Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Deriving health utility weights for infants with Respiratory Syncytial Virus (RSV) Roy, Lilla MC 2013

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

Item Metadata

Download

Media
24-ubc_2013_fall_roy_lilla.pdf [ 2.06MB ]
Metadata
JSON: 24-1.0074259.json
JSON-LD: 24-1.0074259-ld.json
RDF/XML (Pretty): 24-1.0074259-rdf.xml
RDF/JSON: 24-1.0074259-rdf.json
Turtle: 24-1.0074259-turtle.txt
N-Triples: 24-1.0074259-rdf-ntriples.txt
Original Record: 24-1.0074259-source.json
Full Text
24-1.0074259-fulltext.txt
Citation
24-1.0074259.ris

Full Text

DERIVING HEALTH UTILITY WEIGHTS FOR INFANTS WITH RESPIRATORY SYNCYTIAL VIRUS (RSV) by Lilla MC Roy  BScN, Laurentian University, 2008  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Pharmaceutical Sciences)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  September 2013 ? Lilla MC Roy, 2013   ii  Abstract  Background: RSV infects the majority of infants under one year, and is a leading cause of bronchiolitis-related hospitalization, incurring significant cost on the health care system. Economic evaluation guidelines require the use of quality-adjusted life years (QALYs) ? a single measure of health outcome that incorporates both quality of life and length of life; however, there is no established standard for eliciting health utilities in children. Measuring utility in infant RSV presents significant methodological challenges related to proxy response, perspective of the elicitation task, and the temporary nature of the health condition combined with potential long-term outcomes.   Objectives: The objectives of this thesis were to: 1) review the methodological literature surrounding infant utility derivation using a proxy respondent and in temporary health states; 2) derive the health utility for different severities of Respiratory Syncytial Virus (RSV) infection; and 3) determine relative preferences of health state attributes for RSV in infants.  Methods: Using results from a literature review and qualitative study, time trade-off (TTO) and best-worst scaling (BWS) experiments were constructed reflecting four RSV severity levels, using two perspectives ? child and adult. Administered online, respondents randomly received either a questionnaire from either the adult or child perspective, completing both TTO and BWS tasks, allowing elicitation and comparison of both health state utilities and relative preferences for health state attributes from two perspectives   iii  Results: Respondents were willing to trade off more time to avoid more severe health states, indicating greater disutility associated with severe RSV, and suggests respondents understood the task. Responses differed between the adult and child perspectives, with disutilities from the child perspective being slightly greater. BWS revealed that respondents most often chose breathing failure as the worst attribute and 0% risk of wheeze as the best from both perspectives.  Conclusion: Responses differ between adult and child perspectives. This is the first study known to explore utilities for RSV in infants, and can be utilized for the further utility research and economic evaluation of any current or potential treatment for RSV, or other temporary health states in infants.   iv  Preface The work presented in this thesis was conducted and disseminated by the Master?s candidate. The Master?s candidate contributed to the methodological review and design of the studies herein, completed all data analyses, and prepared the manuscripts relating to the contents of this thesis. Versions of Chapter 2 (Literature review: methodology), Chapter 4 (Parent?s experience of their child?s illness with RSV), Chapter 5 (Time trade-off experiment), and Chapter 6 (Best-worst scaling experiment) will be submitted for publication by Lilla Roy. The co-authors of the manuscripts will review and offer critical evaluations, however, the candidate will be responsible for the writing and the final content of the manuscripts. The co-authors of the manuscripts that comprise parts of this thesis made contributions only as is commensurate with a thesis committee or as experts in the specific area is it pertains to the work. The co-authors provided direction and support. The following certificates of approvals were received for conducting the research within this thesis: 1 Parents? Experience of RSV in their Children (RSV focus group). Children and Women?s Research Ethics Board (H11-01723). 2 Valuing Health States in Respiratory Syncytial Virus (RSV Survey). UBC Behavioural Research Ethics Board, the University of British Columbia (H11-03357).    v  Table of contents  Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of contents ........................................................................................................................... v List of tables.................................................................................................................................. xi List of figures .............................................................................................................................. xiii Glossary ...................................................................................................................................... xiv Acknowledgements ..................................................................................................................... xv Chapter 1. Respiratory Syncytial Virus .................................................................................. 1 1.1 Disease course ........................................................................................................... 1 1.2 Short and long-term outcomes of RSV ..................................................................... 3 1.2.1 Short term outcomes.............................................................................................. 3 Hospitalizations ................................................................................................................ 4 Oxygen ............................................................................................................................. 6 1.2.2 Long-term risk of wheeze ..................................................................................... 7 1.2.3 RSV-related mortality ? do infants die from RSV? .............................................. 7 Mortality in otherwise healthy infants ............................................................................. 8 Mortality in children with lung disease. ........................................................................... 9 Mortality in children with heart disease ......................................................................... 10 Mortality in children born prematurely .......................................................................... 10 The overall picture of mortality in RSV......................................................................... 11 1.2.4 Management of RSV ........................................................................................... 12 vi  Chapter 2. Measuring utility in RSV .................................................................................... 14 2.1 Stated preference measures to estimate utility ........................................................ 14 2.1.1 Utility theory ....................................................................................................... 15 2.1.2 Utilities ................................................................................................................ 15 2.1.3 Quality adjusted life years ................................................................................... 16 2.1.4 Visual analogue scale (VAS) and standard gamble (SG) ................................... 17 2.1.5 Time trade-off (TTO) .......................................................................................... 18 2.1.6 Best-worst scaling (BWS) as a method to elicit preferences .............................. 20 2.2 Previous cost-utility/utility studies in RSV............................................................. 22 2.2.1 Cost utility analyses ............................................................................................ 22 2.2.2 Utility studies ...................................................................................................... 32 2.3 Knowledge gap #1 .................................................................................................. 35 2.4 Methodology ? who?s life, who?s perspective, and how do we approach temporary health states? ......................................................................................................................... 35 2.4.1 Challenges in deriving utilities in infants ............................................................ 35 2.4.2 Challenges in deriving utilities in temporary health states ................................. 37 2.4.3 Further modifications of the time trade-off ......................................................... 44 2.5 Knowledge gap #2 .................................................................................................. 45 2.6 Previous qualitative studies..................................................................................... 45 2.7 Knowledge gap #3 .................................................................................................. 47 Chapter 3. Rationale for thesis research ............................................................................... 48 3.1 Knowledge gaps ...................................................................................................... 48 3.2 Justification for thesis research ............................................................................... 48 vii  3.3 Expected implications of study ............................................................................... 49 3.4 Thesis objectives ..................................................................................................... 50 3.5 Thesis overview ...................................................................................................... 51 Chapter 4. Parent?s experience of their child?s illness with RSV ......................................... 53 4.1 Background ............................................................................................................. 53 4.2 Methods................................................................................................................... 54 4.2.1 Recruitment and ethics ........................................................................................ 54 4.2.2 Design and data collection .................................................................................. 55 4.2.3 Analysis ............................................................................................................... 55 4.3 Results ..................................................................................................................... 56 4.3.1 Sample demographics ......................................................................................... 56 4.3.2 Descriptive .......................................................................................................... 56 4.3.3 Key parental experiences with RSV.................................................................... 57 Consistent themes ........................................................................................................... 57 Common themes ............................................................................................................. 59 Occasional themes .......................................................................................................... 60 Additional themes and impressions................................................................................ 65 4.4 Discussion ............................................................................................................... 67 4.4.1 Major findings ..................................................................................................... 68 4.4.2 Limitations .......................................................................................................... 71 4.4.3 Conclusion ........................................................................................................... 72 Chapter 5. Time trade-off experiment .................................................................................. 73 5.1 Background ............................................................................................................. 73 viii  5.1.1 Purpose and advantage of TTO ........................................................................... 73 5.1.2 Challenges in the infant health state of RSV....................................................... 74 5.1.3 Elicitation task ? possible approaches ................................................................. 76 5.1.4 Incorporating qualitative research into choice experiments ................................ 77 5.2 Methods................................................................................................................... 78 5.2.1 Methodology ....................................................................................................... 78 5.2.2 Attributes ............................................................................................................. 79 5.2.3 Recruitment and ethics ........................................................................................ 80 5.2.4 Elicitation design ................................................................................................. 80 Preamble and follow up questions ................................................................................. 80 ?Health profile? versus ?health state? ............................................................................ 81 Elicitation task ................................................................................................................ 81 5.2.5 Pilot study ............................................................................................................ 85 5.3 Analysis................................................................................................................... 85 5.3.1 Calculation of disutility and utility...................................................................... 85 5.3.2 Data exclusion criteria ......................................................................................... 86 5.4 Results ..................................................................................................................... 86 5.4.1 Sample characteristics ......................................................................................... 86 5.4.2 Overall disutilities ............................................................................................... 90 5.4.3 Subgroup analyses .................................................................................................... 90 5.5 Discussion ............................................................................................................... 93 5.5.1 Major findings ..................................................................................................... 93 5.5.2 Strengths .............................................................................................................. 95 ix  5.5.3 Clinical relevance ................................................................................................ 95 5.5.4 Limitations .......................................................................................................... 96 5.5.5 Conclusion ........................................................................................................... 98 Chapter 6 Best-worst scaling experiment ................................................................................. 99 6.1 Background ..................................................................................................................... 99 6.1.1 Origin, purpose, and advantage of BWS ............................................................. 99 6.1.2 Design theory (OMEP and BIBD designs) ....................................................... 101 6.1.3 Challenges of BWS choice experiment ............................................................. 103 6.2 Methods................................................................................................................. 105 6.2.1 Choice task design ............................................................................................. 105 6.2.2 Attributes and levels .......................................................................................... 105 6.2.3 Recruitment and ethics ...................................................................................... 106 6.2.4 Choice task ........................................................................................................ 106 6.2.5 Sawtooth Software .................................................................................................. 107 6.2.6 Pilot ................................................................................................................... 108 6.2.7 Analysis ............................................................................................................. 108 6.3 Results ................................................................................................................... 109 6.3.1 Sample characteristics ....................................................................................... 109 6.3.2 Relative preferences ................................................................................................ 111 6.4 Discussion ............................................................................................................. 115 6.4.1 Major findings ................................................................................................... 116 6.4.2 Clinical relevance .............................................................................................. 