = Limited English proficiency; NEP = Not English proficient; LOS = Length of stay; and CSICU = Cardiovascular surgical intensive care unit.  a Weighted average method. * p < .05.   In summary, the CSICU LOS did not differ between the EP and LEP/NEP groups.  The difference of the two groups’ LOS on the post-operative nursing unit and for the entire post-operative period (from CSICU admission to hospital discharge) was found to be statistically 66  significant when stratified by English proficiency.  The LEP/NEP patients stayed longer than their EP counterparts.  Neither the majority of EP nor LEP/NEP patients undergoing isolated CABG surgery met the targeted LOS established by the surgical clinical pathways.   The association between the patients’ characteristics and their post-operative length of stay. The patients’ post-operative LOS (from time of CSICU admission to hospital discharge) was categorized into quartiles (i.e., 1 to 5 days, 6 days, 7 to 8 days, and 9 or more days).  The patients’ LOS, when categorized, was not associated with their language proficiency or gender (see Table 9).9  These findings are in contrast to the statistically significant results calculated with the Mann Whitney U analysis and reported in Table 8.  Categorizing LOS into quartiles resulted in loss of statistical information.    Age was found to be statistically significantly associated with post-operative LOS; younger patients had the shortest length of stay (35.1% (n = 59) of patients under 60 years of age and 46.7% (n = 64) of patients 60 to 64 years of age had a post-operative LOS of less than 6 days.  Patients with the longest post-operative LOS were aged 75 years or older; 33.9% (n = 39) stayed 7 to 8 days and 30.4% (n = 35) stayed 9 or more days, post-operatively.   Marital status was associated with post-operative LOS (see Table 8).  More married and common-law patients (36.3%, n = 185) stayed less than 6 days compared with 26.3% (n = 15) and 20.0% (n = 16) of single and divorced or widowed patients, respectively.  Employment status was also associated with post-operative LOS (see Table 9).  About one half (40.9%, n =                                                  9 Language proficiency was found to be statistically significantly associated with the postoperative LOS distribution when tested with a Mann Whitney U test.  The association was not statistically significant with the data categorized into quartiles and tested with a Chi-square statistic. 67  97) of the employed patients had stays of less than 6 days compared with 29.6% (n = 99) of the retired patients and 27.3% (n = 15) of unemployed or disabled patients.   Table 9   Post-operative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patient demographics   Number of days    1 to 5 6  7 to 8 ≥ 9  Characteristic N N (%) N (%) N (%) N (%) Statistic Language proficiency           x2 (3) = 7.20 EP 588 203 (34.5) 119 (20.2) 144 (24.5) 122 (20.7)   LEP/NEP 103 23 (22.3) 24 (23.3) 26 (25.2) 30 (29.1)                Gender          x2 (3) = 5.60 Male 584 200 (34.2) 120 (20.5) 143 (24.5) 121 (20.7)   Female 107 26 (24.3) 23 (21.5) 27 (25.2) 31 (29.0)                Age          x2 (9) = 53.77* < 60 years  168 59 (35.1) 54 (32.1) 30 (17.9) 25 (14.9)   60 to 64 years 137 64 (46.7)  25 (18.2) 31 (22.6) 17 (12.4)   65 to 74 years 271 79 (29.2) 47 (17.3) 70 (25.8) 75 (27.7)   75+ years 115 24 (20.9) 17 (14.8) 39 (33.9) 35 (30.4)                Marital status          x2 (6) = 24.44* Single 57 15 (26.3) 12 (21.1) 16 (28.1) 14 (24.6)   Married/common-law 510 185 (36.3) 112 (22.0) 120 (23.5) 93 (18.2)   Divorced/widowed 80 16 (20.0) 10 (12.5) 23 (28.8) 31 (38.8)                Employment status          x2 (6) = 25.15* Employed 237 97 (40.9) 53 (22.4) 51 (21.5) 36 (15.2)   Retired 334 99 (29.6) 51 (15.3) 89 (26.6) 95 (28.4)   Disability/unemployed 55 15 (27.3) 16 (29.1) 13 (23.6) 11 (20.0)   Note.  EP = English proficient; LEP = Limited English proficiency; NEP = Not English proficient. * p < .05.  The patients’ health characteristics that were statistically significantly associated with English proficiency (i.e., BMI, hypertension, diabetes, and smoking) were cross tabulated with 68  post-operative LOS (see Table 10).  Of these health problems, diabetes was the only one that was statistically significantly associated with post-operative LOS.  About one quarter (26.8%, n = 81) of the patients with diabetes had a LOS of 9 or more days compared with 18.0% (n = 70) of those without diabetes.    Table 10   Post-operative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patients’ pre-operative characteristics   Number of days    1 to 5 6  7 to 8 ≥ 9  Characteristic N N (%) N (%) N (%) N (%) Statistic BMI  in kg/m2       x2 (9) = 5.14 < 25.5 157 47 (29.9) 31 (19.7) 39 (24.8) 40 (25.5)  25.5 to 27.4 180 59 (32.8)  43 (23.9) 41 (22.8) 37 (20.6)  27.5 to 30.4 165 61 (37.0) 33 (20.0) 38 (23.0) 33 (20.0)  30.5+ 168 50 (29.8) 33 (19.6) 46 (27.4) 39 (23.2)         Hypertension      x2 (3) = 4.81 No 66 28 (42.4) 13 (19.7) 10 (15.2) 15 (22.7)  Yes 625 198 (31.7) 130 (20.8) 160 (25.6) 137 (21.9)         Diabetes      x2 (3) = 8.08* No 388 137 (35.3) 83 (21.4) 98 (25.3) 70 (18.0)  Yes 302 89 (29.5) 60 (19.9) 72 (23.8) 81 (21.8)         Smoking      x2 (6) = 5.79 Never 311 108 (34.7) 69 (22.2) 67 (21.5) 67 (21.5)  Former 63 102 (32.4) 58 (18.4) 85 (27.0) 70 (22.2)  Current 315 15 (23.8) 15 (23.8) 18 (28.6) 15 (23.8)  * p < .05.  No intra-operative characteristics were examined in relation to LOS because none was statistically significantly associated with English proficiency.  The patients’ post-operative characteristics that were statistically significantly associated with English proficiency (i.e., 69  infection, reintubation, and long-term transfer delay) were examined with respect to post-operative LOS.  As Table 11 shows, all these characteristics were statistically significantly associated with post-operative LOS.   About one half (48.4%, n = 62) of the patients that acquired a post-operative infection stayed 9 or more days in hospital following their surgery compared with 16.0% (n = 90) of those without an infection.  Conversely, more than one third of patients without an infection had a post-operative stay of less than 6 days (38.3%, n = 215) compared with 8.6% (n = 11) of those who acquired infections.   Patients that required reintubation also had statistically significantly longer post-operative LOS.  Almost all (92.3%, n = 12) of the re-intubated patients stayed 9 or more days post-operatively compared with 20.7% (n = 140) of patients not re-intubated (see Table 11).  Prolonged intubation (i.e., 24 hours or more) also was associated with longer LOS.  Two thirds (65.6%, n = 21) of the patients with prolonged intubation stayed 9 or more days compared with 19.9% (n = 131) of patients extubated within a day of surgery.   The patients that were required to wait for a transfer to long-term care obviously had longer LOS; 68.0% (n = 17) of these patients stayed 9 or more days compared with 20.2% (n = 134) of the patients that did not have a long-term care transfer delay.  In summary, several characteristics of the patients were associated with both their English proficiency and post-operative LOS, including their age, marital status, employment status, and history of diabetes.  Post-operative characteristics that were associated with language proficiency and LOS included infection, reintubation, prolonged intubation, and a delay in transfer to long-term care.  These variables were subsequently treated as possible confounders or intermediate variables in the association between language proficiency and post-operative LOS.  70  Table 11   Postoperative length of stay (from cardiovascular surgical intensive care unit admission to hospital discharge) by patients’ post-operative characteristics   Number of days    1 to 5 6  7 to 8 ≥ 9  Characteristic N N (%) N (%) N (%) N (%) Statistic Infection acquired      x2 (3) = 78.66* No infection 561 215 (38.3) 120 (21.4) 136 (24.2) 90 (16.0)  Yes, infection acquired 128 11 (8.6) 22 (17.2) 33 (25.8) 62 (48.4)         Reintubation      x2 (3) = 27.11a* No 677 226 (33.4) 143 (21.1) 168 (24.8) 140 (20.7)  Yes, reintubated 13 0 (0.0) 0 (0.0) 1 (7.7) 12 (92.3)         Prolonged intubation       No 659 226 (34.3) 140 (21.2) 162 (24.6) 131 (19.9) x2 (3) = 42.09* Yes, ≥ 24 hours 32 0 (0.0) 3 (9.4) 8 (25.0) 21 (65.6)         Long-term care transfer delay      x2 (3) = 33.34* No 663 224 (33.8) 140 (21.1) 165 (24.9) 134 (20.2)  Yes, delayed 25 1 (4.0) 3 (12.0) 4 (16.0) 17 (68.0)  a Fisher’s exact test (2-sided) applied. * p < .05.  Multivariate Analyses: English Proficiency and Length of Stay In the multivariate analyses, variables that were not bivariately statistically significantly associated with language proficiency and post-operative LOS were excluded.  Reintubation and prolonged intubation, although statistically significantly associated with English proficiency and LOS were not included in the multivariate model because the number of patients experiencing these events were too small to estimate reasonable standard errors and thus the model was unstable in their presence.  The variables that were found to be associated with language proficiency and post-operative LOS were entered into the model and then removed if not 71  statistically significant in the multivariate model (i.e., diabetes, employment status, and long-term care transfer delay).    Table 12 summarizes the findings of the final polynomial regression model.  This model estimated the LOS (in quartiles) from the time of CSICU admission to hospital discharge, regressed on English proficiency and adjusted for all other statistically significant variables: age, marital status, and post-operative infection.    In the unadjusted model, the LEP/NEP patients were found to be twice as likely (OR = 2.17, 95% CI [1.21, 3.91]) to stay 9 or more days rather than 5 or fewer days, compared with the EP patients (see Table 12).  After adjusting for differences in age, marital status, and post-operative infection, language proficiency was no longer statistically significantly associated with post-operative LOS.    The multivariate model was significantly different from an intercept only model (x2 = 140.22, df = 21, p < .001).  The Nagelkerke R2 was 20.9%.  Those likely to have a LOS of 9 or more days were older, not married, and had a post-operative infection.  Relative to having a stay of less than 6 days, patients aged 75 or more years were 3.71 times more likely to have a 9 or more day LOS (95% CI [1.69, 8.14]) compared with patients under 60 years.  They were 3.45 times more likely to have a stay of 7 to 8 days (95% CI [1.70, 7.01]).   Marital status was also statistically predictive of post-operative LOS in the multivariate model.  Relative to having a LOS of less than 6 days, single and divorced or widowed patients were 2.5 times more likely of having a 9 day or longer post-operative stay (95% CI [1.22, 5.15]) compared with married or common-law status patients.    Post-operative infection was the strongest predictor of LOS.  Relative to having a LOS of less than 6 days, patients with infections had odds of staying longer as follows: adjusted OR = 72  3.30 (95% CI [1.51, 7.19]) for 6 days; adjusted OR = 4.37 (95 % CI [2.09, 9.14]) for 7 to 8 days; and adjusted OR = 12.4 (95 % CIs [6.03, 25.48]) for 9 days or more. In summary, after adjusting for age, marital status, and post-operative infection, language proficiency was not a statistically significant predictor of LOS.  These predictors remained statistically significant in the final polynomial model, with post-operative infection being the strongest predictor of LOS.  It appears that the LEP/NEP patients were older, which may explain their relatively longer LOS.  However, they were also more likely to develop post-operative infections, which meant that they were 12.4 times more likely to have a 9 or more day LOS, rather than a LOS of fewer than 6 days, compared with EP patients’ odds of a 9-day LOS.  The differences in age, marital status, and infection fully attenuated the relationship between English proficiency and LOS; post-operative infection is likely an intermediate variable in the pathway between language proficiency and LOS.  Table 12   Multivariate polynomial regression model of post-operative length of stay on language proficiency and other patient characteristics   Unadjusted  Adjusted  Number of Days Predictors OR (95% CI) p-value OR (95% CI) p-value 6a Language proficiencyb      LEP/NEP 1.78 (0.96-3.29) .07 1.52 (0.77-2.98) .23        Age (in years)c      60 to 64    0.39 (0.21-0.74) .01*  65 to 74    0.64 (0.38-1.10) .11  75+   0.79 (0.37-1.67) .53        Marital statusd      Single   1.18 (0.51-2.71) .69  Divorced/Widowed   0.96 (0.41-2.22) .92       73    Unadjusted  Adjusted  Number of Days Predictors OR (95% CI) p-value OR (95% CI) p-value  Postoperative infection   3.30 (1.51-7.19) .01*       7 to 8a Language proficiencyb      LEP/NEP 1.59 (0.87-2.91) .13 1.55 (0.81-2.97) .19        Age (in years)c      60 to 64    1.02 (0.54-1.94) .95  65 to 74    1.82 (1.02-3.24) .04*  75+   3.45 (1.70-7.00) .01*        Marital statusd      Single   2.08 (0.96-4.48) .06  Divorced/Widowed   1.68 (0.83-3.41) .15        Postoperative infection   4.37 (2.09-9.14) .01*       9+a Language proficiencyb      LEP/NEP 2.17 (1.21-3.91) .01* 1.85 (0.94-3.67) .08        Age (in years)c      60 to 64    0.67 (0.30-1.48) .32  65 to 74    2.49 (1.31-4.71) .01*  75+   3.71 (1.69-8.14)  .01*        Marital statusd      Single   2.57 (1.11-5.94) .03*  Divorced/Widowed   2.50 (1.22-5.15) .01*        Postoperative infection   12.4 (6.03-25.48) .01* Note.  OR = Odds ratio; CI = Confidence Interval; EP = English proficient; LEP = Limited English proficiency; NEP = Not English proficient. aReferent = 1 to 5 days. bReferent = EP.  cReferent = < 60 years. dReferent = Married/Common-law. *p < .05.    74  Chapter Five: Discussion This chapter summarizes the purpose of this study and its key findings, which are compared with those reported in the published literature.  The limitations of the study are discussed and the chapter concludes with some implications for clinical practice, policy, and future research. The multilingual landscape of BC creates a dynamic and challenging environment for the provision of accessible and quality health care for all patients.  This study explored the relationship between English proficiency (EP) and the length of patients’ hospital stays (LOS) following isolated cardiac bypass surgery.  The study was designed to test the hypothesis that there is a relationship between EP and LOS; that is, patients with limited English proficiency (LEP) were postulated to have prolonged LOS.  This study also endeavoured to determine whether LEP and EP patients met the coronary artery bypass graft (CABG) clinical pathway targets specified by the study hospital.  To review, the research hypotheses of this study were: 1. On average, patients with LEP have a longer LOS after isolated CABG surgery when compared with EP patients. 2. Relatively fewer LEP patients undergoing isolated CABG surgery meet the established targeted LOS of the CABG surgery clinical pathway compared with EP patients. Overview of the Findings The findings demonstrated a statistically significant relationship between post-operative LOS (from cardiac surgical intensive care unit [CSICU] admission to hospital discharge and from post-operative ward to discharge) and English proficiency.  The “limited or not English proficient” (LEP/NEP) patients had a median post-operative stay that was one day longer than 75  that of the EP patients (7 days versus 6 days, p = .007).  Therefore, the results of this study provide evidence for the first hypothesis.  The LOS was also associated with the patients’ age and marital status, although they did not attenuate the relationship between English proficiency and LOS.  Developing a post-operative infection was also positively associated with both EP status and LOS.  The LEP/NEP patients had greater odds of developing post-operative infections compared with the EP patients’ odds of post-operative infection.  Post-operative infection was the strongest predictor of LOS and was found to mediate the relationship between English proficiency and LOS.   Regarding the second hypothesis of this study, the findings revealed that the median LOS of all patients, irrespective of their English proficiency, exceeded the targeted goal of five days specified in the hospital site’s CABG surgery clinical pathway.  The median LOS stay for all patients from CSICU admission to hospital discharge was six days.  About one half of the EP patients met the post-operative LOS target.  The LEP/NEP patients’ median stay was 7 days, 1 day longer than that of the EP patients.    The Findings in Relation to the Literature Differences in the patients’ characteristics.   The sampling strategy of this study was consecutive sampling of all isolated CABG surgery patients (elective and urgent) between January 1, 2011 and December 31, 2012.  This form of sampling included all accessible subjects and thus should provide a reasonable representation of the population.  It should be noted that the studies selected for comparison to this study’s sample did not have the same study purpose nor necessarily the same sampling strategy; only one previous study examined LEP patients in the CABG population (Baptiste et 76  al., 2004).  Thus, differences between sample characteristics were expected.  The purpose of comparing the samples is to provide clarity and guidance to the degree to which the current study results can be generalized to other isolated CABG patient populations or other reasons for hospitalization.   Some of the characteristics of the study sample were different from CABG bypass surgery patients described in the literature.  The average age (65.7 years) in this study was comparable to other CABG-related studies.  In Pearson et al.’s (2008) study of clinical pathway CABG patients, the sample had an average age of 67.7 years.  Filon, Pilote, Rahme, and Eisenberg’s (2008) study included CABG patients that were slightly older, at 66.2 years on average.  In another study, Eisenberg et al. (2005) compared the LOS and hospital costs of Canadian and American CABG patients; their Canadian patients were 63.7 years of age, on average, which is somewhat younger than the current study’s sample.  Thus, the average age of this sample was neither considerably younger nor older than other studies conducted with the CABG population.   The EP patients were younger than the LEP/NEP patients in this study.  A potential explanation for this occurrence is the healthy immigrant effect.  Research has observed and documented a phenomenon whereby immigrants' health is generally better than that of Canadian-born people (Ng, Wilkins, Gendron, & Berthelot, 2005).  Even though the healthy immigrant effect diminishes with years since immigration to Canada, these effects remained present when measured 20 years later (Ng, 2011).  Thus, the healthy immigrant effect may be responsible for LEP/NEP patients presenting with heart disease at a later age compared with their EP counterparts.  Alternatively, as described in Chapter two, there is substantial research describing how LEP/NEP patients have less access to health care (Bernstein et al., 2002; Carrasquillo et al., 77  1999) and are less satisfied with their medical care (Carrasquillo et al., 1999; Morales et al., 1999; Ngo-Metzger et al., 2009).  If this is the case, then LEP/NEP patients may have faced significant barriers to care, which may have delayed receiving attention for their disease, resulting in later presentation and management.  However, they did not appear to be more severely ill in the pre-operative phase of their surgery.   The ratio of women to men in this study reflects the population of patients receiving CABG surgery in BC.  In 2011, 17.6% of CABG patients were women (Cardiac Services BC Annual Report, 2011).  Compared with other studies, there were a smaller percentage of women in this sample; women composed 15% of the sample and in other studies have made up 19% to 43% of the sample (Cowper et al., 2006; Eisenberg et al., 2005; Johnston et al., 2004; Pearson et al., 2008; Peterson et al., 2002).  This difference is important because being female is a significant predictor of prolonged LOS after CABG surgery (Johnston et al., 2004; Rosen, 1999; Sawatzky & Naimark, 2009).  Previous studies have concluded that women have a higher risk of complications when undergoing CABG surgery because they tend to be older in age when they develop coronary artery disease, have more comorbidity, and are of lower socioeconomic status compared with men.  All of these factors contribute to prolonged LOS.  Of interest, being female was not associated with a longer LOS in this study.   The patients’ pre-operative health characteristics, including hypertension, hyperlipidemia, and diabetes, were all more prevalent in the study sample compared with samples described in the literature (Eisenberg et al., 2005; Fowler et al., 2005; Johnston et al., 2004).  It should be noted though, that the CABG related study samples described here were acquired from multicentre sites and their purposes for sampling were not similar to this study’s.  78  A history of diabetes was very high in this study at 43.8% when compared with 25.5% in the profile of Eisenberg et al.’s (2005) Canadian study and 34.9% in Fowler et al.’s (2005) American study of CABG patients.  There may be several explanations for this difference.  First, the authors of these studies did not outline their criteria for diabetes diagnosis and disagreement between classifications of diabetes diagnosis may exist depending on which guidelines are used (DECODE study group, 1998).  Second, the current study classified patients as having diabetes as long as a physician’s consultation note documented the diagnosis in the medical record.  Thus, the current study may have been more sensitive to capturing a diabetes diagnosis because we considered any documentation of diabetes in the medical records as positive for disease.  Third, the sample in this study may have had a larger representation of people of certain ethnicities with greater risk of diabetes and other cardiovascular risk factors compared with previous studies where the predominant race/ethnicity of the samples was white or of European origin.  Whether attributed to genetics or culturally-related behaviour, South Asians are known to have higher prevalence rates of diabetes and other risk factors for cardiovascular disease (Anand et al., 2000).  South Asians represent the largest minority group in Canada, accounting for one-quarter of the total minority population and 4.8% of Canada’s total population (Statistics Canada, 2013).  Vancouver has the second largest percentage of South Asians in BC.  Because this study was based in Vancouver, there is a likelihood of greater representation of the South Asian population.   The LEP/NEP patients in this study had a greater rate of hypertension and diabetes compared with the EP patients.  The EP group, however, had a greater proportion of current and former smokers and a higher average BMI.  No published studies describing differences in cardiovascular risk factors stratified by English proficiency were found.  In a Canadian study comparing the cardiovascular risk factors of various ethnic groups, Chiu, Austen, and Manuel 79  (2010) reported that the “white” population had a statistically significantly higher prevalence of smoking and obesity, whereas the South Asian (i.e., Indian or Pakistani descent) and “black” (i.e., African or Caribbean descent) populations had higher prevalence rates of diabetes and hypertension.  In the current study, 41.7% of the LEP/NP patients spoke Punjabi, which may explain the rates of diabetes and hypertension in the sample.   Within the study, the EP patients had a higher average BMI and yet lower rates of diabetes and hypertension than the LEP/NEP patients.  This difference may be explained by ethnicity because the majority of the LEP/NEP patients in this study came from South Asian and Asian backgrounds.  Compared with the national BMI averages, adult mean BMI levels of 22-23 kg/m2 are found in Africa and Asia, whereas BMI levels of 25-27 kg/m2 are prevalent across North America and Europe (World Health Organization, 2003). With respect to the other pre-operative characteristics, the EP and LEP/NEP groups were reflective of the samples described in other LEP studies.  Similar to Baptiste et al.’s (2004) study, the LEP/NEP patients were older, more likely to be women, and more likely to be married compared with the EP patients.  Baptiste et al.’s (2004) study reported that their sample’s LEP patients had lower incomes (based on patients’ residential postal codes) than their EP counterparts.  Income levels were not collected for the current study, but the LEP/NEP group was more likely to be retired, unemployed, or collecting disability insurance than were the EP patients, suggesting that the LEP/NEP patients had relatively lower incomes. Several factors may explain the differences in employment between the EP and LEP/NEP patients.  As discussed earlier, the LEP/NEP patients were older than the EP patients.  