117 6.4.3 Limitations ........................................................................................................ 118 x  6.5 Conclusion ............................................................................................................ 119 Chapter 7 Discussion .............................................................................................................. 120 7.1 Important findings and implications ............................................................................. 120 7.2 Limitations ............................................................................................................ 123 7.3 Future directions for research ............................................................................... 125 7.4 Conclusion .................................................................................................................... 127 References .................................................................................................................................. 128 Appendix A: Medline searches for literature review ............................................................. 142 Appendix B. Focus group guidelines ....................................................................................... 146 Appendix C: Qualitative coding scheme ................................................................................. 153 Appendix D: BWS design statistics ......................................................................................... 158   xi  List of tables  Table 1.1 Lower Respiratory Tract Infection score ........................................................................ 4 Table 1.2 Working definition of RSV-associated mortality ........................................................... 8 Table 2.1 Systematic reviews of cost utility analyses ................................................................... 25 Table 2.2 CUA design summary ................................................................................................... 27 Table 2.3 CUA results and quality summary ................................................................................ 29 Table 2.4 CEA registry utility values ........................................................................................... 34 Table 2.5 Direct elicitation utility studies for  temporary health states valued by proxy respsondents .................................................................................................................................. 42 Table 4.1 Important themes by frequency (6 or more interview) ................................................. 57 Table 4.2 Additional aspects of parents' experience of their child with RSV .............................. 66 Table 4.3 Themes with perceived high sense of importance ........................................................ 67 Table 5.1 Health profile descriptions ............................................................................................ 80 Table 5.2 Canada time trade-off: respondent characteristics ........................................................ 88 Table 5.3 Canada time trade-off: Mean disutility for adult and infant perspective ...................... 90 Table 5.4 Canada time trade-off: Mean disutility for adult and infant perspective by gender ..... 91 Table 5.5 Canada time trade-off: Mean disutility for adult and infant perspective by whether respondents have children ............................................................................................................. 91 Table 5.6 Canada time trade-off: Mean disutility by age for adult perspective ........................... 92 Table 5.7 Canada time trade-off: Mean disutility by age for infant perspective .......................... 93 Table 6.1 Best-worst scaling attribute levels .............................................................................. 106 Table 6.2 Characteristics of BWS respondents........................................................................... 110 xii  Table 6.3 Relative utilities (conditional logit model) ................................................................. 112   xiii  List of figures  Figure 2.1 Time trade-off method ................................................................................................. 20 Figure 2.2 Results of literature search .......................................................................................... 24 Figure 5.1 TTO elicitation task (adult) ......................................................................................... 83 Figure 5.2 Reasons for exclusion from final analysis ................................................................... 87 Figure 6.1 BWS sample choice task ........................................................................................... 107    xiv  Glossary  Abbreviation Definition BIBD Balanced Incomplete Block Design BIPAP/CPAP Biphasic Positive Airway Pressure/Continuous Positive Airway Pressure BPD Bronchopulmonary Dysplasia BWS Best-Worst Scaling CHD Congenital Heart Disease DCE Discrete Choice Experiment GA Gestational Age IV Intravenous LCA Latent Class Analysis LRI Lower Respiratory Infection NG tube Nasogastric tube QALYs Quality Adjusted Life Years SG Standard Gamble TTO Time Trade-off VAS Visual Analogue Scale    xv  Acknowledgements  I am truly grateful for the people who, through their innate strengths and virtues, have given me support and inspiration during my studies and day to day life. In particular, the continual guidance and mentorship from my supervisors, Dr. Larry Lynd, Dr. Carlo Marra, and Dr. Nick Bansback, is deeply appreciated. I would also like to thank the members of CORE for their unwavering confidence, camaraderie, and assistance during this program, and AllerGen NCE for their professional and financial support.  Thanks also go unconditionally to the friends and family who have supported me in this endeavor. Thank you for your patience, your acceptance, and your faith.  1  Chapter 1.  Respiratory Syncytial Virus  1.1 Disease course Respiratory syncytial virus (RSV) is a paramyxoviridae pneumovirus, discovered in 1956, with two subtypes ? RSV-A and RSV-B, of which it is unclear whether one is more virulent than the other (2). Respiratory syncytial virus is transmitted from respiratory secretions by close contact with infected individuals or contact with contaminated surfaces. The virus carries a seasonal, predictable pattern, with outbreaks occurring throughout 4-6 months of the year, from the late fall to the early spring(3,4). The virus has an incubation period of 2-8 days, and symptoms usually last up to 2 weeks (5). Its peak virulence is historically between 5 and 8 days (6). Reinfection with RSV in the second year of life is approximately 76%; however, the severity of the disease generally decreases with repeat infections (7).   Respiratory syncytial virus is the most common cause of viral respiratory infection in infants and young children worldwide (4,6,8,9), and is the leading cause of hospitalizations in infants in the United States (1,8). The prevalence of RSV is widely understood to be very high among children under 2 years old (3,10), with all children by the age of three having been infected by the virus at least once (1,6,11), and 70% of infants are infected in their first year of life (6,12,13). Approximately one half of those who are infected develop a lower respiratory tract infection (LRTI) (a potentially severe infection of the lower lungs), and about 2.5% of these infants require hospitalization, the majority of which  are under 6 months of age (6,13,14).   2  RSV-related morbidity is highest in the very young and the elderly (15). Many patients admitted to hospital with RSV also have an existing underlying illness (16?19). A multi-city Canadian study in 1997 reported 52.6% of children had a pre-existing comorbidity (e.g. congenital heart disease, chronic lung disease, <36wGA, <6wold, early hypoxia) (17). Although the risk of mortality due specifically to RSV is relatively low, the risk is increased, and cannot be ignored, for compromised individuals (see section 1.2.3 for further mortality discussion).   Severe disease is associated with the following risk factors: male gender, prematurity, chronic lung disease (CLD), congenital heart disease (CHD), cystic fibrosis or other lung diseases, acquired and congenital immune-deficiencies, tobacco exposure, and crowding (1). A Canadian Arctic study also reports Inuit race as a risk factor, although admittedly this may be related to socioeconomic factors such as poverty, accessibility to health care, and environmental factors rather than genetics or immunity (20).  The long term impact of RSV is unclear. RSV can be detected by a diagnostic test up to 100 days after infection. This has been hypothesized to relate to ensuing airway hypersensitivity and recurrent wheeze in the first 13 years of life (1,15). The association of RSV with childhood wheeze has been demonstrated by multiple studies (21?25). Infants and children infected with RSV have increased odds of frequent childhood wheeze compared to the odds of those without RSV up to the age of 13 (age 11: OR=2.4(CI 1.3-4.6)) (24). After this point, there has been no clear association shown.  3  1.2 Short and long-term outcomes of RSV In order to have a comprehensive understanding of RSV, and to create the foundation for the construction of an appropriate utility tool, it was necessary to describe the attributes and aspects of RSV illness which are central to the disease. Initiating from the outcomes used in the RSV trials, several key attributes of RSV are summarized in this section, and were later used for the development of vignettes and attribute selection. 1.2.1  Short term outcomes The severity of RSV can be classified in different ways. For example, it can be broken down to mild infection (upper respiratory tract infection (URTI)), moderate infection (lower respiratory tract infection (LRTI)), and respiratory failure (LRTI) (1). The Lower Respiratory Infection (LRI) Score (Table 1.1) has been consistently used in RSV research since its development in 1990, and it, contains 6 clinical levels, from 0 (no LRI) to 6 (mechanical ventilation) from four different domains (26?29). The score uses three components of respiratory status ? oxygen saturation, respiratory rate, and physical findings (adventitious sounds and retractions) as differentiating criteria. Authors have adopted slightly modified approaches using the highest score, a mode, or a mean of the three components of the score. The LRI score for each child is based on a clinical expert?s opinion on the following criteria.   4  Table 1.1 Lower Respiratory Tract Infection score Score Oximetry (% oxygen saturation) Respiratory rate Retractions and adventitial sounds    Retractions Adventitial (wheezing, rales) 0 ? well (baseline) Baseline Baseline (no URTI) No change/none (no URTI) No change/none (no URTI) 1 ? URTI (nasal congestion, rhinorrhea) Baseline (>/= 95%) Baseline (URTI) <45/min No change/none/ minimal No change/none/ minimal 2 ? Mild LRI Fall <5% (91-94%) Rise <15% (1-14/min) 45-59/min Intercostal Mild increase 3 ? Moderate LRI Fall of 5%-10% (86-90%) Rise of 15%-30% (15-30/min) 60-74/min Intercostal and subcostal Moderate increase 4 ? Severe LRI Fall >10% (</=85%) Rise <30% (<30/min) 75+ All of above with seesaw chest motion Severe increase 5 ? Respiratory failure Mechanical ventilation Mechanical ventilation Mechanical ventilation Mechanical ventilation *adapted from Groothuis et al, 1990; Groothuis et al 1993; Rodriquez 1997  Hospitalizations Reported hospital admission rates in North American preterm infants infected with RSV vary from 2.7% to 37% (26,30?34). In Canada, Law et al (31) reported that 3.7% (n=69/1860) of infants born between 33-35 weeks gestational age (GA) were admitted with RSV related LRTI  in their first RSV season. Similarly, the IMPACT-RSV (26) trial reported RSV hospitalization rates in infants born <35 weeks GA of 4.8%  (n=48/1002 intervention arm) and 10.6% (n=53/500 control arm), and in the UK, a study of premature infants born less than 32 weeks GA showed an admission rate in their first year of life of 4% (n=3/82) for RSV LRTI. In the Canadian Arctic, these rates are much higher, with a reported 48% of infants admitted to hospital for a LRTI, the highest recorded rates in the world, with RSV (as a stand-alone infection, not including co-infections) responsible for nearly 40% of these LRTI admissions (20). In infants who are high risk, such as preterm infants, RSV hospital admission have been reported as high as 37% (35).   5  Children are admitted to a particular hospital ward based on the severity of their illness. A child can be admitted for precautionary monitoring reasons only, or their situation could be severe enough to require respiratory resuscitation and consequently, mechanical ventilation. An infant with RSV could, in short, have one or more of three experiences in the hospital ? they could be seen in an emergency department (ED) and sent home, admitted to a general medical ward, or admitted to the intensive care unit (ICU). A patient could experience all three throughout the course of an illness, if they have an ED visit prior to or after their principle hospital admission. If admitted to the ICU, they are likely to have a subsequent stay on the general ward. This relationship is not reciprocal, as all those admitted to a general ward will not necessarily have an ICU stay.  Decreasing length of stay has been a focus of therapeutic intervention. The length of stay of those infants with pre-existing illness were in hospital significantly longer compared to those who were premature or younger than 6 weeks of age, who in turn spent longer than those without any known  risk factors (16). A Canadian study also assessed health care utilization of preterm infants related to RSV after discharge home after birth. Of this preterm group of 1832 infants, 3.6% were re-hospitalized with confirmed RSV related respiratory infections, and another 3.1% had a confirmed RSV related emergency room visit (31). An additional 31% had an MD or clinic visit that was not confirmed to be RSV-related, although it is likely that RSV was related to a portion of these visits and just not severe enough to warrant testing.  6  Oxygen Lower respiratory infections result in decreased ability for the lungs to functionally oxygenate the body. Oxygen levels drop, demonstrated by signs of hypoxia and decreased oxygen saturation (i.e. O2 saturations) thus requiring supplementary provision of oxygen to de-compensating infants. Patients receive oxygen in a variety of methods ? nasal prongs, blow-by oxygen, face mask, high-flow oxygen (mask devices that deliver high concentrations of oxygen at a measureable rate), head boxes (boxes over the head of a sleeping infant), BIPAP/CPAP (mechanically assistive devices that respond to the body?s own willingness to breathe), and mechanical ventilation (in which the body is sedated and a machine determines the rate and depth of inspiration). In earlier, smaller studies, up to 24% of high risk children admitted to hospital required mechanical ventilation for respiratory failure due to RSV infection (16,34,36,37). Larger, more recent randomized control trials show much lower rates, up to 0.7% (n=1502) in preterm infants (26) and 2.2% (n=1287) in infants with CHD (30), although one 1997 Canadian observational study of 1425 high risk children reports a rate up to20% (17). This difference may have been due to the nature of the strict inclusion criteria in the randomized control trials, limiting high risk children with other comorbidities besides prematurity and CHD, such as lung disease.  The requirement for supplementary oxygen is both a factor in the length of hospital stay and health care costs, as acutely ill infants can generally only receive oxygen in hospital. Consequently, the number of days on oxygen is a common endpoint in research. The vast majority of infants admitted require oxygen and reported days with increased oxygen varied between 2-9 days. (16,26,30)  7   1.2.2 Long-term risk of wheeze The potential association between RSV and long term wheeze has been well documented (3,38). Reviews reported a prevalence of up to 76% of recurrent childhood wheeze following an RSV infection. Odds of frequent wheeze and infrequent wheeze by age 6 in those who have had RSV lower respiratory tract infection (LRTI) in early childhood were greater than those who had not had RSV LRTI infection (OR 4.3 (2.2-8.7) and 3.2 (CI 2.0-5.0) respectively), with risk of wheeze decreasing as the child gets older, to the age of 13. After this point, there was little risk of atopy and wheeze in those who have had RSV compared to those who have not (24). Further supporting the long term respiratory implications, a 2012 study reported an increased respiratory related health care utilization in the first two years of life of infants who have had RSV hospitalizations compared to infants who have not had RSV hospitalizations (respiratory emergency visits, respiratory family physician visits, and respiratory outpatient visits) (39). Recurrent hospitalization also occurred in a small percentage of infants very close to discharge (40).  1.2.3 RSV-related mortality ? do infants die from RSV? The question of whether death can be attributed to Respiratory Syncytial Virus (RSV) is important to decision makers and consumers alike. The association of RSV with mortality is clear (3,4,19,41). Case fatality rates of RSV are relatively low in the general population of otherwise healthy infants, older children, and adults, however, the morbidity and mortality is more substantial in children and infants with underlying illnesses and compromised health. Existing illness with congenital heart disease (CHD), chronic lung disease (CLD), 8  bronchopulmonary dysplasia (BPD), and prematurity (generally defined as less than 36 weeks gestational age (wGA)) were the most common conditions which potentiate poor outcomes of infection with RSV. Estimates for mortality related to RSV in these particular subgroups have approached 23% (42).  