Plausibly then, the LEP/NEP patients were closer to the age of retirement.  In addition, a proportion of the LEP/NEP patients could have been ‘family class immigrants,’ parents or grandparents sponsored 80  by relatives to immigrate to Canada (Vanderplaat, Ramos, & Yoshida, 2012).  Because family class immigrants are typically older and often not able to speak English, they have fewer opportunities to work outside the home.  In addition, cultural expectations and norms may be at play.  The majority of the sample’s LEP/NEP patients spoke Asian and South Asian languages.  These cultures are known to have a highly supportive network of extended families (Vanderplaat et al., 2012).  It is plausible that the ‘retired’ LEP/NEP patients chose to stay at home and to contribute their labour in other forms (e.g., child care) (Vanderplaat et al., 2012).  There are several noteworthy differences between this study and the current LEP literature.  Among the 103 LEP and NEP patients of the current study, the most common language spoken was Punjabi.  One quarter of the LEP/NEP patients spoke Cantonese.  The majority of the comparison studies of LEP patients were conducted in the US, with the predominant comparison being between patients that spoke Spanish and those that spoke English (Bernstein et al., 2002; Cohen et al., 2005; Divi et al., 2007; Elderkin-Thompson, 2001; Grubbs et al., 2008; Levas et al., 2011; Wilson et al., 2005).  The most common language spoken in Baptiste et al.’s (2004) Canadian study was Portuguese followed by Italian.  In addition, other than a CABG subgroup in Baptiste et al.’s (2004) study, and the MI patients in Grubbs et al. (2008) study, the admission diagnoses in the other LEP studies were not related to CABG surgery.  Thus, comparing the results of such LEP studies with the current study is difficult.   The differences in the various studies’ sample characteristics are important because they speak to the current sample’s specific population and to the unique make-up of BC’s inhabitants.  Compared with previous studies about CABG surgery patients, this study sample was composed of relatively fewer women, had relatively higher rates of selected comorbidities (including hypertension, hyperlipidemia, and diabetes) and had a higher proportion of specific ethnic groups 81  (e.g., South Asian and Chinese people).  This study sample is also not similar to the typical profile of study participants in other LEP/NEP studies because the majority of previous LEP/NEP studies focused on different ethnicities and varying reasons for hospitalizations.  These factors are important to consider when comparing the results to those of other studies and in generalizing the results of this study to other populations.      Factors affecting the relationship between English proficiency and length of stay.  The findings of this study support the literature that shows that limited English proficiency results in longer hospital LOS (Baptiste et al., 2004; Levas et al., 2011; MacDonald et al., 2010).  As discussed in chapter two, Baptiste et al.’s (2004) Canadian study demonstrated that the 746 LEP patients having CABG surgery had odds of staying longer post-operatively that were 1.07 times (95% CI [1.03, 1.12]) greater than EP patients’ odds.  Overall, their study found a 6% (0.5 days) increase in the LOS of LEP patients compared with EP patients.  The current study found a 1.0 day increase in LOS of LEP patients compared with EP patients.   Of interest, Baptiste et al. (2004) noted that the effect of English proficiency on LOS was greater with other diagnoses and procedures (i.e., stroke admission).  The researchers suggested that various contextual factors (i.e., quality of treatment, time to procedure, complications) not measured in their study may have also influenced the relationship between English proficiency and LOS.  Unlike Baptiste et al. (2004), who did not account for the hospital complications of their study participants, the current study explored such contextual factors and found that post-operative infection was an important mediator of the relationship between English proficiency and LOS.  Thus, investigating the mechanisms by which English proficiency is mediated is vital in accurately revealing its relationship with LOS.    82  Another influencing contextual factor was observed in Grubbs et al.’s (2008) study of MI patients.  They found a longer LOS in NEP patients compared with EP patients in their unadjusted analysis (OR = 3.9%, 95% CI [0.7%, 7.1%]) although the relationship was no longer statistically significant after controlling for history of a cardiac procedure after admission (i.e., catheterization coronary artery stent or bypass surgery) (OR = 2.8%, 95% CI [-0.6%, 6.2%]).  Thus, Grubbs et al.’s (2008) findings suggest that hospital-level characteristics, such as access to quality level services and staff may contribute to differences in patients’ LOS.  Like Grubbs et al. (2008), the current study controlled for a cardiac procedure because all of the enrolled patients received the same type of surgery.  But the study was not able to adjust for other characteristics (e.g., emergency versus elective surgery) that may have yielded differences in processes of care and thus have potentially influenced the relationship between English proficiency and patients’ LOS.  The importance of hospital-level characteristics may explain why there were no differences found in the LOS in the CSICU when stratified by language proficiency.  Patient care in the CSICU is highly driven by the patient’s peri-procedural course and acute medical condition (which is monitored continuously) and by standard algorithms of care.  While the patient is in the CSICU, the focus is on managing intra-operative complications.  The goals of care in the CSICU focus on the stabilization of the patient, rather than rehabilitation and transition to discharge on the ward.  Rehabilitation and recovery from surgery is dependent upon active patient participation in care.  These circumstances would mitigate the reliance on verbal communication between patient and clinician.  It is then conceivable that such hospital-level characteristics minimized the influence of English proficiency on LOS in the CSICU to the point where little to no relationship was seen.  83  The theoretical framework of this study recognizes how patient-provider language concordance facilitates health communication (Northouse & Northouse, 1998).  Other hospital-level factors, such as the availability and use of both formal and informal interpreters, are important elements that were not measured or controlled in this study.  Evidence of these supports may have had a moderating effect on the relationship between language and LOS (Flores, 2006; Hyman, 2009; Ngo-Metzger et al., 2007; Shi, Lebrun, & Tsai, 2009; Wilson et al., 2005).   The literature strongly supports the assertion that socioeconomic factors are predictors of LOS (Coevoet, Fresson, Vieux, & Jay, 2013).  Socioeconomic characteristics affect a patient’s access to care, the quality of care received, and the ability to recover from a health condition (Coevoet et al., 2013; Ponce et al., 2006; Shi et al., 2009).  This study found that employment status was statistically significant in the bivariate analysis with English proficiency and with LOS.  However, in the multivariate analysis of LOS regressed on the selected predictors, employment status was not found to be statistically significant.  This may have occurred because it was confounded with age.  Because this study was not able to consider the race/ethnicity, education level, or income level of the patients, the full effects of these additional socioeconomic influences on English proficiency and LOS could not be examined.   Although there is a paucity of studies that have examined the association between LEP and hospital LOS, as discussed in Chapter two, there is substantial evidence that patients with language barriers are disadvantaged in the occurrence of adverse events when compared with their EP counterparts (Bartlett et al., 2008; Divi et al., 2007; Hyman, 2009; Karliner et al., 2010).  Of the limited post-operative adverse events recorded in this study, reintubation, prolonged intubation, delayed transfer to long-term care, and post-operative infection (which are all 84  considered adverse events) occurred more frequently in the LEP/NEP group compared with the EP group.  Collecting and analyzing all recorded adverse events would have provided additional information about the differences in LOS and English proficiency.  These findings warrant further investigation and better understanding of the phenomenon of interest.   English proficiency, post-operative infection, and length of stay. When included with English proficiency in the polynomial regression model, post-operative infection was found to be the strongest predictor of LOS.  The patients with post-operative infection had a 12-fold increased odds (95% CI [6.03, 25.48]) of staying nine or more days, rather than 5 or fewer days.  Previous research has demonstrated that post-operative infections are a major cause of poor patient outcomes and prolonged LOS (Cristofolini, Worlitzch, Wienke, Siber, & Borneff-Lipp, 2012; Fowler et al., 2005; Lapsey & Vogels, 1998; Peterson et al., 2002).  An important distinction between this study and previous research is that the majority of the literature is dedicated to studying post-operative major infections and their effects on LOS.   Fowler et al.’s (2005) review of 331,429 CABG procedures in the US between 2002 and 2003 revealed that 3.51% of patients had major post-operative infections (i.e., surgical site infections or septicemia).  Although less frequent than other complications that arise following CABG surgery, major infections can account for up to 50% of hospital stays greater than 14 days (Lapsey & Vogels, 1998; Peterson et al., 2002).  For example, Cristofolini et al.’s (2012) study of CABG patients (in combination with all surgical procedures) reported that the median LOS for patients with surgical site infections was 37 days compared with 13 days (p < .001) in patients without surgical site infections.   85  The study sample’s overall infection rate was 18.6%.  This higher rate may be partially explained by the coding scheme applied, which did not differentiate between minor and major infections (i.e., oral candidiasis versus surgical site infections).  Infections were coded to be present when a medical record was noted to have a positive culture or diagnosis in the physician’s notes.  Accordingly, the lack of distinction between minor and major infections should be considered in the interpretation of this study’s results.  Nonetheless, of the 128 patients with infections in this study, major infections, such as pneumonia (n = 34) and sternal wound infections (n = 14), accounted for 58.6% of all the infections or 7% of the total sample.  Oropharyngeal candidiasis, a superficial infection, was the most common type of infection noted in the patients’ charts.  Although not life threatening, this nosocomial infection is resistant to antifungal agents and poses significant problems for hospitalized patients (Pfaller, 1999).  Thus, any form of post-operative infection in the CABG surgery population is an important finding because of the implications for patients’ recovery and for the associated healthcare costs (Fowler et al., 2005; Gahli, Hall, Ash, & Moscowitz, 1999).  These findings reflect the study site’s current struggle with hospital-related infections.  Urinary tract infection, surgical site infection, and pneumonia rates within the cardiac surgical program were a primary focus of the hospital at the time of writing.  The cardiac surgical program’s surgical site infection (inclusive of any incisional wounds such as sternum, leg, or arm) and pneumonia rates were 5.2% and 5.9%, respectively, between October 2012 and July 2013 (BC Patient Safety and Quality Council, 2013).10  At the time of this study, these infection rates had resulted in an initiative to reduce surgery-related infections and to improve surgical cardiac care (BC Patient Safety and Quality Council, 2014).                                                   10 The cardiac surgical program includes surgery for valves, CABG (open chest and robotic assisted), aortic and thoraco-abdominal aneurysms or dissections (open chest and minimally invasive endovascular grafts).  86  The findings of this study are clinically significant in many aspects.  It is widely accepted in the literature that post-operative infections are responsible for prolonged length of stay after surgery and this point was demonstrated in this study.  