Patients with pre-existing comorbidities, such as lung disease, are faced with the added challenge of fighting the RSV virus with compromised abilities. For this reason, when death occurs, it is difficult to tell what may be the cause of death ? was it attributable to RSV, or to the pre-existing condition? The following sections were developed to establish an understanding of the published mortality risk for different subgroups of children with RSV - CHD, CLD, prematurity, and overall/low-risk (search updated July 4, 2012), chosen because these subgroups are dominant in the literature. This section summarizes reported mortality rates and whether RSV-association was indicated (RSV-association defined per Table 1.2). Table 1.2 Working definition of RSV-associated mortality Indication Definition Yes source explicitly stated that mortality was associated with RSV (listed as cause of death, or statement of rationale for association found in paper) Some source stated that some cases were RSV-associated and others weren?t, but breakdown of subgroups not provided Unclear no statement explaining whether cause of death was associated with RSV infection, chronic illness, or simply an observation after the fact (e.g. total mortality) No Source states it is not related to RSV  Mortality in otherwise healthy infants  Although the consensus is clear that mortality in otherwise healthy infants is low, there were fewer studies reporting rates of death in this infant population than in high risk populations. Overall estimates may have included infants that are at high risk or infants who have pre-existing conditions. ?Overall? mortality ranged from 0% - 8.6% (18,43?51), while estimates in healthy 9  children have ranged from 0.1% to 1.7% (16,18,19,36,50,52?54). In the papers reviewed, none of the mortality reported occurred in children without prior known risk factors (16,44,48,54,55).  Mortality in children with lung disease.  Studies of high risk children with lung disease reported mortality rates from 0% to 3.5%. Although all listed studies included lung disease in their inclusion criteria, many were mixed with other subgroups (i.e. prematurity), and the possibility of interaction between these groups should be carefully considered when describing mortality in children with lung disease. In addition, evaluation of mortality was not the primary purpose and therefore these studies did not have enough power to evaluate mortality specifically. Only four of the twelve studies specifically attributed death to RSV, four studies were unclear whether death was attributable to RSV, and two stated the case fatalities were not related to RSV. The total death rate which is reported after follow-up without cause of death makes it difficult (if not impossible) to attribute death to RSV. This kind of overall estimate is unable to be used in many decision-analytic models as they require a specific death rate for those admitted with RSV infection (56).   An early 1990s Canadian study (37) indicated a mortality rate in children with CLD of 3.5%, assigning an arbitrary time window of death within two weeks to indicate RSV-association. As the course of RSV is approximately 10-14 days, this may be a reasonable time frame, however, it may be have been an over-estimation, or under-estimation of mortality, as no further information on cause of death was given. We should also consider that this upper estimate of 3.5% may have changed, as this study was conducted twenty years ago, and prior to the use of newer treatments for RSV.  10   Mortality in children with heart disease  There is a large range of estimates for mortality risk in infants with heart disease, and appear to be at higher risk of severe RSV than other groups. In the papers reviewed, reported rates of death have ranged between 0% and 37% for children with CHD. The largest value, 37% (n=27), was from an early 1980?s study and is dramatically higher than any conducted since then (57). Four studies (16,29,36,37) conducted in the early 1990?s report up to 3.4% mortality in children with CHD who have RSV, however, only one of these studies reported RSV-associated mortality (3.4%) (37). The largest RCT thus far investigating palivizumab use in infants with heart disease and RSV reported an all-cause mortality rate of 3.3% (intervention arm) and 4.2% (placebo arm), and RSV-attributable mortality rate was 0.31% (intervention arm) and 0.62% (placebo arm) (30). A meta-analysis reported 3.72 as a weighted average. Based on the papers reviewed, it appears that the risk of death in infants with heart disease is still slightly higher than the estimates for infants with lung disease, around 4%.  Mortality in children born prematurely  The reported range of mortality for infants born prematurely is from 0% to 8.1%. The high estimate was from a large study (n=2415) (40), which specifies cause of death in the RSV cohort as ?known? (1.97%) or ?sudden unknown cause? (6.13%), however, the ?known? cause of death was not explained. The sudden unknown death rate was attributed to Sudden Infant Death Syndrome (SIDS), and therefore should not be attributable to RSV (40,56). The large scale IMPACT-RSV palivizumab trial also reported overall mortality during the RSV hospitalization without cause of death, and similarly to other subgroups already discussed, many studies do not 11  specifically state the cause of death, others are unclear, and only five of those reviewed deemed mortality as RSV-attributable. As infants born pre-term and with other illnesses can have complications due to their condition, and often can be cross-classified (i.e. premature infants often have lung problems) the use of these rates require serious consideration before assuming that a death was due to RSV rather than their underlying condition.  The overall picture of mortality in RSV  Four reviews have also been published between 2003 and 2010 that provide summaries of mortality estimates related to RSV -  two from the UK, and two from the USA (3,4,42,56). Mortality estimates range from 0% to 33% depending on the subgroup in question, and echo the paucity of cause of death information (42). Even when death is reported as RSV-associated, how RSV association was ascertained was rarely clear (16,30,34,37,58?60). The best mortality estimates may be those which are reported specifically as being RSV-associated, however, the downside of using data from only those studies that reported RSV-association is that it limits the use of larger study data, which have, thus far, not specified if deaths are RSV-associated (26,40). Overall, there is strong evidence that the overall mortality rate of RSV is less than 1%, and that the risk of mortality is significantly higher in children with comorbidities such as lung disease, congenital heart disease, and those born prematurely. Using only the RSV-associated rates presented here, this would mean that in these subgroups, there would be a mortality rate of approximately 1.8% in prematurely born children, 4.2% in children with CHD, and 3.5% in children with lung disease.  12  1.2.4 Management of RSV Current treatment involves prophylactic measures and symptom management. Primary prophylactic strategies include general and hand hygiene to decrease the spread of infection (6).  Vaccine development has been researched for many years, however, there is only one prophylactic pharmacologic agent that is currently approved in multiple countries, including Canada and the UK - palivizumab (Synagis?), a humanized monoclonal antibody specific for RSV.   According to the pivotal randomized control trial, the IMPACT-RSV study, intramuscularly administered palivizumab can decrease rates of hospitalization and admission to ICU in high risk children (26). Treatment is initiated prior to infection, with one weight dependent dose given monthly during the RSV season, to a maximum of 5 doses. The clinical trials were conducted in preterm infants under six months old or children with bronchopulmonary dysplasia (BPD), and resulted in an overall 55% reduction in RSV hospitalization (78% reduction in premature subgroup, 39% reduction in BPD subgroup) (26). In congenital heart disease (CHD), the trial showed a 45% reduction (30). Although the methods of these trials were not closely examined for this thesis, it is important to note that the results were used differently across economic studies informing current practice guidelines in Canada (details will be provided in the next section).   Palivizumab is approved for administration for prophylactic treatment of RSV infection in high-risk children and infants (1,61). In Canada, this includes infants born prematurely, those with clinically significant congenital heart disease (CHD), and bronchopulmonary dysplasia (BPD). 13  The cost of palivizumab, however, is very high. A single 50mg vial can cost over CAN$750 (a 2010 estimate). If the recommended dose of palivizumab is 15mg per kg body weight, a 6 month old 7.5kg infant, it would cost CAN$1687.50 for one dose, without considering wastage and cost of administration (62).  During active disease, there is no single successful treatment option. Ribivirin, a nucleotide analogue, and intravenous immunoglobulin (IVIG) have been used in high risk populations, but their effectiveness is debatable (15). Oxygen therapy is the primary treatment, as well as nasopharyngeal suctioning. Nasopharyngeal suctioning is particularly beneficial to infants as they are nose breathers, and secretions are largely in the upper airways. Adrenaline/epinephrine, bronchodilators, corticosteroids, surfactant, inhaled nitric oxide, and heliox have shown little overall effect. Antibiotics are used in the case of co-existing infections.(6).  Given the significant morbidity burden due to RSV in young children, the increased risk of serious morbidity and mortality in children with pre-existing conditions, and the substantial financial investment required for prophylaxis using palivizumab, it is essential to conduct careful evaluations of existing and future treatment options in order to reduce the health care burden of this illness. Information on outcomes of RSV is required in order to effectively evaluate treatments for RSV. Canadian guidelines specify that one of the outcomes required for economic evaluation of treatments for RSV is a quality adjusted life year, a measure of quality and quantity of life. Calculation of a QALY requires the use of a health state utility value. The next section describes stated preference measures to estimate utility, and previous cost-utility research in RSV.  14  Chapter 2.  Measuring utility in RSV  2.1 Stated preference measures to estimate utility Funding decisions are no longer solely reliant on measures of cost and effectiveness. Decision makers are now more accountable for making evidence-based decisions, and for ensuring their decisions use patient centered well-being and health outcomes as a focus. As budgets get tighter, patients and taxpayers are more interested in how their health care dollars are spent, and want to voice opinions about decisions. This has resulted in an environment where a quality of life component of economic evaluations is required. The quality of life impact of different conditions are measured through valuation exercises of these health conditions to determine the utility ? or quality of life impact ? of that health condition. The utility is then used to calculate a quality adjusted life year (QALY), a measure of quality and quantity of life, which is used in economic evaluations. By completing valuation exercises, patient, individual, and social preferences around health conditions are able to be incorporated into policy and government level decisions. Study results may indicate that a treatment or program is effective, but if its target or delivery method conflicts with social or individual values, then uptake or adherence may suffer. By involving preferences, researchers and governments improve the likelihood of having target populations take up a service or treatment, and results in better health and cost outcomes. However, consideration must also be given to whether the utility was derived based on embedded societal preference or individual preferences, as these are two different types of preferences, and they will influence uptake differently.  15  2.1.1 Utility theory Utility theory was explored and developed by a mathematician, von Neumann, and an economist, Morgenstern, in the 1940?s, (von Neumann-Morgenstern, or vNM, utility theory) (63). It is based on a theory of rational decision making (64) which describes choices and decisions that should be made by humans under uncertainty. Von Neumann-Morgenstern utility theory forms the foundation for game theory and expected utility theory, and has been used in the formation of preference based health related quality of life measurements such as the standard gamble. These theories do not necessarily represent how humans actually make decisions (63?65), and the axioms which describe what rational decisions look like ? transitivity, continuity, and independence - have been challenged incessantly over the years (65,66). Whether the axioms hold or not, and despite the presence of other theories (e.g. random utility theory), utility theory is still commonly used, and there is still insight to be generated (67).   2.1.2 Utilities The preference weight generated from health related quality of life measurement tools based on vNM utility theory (e.g. standard gamble) is a utility. A utility is a measure of preference and quality of life for a particular disease state relative to a (hypothetically) otherwise perfect health state. Any given health state can be represented by a cardinal value ranging from 0 to 1, where 0 is equivalent to death, and 1 is considered to be perfect health (68). Utilities are important when quality of life is an important outcome or if a common unit incorporating both quality and quantity of life is desired (68). The assumption is that the most desirable alternative is that which offers the highest utility (67).  16  Utility can be measured using ordinal or cardinal methods. Ordinal utilities are derived using ranking exercises, are simple and easy, and are often the preferred approach. An ordinal utility requires an indication of preference for one health state over another (?which do you prefer?), however, they do not provide information on magnitude or strength of preference, and for this reason are often not considered to be adequate information on their own. Cardinal utilities provide information on strength of preference between health states, and are required to calculate a QALY. A health state utility value of 0.25 is considered to be half as desirable as a health state of 0.5. On the assumption that death (utility of 0) and perfect health (utility of 1) is the same across respondents, cardinal utilities can also be aggregated using an arithmetic mean, giving each respondent equal weight in the mean, and creating a social utility. For this reason, cardinal utilities are often preferred  (68).  2.1.3 Quality adjusted life years The combination of quality of life and quantity of life is measured using the quality adjusted life year (QALY). QALYs incorporate both quality of life (utility) and length of life into one measure. It is currently the only unit that is used in cost-utility analysis that permits consideration and integration of both quality of life and length of life  (68). Utility, as an indicator of the magnitude of quality of life impact, is applied to a measurement of time to produce a QALY.      ?      (63,64). For example, one year in a particular health state with a utility of 0.5 is equal to 0.5 QALYs. If an intervention was to increase the utility for one year to a 0.75, the QALYs gained would be 0.25 (i.e. 075-0.5). Likewise, if a person were to live one extra year at a utility of 0.5, they would generate an additional 0.5 QALYs (63,68).  17  Assumptions surrounding time can cause issues in the calculation of QALYs. The QALY makes assumptions on risk neutrality, additivity, and aggregation across individuals (i.e. a QALY is a QALY is a QALY) (69). In an issue coined maximal endurable time (MET) preference, the QALY model (V = L x Q) does not apply because time and utility are not proportional (70). Maximum endurable time preference occurs when shorter time in higher QALY is preferred nonlinearly. This can result in health states that are ?worse than death? (71). The QALY also assumes constant proportional trade off (described in more detail later in this section) (72).  2.1.4 Visual analogue scale (VAS) and standard gamble (SG) The visual analogue scale (VAS) is a simple ranking tool often used in conjunction with a traditional quantitative utility elicitation method (e.g. standard gamble or time trade off). The respondent is faced with an instrument, e.g. a line, with both ends of the line clearly marked with two health states (e.g. death and perfect health). The respondent simply marks on the line the value they attribute to the health state in question. The primary disadvantage of this is that this is not based on utility theory, and therefore true vNM utilities cannot be elicited from the VAS.   The standard gamble (SG) is considered the ?gold standard? of the traditional methods of utility elicitation (63). In the standard gamble, respondents are faced with a choice between two outcomes ? a particular health state for time t, or a probability of perfect health (P) for time t with a probability of immediate death (1 - P). Probabilities are varied until the respondent reaches a point of indifference. At the point of indifference the preference for the described health state (or the one the person is in) is equal to the vNM utility for that health state (63). It has been suggested that utilities derived from a time trade off (TTO) may better reflect individual 18  preferences than utilities derived from a standard gamble (73), perhaps due to the complexity of the standard gamble task. The limitations of the SG may mean that the TTO is a better alternative in certain situations such as where gambling is too complex, not practical, not feasible, or where respondents are risk averse or don?t understand probabilities (74).  