Post-operative infections are not only costly to the healthcare system, but have negative (and potentially devastating) effects on patients’ health and recovery.  The high infection rates in this study call for re-evaluation of the delivery of care that may be compromising patient health.   This study also demonstrated that LEP/NEP patients were at a higher risk for developing post-operative infections, which further supports the importance of considering other factors that contribute to an outcome.  The study site’s current interventions to reduce surgical-related infections stem from evidenced-based practices (BC Patient Safety and Quality Council, 2014), but this study’s findings also highlight the importance of non-conventional factors that could be affecting infection rates.  Reflective of previous studies, age, diabetes, and increased BMI were found to be associated with prolonged LOS (Collins et al., 1999; Fowler et al., 2005; Johnston et al., 2004).  These variables are also determinants of post-operative infection complications (Collins, Daley, Henderson, & Khuni, 1999; Fowler et al., 2005).  Of these variables in the current study, only age remained statistically significant in the final multivariate polynomial regression model.  The LEP/NEP patients in this study were older than the EP patients, which may have partly explained their increased risk of post-operative infections.   The LEP/NEP patients were found to be at greater risk of developing post-operative infections than were the EP patients.  These findings suggest that post-operative infection was an intermediate variable in the relationship between English proficiency and LOS.  English proficiency influenced the likelihood of developing a post-operative infection, which was 87  associated with a longer LOS.  Accordingly, knowing that a patient is LEP/NEP has particular clinical value.  Like aging, having diabetes, and being obese, limited English proficiency may be an additional identifiable characteristic of patients at high risk for post-operative infections. There has been no research specifically linking the association between English proficiency and post-operative infection rates.  There is existing literature demonstrating that LEP patients have more adverse events while in hospital (Bartlett et al., 2008; Cohen et al., 2005; Divi et al., 2007; Jacobs et al., 2006; Karliner et al., 2010).  In this study, adverse events that were bivariately statistically significantly associated with LEP/NEP status and LOS included reintubation and prolonged intubation.  These events are known risk factors for post-operative infection (i.e., ventilator acquired pneumonia).  Unfortunately, we could not enter reintubation and prolonged intubation in the multivariate polynomial model because there were so few cases; their overall impact in the model was not examined.  Notwithstanding this limitation, these data support the literature that LEP patients experience more adverse events compared with EP patients.  Coupled with the evidence that the majority of adverse events in LEP patients are attributable to some failure in communication (Bartlett et al., 2008; Divi et al., 2007), this finding is compelling.  This study raises questions about whether patient-provider language barriers are responsible for higher post-operative infection rates in LEP/NEP patients, thus, influencing their outcomes and LOS.  If so, it is conceivable that language barriers may have inhibited the clinicians’ ability to elicit symptoms, resulting in adverse events such as the need for reintubation and prolonged intubation.  The results of this study support the study’s theoretical framework proposing that barriers to communication may adversely affect patients’ understanding of post-operative educational offerings and the importance of healthcare providers’ capacity to tailor education to individual learning needs.  Any patient with poor 88  understanding of the requirements to achieve a successful post-operative recovery would be at risk for poor adherence to therapies.    From the study findings, it is interesting to note that, given the higher proportion of current smokers in the EP patients versus the LEP/NEP patients, the EP patients were not at greater risk of requiring reintubation, having prolonged intubations, or developing post-operative infections.  The literature has shown that current smokers have an increased risk of pulmonary complications, including prolonged intubation, reintubation, and pneumonia (Al-sarraf et al., 2009; Saxena et al., 2013).  Current smokers are also at risk for wound healing complications, including dehiscence and surgical site infections (Sorensen, 2012).  One would anticipate, then, a higher incidence rate of post-operative pulmonary complications, including infection in the EP group.  The findings revealed, however, that LEP/NEP patients had higher post-operative infection rates.  These patients also experienced higher rates of reintubation and prolonged intubation.  These results are clinically important because they raise questions about the relationship between post-operative respiratory support (e.g., intubation), infection, and language proficiency.  If language barriers contribute to higher reintubation and prolonged intubation rates, then it is conceivable that such events are also responsible for more post-operative infections in LEP/NEP patients.  Such speculation requires further exploration to be substantiated.     Other predictors of length of stay. Marital status and age were significant independent predictors of LOS.  These associations are well supported.  Being of single marital status is an independent risk factor for several important hospital outcomes possibly because it indicates a lack of social support, which 89  is known to influence recovery and long-term outcomes (Barbash et al., 2013; Gordon, 1999).  Barbash et al. (2013) found that being married leads to better short- and long-term outcomes for patients undergoing elective or urgent coronary revascularization.  Another study of patients undergoing cystectomy for bladder cancer, found that married patients had relatively shorter hospitalizations (Pruthi, Lentz, Sand, Kouba, & Wallen, 2009). As discussed earlier, age was also a predictor of LOS in this study.  Age has been established in the literature as a risk factor for poor patient outcomes (Collins et al., 1999; Fowler et al., 2005; Johnston et al., 2004; Rosen et al., 1999).  Rosen et al.’s (1999) study of 3,605 CABG patients demonstrated that a 10-year increase in age from 70 to 80 years extended the LOS in hospital by 1.35 days.  In another study, Peterson et al. (2002) found that patients over 75 years of age had significantly longer LOS than patients less than 65 years.  The older patients were three times more likely to have a hospital stay of more than 14 days compared with those aged less than 65 years (OR = 1.62, p < .05) (Peterson et al., 2002).  This finding is of clinical importance to policy makers and healthcare leaders.  Coupled with the knowledge of current patterns of immigration and globally increasing life expectancies, knowing that older patients are more at risk for a longer LOS after CABG surgery adds to the growing pressure for adequate healthcare planning and allocation of resources for the aging population.    Clinical pathway targets and length of stay. As outlined in Chapter two, the expected LOS of the study site’s clinical pathway for CABG surgery is 5 days and thus, prolonged LOS was defined as being any stay greater than 5 days.  With this definition, the LEP/NEP had greater odds of having a prolonged stay of 5 or more days compared with EP patients (OR = 1.83, 95% CI [1.12, 3.00]).  Not anticipated, 90  though, was the finding that both the EP and LEP/NEP patients’ median LOS was longer than the 5-day clinical pathway target.  The sample’s overall median LOS was 6 days.  Although the EP group had a median LOS of 6 days, the LEP/NEP group exceeded this average by 1 day.   Failure to meet the pathway target LOS for CABG patients suggests that the pathway should be revisited to take into account non-modifiable risk factors.  It is generally accepted at the study site that the LOS goal of the clinical pathway is not always applicable for patients with complex medical histories or those with advanced age.  Although this understanding is based on the subjective experiences of the hospital staff, the literature reflects this belief.  Age, gender, marital status, severity of disease, and comorbidities are key determinants of LOS after CABG surgery (Cowper et al., 2006; Ghali et al., 1999; Pruthi et al., 2009).  The clinical importance of not ‘achieving’ pathway targets is minimized by the fact that such risk factors for prolonged LOS are not modifiable.  Therefore, the current CABG clinical pathway may not be well suited to the study site’s population or there may be systemic issues throughout the trajectory of care reflected in a prolonged LOS. There is conflicting opinions regarding the indiscriminate determination of LOS identified in clinical pathways and the efficacy of clinical pathways (Pearson et al., 2001).  Several researchers have found considerable unexplained inter-hospital variability in the post-operative LOS of CABG patients (Cowper et al., 2006; Peterson et al., 2002).  For example, the downward trend of LOS after CABG surgery during the 1990s has been attributed to the implementation of ‘fast-track’ clinical pathways.  But the change may have resulted from an increase in the number of patients transferred to non-acute facilities following surgery, rather than staying in acute care (Cowper et al., 2006).  Thus, although clinical pathways appear to have 91  contributed to shorter LOS post-operatively, streamlining care delivery also has shortened LOS (Cowper et al., 2006).    Of the two studies that discussed the effects of  transferring patients to outlying hospitals on LOS, the researchers either acknowledged that they did not account for such differences (Grubbs et al., 2008) or excluded cases that were transferred to or from another acute care facility (Baptiste et al., 2004).  This current study attempted to account for patients that were ready for discharge to long term care, but who were delayed because of bed unavailability.  This study did not, however, account for patients that were not necessarily fit for discharge home, but were transferred to their referring hospital for further recovery.  This would have been an important factor to consider because, at the time of the study, patients from outside of Vancouver were often transferred from other hospitals that did not perform CABG surgeries or were unable to accommodate the surgery.  The exclusion criteria of the various studies could have affected the homogeneity of the samples and complicated any comparisons made.  Measurement of English proficiency.  Previous studies have utilized various measures of English proficiency, some of which were not validated.  Several studies have measured patients’ language ability through self-reports of English proficiency (Lebrun et al., 2012; Levas et al., 2011; Lindholm et al., 2012; Ponce et al., 2006) or preference of language, whereas other researchers have relied on the information recorded on admission forms in medical records (Baptiste et al., 2004; Grubbs et al., 2008).  Some researchers have not reported how they measured English proficiency (e.g., Divi et al., 2007; McDonald et al., 2010).  In addition, Gee et al. (2010) documented how concepts of language preference and English proficiency were often incorrectly interchanged in studies.   92  Because of the challenges in accurately measuring patients’ English proficiency in research, this study strived to apply an objective form of measurement with the use of the LEP screening tool.  The use of the tool strengthened the validity and reliability of correctly identifying those patients with LEP.  Of note, despite reaching the sample size target for this study, only 56 (8.0%) of the 691 patients were classified as having LEP.  The 47 (6.8%) NEP patients were combined with the LEP group; the combined sample size was 103 (14.9%).  In combining the two groups, there may have been essential differences in the characteristics of the LEP and NEP patients that were missed or obscured.  Combining these two groups could have potentially influenced the conclusions made in this study.    Methodological Limitations  A retrospective chart review was used to examine the relationship between LOS and English proficiency after isolated CABG surgery.  This study design has a number of limitations.  It is difficult to ascertain cause-and-effect relationships in retrospective studies.  