2.1.5 Time trade-off (TTO) The TTO is a traditional, well-established method of deriving utilities, developed in the last 40 years, which adheres to the principles of expected utility theory, although it lacks the theoretical underpinnings of the standard gamble. Using a time trade-off, time in an undesirable health state is traded off for healthy years of life for some shorter period of time. The amount of years traded off theoretically proportionally represents the perceived severity of the health state ? the more years traded off, the more severe the health state is, and the more impact it should have on quality of life. Typically, the subject is faced with a choice: to live in a described health state for 10 (t) years or full health for x years, with x being less than 10 years (68). Trade-offs can be assisted using a visual board to determine the number of years in full health that are equivalent to 10 years in the specific health state presented. The utility itself is calculated as u=x/t, where x is the time traded, and t is the time available (in this case, it is 10 years) (Figure 2.1).  This point, the utility, is then applied to the time in a health state to create a Quality Adjusted Life Year (QALY). QALYs are a useful, quantifiable measure of quality of life that is used in economic evaluations and policy decision making.   Several issues are inherent in the time trade off that are relevant to infant health states. QALYs can be influenced by time issues. The TTO does not detect maximum endurable time preference 19  due to violation of the assumption of constant proportional trade off (CPTO). People value time differently ? older adults are less willing to give up large portions of their life to avoid certain health states, as they have less time left to live. This is discussed in more detail in section 2.4.2.   The TTO also can pick up states that are worse than dead (WTD) (68,74). The utility scale (of 0 to 1) assumes that death is the worst possible health state, however, some health conditions respondents consider to be ?worse than dead?. How to incorporate these utilities into economic evaluations and how to appropriately measure and interpret them is still under debate (70,75?77). Patrick (78) recommended the inclusion of states worse than dead, while others often excluded any states worse than dead, assigning them a utility of 0. (70,74,78). More often than not, studies do not include states worse than dead, and those that do, use different methods for states worse than dead and states better than dead (79). Consensus on how to address states worse than dead is lacking.  The framing of the elicitation task must be carefully considered as it can induce bias. For example, the utility of a health state can be affected by its duration, so it is imperative to specify time frame within the elicitation task(68)(94)(93). It is also important for the TTO to specify certainty surrounding what outcomes may or may not follow the health state in question, as this could affect responses from one individual to the next. For example, a health state lasting for 1 year with no long term consequences versus a chronic health state (lasting for 10 years or more) will have different meanings for respondents. This may be particularly true if respondents are close the end of their life, or if they expect to have most of their life left to live. Framing bias can also be introduced through the absence of time parameters such as age of onset, and the way the question is described. In addition, many studies have included probabilities within health state 20  attributes - it is important that these studies should consider the fact that people have different attitudes toward probabilities (i.e. risk aversion versus risk seeking). (68)  Figure 2.1 Time trade-off method   2.1.6 Best-worst scaling (BWS) as a method to elicit preferences It cannot be denied that there are issues with both utility theory and the use of QALYs (80?84). The principles above enlighten the challenges of expected utility theory and direct elicitation methods. In the context of health states and many health care decisions, revealed preferences are not easy to measure - you cannot choose to experience a health condition, and report the impact it has on you. Therefore, although empirical evidence shows differences between stated and revealed preferences (85), the best option available is for researchers to use stated preference methods for valuation of health states. In recent years, variations of stated preference tools have developed other methods such as discrete choice experiments (DCE), and specifically, best-worst scaling (BWS) experiments. 21   The discrete choice experiment model was originally used in marketing, but is being increasingly applied to healthcare research (86?89). The DCE presents respondents with two or more treatment options that are comprised of the same treatment attributes, but different levels of each attribute, from which they select the option they prefer for their individual circumstance or condition. A DCE has a number of implied assumptions ? that each health condition (i.e. RSV) can be broken down into specific characteristics (e.g. risk of death) that are can be defined by levels (e.g. risk of death = 0.05%, 0.1%, 1%), that each level of each attribute has a unique utility, and that these utilities can be combined across attributes. That is, if the sum of the patient?s utilities for attributes of treatment ?A? are greater than the sum of the utilities for the attributes of treatment ?B?, then given the choice, the patient should prefer treatment A to treatment B.  Best-worst scaling is a variation of DCE that provides relative preferences for different attributes and levels of a particular program or health condition. These preferences can indicate specifically which aspects of an illness or treatment influence and sway patient or consumer choices. Best-worst scaling works by presenting the respondent with a list of attribute levels, or characteristics of a treatment or program. Respondents are asked to choose which of the levels is the worst attribute level, and which is the best. Presumably, the respondent (consciously or subconsciously) looks at each attribute level, and compares it to each possible attribute level pair, and chooses the pair (one best, one worst) which demonstrate the maximum utility difference of all the choices. For this reason, BWS is sometimes called Maximum Difference Scaling. BWS tasks are considered easy to complete, and significantly less complex than DCE or traditional 22  preference elicitation methods. The primary disadvantage of choice experiments such as BWS is that the utilities are not on the vNM utility scale of 0-1.  One of the primary advantages of a best-worst scaling experiment is that it provides part worth utilities for all attribute levels save one. In combination with a time trade-off, these two methods provide different information on preferences. It is then possible to map the part-worth utilities to the utility for a health state valued using TTO. This enables an understanding of which attribute levels influenced respondents? preferences when completing the overall health state valuation task. Also, as it is impossible to value every possible health state (combination of attributes) using a TTO elicitation task, the part-worth utilities of the BWS can be used to elucidate utilities for different combination of attribute levels that may not have been valued in the TTO exercise. (90)  2.2 Previous cost-utility/utility studies in RSV The measure of utility ? quality of life - is used in the calculation of a quality adjusted life year to incorporate quantity with quality of life, which is required by decision makers and policy makers for cost-utility evaluations. There have been several cost-utility studies conducted worldwide in RSV. The next section describes the cost-utility research conducted to date.   2.2.1 Cost utility analyses Respiratory syncytial virus (RSV) has been of interest to researchers for many years. In 1998, a therapeutic agent ? palivizumab ? was approved as a prophylactic treatment for RSV (26). Two large scale randomized control trials, the IMPACT-RSV trial and the Feltes trial were pivotal in 23  providing evidence of safety and efficacy necessary to enable formulary approval (26,30). Since this time, there have been several economic evaluations conducted pertaining to the cost-effectiveness and cost-utility of palivizumab in RSV (62).   In Canada, recent estimates of cost approach $1,500 per 100mg vial, or approximately $7,900 per child per season (62,91,92). Palivizumab has been shown to reduce RSV hospitalization, risk of ICU admission, length of hospital stay, and the number of days on supplemental oxygen (26,37,93). The costs associated with prophylaxis using palivizumab have been heavily evaluated. A literature review was conducted to identify the frequency and integrity of utility use in cost-utility evaluations of RSV.   A database search was conducted using the following search terms: respiratory syncytial virus, RSV, cost analysis, cost utility, cost effectiveness, cost benefit, pharmaceutical economics, cost, budget, and economics (Appendix A: Medline searches for literature review). Variations of this search were executed in the following databases: Medline, Embase, Cochrane Database of Systematic Reviews, International Pharmaceutical Abstracts, and Health Technology Assessment. Inclusion criteria specifications included publication in English, any cost utility study whose outcome included quality adjusted life year (QALY), and any study deriving a utility specifically for RSV. After excluding duplicates, 19 articles met the inclusion criteria ? 12 cost-utility analyses, 3 systematic reviews, and 4 poster session references or published abstracts discussing cost-utility analyses (Figure 2.2). Two of the systematic reviews contained additional economic evaluations within them, giving a total of 18 cost-utility analyses (56,94). The 24  systematic reviews were primarily descriptive; only one review contained a meta-analysis identified parameter estimates for risk factor input data for the subsequent economic model. (94)          Three systematic reviews were retrieved involving economic evaluations of palivizumab for the prophylaxis of respiratory syncytial virus. In 2008, the first of the three systematic reviews originated from the UK, and reported 18 primary economic evaluations (56). Of these, only 3 were cost-utility analyses (CUAs) conducted between 2004 and 2007, and 15 were cost-effectiveness or cost-benefit evaluations conducted between 1999 and 2007. Another systematic review in 2011 revealed an additional 8 CUAs (62). This reflects the increasing push for consideration of quality of life measures in policy decision making, which, in health economics, is the (QALY).  Based on review of the articles in the systematic reviews (Table 2.2, 2.3), there was a clear inconsistency between economic evaluations of RSV. Depending on the study population, the incremental cost-effectiveness ratio (ICER), in this case measured as the incremental cost per QALY ranged from CAN$13,029 ? CAN $30,618, and ?6,664 ? ?454,100 (base case ICERs). In sensitivity analyses, the ICERs ranged from CAN$6,618 - $50,701, and ?5,508 - ?3,905,500. 1232 citations retrieved After reviewing paper & references = 19  After title/abstract review & duplicates removed = 113 Systematic reviews = 3 Cost-utility analyses = 16 Figure 2.2 Results of literature search 25  When converted to Canadian dollars, results from other European and USA studies were as high as CAN$1,038,231/QALY, although most of the ICERs were under CAN$50,000 (62).   Study populations typically included ?normal? children, children born prematurely, children with chronic lung disease (CLD) and children with congenital heart disease (CHD). One study included subgroup analyses for cyanotic and acyanotic CHD, as well as all combinations of risk factors (94). The reviews unanimously concluded that the cost of prophylaxis using palivizumab is not worth the potential savings in hospital costs when considering the entire population for which vaccination is recommended. However, in certain high risk groups, the incremental cost per QALY gained as a result of prophylaxis using palivizumab was below the historically used cost threshold (Canada = $50,000 - $75,000/QALY, UK = <?30,000/QALY). (56,95). These reviews maintain that palivizumab is cost effective in high risk infants only, including prematurely born infants, those with CLD, and those with CHD. The variance in ICERs reported in the reviews is purportedly driven primarily by the mortality rate assumed by each CE study. Mortality rates in sensitivity analyses varied between 0.05% - 8.11%. (56) Table 2.1 Systematic reviews of cost utility analyses Author Year Objective # studies identified Study quality assessment Conclusion Smart et al (Canada) 2010 Identify important cost and outcome determinants 21  (11 CUA, 12 CEA) Quality of Health Economic Studies (QHES) *CE in specific groups of high risk infants *CE$ estimates vary between populations, more positive in those at highest risk for RSV hospitalization Wang et al  (UK) 2008 Review effectiveness and CE of palivizumab and examine prognostic factors to determine whether subgroups can be identified 23  (2 RCT, 3 SR, 3 CUA, 15 CEA) None *Clinically effective for reducing risk of serious LRTI and requiring hospitalization in high risk children *Assumed mortality rate was most important driver of ICER in economic evaluations $ CE = cost effective  26  The characteristics of the cost utility reviews found in the literature are summarized in Table 2.1. There were four studies from Canada, three from the USA, four from the UK, and seven from Europe. The sample studied was infants and children in all cases, with special populations including infants and children with CLD, CHD, and preterm infants. Preterm infant groups ranged from 29 to 35 weeks gestational age (wGA). All except two studies found that palivizumab was cost-effective in certain high risk populations, however, the threshold for cost-effectiveness differed among studies (94?105) or, in some cases, was unclear (106?109). As discussed above, the incremental cost per QALY varied between studies. Most studies clearly identified the processes and limitations of their approach to the economic evaluation. Only one study chose not to follow a full lifetime time horizon.  The issue of long term consequence from RSV, such as recurrent wheeze and asthma, is controversial. This particular outcome has been shown to be associated with RSV, although the direction and mechanism of causation is unknown (110?112). Interestingly, although several authors did include asthma or wheeze as an outcome in their model (56,100,102?104,106,107), several authors did not include any implications of long term morbidity due to recurrent wheeze and asthma in their analysis.27  Table 2.2 CUA design summary Author Year Country Population Design Intervention Time horizon Perspective Main outcome Resch et a. (109) 2012 Austria Infants: Premature (33-35wGA), BPD, CHD CUA Palivizumab Lifetime Health care system Societal Cost/ QALY Cost/ LYG Bentley et al (poster session) (101) 2011 UK Infants: CLD, and premature (<29, 29-32, & 33-35wGA) CUA Palivizumab Lifetime NHS Cost/ QALY Neovius et al (102) 2011 Sweden Infants: Preterm (<29wGA) CUA Palivizumab Lifetime Societal Cost/ QALY Wang et al (94) 2011 UK Children: with/without CLD/CHD, acyanotic / cyanotic CHD,  and 64 combinations of risk factors CUA Palivizumab Lifetime UK NHS Cost/ QALY Weiner (abstract) (107) 2011 USA Preterm infants <32wGA, 32-35wGA (>2risk factors), 32-35wGA (<1 risk factor) CUA Palivizumab Unknown Societal Cost / QALY Human et al (abstract) (108) 2010 Canada Children <2 years with hemodynamically significant CHD CUA Palivizumab Unclear Provincial (BC) Cost/ LYG Cost/ QALY Nuijten & Wittenberg (105) 2010 Spain Preterm children <32wGA CUA Palivizumab Lifetime 78.05y Spanish NHS 2006 Cost/ QALY Smart et al (62) 2010 Canada Premature infants (32-35wGA) CUA Palivizumab Lifetime public health care system Cost/ QALY Chirico (103) 2009 Italy Preterm infants (<33wGA, 33-35wGA) with or without BPD CUA Palivizumab Lifetime Italian NHS Cost/ QALY Cost/ LYG Nuijten et al (97) 2009 Germany Hypothetical infants with hemodynamically significant CHD CUA Palivizumab Lifetime  Societal Payer (3rd party) Cost/ QALY Nuijten et al (96) 2009 Netherlands Preterm infants, infants preterm/ BPD, children with CHD CUA Palivizumab RSVIG Lifetime  80.5y,  <80.5y (CHD) Societal Cost/ QALY Cost/ LYG Tam et al (98) 2009 Canada Inuit children of all gestational age CUA Palivizumab Lifetime Payers (publicly funded, base) Societal Cost/ QALY 28   Author Year Country Population Design Intervention Time horizon Perspective Main outcome Lanctot et al (95) 2008 Canada Premature infants born 32-35wGA CUA Palivizumab Lifetime 78.65y  Payer  Societal Cost/ LYG Cost/ QALY Resch (104) 2008 Austria Premature infants </= 35wGA, with BPD, and CHD CUA Palivizumab Lifetime Health insurance (payer?) Societal Cost/ QALY Wang et al (56) 2008 UK Preterm infants without CLD, children with CLD, and children with CHD Review CUA Palivizumab Lifetime 76.8 (preterm or CLD) 76.1 (CHD) UK NHS Societal Cost/ QALY Cost/ LYG Nuijten et al (99) 2007 UK Preterm infants (<35wGA), children with BPD or CHD CUA Palivizumab Lifetime (preterm/BPD= 77.5y; CHD=62.05y) UK NHS (payer?) Societal hospitalizations avoided Cost/ QALY El hassen et al (100) 2006 USA 26-32wGA CUA Palivizumab 1 year (no asthma) 8 years (asthma) Societal Cost/ QALY Yount & Mahle (106) 2004 USA Children with hemodynamically significant CHD CUA Palivizumab Lifetime 76.65y Societal Cost/ QALY Cost/ LYS   29  Table 2.3 CUA results and quality summary Author Dollar Discount rate Sensitivity analysis Items varied: Utility  source Threshold (/QALY) ICER (/QALY)  (base case) ICER (/QALY) discounted Cost-effective? Comments Resch et al 2012 (109) euro ? 2012? 