Correlations between variables must be cautiously analyzed and interpreted.  This chapter has provided some discussion of the potential for multifaceted contextual factors influencing the relationship between English proficiency and LOS.  Thus, unanticipated confounding factors not accounted for in the study may have affected the results, leading to inaccurate conclusions.  For example, this study did not control for urgency of surgery, which is a marker for pre-procedure education and readiness, and complexity (Cardiac Services BC, 2008).  With the knowledge that LEP patients have poorer access to health care than do EP patients (Bernstein et al., 2002; Carrasquillo et al., 1999), data identifying whether language proficiency was associated with being an emergency rather than elective case would have been valuable information.  It is highly 93  plausible that the emergency surgery patients would not have received the same pre-operative education nor the opportunity to digest the information for their procedure as would have had elective surgery patients.  This difference may have been a confounding variable that influenced patients’ ‘readiness’ for surgery and, thus, their recovery process and LOS.   Another important confounding variable this study did not measure was the patients’ health literacy.11  In the US, the literature reports that individuals whose primary language is not English have poorer health literacy than first language or fluent English speaking individuals (Berkman et al., 2011; Nielsen-Bohlman, Panzer, & Kindig, 2004).  It is also known that limited health literacy contributes to disparities in access to health care and is associated with poorer health outcomes, such as limited knowledge of chronic disease management, higher rates of hospitalization, and poor self-rated health (Berkman et al., 2011).   It is unknown whether health literacy and English proficiency interact in their influence on LOS.  As mentioned earlier, other social determinants of health, such as race/ethnicity and socioeconomic status were not measured.  Although this study attempted to capture socioeconomic factors such as marital and employment status, other data such as income and education may have been associated with EP and LOS (Levas et al., 2011).  The potential synergistic effects of these variables are unknown.  All of the data in this study were gathered from patients’ medical records.  The charts were not designated for the purposes of conducting research.  Accordingly, the investigator was dependent upon the medical record information being available and accurate.                                                    11Health literacy is defined as “the degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions" (Nielsen-Bohlman et al., 2004, p. 32).  Health literacy is dependent upon understanding health information that requires basic skills, such as reading, writing, and numeracy. 94  The retrospective study design posed particular obstacles in the ability to accurately determine the English proficiency of the patients.  Although the researchers developed and utilized a validated screening tool, this tool was subject to measurement error.  For example, during the data collection phase, conflicting or missing documentation made it difficult to ascertain the English proficiency status of some patients.  Inaccurate determination of English proficiency status may have resulted in erroneous data and conclusions.  Ideally, a prospective study would allow the researcher to collect all data relevant to the determination of a patient’s English proficiency.      This study was set in one major tertiary hospital.  Thus, the sample may not be representative of other communities.  Unique to Vancouver, Punjabi, Cantonese, Chinese (not specified), and Mandarin accounts for more than one half (57.7%) of the languages spoken in the city (Statistics Canada, 2013).  In this sample, the largest language groups were Punjabi-, Cantonese-, and Mandarin-speaking individuals.  It is conceivable that there are specific ethnic and cultural aspects of these communities that influenced their LOS, rather than the explanatory variable of interest.  These factors may reduce the internal validity of the study.  As previously mentioned, this study recruited all consecutively eligible patients, thus minimizing the possibility of selection bias.  But this sampling design also may have resulted in using the most conveniently available subjects as participants as opposed to a randomized selection of patients.  The lack of randomized selection may have resulted in a phenomenon where the charts selected were atypical of the real population (Polit & Beck, 2012).   Finally, in hindsight, there are two issues of the data related to reintubation and prolonged intubation.  First, the definition of prolonged intubation in this study was not consistent with clinical practice.  The current CSICU targeted extubation goal is within six hours of admission.  95  Thus the study’s definition of prolonged intubation (intubated greater than 24 hours from time of surgery) may have lacked sensitivity.  Although the occurrence of this event was associated with both LEP/NEP status and prolonged LOS, the broad definition of prolonged intubation may have inadvertently missed a number of patients clinically considered as having prolonged intubation.  Second, collecting data regarding the direct cause of reintubation or prolonged intubation could have provided additional information for differences in such events between the two groups.  Of the data collected, this study did not find differences in the rate of return to the operating room when stratified by language proficiency.  A statistically significant difference in the rate of return to the operating room between the two groups could have explained the disparities found in the reintubation rates because these patients would have been reintubated for the procedure and would have required longer post-operative recovery.   Clinical Implications Investigating the effects of English proficiency on LOS in hospital is pertinent to today’s ‘global’ clinical practice, yet there is a dearth of literature on this topic.  At the time of writing, only two research articles were found that explored the effects of English proficiency of inpatients while recovering from cardiac procedures.  The findings of this study have important implications for nursing practice, education, and policy.    Nursing practice.  Within the busy environment of the hospital, nurses faced with linguistic barriers of communicating with their patients are well versed in ‘getting by’ (Villaruel, 1999).  The literature has provided evidence that LEP patients have barriers to accessing health care and are 96  less satisfied with the quality of their care than are EP patients (Carrasquillo et al., 1999; Karliner et al., 2012; Morales et al., 1999; Ngo-Metzger, Phillips, & Sorkin, 2009).  The study results highlight the importance of investigating the impact of language barriers on the ‘patient-provider relationship’ and the subsequent effects on patient outcomes as described by Northouse and Northouses’s (1998) and Martinez’s (2010) theoretical frameworks.   Patients with a LOS of more than five days were older and more likely to be unmarried.12  These findings should encourage nurses to anticipate and address the potential barriers to recovery for this vulnerable population.  First, older patients who may have limited social networks of support should be recognized prior to surgery.  The literature emphasizes the importance of a thorough pre-operative assessment and ensuring that the patient is in the 'optimal’ condition for surgery (Dhesi, 2010).  But nurses in the anesthetist-led pre-operative clinics may have little experience in care for the elderly.  Nurses assessing patients in the pre-operative interviews should be formally trained in recognition of the high-risk, older surgical population.   Second, other involved health professionals should be trained in recognizing and optimizing relevant issues, from formal risk assessment to appropriate peri-operative planning of high-risk, older surgical patients (Dhesi, 2010).  This model of care should allow the full scope of pre-operative assessment to support high-risk, older patients to achieve successful post-operative recovery.   The findings that post-operative infections are the strongest predictor of LOS should continue to encourage nurses to practice especially clean and sterile techniques when caring for these patients.  In particular, the findings of high rates of oral candidiasis, sternal wound infections, and pneumonia raise concerns and questions about why these infections are occurring                                                  12 These patients were more likely to have a length of stay of nine or more days. 97  in CABG patients, at this site.  Globally, hospital-related infections have been a growing concern and ongoing investigations and interventions to address oral candidiasis along with surgical site infections, urinary tract infections, and pneumonia are currently being implemented at the study site.  Again, the nursing staff can help combat such infections through a multitude of interventions, including adhering to regularly scheduled oral hygiene, particularly in intubated and critically ill patients, practicing sound aseptic techniques with wound care, removing urinary catheters soon after insertion (when appropriate), and early mobilization of patients.  With any proposed intervention, nurses should also strive to enact evidenced-based practices and demand that mandated interventions be firmly supported by evidence.    LEP/NEP patients were more prone to infections in this study.  Although further evidence is needed to support changes in practice, nursing staff should be cognizant of the potential consequences of inadequate communication with their LEP/NEP patients.  Nurses should employ strategies to ensure that these patients fully understand their post-operative teaching related to reducing post-operative infection, such as regular pulmonary hygiene and early ambulation.  Karliner et al.’s (2012) study regarding discharge teaching demonstrated that LEP/NEP patients had poorer understanding of their medications compared with EP patients. They concluded that having accessible processes, such as improved discharge materials, as well as asking patients to reiterate what was taught to them, could assist in comprehension.  Nurses play a crucial role in changing ‘accepted’ practices related to communicating with LEP/NEP patients.  This study highlights how the effects of ‘getting by’ may have significant consequences for LEP/NEP patients’ outcomes.  Nurses should be documenting, communicating, and advocating for these patients to have family translators or preferably, professional interpreters if they observe that language barriers are impeding their patients’ 98  quality of care.  Of all healthcare professionals, frontline nurses spend the most time with patients and are poised to recognize the need for, and expedite the availability of, proper translators.   The findings in this study also demonstrate that nursing practice should utilize clinical pathways with the understanding that they are ‘guides’ to recovery and streamlining of care, but are not a measure of nursing ‘success.’  These findings call into question the value of the current clinical pathway LOS target, especially considering the aging cardiac surgery clientele.  When patients do not meet expected pathway goals, there are a multitude of reasons to consider.  This study validated that non-modifiable risk factors such as age and marital status can inhibit the ability to meet pathway targets.  The study results also showed that the majority of the patients did not meet the target pathway of five days’ stay.  These findings render the pathway’s LOS goal of being less clinically important when caring for patients with these non-modifiable risk factors.  This study should support nurses in challenging the utility of current CABG clinical pathways. Considerations should be given to the development of risk-stratified pathways to further support better practices.    Policy. The analyses served to identify a specific group of vulnerable individuals that have the greatest risk of prolonged LOS after CABG surgery.  Ensuring policies support patients that are older and have less social support is important to preventing prolonged LOS. In light of the study’s findings, the current efforts at the study site to re-examine policies of infection control are warranted.  