5% Deterministic Discount rate (3-10%) Vials Inpatient cost LOS Utilities Greenough 2004 Unknown 7,631 (BPD) 2,303 (CHD) 7,550-8,090 (<36wGA) 24,654 (BPD) 8,484 (CHD) 24,392- 26,292 (<36wGA) (discounted) Yes *No probabilistic SA *No reference data for asthma disutility Bentley et al 2011  (101) pound ? Unclear Yes  Unknown ?30,000  ?19,168 (CLD) ?18,174 (<29wGA)  ?1,185 (29-32wGA) Yes *Abstract only Neovius et al  2011 (102) SEK 2009 3% Multiway Probabilistic mortality (1.0 - 3%, base = 1.1%) asthma mortality and asthma wheezing Greenough 2004; assumed decrements for hospitalization and ICU 500,000 SEK  34,197 - 8,856,829  Yes  *Local data *Assumption of causal link with asthma/mortality *base case assumes no mortality or asthma effect Wang et al 2011 (94) UK ? 2006 3.50%   Greenough 2004 (hospitalization 0.88, no hospitalization 0.95) CHD = 0.71 30,000  Huge range... All combinations  Yes  *Risk factor estimates unreliable *No interaction assessed with risk factors Weiner  2011 (107) USD Unknown Yes recurrent wheeze (yes/no) Unknown Unknown 3,791 (HR) 22,690 (LR) (without RW) 26,290 (HR) 231,784 (LR) Yes  *Abstract only  Human et al 2010 (108) CAN Yes Unclear  Unclear Unclear  $111,118/LYG $126,621/QALY Unclear *Abstract only Nuijten & Wittenberg 2010 (105) euro ? 2006 3% Univariate   Discount rate (0-6%) Vials Hospitalization costs LOS Utility (0.95, 0.88, 1) Greenough et al 2004 ? 30,000 ?6,142 - 12,814 ?7,475 - 18,872/LYG    Yes  30  Author Dollar Discount rate Sensitivity analysis Items varied: Utility  source Threshold (/QALY) ICER (/QALY)  (base case) ICER (/QALY) discounted Cost-effective? Comments Smart et al 2010 (62) CAD 2010    modified from Lanctot   $20,814 Yes  *Does not include asthma Chirico 2009 (103) Euro (? ) 2007 3% Univariate Threshold Dose Mean cost recurrent wheeze Probability recurrent wheeze (within 24 months) Greenough et al 2004 ? 50,000.00   ?2,731 - 14,937  ?4,332 - 28,417 Yes *Assumption risk wheeze same for all Nuijten et al 2009 (97) euro ? 2005 5% Probabilistic Discount rate (3-10%) Greenough et al 2004 ? 30,000 ?2,675 - 10,116/ LYG ?2,615 - 9,529/QALY   Yes *Conservative assumptions; outside data Nuijten et al 2009 (96) euro ? 1.5-4% Yes Discount rate   Utility (0.95 base case, Range 0.88-1.0) Greenough et al 2004 ? 30,000 ?12,728 - 20,236/QALY ?14,701 - 24,977/LYG   Yes *Cost based on single study *Appear to underestimated costs versus overestimate Tam et al 2009 (98) CAD$ 2007 5% Probabilistic Mortality Discount rates (0,3%) Utilities LOS (hospital & ICU) Greenough et al 2004 adapted for Lanctot et al $50,000  39,435 (all) 152,145 (Iqaluit) 24,750 (rural) 334 ( high risk) 10,190 (Baffin <6mo) 103,235 (Iqaluit <6mo) Yes  Lanctot et al 2008 (95) CAD 2007 5% Probabilistic Mortality (<8.1%) Utility (0.88, 0.95) Discount rate (0%, 3%) Administration cost Vial-sharing Greenough et al 2004 $50,000 - 70,000  $13,029 - 30,618 $18,825 - 44,237/LYG   Yes *No country specific utility *No local mortality data (few studies powered)  31  Author Dollar Discount rate Sensitivity analysis Items varied: Utility  source Threshold (/QALY) ICER (/QALY)  (base case) ICER (/QALY) discounted Cost-effective? Comments Resch 2008  (104) euro ? 2006 3% Yes Discounting (3, 10%) LOS Utility Vial use (10, 20%) Greenough et al 2004 ? 30,000 6429 (preterm) 9881 (BPD) 3115 (CHD) 20,704 (preterm) 31,867 (BPD) 11,390 (CHD) Yes *When including asthma ICERs were lower Wang et al 2008 (56) pound (? ) 2006 3.50% Deterministic Preterm: mortality rates (0.0005, 0.0013, 0.0043, 0.0073, 0.0811)  CLD: mortality rates (0.03, 0.04, 0.05) Hospitalization (0.5-2.5%) (CHD) Greenough et al 2004 (hospitalization 0.88, no hospitalization 0.95) CHD = 0.71 ?30,000  NHS perspective: 454,100 (preterm no CLD) 63,800  (CLD)  79,800  (CHD) Societal: 475,600 (preterm no CLD) 66,900 (CLD) 83,200 (CHD) Yes  *Did not consider wheeze in base case Nuijten et al 2007 (99) pound (? ) 2003 3.50% One-way Probabilistic Utility (0.95, 0.88 - 1) Greenough et al 2004 ? 25,000 ?7,042 ? 16,720 ?8,141 ? 22,826/LYG  (CHD) 2,427 ? 6,664 2,483 ? 7,002/LYG   Yes  El hassen et al 2006  (100) 2002 USD 3% Yes Years (<10) QoL asthma (1 SD) by HUI (symptomatic asthma with and without) $200,000  (Base case)  $675,780 - $1,855,000 Base case: No Yes (26- 27 wGA) *Assumptions: cost and LOS same for all GA *Long term implications changed ICER Yount & Mahle 2004 (106) 2002 US$ 3% Yes Mortality (1-5%) Utility of CHF (0.5-1.0) 1993 Beaver Dam Study adult CHF utility Unclear $100338/LYS $114337/QALY $100338/LYS $114337/QALY No  *Studies do not incorporate outpatient costs *Indirect costs poorly defined *32  2.2.2 Utility studies The choice of utilities used in the CUAs completed to date is of particular interest to this project. According to the cost-utility studies found in this search, a total of 18 CUAs have been done, three of which are only available in abstract form. Upon review, all but five (3 unclear, 2 another source) of these studies used a utility for RSV derived by Greenough et al. (113). This UK study describes the utility of 5 year olds with a history of premature birth, infant diagnosis of Chronic Lung Disease (CLD), and at least one confirmed RSV hospital admission. The utility was determined by a parent proxy, who completed the Health Utilities Index 2/3 health related quality of life questionnaire reflecting the 5 weeks immediately prior to the time the questionnaire was filled out. The median HUI 2 score of children with proven RSV was 0.88 (range 0.16-1.00), and non-RSV-proven children was 0.95 (range 0.03-1.0). The median HUI 3 score was 0.93 (-0.05-1.0) and 0.97 (0.32-1.0) respectively. In many studies, the assigned utility for those requiring hospitalization was 0.88, and those not requiring hospitalization was 0.95. Beyond 16 years of age, all utilities were considered to be 1.0. (95?99,105)  The limitation of using this utility for the calculation of infant QALYs lies in that it is not a utility of the illness itself, rather, it is a utility derived from parents describing the current condition of 5 year old children with a history of RSV as infants, who also have a respiratory co-morbidity (CLD). Not only does it describe a utility of older children and not infants themselves, but it also describes the outcomes of children who have a co-morbid condition. The impact of RSV on those with co-existing illness would not be expected to be the same for those who do not have co-existing illness. This, however, has been the only utility study found conducted to date for respiratory syncytial virus, and thus, has been the primary source used. 33   Two CUAs (100) used utility measures of other diseases as a proxy for utility of RSV. Yount & Mahle (106) was published prior to Greenough?s study, and used a utility representative of adult heart failure it their analysis (0.71). This means, in the absence of a utility specific to CHD, the utility of adult heart failure (derived through TTO from 28 respondents) was assumed to be equivalent to the utility of an infant with CHD and long term effects on the heart (114). This is a significant assumption, as not only does this assume that utility in an adult is the same as utility in a child, it also assumes the two conditions have the same impact on societal preferences. Given that utility research is built upon the theory that each health state is qualitatively different than another, this study could have been improved with the existence of a utility specifically for RSV.   El hassen (100) conducted a cost-utility analysis for a health state which followed the assumption that asthma was a possible long term outcome of RSV, and assigned a utility (derived using the Health Utility Index) reflective of children with and without symptomatic asthma. For those who have RSV and later develop asthma, this may be a valid estimate of the long-term effects of living with asthma, however, this utility only considers the long term effects of RSV, and does not take into consideration the qualitative effect of the acute phase of RSV (for example, a  temporary, acute, 10 day event during infancy). It would make sense to think that there would be additional disutility associated with RSV if the main course of disease is considered.  The Cost Effectiveness Analysis Registry (115) lists 5 different utility values for respiratory syncytial virus (Table 2.4). The utility values range from 0.88-0.97, as shown below. The origin of these values originates from Greenough?s (113) study.  34  Table 2.4 CEA registry utility values Site reference First author Health State Utility Value Utility derivation method 2012-01-09066 Meijboom (116) Children with confirmed respiratory syncytial virus (RSV) infection 0.88 HUI 2/3$   Respiratory syncytial virus (RSV) hospitalization 0.96    General Practitioner-treated respiratory syncytial virus (RSV) bronchiolitis 0.99  2011-01-07910 Resch (109) Children with respiratory syncytial virus (RSV) infection not requiring hospitalization 0.95 HUI 2/3$   Children with respiratory syncytial virus (RSV) infection requiring hospitalization 0.88  2009-01-05709 Nuijten (96) No respiratory syncytial virus hospitalization 0.93 HUI 2/3$   Respiratory syncytial virus hospitalization 0.93    Non-respiratory syncytial virus hospitalization 0.95    Respiratory syncytial virus hospitalization 0.88    Non-respiratory syncytial virus proven 0.97    Respiratory syncytial virus proven 0.93    Non-respiratory syncytial virus proven children 0.95    Respiratory syncytial virus proven children 0.88  2009-01-04869 Chirico (103) Respiratory Syncytial virus hospitalization 0.88 HUI 2/3$   No respiratory syncytial virus hospitalization 0.95  2008-01-04422 Lanctot (95) infants without previous respiratory syncytial virus infection 0.95 HUI 2/3$   infants with previous respiratory syncytial virus infection 0.88  2008-01-03706 Resch (104) High risk children with RSV infection but no admission 0.93 HUI 2/3$   High risk children hospitalized with RSV 0.88    High risk children without RSV 0.95  2007-01-03544 Nuijten (99) RSV-Infected children 0.88 HUI 2/3$   High risk children for RSV infection 0.95  *(modified from Cost effectiveness Analysis Registry) $HUI 2/3 utility derived from study by Greenough et al (2004) of 5 year old children with history of RSV as infants  The Cost Effectiveness Analysis Registry reports utilities from within journal publications that have been used for RSV health states. The utilities listed for different severities of RSV infection are between 0.608 and 0.99, ranging from mild upper respiratory infections (i.e. general practitioner treated RSV, utility = 0.99) to children with RSV requiring hospitalization (utility=0.88). The registry does not include values specific to the most severe case of RSV, in which infants require intensive care unit admission and intubation.  The utilities were pulled from seven different studies, and whose limitations have been discussed previously. Review of this registry list highlights how the same utility values have been applied differently by different authors to the range of severities of RSV. There is a need to be able to measure the RSV-35  attributable disutility given that the morbidity is often associated with the presence of co-morbidity.  2.3 Knowledge gap #1 Cost-utility studies have been used for economic evaluations of palivizumab in RSV. Several different utility values are being used to conduct cost-utility evaluations for RSV in different populations of young children, however, none of the utility estimates used are specific to RSV in infants. The next section explores the methodology related to deriving health state utility values for RSV in infants.  2.4 Methodology ? who?s life, whose perspective, and how do we approach temporary health states? Utility theory and derivation methods are typically founded in chronic health conditions, and are based on rational decision making. In the context of RSV, however, several significant issues present themselves. Not only does utility elicitation of an infant health state require a proxy, RSV is a short-term health state with potential long-term outcomes. Framing the elicitation question becomes extremely challenging. The next section reviews the literature on how to approach measurement of temporary health states in infants requiring proxy response.  2.4.1 Challenges in deriving utilities in infants Determining health utility values for infants is challenging, and many authors have explored the problems related to deriving utilities in children (81,82,117). Infants do not have the cognitive ability to communicate their preferences, and so we rely on proxies - usually parents or 36  caregivers - to complete the valuation or measurement tasks needed to elicit health utilities. The use of a parent presents its own issues, such as the risk of an interaction between the parent's utility and that of the child's, as the effect of the child?s illness on the parent?s own quality of life can influence their responses to quality of life questionnaires or valuation tasks (i.e. TTO or SG) (83).  It is important to define perspective when designing valuation tasks. Questions are frequently framed from the ?consumer?s? perspective, or whoever is experiencing or paying for the product or program. In the case of RSV, however, the consumer is not clear ? is it the parent, or the child, or the taxpayer, or a combination of the three? (118). Should the elicitation task be from the perspective of the adult or the child? If it is the child, how do we approach it? Valuation of infant health states present the unique challenge of respondents also not being able to remember what it was like to be an infant, nor ?to imagine? what it would be like to be an infant (81).  The standard for health utilities requires a societal perspective (119). When choosing a proxy for infant health states, simply using parents may not give a representative "societal" perspective. Although it may be argued that parents, professionals, populations with similar health conditions, and those with expertise/experience in the disease may be most reliable, at other times it may be relevant to seek the preferences of people indirectly affected by the decision, or of those who are paying for the potential intervention. When choosing a proxy in the context of infant utilities, it is important to consider potential differing preferences among groups. For example, parents with children may have different preferences than members of the lay public with no children. In previous studies, interestingly, although some subgroups (e.g. respondents experienced in the 37  disease versus respondents who were not) show differences (120?122), in some cases relatively small differences have been found between some groups that are commonly considered to have baseline differences (i.e. SES or income level). (68)  Methodologically, there is currently no established standard for eliciting health utilities for children. Health status instruments, such as the HUI-3 and SF-36, have not been validated in children less than 5 years of age (83). Direct valuation using TTO or standard gamble is suggested as the preferred method as it is possible to include age specific descriptions of important attributes. In pilot studies using direct valuation, researchers have found that parents are often unwilling to gamble their child?s life in order to avoid a temporary condition (123). Similarly, using the time trade-off technique, studies have also found that parents are sometimes unwilling to trade years of their child?s hypothetical life expectancy to avoid temporary health states (124?126). Nonetheless, these authors have had success is asking parents, and the public, to trade years of their OWN life to avoid a condition in their hypothetical child.   2.4.2 Challenges in deriving utilities in temporary health states Respiratory syncytial virus presents a unique challenge as it involves eliciting preferences for a special population ? infants - as well as capturing the temporary nature of the disease. Traditional direct valuation methods of eliciting utilities pose problems when applied to temporary conditions. Although generally utility theory is considered robust, its application to temporary health states and to children has been highly criticized (81,83,127,128). There is no current validated and supported recommendation for methods to elicit health utilities for infants nor an ideal method for deriving utilities for temporary health conditions. 38   One of the principles of utility theory that is violated in temporary health states is that of constant proportional trade off (CPTO), an issue central to time trade off and application to QALYs (72). CPTO implies that the proportion of time traded to retain a certain quality of life is independent of the number of years of life remaining. Conversely, Bala (128) reports that with a shorter illness, the focus turns to death, and respondents are less willing to trade any time in order to improve quality of life. In other words, although the experience of the disease may negatively affect quality of life, because it is temporary, people are not willing to give up years of life in order to avoid the health state. It is reasonable therefore to foresee that temporary health states, if measured using a TTO, may have a very low impact on a Quality Adjusted Life Year (QALY), as respondents may not be willing to trade time from their life for a very brief period of better quality of life. RSV, interestingly, has potential long-term outcomes such as asthma and wheeze. This presents the challenge of needing to consider both temporary and long-term aspects in the elicitation task. The value of the instantaneous health state becomes no longer simply a valuation of only the acute period of RSV, but also values the potential of long term positive impacts.  In 2009, Wright et al. conducted a systematic review on the subject of temporary health state methodology (127). Direct health status measurement techniques, the standard gamble (SG) and time trade-off (TTO), are widely accepted methods of utility derivation. Traditionally used for chronic conditions, they have been applied to temporary conditions, although they have not been evaluated for this purpose. In response, modified TTO techniques such as a Waiting Trade-off (WTO), a Sleep Trade-off (STO), and a chained approach have been developed (123).  39  The STO asks respondents how much sleep they would be willing to give up over a defined short period of time to avoid a particularly health state. This method may be feasible in the context of RSV and pediatric populations where caretakers can identify with sleep loss due to caring for children. However, it echoes the conflict around confounding, and the difficulty identifying whether the utility derived is that of the parent or that of the child?s. The sleep trade-off, however, is anchored between the temporary sleep-deprived state (0) and current or perfect health (1), which would require further scaling to return to the 0-1 (death-perfect health) scale of the traditional TTO. (129,130)  The WTO was developed for the valuation of preference surrounding diagnostic testing. The traditional time trade-off presents similar unrealistic scenarios as short term health states ,as it would require the time spent undergoing a test to be varied (131). Similarly to RSV, it also involves weighing the risk of potential outcomes, in this case related to test results. The WTO is used in conjunction with a 0-10 rating scale (death to full health), on which a respondent rates the health state they are in (anchor state). They then are asked how many days in the anchor state they would be willing to endure to avoid the testing experience (essentially a two-step chained method). A different version of the WTO asks respondents to trade off time waiting for an ?ideal? test (the same test but with no side effects) or a more uncomfortable test but immediately. Although there are similarities to some of the challenges in pediatric temporary health state valuation, the WTO was originally designed for diagnostic testing, and the concept of ?waiting? is less intuitive for a health state than it is for diagnostic testing(127). In our scenario, this method does not provide a way to incorporate the long term risk of wheeze associated with RSV.  40  Because temporary health states are often considered to be less serious (since respondents recover back to their normal health), respondents are less willing to trade off death in order to avoid them (128). The chained approach allows a ?worst case? or case worse than the health state to be valued against death. The health state of interest is then traded off in relation to this previously valued ?worst? health state instead of death. This allows an indirect comparison with death that may be easier for the respondent to comprehend, and prevents additional complications associated with varying life expectancy (132). However, the chained approach is subject to framing effects due to the possible worse-than-death health state (133).It can also be influenced by the intermediate, more severe, health state. The two-step process also introduces increased variability (due to the requirement of two questions) and potential imprecision, and is more labor intensive. Previous studies have shown inconsistency between conventional methods and chained methods (129,134?136). (127).   As the review conducted by Wright et al (127) was published in 2009, a literature search was conducted from 2008 to 2012 to assess how researchers have approached time trade off elicitation for temporary health states in addition to the traditional TTO, STO, WTO, and chained trade off described previously. Using the same search terms as in the Wright et al paper(see Appendix A: Medline searches for literature review), abstracts titles and abstracts were reviewed for inclusion. Papers were considered if they used a TTO direct elicitation method (time trade off or variant thereof), were in English, used a proxy respondent, and evaluated a temporary, non-chronic health state. Results of this search are shown in Table 2.5. The objective of the search was to assess how, when using a proxy respondent, from whose perspective the elicitation task was framed (whose life was traded), the time frame used, and whether it was 41  framed as a temporary or chronic state. The search did not result in further approaches to trade off, however, it did provide a broad sample of the form of time trade-off questions have been used in recent years.  42  Table 2.5 Direct elicitation utility studies for temporary health states valued by proxy respondents  Valuation task Who's life Time frame Framed as temporary or chronic state^ Health state Primary outcome Utility calculation Fennessy 2011 TTO own life (for own outcome) years from next 10 (followed by death) chronic uterine fibroids utility WTO (QALweeks) U = indifference/10 Prosser 2011 TTO traded own life to prevent condition in child years from end of remaining life started with 10 days (if yes, if no) Influenza time traded n/a Basu 2010 TTO (chained) traded own life to prevent spillover effects any years from life expectancy chronic prostate cancer utility  Eldabe 2010 TTO own life (for own outcome) time frame 5 years (r/t life expectancy) temporary (due to mortality) chronic pain (MP, NMP)   Carroll 2009 TTO traded child's life to prevent condition in child unclear how time was framed temporary and chronic health states hospitalization 10 day  ICU 10 day mild/mod/severe persistent asthma utility u = x/t t = life expectancy of child for both long and temp states Biljenga 2009 TTO traded own life to prevent condition in child AND self how much time of life in full health - maximum of 10 years, not stated 'where' in life. temporary (1 year) health states related to pregnancy/birth/neonatal period  utility u = x/t where t = 10 years Happich 2009 TTO own life (for own outcome) years from life expectancy temporary tinnitus   Lieu 2009 TTO traded own life (for own outcome) ?portion of life? (did not state end of life) to avoid condition now  temporary herpes zoster   Lieu 2008 TTO traded own life (for own outcome) life expectancy; unclear 'when' life was traded temporary herpes zoster utility disutility = time traded/duration of health state (both in days) Prosser 2005 TTO traded own life to prevent condition in child years from end of remaining life temporary influenza and vaccine adverse events time traded Note: <50% would trade for mild health states Lee 2005 TTO traded own life to prevent condition in self; then in child years from end of life using life expectancy long term (lifetime) and short term (8 weeks) Pertussis: local reaction, systemic reaction, mild cough, severe cough, pneumonia utility Disutility = time traded/LE of adult or infant  43   Valuation task Who's life Time frame Framed as temporary or chronic state^ Health state Primary outcome Utility calculation Prosser 2004 TTO traded own life to prevent condition in child (state this incorporates both child and adult preferences) remaining life  pneumococcal vaccine preventable states (otitis media, pneumonia, bacteremia, meningitis) time traded When applied to an economic model, 7 days traded gave 0.02 qalys Gu 2009 TTO traded own life (for own outcome) AND friend/patient perspective& years from max15 temporary post-operative hip outcomes; chained    44  2.4.3 Further modifications of the time trade-off Basu et al. (124) used a new approach for valuing utility in prostate cancer. Basu argues that the spillover effect of a condition, the impact on the quality of life of a close family member, can very closely approximate the utility of the health condition in the person with the health state themselves. Basu et al. tested this hypothesis using couples in which one partner had been diagnosed with prostate cancer. The cancer-free partner valued the health state based solely on the impact that this health state had on them, not including the impact it had on the patient. The spillover impact was shown to closely approximate the value of the patient gave for impact of the cancer on their own health-related quality of life.  Another modification to the TTO is to vary the time horizon of the time trade-off question. The time horizon can be limited to the length of the temporary health state, or correspond to the respondent?s life expectancy (117,126,137,138). Limiting the length of the health state avoids time preference issues; however, it does not allow the health state to be valued as worse than dead. Allowing the time horizon to be equal to the life expectancy introduces variation between respondents, and is clearly influenced by the length of time that a respondent has left in their life. This can be adjusted for similarly to chronic health states by discounting (127).  Lee et al (126) derived utilities from adult proxies for adolescent, child and infant pertussis health states using a modified time trade-off.  They asked an adult respondent to trade their own life in order to prevent a hypothetical condition in themself, and compared this to the same respondent trading years of their own life to prevent a hypothetical condition in their hypothetical infant. They found that respondents consistently traded more for infant versus 45  adolescent or adult health states. This is echoed in another study which found that respondents value child health states more (greater disutility) than adults (138). The benefit of the second approach (trading your own life to avoid a condition in your child) has the benefit of preventing the proxy from trading years from an infant?s hypothetical life expectancy. This has been a successful approach to eliciting trade-offs(139) in groups who have previously shown refusal to trade any years from their child?s life, or have refused to gamble with any level of risk death in their infant (123,125).  2.5 Knowledge gap #2 Several different approaches to short-term health state valuation have been explored, as well as alternative approaches to preference elicitation (i.e. best-worst scaling), however, there is no consensus on the best approach to measuring infant utility in temporary health states. There is a clear need for methodological research in order to derive utilities for conditions such as RSV. The development of a utility elicitation tool requires information on which attributes are important in the context of illness of RSV. A review of the literature for previous qualitative studies is summarized in the next section.  2.6 Previous qualitative studies Few qualitative studies have been conducted describing the experience of families who have had an infant or child hospitalized with RSV(140,141), although hospitalization of a family member ? child, spouse, or parent ? has an impact on the rest of the family members (124). Nearly all parents who have had a child with severe illness experienced some level of distress; one of the common factors being the inability to console their child while they were getting treatment 46  (140). Leidy cites that there is reason to believe that impacts on daily routine and quality of life extends past discharge. Liedy (141).confirmed that child hospitalizations evoke concerns in parents such as financial concerns, parenting concerns, concerns about the child?s experience, and the child?s future. Functional health status reports by parents were less in post-RSV children than healthy children in the 60 days after discharge, an indicator of the impact of the illness past returning home from the hospital (141).   Two studies were found evaluating quality of life in children related to RSV. They have reported child health and functional status or family function using the visual analogue scale (health on a scale of 1-100) (141), and measures of parental and infant distress (140). Robbins et al. (140) reported anxiety related to perceived child distress, child pain and the inability to console their child, parental distress while watching procedures, and anxiety watching their child?s behaviour responses (turning away, crying). Both of these studies included a proxy measure, in which parents rated their child?s health.  No studies were found asking open ended questions about the experience of children and parents during an illness with RSV, and the aspects of the disease that are most concerning to parents.   Although not specific to RSV, qualitative studies have been completed investigating characteristics and impacts on parental and child quality of life related to child illness and hospitalization. Lower respiratory tract infections have been characterized by parents with difficulty feeding (142), apnea, and wheeze (93). Parents describe stressors of the hospitalization experience as the role changes, the sights and sounds of the unit, and changes in infant behaviour and appearance (143). Their concerns upon admission include survival of the child, the 47  possibility of chronic/long term impairment, the child?s diagnosis, and the amount of pain experienced by the child (144). Parents also report challenges with missing work (145), the length of time their child was in hospital (1), the processes associated with isolation procedures in which visitors in hospital are required to wear personal protective gowns, masks, and gloves to prevent the spread of infection  and reliance on others for child care (145).   Although not specifically RSV, in a recent UK study, parents who had children admitted to hospital with severe respiratory infections were interviewed. The parents identified ?missed opportunities for timely treatment? (p.154) (146), in which parents returned to health services multiple times before their child eventually requiring hospitalization. Parents described problems assessing the severity of their child?s illness and conveyed beliefs about accessing services (fear of being seen as neurotic, fear won?t be taken seriously, powerless to challenge clinical authority) (146).   2.7 Knowledge gap #3 Information is rare, but present, on health related quality of life of families with infants hospitalized with RSV. Although qualitative research exists on aspects of illness that are similar to RSV, such as hospitalization and medical tests, no qualitative research on RSV illness in infants was found. This information is essential for the design of appropriate vignettes for time trade-off tasks.   48  Chapter 3.  Rationale for thesis research  3.1 Knowledge gaps There were significant knowledge gaps identified from the literature surrounding health utility derivation in pediatric temporary health states. There is a need for economic evaluation surrounding RSV treatments, which requires a QALY calculation. Despite this, there is no existing RSV-specific health state utility value. In addition, there is no consensus on the best methodological approach to measuring health utility in an infant population temporary health state, particularly considering the potential for long term consequences. Lastly, there is no qualitative evidence of which aspects of RSV are most important to families with an infant with RSV.  3.2 Justification for thesis research The use of health utilities as a quality of life measure has routinely been used in economic evaluation to support decision making, and is required by government and policy decision makers. Given the burden on health care services for this age group, and the room for development of therapies to treat RSV, and the high cost of the available drugs, the need to derive a utility for this particular age group with this health condition is integral to spending health care dollars in a manner informed by the social preferences of the paying population.   Having been historically used in chronic disease, health utilities for short term health conditions present multiple challenges to the theories and practical application of traditional elicitation methods. RSV provides additional challenges in that it is not only a short term health state, but 49  has potential long term negative outcomes such as wheeze. Given that it is an infant health condition, valuation of this health state also requires the use of a proxy respondent, adding another level of complexity to the elicitation task. Whether the question should be asked from the adult or child?s perspective, whose life years should be traded off, and interaction between the utility of the respondent and the utility of the child all need to be considered. There are currently no established standards for deriving utilities in infants, leaving decisions for these populations vulnerable to poorly informed evaluations as they lack an appropriate utility measure. This highlights the need for development of appropriate and novel tools for measuring infant utility and quality of life.  Given that the unique nature of temporary pediatric health states, this leaves room for exploration of new approaches to valuation, and exploration of adaptations to traditional utility elicitation methods. In a disease in which essentially no preference work exists, and for which treatment options are limited, it is important to be able to tailor interventions to the aspects of the disease which patients and families value most. A recent adaptation of discrete choice experiment ? Best-Worst Scaling (BWS) ? provides another lens through which one can gain insight into preferences regarding RSV. The combination of these two methods ? new approaches to valuation and best-worst scaling experiments in the RSV context - can provide the foundation for development of future valuation methods for temporary health states in infants.  3.3 Expected implications of study It is anticipated that the results of this research will be of significant interest to policy makers, government decision makers, pharmaceutical research, and health care practitioners. 50  Understanding the attitudes toward short term, acute illness in children facilitates the development of appropriate and meaningful policies, treatments, and patient-professional communication in the health care setting.  The results of the study will set the impact on quality of life of RSV in perspective relative to other health states on the utility scale of 0 to 1, allowing comparison using a consistent metric.. Although the addition of an accurate utility to the economic research may not change the existing decisions regarding palivizumab, it has significant potential to impact future decisions on newer treatments and drugs.  