The infection rates were alarmingly high with the majority of infections being oral candidiasis, sternal wound infections, and pneumonia.  Evaluation of the 99  source of such infections and implementing evidenced-based interventions should be supported.  Current policy changes are underway because the provincial government has mandated a practice and safety council to assess and address the infection rates of all hospitals (British Columbia Patient Safety Quality Control, 2014).  The current study site is collecting multi-sourced data to track and follow its post-operative infection rates.  They have implemented several interventions to prevent and reduce infection.  Hence, it is timely that the study results have identified another potential determinant of infection.  Strategies to combat post-operative infection should be multidisciplinary and address all components of the patient’s trajectory.   Although this study has not provided unequivocal evidence that a new policy for the care of LEP/NEP patients should be implemented, a policy should be focused on identifying LEP/NEP patients better within the healthcare system.  Evidence-based policy begins with accurate documentation of LEP status in medical charts and reporting systems.  A prerequisite for healthcare organizations to address these language barriers is to identify those patients most likely to benefit from language assistance (Karliner et al., 2012).  At the study site, identification and documentation of patient language preferences on the registration and admission assessment forms were at times missing or conflicting with other documentation in the chart.  Thus, health organizations and clinical leaders should ensure that all patients receive standardized assessments of their English proficiency.  Precise and complete documentation of language on admission could immediately mitigate the potential barriers to communication.   In addition, accurate documentation of patients’ language status would allow for data collection related to the incidence and impact of preventable adverse events in this special population.  Methods to link such data have proven problematic because the inability to clearly differentiate data pertaining to incidents (in consideration of language ability) creates difficulty 100  in identifying LEP-related adverse events (Divi, 2007; McKee & Paasche-Orlow, 2012).  Thus, the policies related to the documentation of patients’ English proficiency should be revisited with an eye to developing reliable methods of quality data collection. Because the literature strongly supports the provision of professional interpreters to address healthcare quality, efficiency, and costs related to language barriers, current policies should continue to support streamlined language access services for LEP/NEP patients (Hyman, 2009).  Systematic processes should be instituted or be in place so that it would prove difficult for attending healthcare professionals to not employ the services of professional interpreters for LEP/NEP patients.  Such processes include uncomplicated referral and booking systems for language services, having readily available onsite health interpreters and bedside two-way telephone sets for telephone interpreting, employment of bilingual health professionals (although professional interpreters are preferred), and training of staff regarding such services (Lindholm et al., 2012; Schenker et al., 2011).  Documentation of language services provided during a patient’s stay is also important in assisting hospitals to examine their processes and then initiate efforts to address any disparities in outcomes for LEP patients (Divi, 2007).   Recommendations for Future Research  Further research is required to further elucidate the study’s findings.  First, the high overall infection rates in this study suggest that further research is warranted to examine the mechanisms through which post-operative infections are acquired in the study site’s CABG patient population, and especially by LEP/NEP patients.  Current actions at the study site are being implemented to address and curtail nosocomial infections (i.e., new surgical wound care protocols).   101  Second, post-operative infection was found to be a mediator between English proficiency and LOS, but key questions regarding the exact relationship between English proficiency and post-operative infection remain.  Questions remain about why the LEP/NEP patients in this study experienced more adverse events, including post-operative infection, delays in transfer to long-term care, the need for reintubation, and prolonged intubation.  Future research should be directed at establishing a causal link between limited English proficiency and post-operative infection.  Research should also focus on describing the mechanisms by which the effects of English proficiency on LOS and other outcomes are mediated.    This study attempted to isolate the connection between English proficiency and LOS, but in doing so, the complexity of the challenges that LEP patients face created more questions when interpreting the results.  Clarity is needed to answer questions such as, “Why did the LEP/NEP patients experience more intubations and prolonged intubation than the EP patients?  Could the lack of communication related to surgery expectations and home preparations pre-operatively have affected post-operative recovery?”  Such questions demand further in depth investigation and understanding. A robust research plan aimed at establishing the casual link between limited English proficiency and poor patient outcomes should be implemented to provide evidence for key policy makers regarding the implications of limited English proficiency on patient outcomes.  Only ample and targeted research will support informed decisions and evidence-based practice.   Third, development of a standardized LEP screening tool of patient medical records is essential for enhancing clinical screening, research and interpretation.  Attention was paid to constructing a reliable and valid measurement tool to assess English proficiency status through patient medical records.  Further testing and refining of this tool would contribute to utilizing a 102  ‘gold standard’ in assessing the English proficiency of patients through chart auditing.  Karliner et al. (2012) found that of the 21 articles they reviewed regarding professional interpreters’ benefits for LEP patients; only one third of the studies used a standardized measure to identify the participants’ need for an interpreter.  Of those studies that used a standardized measure, several ascertained the patients’ primary language and did not establish English language proficiency.  A standardized LEP assessment tool would allow for better comparison of studies and better conclusions (Jacobs, 2006).  Investigators also need to be sharper in their definitions of ‘interpreter’ and who they define as having LEP.  In reviewing the studies available, the lack of clarity of such definitions made comparing studies difficult and the conclusions ambiguous (Karliner et al., 2012).   Subsequent studies that explore the association between limited language proficiency and LOS should include a prospective study design.  A prospective study would reduce the chance of random and systematic errors.  A prospective study would also allow for the examination of contextual effects such as hospital characteristics, socioeconomic status, race/ethnicity, health literacy, and acculturation, which may have interacting and compounding effects on language barriers.  In addition, investigating readmission rates of CABG patients stratified by English proficiency would add value to understanding the potential impact of limited English proficiency on longer-term patient outcomes. Lastly, the study findings show the importance of validating clinical pathways, their purpose, and their target population.  Constantly evolving technologies for CABG surgery and care combined with increasing comorbidities of patients influence the validity and applicability of such pathways.  Health clinicians and hospital executives employing such pathways as a standard to measure hospital success and efficiencies would be quick to deduce that a hospital is 103  not meeting its targets.  The current CABG clinical pathway does not encompass the multiple layers of patient and hospital characteristics that influence LOS.  Research on creating and validating an updated risk stratified CABG clinical pathway would strengthen its use and applicability for practice.      Conclusion  This retrospective chart review of isolated CABG surgery patients endeavoured to demonstrate the effects of limited English proficiency on post-operative LOS.  LEP/NEP patients had a longer length of stay when compared with EP patients.  This study further revealed that LEP/NEP patients were more prone to post-operative infections than were their EP counterparts and that post-operative infections significantly contributed to prolonged LOS.  But the complexity of the relationship between English proficiency and LOS must not be underestimated.  Consideration of other unanticipated and unexplored confounding factors warrants further research to provide clarity to the situation for patients, healthcare providers, and policy makers.     104  References  Al-Sarraf, N., Thalib, L., Hughes, A., Tolan, M., Young, V., & McGovern, E. (2008). Effect of smoking on short-term outcome of patients undergoing coronary artery bypass surgery.  Annals of Thoracic Surgery, 86(2), 517-523. doi: 10.1016/j.athoracsur.2008.03.070 Anand, S. S., Yusuf, S., Vuksan, V., Devanesen, S., Teo, K. K., Montague, P. A., Kelemen, L., Yi, C., Lonn, E., Gerstein, H., Hegele, R., & McQueen, M. (2000). Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: The study of health assessment and risk in ethnic groups (SHARE). Lancet,356 (9226), 279-284. doi: 10.1016/S0140-6736(00)02502-2 Baker, G. R., Norton, P. G., Flintoft, V., Blais, R., Brown, A., Cox, J., Etchells, E., Ghali, W.A., Hebert, P., Majumdar, S. R., O'Beirne, S., Palacios-Derfliinger, L., Reid, R. J., Sheps, S., & Tamblyn, R. (2004). The Canadian Adverse Events Study: The incidence of adverse events among hospital patients in Canada. CMAJ, 170, 1678-1686. doi: 10.1503/cmaj.1040498 Barbash, I. M., Gaglia, M. A., Torguson, R., Minha, S., Satler, L. F., Pichard, A. D., & Waksman, R. (2013). Effect of marital status on the outcome of patients undergoing elective or urgent coronary revascularization. American Heart Journal,166(4), 729-736. doi: 10.1016/j.ahj.2013.07.018; 10.1016/j.ahj.2013.07.018 Bartlett, G., Blais, R., Tamblyn, R., Clermont, R. J., & MacGibbon, B. (2008). Impact of patient communication problems on the risk of preventable adverse events in acute care settings. CMAJ, 178, 1555-1562. doi: 10.1503/cmaj.070690 105  Berkman, N., DeWalt, D., Pignone, M., Sheridan, S., Lohr, K., Lux, L., Sutton, S. F., Swinson, T., & Bonito, A. (2004). Literacy and health outcomes. Rockville, MD: Agency for Health Care Research and Quality.  Bernstein, J., Bernstein, E., Dave, A., Hardt, E., James, T., Linden, J., Mitchell, P., Oishi, T., & Safi, C. (2002). Trained medical interpreters in the emergency department: Effects on services, subsequent charges, and follow-up. Journal of Immigrant Health, 4, 171-176. doi: 10.1023/A:1020125425820 Bowen, S. (2001).  Health Canada: Language barriers in access to health care (Report No.  H39-578/ 2001E). Ottawa: Health Canada.  BC Patient Safety and Quality Council. (2013, October 17). Improving cardiac surgery care. [PowerPoint slides]. BC Patient Safety and Quality Council. (2014) Retrieved from http://bcpsqc.ca/about-the-council/ Campeau L. (1976). Grading of angina pectoris [letter]. Circulation, 54, 522–523. Carrasquillo, O., Orav, E. J., Brennan, T. A., & Burstin, H. R. (1999). Impact of language barriers on patient satisfaction in an emergency department. Journal of General Internal Medicine, 14(2), 82-87.  Cardiac Services BC, Provincial Health Services Authority.  (2008). Cardiac Services BC Cardiac Surgery Waitlist Management Reference Manual.  Vancouver: Provincial Health Services Authority, BC.  Cardiac Services BC, Provincial Health Services Authority.  (2010). Annual report 2010. Retrieved from http://www.phsa.ca/NR/rdonlyres/313D37DA-7368-48AC-ADED-6C09F5313E95/53773/CSBCAR2011.pdf 106  Cardiac Services BC, Provincial Health Services Authority.  (2011). Annual report 2011.  Vancouver: Provincial Health Services Authority, BC.  Cardiac Services BC, Provincial Health Services Authority. (2013) OHS and pure CABG volume summary for 2012.  