The application of new methodological approaches to utility derivation will demonstrate the strengths and limitations of different elicitation methods in this context, which has not been done previously. Significant time was placed on the design of the elicitation task, and the following chapter describes the literature base used for the development of health utilities and preferences for this study in infants with RSV.   3.4 Thesis objectives The objective of this thesis was to derive health utilities for RSV and relative preferences for attributes associated with RSV infection in infants. This thesis explores methodology in preference elicitation for both infants who require proxy respondents, and for temporary health states. A detailed methodological assessment was conducted surrounding proxy response in temporary health states and possible approaches to health state valuation in the context of RSV. Using attributes validated through qualitative assessment and expert opinion, three approaches to 51  preference elicitation were explored for infants in a temporary health state: a time trade-off experiment using a proxy perspective, a time trade-off experiment using self-perspective, and a best-worst scaling experiment. The product of these approaches demonstrates the complexity of utility assessment in infant temporary health states, and outlines the feasibility and advantages of the methods used.  3.5 Thesis overview The work in this thesis emphasizes the questions and issues surrounding the derivation of the health utilities in non-chronic conditions and in health states in which proxy respondents are necessary. This first chapter introduces Respiratory Syncytial Virus - the health state of interest, its disease course and management, and the directions for research.    The second chapter considers key epidemiological attributes of RSV with a detailed discussion of the associated mortality risk. It then examines previously conducted cost-utility research and the strengths and limitations of the utilities used in those studies. Lastly, the methodological literature is reviewed around preference elicitation in infants and temporary health states, identifying which approaches have been used, and their potential application in the context of RSV.  In chapter 3, through qualitative research, I explored aspects of a child?s RSV hospitalization that were important to parents. Those aspects that were most important are highlighted, and linked to areas of strength or potential areas for development in the clinical setting, in the community, and in future research. 52   In chapter 4, using the information gained through the focus groups, I assessed the preferences surrounding RSV, and conducted a time trade-off experiment from two perspectives. Specifically, this resulted in the derivation of health utilities for a 6 month old infant for different levels of RSV severity - mild, moderate, severe, and respiratory failure.  In chapter 5, I estimated the relative importance of attributes describing the experience with RSV using a Best-Worst Scaling (BWS) choice experiment. This derives relative preferences for location of care, breathing at worst point, oxygen support, tubes, and risk of long term wheeze. In chapter 6, the collection of works is evaluated in terms of the strengths and limitations of the methodology used for preference elicitation in infant temporary health states, and future directions for research in this area.    53  Chapter 4.  Parent?s experience of their child?s illness with RSV  4.1 Background RSV is a respiratory illness common in infants and children under 2 years old. Although RSV can occur in adults, the vast majority of the health care burden is in young children; 70% of children under the age of 1 year have had RSV, and virtually all children have been infected at least once by the age of 3 (6) . RSV can range from a mild upper respiratory tract infection (URTI) similar to the common cold (e.g. cough, rhinitis, low grade fever]), to a lower respiratory tract infection (LRTI) (difficulty feeding, dyspnea, accessory muscle use, wheeze) and at worst, respiratory failure (93). The incidence of severe illness is high, with 2.5% of children with LRTI requiring hospitalization (14,24). In the United States, RSV is a leading cause of infant hospitalization (1). In Canada, it has been estimated that up to 40% of children less than 2 years old hospitalized per year with LRTI is due to RSV (20).   The measurement of quality of life of infants and small children requires the use of a proxy respondent. However, it is well documented that the child health experience is different from an adult (147), and that the impact of illness of any family member affects the rest of the family unit (124,143,148). Siblings report reactions to their sibling being hospitalized that their parents did not pick up on [Craft 1985]. Parents? stress levels increase with a child?s illness (148), and the impact on family members? quality of life (i.e. anxiety, work) is shown to possibly extend up to several weeks after critical childhood illness (149). For this reason, reliance on a proxy is controversial due to the interaction between the quality of life of the parent and that of the child 54  (150?152). As there is no immediate solution to this in children too young to talk, researchers must acknowledge the challenges when interpreting proxy responses (152,153).  Research is needed in evaluation of pediatric health related quality of life in order to improve the quality of care in this population (147,154).The objective of this study was to assess, qualitatively, the experience of parents who have had a child admitted to the hospital with RSV. The primary outcome of interest was to identify principal aspects of the disease that are most important to parents during the time their child had RSV. Secondary outcomes of interest included exploring the smaller sub-themes of these attributes and reporting their relationship to the parents? experience. The identification of these attributes was then used to inform a preference based survey that will determine the health utility of RSV.  4.2 Methods  4.2.1 Recruitment and ethics 400 parents of children admitted to British Columbia Children?s Hospital within 5 years of the study (to decrease recall bias) were invited by mail to participate in focus group sessions. 12 to 20 participants was the targeted number to interview. Parents were eligible to participate if they were over the age of 18, and able to communicate (written and spoken) in English. Participants agreed to participate in an interview (either individually or with a larger group), and were offered a CAN$25.00 honorarium to help cover costs such as travel, parking, and time. Interviews were no more than 90 minutes long, and located in a variety of settings. All sessions were audio recorded, transcribed by the same trained transcriptionist, and reviewed by a primary researcher 55  (LR). Interviews were conducted by the same researcher, ensuring consistency between interviews. The study was overseen by an experienced qualitative researcher, and ethics approval was obtained from the British Columbia Children?s Hospital Research Ethics Board.  4.2.2 Design and data collection Interview guidelines were constructed aimed to explore both child and adult related experiences. Open ended questions were developed from both the parent and child (proxy) perspective (145,155). Topics included (from the parent perspective) emotional and physical feelings, what was most important to parents, impact on parents? quality of life, what bothered parents to see in their child, and (from the child perspective) child coping challenges, relief mechanisms for child, and how the parent assessed how the child was feeling. Probing questions were outlined, such as the presence of financial concerns and concern about child?s future, however, dynamic, unstructured conversation was encouraged (refer to Appendix B: Focus group Guidelines)  4.2.3 Analysis Transcriptions were analyzed using NVivo software (version 9.0). The data were coded by one person (LR), and reviewed by an experienced qualitative researcher (NH). Parents? thoughts and impressions were explored related to major themes around parents? experience (e.g. emotions, daily life effects), and the parents? perceptions of the child?s experience (e.g. challenges, coping mechanisms). Themes were defined according to child and adult perspective, and the frequency they were referenced ? consistent (>15 references), common (10-15 references), and occasional (<10 references). Frequency of reference was considered more reliable than the frequency of interview occurrence it is less likely to be influenced by probing questions when referenced more 56  than 8 times (the total number of interviews). Additional major themes and subthemes were added as they emerged. The coding scheme can be found in Appendix C.   4.3 Results  4.3.1 Sample demographics 12 parents participated in the interviews; 11 participants were female, and 1 participant was male. Participants had children who were currently living with co-morbidities and children who were considered ?otherwise healthy? (the exact ratio of these was not known, as health history was not collected intentionally, only revealed through conversation).  4.3.2 Descriptive The data resulted in 18 major themes, and an additional 54 subthemes. The complete list of themes and subthemes can be found in Appendix C. Table 4.1 provides a frequency summary of sources and references for the most commonly referenced themes. Of all the themes, parents referenced fear/nervousness/worry, breathing, and going home the most. In total, only 11 themes arose in 6 or more interviews (of a total of 8 interviews). The themes with 6 or more parents referencing the same topic were:   57  Table 4.1 Important themes by frequency (6 or more interview) Major theme Subtheme # of sources  # of references Probed question? After hospitalization Going home 6 19  Child symptoms Breathing Feeding changes 8 8 24 14  Opinion of hospital Negative 6 12  Parent feelings      Anger Fear, nervousness, worry 6 7 5 28  Parental daily life impacts Childcare Siblings? reactions Support     Food     Hygiene 6 7 6 3 2 10 10 13 6 2 Often Yes Parents? hospital experience Isolation 6 8 Yes Presence  7 9  Presence of chronic illness  7 11 Yes Reason for seeking medical attention  6 8 Yes 4.3.3 Key parental experiences with RSV  Consistent and common themes are described below, in order of reference frequency.   Consistent themes Parent feelings ? anxiety (parent perspective). Parents recalled a variety of feelings and emotions associated with their memory of their child?s illness with RSV, and spoke frequently about these emotions. The absolute most frequently mentioned memories of parents (in all interviews except one) were feelings of fear, nervousness and worry. Parents also recalled feelings of anger, and a varied amount of guilt, relief, frustration, confusion, sadness, and stress.   After hospitalization ? going home (parent perspective). As parents recalled the time following hospital discharge, they also consistently expressed concern about future health consequences as a result of their child?s illness with RSV. One parent was concerned about the effect of RSV in combination with their child?s chronic illness, and several parents were concerned about long term consequences of RSV. This included both physical health ? ?she may have asthma for the rest of her life?, ?will this mean anything later on??, ?there could be the possibility of something 58  else?, and the psychological effect - ?I wondered what effect it would have on him? being poked and prodded? so much in the first few days?.   In the transition to home, many parents reported fears, changes in behavior, and long term effects on their children. Some parents reported a trend of ill-health in their child in the subsequent months and years. One said ?he?s always been our frail little boy... whenever anyone gets sick, it?s him?had to take him to the ER twice for asthma? he has eczema? allergies? ear infections? ?, another says ?she?s never been (back to normal)?, ?she was colicky after that?, and ?I have a hospital bag packed on the dryer still because that?s how often (we go)?. Most frequently, parents expressed worry about recurrence as they returned home or when their child became ill again ? saying ?constantly on watch and listening to them breathing and lifting their shirt up? state of alarm every time they get sick?, ?I didn?t want to go out anywhere... was she going to relapse??, ?I was worried? always watching his breathing?, and ?I was scared something else was going to happen?.  Although there were a few parents who felt that they adapted well to the post-hospital situation, two people verbalized concerns about the discharge medications prescribed, confusion about how to administer the medications, uncertainty about what the long term effects might be, and articulated the importance of being the one to administer the medications themselves. As one participant said, the ?most important thing at the time was that I... give him the medicine myself?. Another pointed out that ?in hindsight, I was very concerned about putting her on medication ? is she going to be on this for the rest of her life??.   59  Child symptoms ? breathing (child perspective). The most consistently emphasized symptom that was described was changes in the infant?s breathing. Some parents used sounds instead of words, making gasping sounds with rapid, shallow breathing. When illustrating verbally, breathing was said to be ?very short?, ?noisy?, ?snerty and gurgley? and ?rattle, rattle, rattle?. Infants took ?great gulping breaths?, with ?little ribs squeezing?. Parents reported congestion, wheeze, coughing, grunting, and ?struggling? to breathe. Wheeze was described repeatedly just as ?? and the wheeze. The wheeze!?.  Common themes Support by partners and family members was mentioned frequently. Parents reported family members being present upon admission to the hospital, and staying for periods with the child in hospital. They indicated that support was provided by the constant presence of nurses in the ICU and by doctors checking in. Lack of support was shown in statements such as ?my husband was still (overseas)... I?m the only one here... it was just me?, ?I would clean my underwear in the sink because I had no one. My husband wasn?t able to get off work. And he was caring for my daughter?, and ?it was stressful. I don?t really have family in town?. Some parents commented that family members brought in food and clothing, as they found it difficult to find food, and time to obtain clean clothes and toiletries while at the hospital.  Parents also reported a change in feeding routine, the presence of fever, changes in sleeping patterns, and lethargy as symptoms of RSV. Children had decreased appetites, would refuse to feed altogether, vomit, or would be unable to feed due to breathing problems and congestion. Fever was less commonly reported, with only two parents recalling this as a primary symptom. 60  Sleeping changes included not waking to feed, sleeping a lot and being constantly awoken with coughing fits - ?he?d cough for 10 minutes and then he?d sleep for 10 minutes? all night long?. Participants reported lethargy separate from sleeping changes, using the words ?weakness?, ?sleepy?, and reporting an inability to wake the child.  Parents were asked how their child?s illness with RSV affected their own quality of life. Parents reported that finding people to take care of their other children was stressful and ?tricky?. Most families relied on either extended family (e.g. the infant?s grandparents, aunt), to provide the last minute child care or the father took over the extra duties for the children at home while the mother stayed with the child in hospital. In the latter case, one person describes it as ?divide and conquer ? he took our son and I took our daughter...?.  In order to meet these needs, most parents felt the need to take time off work.  Occasional themes Many participants spoke about ?presence?, and being with their infant while they were sick. Statements describe this as ?I can?t imagine not being there?, ?you just want to be within reach?, and ?his mother was with him the whole time, usually holding him?. Parents also found it difficult when they were unable to be as present as they normally would be. Many parents reported conflicting emotions of not being able to hold their child. The parent of an intubated child expressed the desire to be holding her child and ?I wasn?t able to hold her very well?, while another parent voiced fear as a barrier - ?I was petrified to hold him?.  61  The participants clearly demonstrated the existence of an un-named, instinctual ability to ?just know? something was wrong with their child. For example, parents said they needed to ?trust their gut?, that they repeated to themselves at the time of the illness over and over ?I know I?m right?, that ?something is different?, that ?something was wrong?, and ?I could just tell?. One parent describes their child as not being present ? ?he?s not there? he?s looking at you. He?s awake, but he is not present. My baby is not present?.  During the hospital stay itself, frequent assessment by health care professionals was found to be disruptive, annoying, and repetitious. The stay in a ward (versus the ICU) materialized in feelings of being more alone, as well as fear and stress. For example, ?I didn?t realize specifically that you?re the only person around to watch [the child]?, ?I wish they had safer-feeling beds so you could feel like you could leave the room every once in a while?, ?bit of a shock coming from the ICU, where you have somebody all the time, to that other ward where you see somebody every 4 hours?, and ?upstairs was a bit scarier because you?re sort of on your own?.   