Unpublished raw data. Cheng, E. M., Chen, A., & Cunningham, W. (2007). Primary language and receipt of recommended health care among Hispanics in the United States. Journal of General Internal Medicine, 22(Suppl 2), 283-288. doi:10.1007/s11606-007-0346-6 Cheng, E., Chen, A. H., Grumbach, K., Wang, F., & Fernandez, A. (2005). Effects of limited English proficiency and physician language on health care comprehension. Journal of General Internal Medicine, 20, 800-806. doi:10.1111/j.1525-1497.2005.0174.x Chiu, M., Austin, P. C., Manuel, D. G., & Tu, J. V. (2010). Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. CMAJ, 182, E301-E310. doi:10.1503/cmaj.091676 Cioffi, J. (2003). Communicating with culturally and linguistically diverse patients in an acute care setting: Nurses’ experiences. International Journal of Nursing Studies, 40, 299-306. doi:10.1016/S0020-7489(02)00089-5 Cohn, L. H., Rosborough, D., & Fernandez, J. (1997). Reducing costs and length of stay and improving efficiency and quality of care in cardiac surgery. Annals of Thoracic Surgery, 64(6, Suppl 1), S58-S60. doi:10.1016/S0003-4975(97)01158-2 Collins T.C., Daley J., Henderson W.H., & Khuri S.F.  (1999). Risk factors for prolonged length of stay after major elective surgery. Annals of Surgery , 230, 251–259 107  Coevoet, V., Fresson, J., Vieux, R.,  & Jay, Nicolas. Socioeconomic deprivation and hospital length of stay: a new approach using area-based indicators in multilevel models. Medical Care, 51(6), 548-554. Cowper, P. A., DeLong, E. R., Hannan, E. L., Muhlbaier, L. H., Lytle, B. L., Jones, R. H., Holman, W.L., Pokorny, J.J., Stafford, J.A., Mark, D.B., & Peterson, E. D. (2006). Trends in postoperative length of stay after bypass surgery. American Heart Journal, 152(6), 1194-1200. doi: 10.1016/j.ahj.2006.07.017 Cristofolini, M., Worlitzch, D., Wienke, A., Siber  R.E., & Borneff-Lipp  M.(2012). Surgical site infections after coronary artery bypass graft surgery: Incideince, perioperative hospital stay, readmissions, and revision surgeries.  Infection, 40 (4), 397-404. Cruz-Flores, S., Rabinstein, A., Biller, J., Elkind, M. S., Griffith, P., Gorelick, P. B., Howard, G., Leira, E.C., Morgenstern, L.B., Ovbiagele, B., Peterson, E., Rosamond, W., Trimble, B., & Valderrama, A. L. (2011). Racial-ethnic disparities in stroke care: The American experience: A statement for healthcare professionals from the American heart Association/American stroke association. Stroke: A Journal of Cerebral Circulation, 42(7), 2091-2116. doi:10.1161/STR.0b013e3182213e24 DECODE Study Group on behalf of the European Diabetes Epidemiology Study Group (1998). Will new diagnostic criteria for diabetes mellitus change phenotype of patients with diabetes? Reanalysis of European epidemiological data. British Medical Journal,  317, 371–375. Dhesi, J. (2010). Improving outcomes in older patients undergoing elective surgery. The Journal of Royal College of Physicians of Edinburgh. 40(4), 348-353  108  Divi, C., Koss, R. G., Schmaltz, S. P., & Loeb, J. M. (2007). Language proficiency and adverse events in US hospitals: A pilot study. International Journal for Quality in Health Care, 19(2), 60-67. doi:10.1093/intqhc/mzl069 Eisenberg, M. J., Filion, K. B., Azoulay, A., Brox, A. C., Haider, S., & Pilote, L. (2005). Outcomes and cost of coronary artery bypass graft surgery in the United States and Canada. Archives of Internal Medicine, 165(13), 1506-1513. doi:10.1001/archinte.165.13.1506 Elderkin-Thompson, V., Cohen Silver, R., & Waitzkin, H. (2001). When nurses double as interpreters: A study of Spanish-speaking patients in a US primary care setting. Social Science & Medicine, 52, 1343-1358. doi:10.1016/S0277-9536(00)00234-3 Epps, C. D. (2004). Length of stay, discharge disposition, and hospital charge predictors. AORN Journal, 79, 975-997. doi:10.1016/S0001-2092(06)60729-1 Fleiss, J. L. (1971). Measuring nominal scale agreement among many raters. Psychological Bulletin,76(5), 378–382  Flores, G. (2006). Language barriers to health care in the United States. New England Journal of Medicine, 355, 229-231. doi:10.1056/NEJMp058316 Freedman, J., Guller, U., Benjamin, D. K., Higgins, L. D., Pan, D., Cook, C., & Pietrobon, R. (2006). National trends in health care utilization and racial and socioeconomic disparities in pediatric pyogenic arthritis. Journal of Pediatric Orthopedics, 26, 709-715. doi:10.1097/01.bpo.0000229973.78565.02 Filon, K., Pilote, L., Rahme,. E., & Eisenberg, M. (2008).  Use of perioperative cardiac medical therapy among patients undergoing coronary artery bypass graft sugery, Journal Of Cardiac Surgery, 23(3), 209-215. doi:10.1111/j.1540-8191.2008.00596.x 109  Fowler, V. G., O'Brien, S. M., Muhlbaier, L. H., Corey, G. R., Ferguson, T. B., & Peterson, E. D. (2005). Clinical predictors of major infections after cardiac surgery. Circulation, 112(9 Suppl), I358-65. doi:10.1161/CIRCULATIONAHA.104.525790 Ghali, W. A., Hall, R. E., Rosen, A. K., Ash, A. S., & Moskowitz, M. A. (1996). Searching for an improved clinical comorbidity index for use with ICD-9-CM administrative data. Journal of Clinical Epidemiology, 49(3), 273-278. doi:10.1016/0895-4356(95)00564-1 Gee, G. C., Walsemann, K. M., & Takeuchi, D. T. (2010). English proficiency and language preference: Testing the equivalence of two measures. American Journal of Public Health, 100, 563-569. doi:10.2105/AJPH.2008.156976 Gwet, K. L. (2010, June 28).  Sample size determination. [Web log post].  Retrieved from http://agreestat.com/blog_irr/sample_size_determination.html Gordon, H.S., & Rosenthal, G.E. (1995). Impact of marital status on outcomes in hospitalized patients. Archives of  Internal Medicine, 155(22):2465-2471. doi:10.1001/archinte.1995.00430220129014 Hampers, L. C., & McNulty, J. E. (2002). Professional interpreters and bilingual physicians in a pediatric emergency department: Effect on resource utilization. Archives of Pediatrics & Adolescent Medicine, 156, 1108-1113.  Hudelson, P., & Vilpert, S. (2009). Overcoming language barriers with foreign-language speaking patients: A survey to investigate intra-hospital variation in attitudes and practices. BMC Health Services Research, 9, 187. doi:10.1186/1472-6963-9-187 Hyman, I. (2009). Literature review: The costs of not providing interpretation in health care. Access Alliance, 1-16. 110  Jacobs, E. A., Shepard, D. S., Suaya, J. A., & Stone, E. L. (2004). Overcoming language barriers in health care: Costs and benefits of interpreter services. American Journal of Public Health, 94, 866-869.  Johnston, G., Goss, J. R., Malmgren, J. A., & Spertus, J. A. (2004). Health status and social risk correlates of extended length of stay following coronary artery bypass surgery. Annals of Thoracic Surgery, 77, 557-562. doi:10.1016/S0003-4975(03)01586-8 Johnstone, M. J., & Kanitsaki, O. (2006). Culture, language, and patient safety: Making the link. International Journal for Quality in Health Care, 18, 383-388. doi:10.1093/intqhc/mzl039 Kandula, N. R., Lauderdale, D. S., & Baker, D. W. (2007). Differences in self-reported health among Asians, Latinos, and non-Hispanic whites: The role of language and nativity. Annals of Epidemiology, 17, 191-198. doi:10.1016/j.annepidem.2006.10.005 Karliner, L. S., Auerbach, A., Napoles, A., Schillinger, D., Nickleach, D., & Perez-Stable, E. J. (2012). Language barriers and understanding of hospital discharge instructions. Medical Care, 50, 283-289. doi:10.1097/MLR.0b013e318249c949 Karliner, L. S., Jacobs, E. A., Chen, A. H., & Mutha, S. (2007). Do professional interpreters improve clinical care for patients with limited English proficiency? A systematic review of the literature. Health Services Research, 42, 727-754. doi:10.1111/j.1475-6773.2006.00629.x Karliner, L. S., Napoles-Springer, A. M., Schillinger, D., Bibbins-Domingo, K., & Perez-Stable, E. J. (2008). Identification of limited English proficient patients in clinical care. Journal of General Internal Medicine, 23, 1555-1560. doi:10.1007/s11606-008-0693-y 111  Kelsey, J. L., Wittmore, A. S., Evans, A. S., & Thompson, W. D. (1996).  Methods in observational epidemiology (2nd ed.). Oxford: Oxford University Press. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. Lapsy, H.M., & Vogeuls, R. (1998).  Quality and cost impacts: prevention of post-operative clean wound infections. International Journal of Health Care Quality Assurance  11(7), 222–231 Levas, M. N., Cowden, J. D., & Dowd, M. D. (2011). Effects of the limited English proficiency of parents on hospital length of stay and home health care referral for their home health care-eligible children with infections. Archives of Pediatrics & Adolescent Medicine, 165, 831-836. doi:10.1001/archpediatrics.2011.61 Lindholm, M., Hargraves, J. L., Ferguson, W. J., & Reed, G. (2012). Professional language interpretation and inpatient length of stay and readmission rates. Journal of General Internal Medicine, 27, 1294-1299. doi:10.1007/s11606-012-2041-5 Martinez, G.A. (2010). Language and power in healthcare: Towards a theory of language barriers among linguistic minorities in the United States. In Watzke, J., Chamness, P. & Mantero, M. (Eds.). Readings in Language Studies Volume 2: Language and Power (p. 59-74). St Louis, MO: International Society for Language Studies.  Morales, L. S., Cunningham, W. E., Brown, J. A., Liu, H., & Hays, R. D. (1999). Are Latinos less satisfied with communication by health care providers? Journal of General Internal Medicine, 14, 409-417.  McKee, M., & Paasche-Orlow, M. (2012). Health literacy and the disenfranchised:The importance of collaboration between limited English proficiency and health literacy 112  researchers. Journal of Health Communication, 17(supp l3), 7-12. doi:10.1080/10810730.2012.712627 Nielsen-Bohlman, L., Panzer, A., & Kindig, D. (Eds.). (2004). Health literacy: A prescription to end confusion. Washington, DC: National Academies Press. Northouse, L.L., & Northouse, P.G. (Eds.). (1998). Health communication: strategies for health professionals.  Stamford, CO:  Appleton & Lange Ng, E. (2011). The healthy immigrant effect and mortality rates. Health Reports, 22(4), 25-9.  Ng E., Wilkins R., Gendron F., & Berthelot, J.M. (2005). Dynamics of immigrants' health in Canada: Evidence from the National Population Health Survey. Healthy Today, Healthy Tomorrow? (Catalogue no. 82-618) Ottawa: Statistics Canada. Ng, E., Pottie, K., & Spitzer, D. (2011). Official language proficiency and self-reported health among immigrants to Canada. Health Reports, 22(4), 15-23.  Ngo-Metzger, Q., Phillips, R. S., & Sorkin, D. H. (2009). Healthcare experiences of limited English-proficient Asian American patients: A cross-sectional mail survey. Patient, 2, 113-120. doi:10.2165/01312067-200902020-00007  Oliva, N. L. (2008). When language intervenes: Improving care for patients with limited English proficiency. American Journal of Nursing, 108(3), 73-75. doi:10.1097/01.NAJ.0000312266.17485.b6 Pfaller, M.A. (1996). Nosocomial candidiasis: emerging species, reservoirs, and modes of transmision. Clinical Infections Diseases. 22(suppl 2): S89-94. Pearson, S. D., Kleefield, S. F., Soukop, J. R., Cook, E. F., & Lee, T. H. (2001). Critical pathways intervention to reduce length of hospital stay. American Journal of Medicine, 110, 175-180. doi:10.1016/S0002-9343(00)00705-1 113  Peterson, E. D., Coombs, L. P., Ferguson, T. B., Shroyer, A. L., DeLong, E. R., Grover, F. L., & Edwards, F. H. (2002). Hospital variability in length of stay after coronary artery bypass surgery: Results from the society of thoracic Surgeon’s national cardiac database. The Annals of Thoracic Surgery,74(2), 464-473.  Plichta Keller, S. & Kelvin, E.A. (2013). Munro’s Statistical Methods for Health Care Research (6th ed.). Philadelphia: Lippincott Williams & Wilkins.  Preedy, V. R., & Watson. R. R. (2010). Coronary artery bypass graft.  In V. R. Preedy & R. R. Watson (Eds.), Handbook of disease burdens and quality of life measures (pp. 4179). New York: Springer.  doi:10.1007/978-0-387-78665-0_5394 Pruthi, R.S., Lentz, A.C., Sand, M., Kouba, E., & Wallen, E.M. (2009).Impact of marital status in patients undergoing radical cystectomy for bladder cancer.  World Journal of Urology, (4), 573-576. Rosen, A. B., Humphries, J. O., Muhlbaier, L. H., Kiefe, C. I., Kresowik, T., & Peterson, E. D. (1999). Effect of clinical factors on length of stay after coronary artery bypass surgery: Results of the cooperative cardiovascular project. American Heart Journal, 138(1), 69-77. doi:10.1016/S0002-8703(99)70249-8 Rivadeneyra, R., Elderkin-Thompson, V., Silver, R. C., & Waitzkin, H. (2000). Patient centeredness in medical encounters requiring an interpreter. American Journal of Medicine, 108, 470-474. doi:10.1016/S0002-9343(99)00445-3 Rudisill, P. T., Phillips, M., & Payne, C. M. (1994). Clinical paths for cardiac surgery patients: A multidisciplinary approach to quality improvement outcomes. Journal of Nursing Care Quality, 8(3), 27-33.  114  Rumble, S. J., Jernigan, M. H., & Rudisill, P. T. (1996). Determining the effectiveness of critical pathways for coronary artery bypass graft patients: Retrospective comparison of readmission rates. Journal of Nursing Care Quality, 11(2), 34-40.  Saxena, A., Shan, L., Reid, C., Dinh, D. T., Smith, J. A., Shardey, G. C., & Newcomb, A. E. (2013). Impact of smoking status on early and late outcomes after isolated coronary artery bypass graft surgery. Journal of Cardiology, 61(5), 336-341. doi:10.1016/j.jjcc.2013.01.002 Sawatzky, J. A., & Naimark, B. J. (2009). The coronary artery bypass graft surgery trajectory: Gender differences revisited. European Journal of Cardiovascular Nursing: Journal of the Working Group on Cardiovascular Nursing of the European Society of Cardiology, 8(4), 302-308. doi:10.1016/j.ejcnurse.2009.02.003; 10.1016/j.ejcnurse.2009.02.003 Sorensen, L.T. (2012). Wound healing and infection in surgery. The clinical impact of smoking and smoking cessation: A systematic review and meta-analysis.  Archives of Surgery, 147(4), 373-383. doi: 10.1001/archsurg.2012.5;10.1001/archsurg.2012.5 Statistics Canada. (2011). Immigration and ethnocultural diversity in Canada: National household survey, 2011 (Catalogue No. 99-010-X2011001).  Retrieved from http://www12.statcan.gc.ca/nhs-enm/2011/as-sa/99-010-x/99-010-x2011001-eng.cfm Statistics Canada. (2013). Population by language spoken most often and regularly at home, age groups (total), for Canada, provinces and territories.  Retrieved from http://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/lang/Pages/Highlight.cfm?TabID=1&Lang=E&PRCode=01&Age=1&tableID=403&queryID=1    115  Schenker, Y., Perez-Stable, E. J., Nickleach, D., & Karliner, L. S. (2011). Patterns of interpreter use for hospitalized patients with limited English proficiency. Journal of General Internal Medicine, 26, 712-717. doi:10.1007/s11606-010-1619-z Shi, L., Lebrun, L. A., & Tsai, J. (2009). The influence of English proficiency on access to care. Ethnicity & Health, 14(6), 625-642. doi:10.1080/13557850903248639 The Criteria Committee of the New York Heart Association. (1994).Nomenclature and Criteria for Diagnosis of Diseases of the Heart and Great Vessels.(9th ed.). Boston: Little, Brown & Co. pp.253–256. Vancouver General Hospital Decision Support. (2013). Elective coronary bypass surgeries at Vancouver General Hospital, January 1, 2011-December 31, 2012 [Discharge Abstract Database]. Unpublished raw data. Vancouver Hospital & Health Sciences Centre. (2005). Clinical path – CABG +/- valve surgery: Interdisciplinary discharge planning tool (Report #32).  Vancouver: Author. Weingarten, S. (2001). Critical pathways: What do you do when they do not seem to work? American Journal of Medicine, 110, 224-225. doi:10.1016/S0002-9343(00)00739-7 World Health Organization (2003). Obesity and overweight. Geneva: World Health Organization. Retrieved from: http://www.who.int/dietphysicalactivity/media/en/gsfs_obesity.pdf Woods, S. L., Sivarajan Froehlicher, E. S., & Motzer (Underhill), S. (Eds.). (2010). Cardiac nursing.  Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins.      Appendix A: Clinical path – CABG +/- valve surgery: Interdisciplinary dischargeplanning tool 116117118119120121122123Source: Vancouver Hospital & Health Sciences Centre, 2005124Appendix B: Clinical path  - CABG +/- valve: Interdisciplinary discharge              planning tool for 6+ days Source: Vancouver Hospital & Health Sciences Centre, 2005125  Appendix C:  Limited English Proficiency Screening Tool  Subject Number:                                           Today’s Date (ddmmmyy):                              Assessor’s Name:                                                                         1. Assessment of English proficiency from health record review *(Definitions on page 3)  Documentation   English proficient Limited English proficiency NOT English proficient  UNKNOWN   Patient Registration Form  (Face Sheet)   ☐  Yes Language preferred:   _____________ Language spoken:     _____________ ☐  Not available ☐  Not completed ☐ Ambiguous Nursing Admission or Assessment Form   ☐  Yes Language preferred:   ____________ Language spoken:    _____________  Translator required: ☐  Professional ☐  Family member ☐  Other staff member  ☐  Not available ☐  Not completed ☐ Ambiguous Pre-Operative Checklist ☐  Yes Language preferred:   ____________ Language spoken:    _____________  Translator required: ☐  Professional ☐  Family member ☐  Other staff member ☐ Not available ☐ Not completed ☐ Ambiguous Nurses’ Notes    ☐  Yes Language preferred:   ____________ Language spoken:    _____________  Translator required: ☐  Professional ☐  Family member ☐  Other staff member ☐ Not available ☐ Not completed ☐ Ambiguous Consult Notes ☐  Yes Language preferred:   ____________ Language spoken:    _____________  Translator required: ☐  Professional ☐  Family member ☐  Other staff member ☐ Not available ☐ Not completed ☐ Ambiguous 126   2. Based on your assessment of the chart, what is this patient’s proficiency with English?   ☐English proficient    ☐Limited English proficiency    ☐Not English proficient     Please explain your answer ____________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________  Verification of English Language Proficiency  3. Questions to ask nursing staff about this patient:  i Are you able to converse with … [patient’s name] in English?   ☐ Yes   ☐ No           ☐ Not sure, there is some difficulty    ii Would this patient understand the English version of the discharge video/booklet? (patient does not have to be English literate, booklet could be read to patient in English) ☐ Yes   ☐ No           ☐ Not sure, there might be some difficulty    4. Based on the nurse’s answers to these three questions, what is this patient’s proficiency with English?   ☐English proficient    ☐Limited English proficiency  ☐Not English proficient    Please explain your answer  ____________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________  127  *DEFINITIONS  ENGLISH PROFICIENT (EP).  There is clear indication in the health records that the patient speaks fluent English and is able to understand and converse in English regarding his/her care without assistance (i.e. documented in Nursing admission/assessment form, nurses’ notes, pre-operative checklist, consult notes).   LIMITED ENGLISH PROFICIENCY (LEP).  There is some indication in the health records that the patient primarily speaks a language other than English AND is NOT fluent in English (i.e., requires some assistance with translation/interpretation from another language).  (LEP is defined as being unable to communicate effectively in English because one’s primary language is not English and fluency in the English language has not been developed.)    NOT ENGLISH PROFICIENT (NEP).  There is clear indication in the health records that the patient speaks NO English, is unable to understand and converse in English, and/or requires a translator (informal or formal).   UNKNOWN LANGUAGE PROFICIENCY.  There is no indication in the health records of the patient’s language status (i.e., the record has not been completed, the record is not readable or decipherable, or there is no indication of whether or how the patient communicated with staff).      128  Appendix D:  Data Abstraction Tool VCH Challenge LEP Study: Data Collection Tool (version 1.2) Subject Number: ___________________ Date of CSICU Admission(ddmmmyy): ____________________ Date of CSICU Discharge(ddmmmyy): ______________________________ Date of Hospital Admission(ddmmmyy): ___________________ Date of Hospital Discharge/Transfer to another facility(ddmmmyy): ______________________  Today’s date(ddmmmyy):_________________________ Initials of Data Collector: ___________________________  Pre-operative Conditions:  Age            _________ Gender Female                   Male Marital Status Single                Married/common law       Divorced            Unknown    Widowed         Employment Status Employed                Unemployed Retired          Other        Unknown         Height  ________ft./cm (circle one)  Weight ________lbs./kg (circle one)  BMI  _ _ . _ HTN Yes                          No Dyslipidemia Yes                          No Diabetes Yes                           No Type          I            II             Diet Controlled Respiratory disease (ie. asthma, COPD, emphysema) Yes                           No If yes, what: ___________________ Number of diseased vessels _________ Left Main disease? Yes                          No 129  Pre-op mitral or aortic valve disease Yes                          No Pre-op LVEF (%)               _______ Pre-op CCS Class I                II             III              IV Pre-op NYHA Class I                II             III              IV Pre-op MI Yes                     No If yes, date? ___________   <7 days    1-6 weeks   >6 weeks History of previous cardiac surgery/procedures Yes                          No If Yes, what: ______________________ History of mental health illness  Dementia  Alzheimer’s  Bipolar disorder/Schizophrenia  Depression  Suicidal Ideation  Other:  NONE History of Stroke/TIA Yes                          No History of Substance Abuse or dependence (i.e. IVDU, alcohol, illicit drugs) Yes                           No If yes, what: _______________________ History of tobacco use (Former ) Never                Former                Current Extracardiac disease (PAD, PVD, carotid stenosis) Yes                           No Preoperative use of inatropic agents Yes                          No  Peri- and Post-operative conditions: Number of grafts _______ Pump time (mm) ON:___:___                  OFF:___:___ TOTAL TIME: ___  minutes 130  Return to OR Yes                          No Prolonged intubation (>24 hours) Yes                          No Reintubation Yes                          No Peri/Post-operative stroke Yes                          No Post-operative delirium (requiring psych consult) Yes                          No Post- operative infection (i.e. UTI, pneumonia, surgical site) Yes                          No if yes, specify site : ______________ Discharge delayed as patient awaiting LTC placement Yes                           No Notes/Additional comments: i.e.-OTHER causes that AFFECTED LOS and/or delayed transfer to ward/discharge (post-op arrest, dialysis, IABP, post-op MI, post-op Atrial fibrillation, dysrhythmias resulting in pacemaker dependency) ______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________ Readmission: Did this subject have any readmission within the 1st 30 days of discharge?  Yes                           No date (ddmmmyy):___________ Reason for admission (diagnosis): ______________________________ OUT OF TOWN PATIENT? YES  NO  131  Appendix E:  Canadian Cardiovascular Society grading of angina pectoris and New York Heart Association Functional Classification Canadian Cardiovascular Society (CCS) grading of angina pectoris (Campeau, 1976)   Grade Description I Ordinary physical activity does not cause angina, such as walking and climbing stairs.  Angina with strenuous or rapid or prolonged  exertion at work or recreation  II Slight limitation of ordinary activity.  Walking or climbing stairs rapidly, walking uphill,walking or stair climbing after meals, or in cold, or in wind, or under emotional stress, or only during the few hours after awakening.  Walking more than two blocks on the level and climbing more than one flight of ordinary stairs at a normal pace and in normal conditions  III Marked limitation of ordinary physical activity.  Walking one or two blocks on the level and climbing one flight of stairs in normal conditions and at normal pace    IV Inability to carry on any physical activity without discomfort, anginal syndrome may be present at rest     New York Heart Association Functional Classification (NYHA) (The Criteria Committee of the New York Heart Association, 1994)  Grade Description I Cardiac disease, but no symptoms and no limitation in ordinary physical activity, e.g. shortness of breath when walking, climbing stairs etc.  II Mild symptoms (mild shortness of breath and/or angina) and slight limitation during ordinary activity.  III Marked limitation in activity due to symptoms, even during less-than-ordinary activity, e.g. walking short distances (20–100 m). Comfortable only at rest.  IV Severe limitations. Experiences symptoms even while at rest. Mostly bedbound patients.   


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items