Few people commented on the relationship with other patients and families present, but one in particular, the parent of a child who was intubated and in ICU, put into words a feeling of normalcy that occurred. She said ?when you?re with all these other families... another couple going through the same thing... it started to feel normal... it was always, how is your baby doing today? Kind of... putting each other up, which was nice?.  62  Specific symptoms and aspects of treatment were identified as being particularly hard to observe in their child: breathing, the IV insertion, monitors, suction, tubes, and x-ray. Parents found it scary to watch their child breathing, describing the parts that make it worst ? ?if you watch the child breathe, their whole inside of their stomach almost gets sucked in?, ?she looked like she?s giving up... to see them struggle like that, trying to get air, it?s just the hardest things to watch?, and ?it was really hard to see that, to hear the rattling?.   The IV insertion was difficult for parents to watch when repeated attempts were necessary. Several parents found the monitors particularly stressful. One parent recalls ?I remember thinking... this teeny little baby with all the wires monitoring his breathing... then just staring at the numbers all day long and just making myself nuts?. Another speaks about ?always worrying about the beep, beep every time the oxygen dipped below?, and how, ?...you?re sitting staring at it, and you know what the number is, it?s 70-something. You can?t remember, but there?s a certain number and when it goes below there it beeps, and you?re just looking, going, how long is this going to beep? And do we need to put the oxygen back on??. More parents reported the x-rays as unpleasant, likening it to torture in which ?you?re watching them across the room, with their arms strapped in... terrified and can?t move... someone is doing something to your baby and you can?t help. You have to sit there and watch, it?s worse than if they did it to you?, ?it was heart-breaking?. Similarly, the tubes were difficult for parents to see, as they talked about how they could not hold the infant properly, how it made one parent not want to touch the infant so as not to interrupt ?so many tubes?. One parent remembers ?he had wires. I just thought what a sad sight to see... a baby with all his wires... I didn?t know what they were for... it was hard to see a baby with a bunch of strings coming from him?. A father 63  remembers ?the first time I saw him sitting in bed... all hooked up... and thinking oh God. This is serious?. The use of suction was least frequently mentioned, with only two parents mentioning that they found it hard to watch their child being suctioned.   Parents were asked what they felt was particularly hard for their child to cope with during their illness with RSV, giving rise to several sub-themes. A couple of parents felt that, during the stay in hospital, the frequent assessments and interference by health care practitioners were upsetting for their child ? ?when you interrupt his sleep... even just (to) listen to his heart?. Some parents felt that the hospital stay itself, with the confinement to the hospital bed, difficulty sleeping, and being attached to equipment was particularly difficult for their child - ?his IV was upsetting?, ?she doesn?t like it there. She?d prefer to be at home?, and ?all the cords getting wrapped up?. On the other hand, some parents felt that their children generally had a high tolerance for the procedures associated with hospitalization.  Parents also thought specific hospital procedures (suction, tests, and oxygen) were hard for their child. Oxygen masks were remembered to have been ?the one that bothered him the most? ripped that one right out?, reporting that ?he?d just? tear at his face? and ?she would cry during that (nebulizer)?. Suctioning is often used as a symptomatic treatment for RSV, however, it can be perceived as hard on the child, with one participant using the words ?traumatic? and ?intrusive?. Although dislike for suctioning procedure was common, it was also acknowledged as providing relief afterwards for their child (see next section: Child relief measures). Tests such as the nasopharyngeal swab used to test for RSV, x-rays, and generally the bombardment of people and actions in the emergency room were perceived as difficult for the child ? ?when they 64  put everything and try everything and he just wants to be with you?, ?the team of people who were all trying to do stuff to her? I think during that time it was difficult?.   Parents often spoke about things that helped their child cope with their illness and the stay in hospital. Trends materialized such as the ability to provide entertainment either by cuddling or colouring, and entertainment by hospital volunteers. In addition, a couple parents mentioned the ?nest? of linens that was created by nurses, the suctioning and BIPAP (an machine used to assist in breathing) to relieve the child of their symptoms, and the reassurance of fundamental parent presence.   There was a variety of symptoms described by parents as things that impacted their children, most of which are reiterated in the RSV literature already. In addition to identifying which primary symptoms their child had, parents were asked to describe these symptoms so that researchers could know how they would describe these symptoms in their own words. Apnea was remembered in 3 cases. Parents recalled ?poking her and shaking her and trying to get her to breathe again?, ?kept trying to wake her? don?t always have? consistent breathing?, and ?I would think he?s not breathing that much, and then I?m waiting , I?m waiting? start counting 1, 2, 3, 4? ?. Although not all parents who reported color change reported apnea, color change in the infant was often a symptom that tipped parents off as to the child?s ill-health. One mother describes it as ?an awful grey?. Another mother recalls ?looking at him, and I would think he looks grey? and then he would breathe and his color would be fine. Finally I said, he is turning blue? I?m going in?. The color was noticed on the face, and around the infant?s lips.  65  Education was identified by two people as a factor in their experience with RSV. This mother stated that ?the biggest, most helpful thing was the education?. She described the transition from being afraid to touch her child among the monitors and tubes, to learning how to recognizing signs of distress or improvement, and how to assist her child with interventions such as suctioning. Another mother commented on the learning associated with the terminology ? ?they?ll say is your child lethargic, and the idea of what lethargy means, until you see your baby like that, you don?t quite understand?. Both women felt strongly about being more equipped to explain the child?s condition, and what to do for the infant.  A thread of discussion indicated that RSV often goes misdiagnosed. This was by the parents themselves, who were not familiar with the disease and therefore did not recognize what it could have been ? ?I had no idea if this was a normal cold, if she has asthma?, ?we didn?t really know what it was at first?, and ?I wonder if he?s asthmatic, or maybe he has cystic fibrosis? I never even thought about RSV?. When these parents did seek medical attention, some of them reported that the doctors thought it was something else ? parents report being told they were investigating for pneumonia, rotavirus, urinary tract infection, and asthma. Some of these parents visited medical clinics or staff more than once prior to being admitted or told the correct diagnosis, and were sent home having been told their child was ?fine?, that they felt they had to ?fight to have people look at him?.   Additional themes and impressions Several additional, less consistent, themes and topics arose during the interviews that are not described above. Table 4.2 describes these additional themes. The ?theme? lists the name of the 66  theme as coded in the analysis. An inference as to the potential significance, relevance, or explanation for the reported observations is given in the third column. Table 4.2 Additional aspects of parents' experience of their child with RSV Theme # sources Inference Evidence Need for care plan 3 Importance of continuity of care  Nasogastric tube 2 Negative experience ?I?ve never seen her so uncomfortable in my entire life?  ?It was scary? ?I didn?t like having to do it, but I liked being allowed to be the one to feed her?.  Reason for seeking medical attention 6 Variety of reasons exist  Asthma 3 Recurring comments about asthma as either fear of asthma, or recurrent symptoms. ?Well, they describe it as asthma-like symptoms.? ?We?ve had to take him to the ER twice for asthma attacks? ?She may have asthma for the rest of her life. But, and that makes me sad, but you know . . . she?s not having open-heart surgery, or whatever.? ?The first time was RSV and the question was whether or not she had asthma. Because I?m asthmatic and my son?s asthmatic.? Opinion of hospital   Positive    Negative ?Very caring? ?Gentle... didn?t disrupt? ?ICU nurses are full of information, full of help?  ?Aside from the parking. Oh my god!? ?I know. Expensive.? ?The sleeping arrangements are terrible? ?Some kind of protocol about taking your own medicine? ?I did find I became... annoyed at how many superfluous people were coming in to the room.? ?A lot of the nurses never bothered to learn my name at all. And I would just be called Mum.? ?Too much waiting time.? ?As soon as we walk in the door, they ask about a DNR... And I hate that. I absolutely hate it.? ?Stop talking about things in front of (them).? Isolation   ?And you?re quarantined in there.? ?So you get a little bit of cabin fever. I get cabin fever.? ?I was always worried when I left if I was contaminating everything in the hospital? ?I wouldn?t recognize some of the (staff) when they took them off? ?It didn?t bother me.? ?It was fine.? Parent feelings  Variety of feelings Fatigue, difficulty sleeping Assessment techniques  Variety of assessment techniques used by different people Instinct, physical (symptoms) and behavioral changes (mood changes, irritability, increased ?snuggle demands?) Certain degree of uncertainty ?I am never quite sure if she?s sick or not? I really do rely on our pediatrician? Daily life impacts   Need to take time off work Juggling childcare and being at the hospital. Chronic illness   Many children also had a chronic illness The child ?seemed sicker? since the RSV infection.   67  Table 5 indicates the author?s (LR) sense of importance of themes. The table lists only those that the author perceives as being of ?high? importance, and answers the question of ?how important is this particular aspect to parents of children with RSV??. The indication of ?high? is influenced by the frequency that the theme recurred (see Table 4.3), the frequency with which the item was mentioned first, and the overall tone of interview based on the perceptions of the author (LR). Table 4.3 Themes with perceived high sense of importance PARENTS CHILD ADDITIONAL THEMES Theme Subtheme Theme Subtheme Theme Subtheme Daily life impacts Support Child coping challenges Physical symptoms After hospitalization Going home Assessment techniques Instinct  Suction Presence  Feelings: Emotional Fear, Nervousness, Worry Child relief measures BIPAP/CPAP* Education**   Stress  Presence   Feelings: Physical Fatigue  Suction   Hospital experience Discomfort in ward Child symptoms Apnea   Unpleasant observations Breathing  Breathing    Tubes  Color change    X-ray  Lethargy   *BIPAP: experienced only by one person, but was the single most helpful part of their experience **Education: stated by one person as the single most important aspect of their experience 4.4 Discussion  This is the first study to qualitatively investigate parents? experience of RSV in their child. These interviews give a perspective on the aspects of RSV in infants that have significant impact on a parent?s quality of life. This information can be used to improve the understanding of health care professionals, decisions makers, and the general population of how an infant?s illness with RSV may impact families? quality of life. Qualitative understanding of this experience could change the way that care is delivered. In the context of utility derivation, this provides critical attributes to consider when describing RSV. 68   4.4.1 Major findings One of the next most significant findings was of the impact of the child?s illness on parents? quality of life after the hospitalization, in the time period following the acute illness (RSV lasts approximately 10 days). The recurrence of the symptoms, management of symptoms and medications, and potential long term effects all play a role in the experiences of parents after their child is discharged from hospital. The long-term nature of this observation implies that the household does not simply ?return to normal? and that behaviours and attitudes may be affected for quite some time after the actual illness.   This study identified an important thread of discussion which indicated that RSV often goes misdiagnosed. A potentially more concerning report is that four participants felt their doctor did not recognize the severity of the infant?s symptoms. As the transition to going home from hospital appeared to have a significant impact on parents, with many parents talking about feelings of fear, worry for recurrence, and concern for future health, one could infer that the disruption in trust of professionals could have contributed to the worry felt by the parents after they return home. It may also have impacted parents? confidence in their ability to accurately assess their child?s condition. This could have contributed to the either a confidence in their ?instincts? discussed previously, or it could influence doubts regarding their ability to assess their child.  It was clear that the emotional impact of having a child ill with RSV is significant for parents. Feelings such as fear and anger experienced during a time of acute illness leave lasting 69  impressions. This was indicated by the frequency of comments about feelings of fear, nervousness, and worry throughout the interviews. This was the single most frequently referenced topic in regards to feelings, and was conveyed with emphasis ? pauses, accent, and repetition. However, the incorporation of this aspect into an attribute for application to utility elicitation is difficult, as how people feel about a situation is individual.   One of the most important symptoms that were described was changes in the infant?s breathing. This was frequently mentioned, often with a firm, decisive tone, and parents commonly used sounds instead of words, making gasping sounds with rapid, shallow breathing, visibly affirming well-established signs of respiratory distress (142). Although feeding changes joined breathing problems as the most common symptoms reported by parents, based on the impressions by the interviewee, there are other symptoms that were less common but impacted equally (if not more) on parents: apnea, color change, and lethargy. Breathing, along with these symptoms, often elicited emotional responses in the parent, such as fear, whereas feeding changes did not appear have associated negative emotional effects.  We found that some parents use physical descriptors to assess how their child is feeling, but many parents are unable to specify exactly what it is that tells them their child is unhappy or ill ? we have called this ?instinct?. This tells us that it is hard for parents to pinpoint exactly what they are seeing, and it may imply that they have the capacity to see things that others are not able to see.   70  The presence of the parent was reported as the most common feature that helped relieve the child during their illness.  However, there was limited discussion about the negative impact on children when parents were not there. As it is impossible to have the child?s point of view, and there is little discussion as to how the parent specifically made this observation, this may imply that this was more of a comfort from the parent?s perspective, or it may not have been captured by the previously mentioned ?instinct? and ?how they know?. Based on frequency, it can also be inferred that parents value highly the ability to stay with their sick child. The ability to be there with the sick infant was repeated throughout interviews. Many parents spoke of holding their child, and the conflicting emotions of not being able to, indicating that holding the infant was an important, and positive, action, and should be facilitated whenever possible.  Support systems are an essential component within reports of positive aspects of the parents? experience. Comments about the absence of support reveal the importance of these support networks and the feelings of being alone when these supports are absent. Having support from the family and health team members, in the form of time, concern, and education, was important. Education, in particular, was paramount to one particular individual, shaping much of the interview, and providing perspective on how this person managed to cope. Although only specified in one individual, the clarity and passion from which they spoke would indicate that others may benefit from this same approach.  We also found that, generally speaking, the hospital stay itself for parents was not overly traumatic. Parents recounted different aspects of the hospital experience - from the need to find and pay for parking, the ability to easily find food for themselves, to the relationship they had 71  with other patients and families. Rarely, parents also expressed negative opinions of some of the hospital proceedings, including frequency of assessments, protocols relating to asking about Do Not Resuscitate (DNR) status, and too much waiting time. Although these comments are not generalizable to all RSV patients, an interesting observation, particularly for health care professionals, surrounds the general feeling of discomfort in the ward setting. This feeling also occurred outside of those who had children that had also been admitted to ICU, and indicates an avenue for improved education or practice on the ward.  Previous studies in RSV and child hospitalization have found some similar findings. Leidy et al (141) cited that there is reason to believe the impact extends past discharge, and also that child hospitalizations evoked concerns in parents such as financial concerns, parenting concerns, concerns about the child?s experience, and the child?s future. Oliveira et al. (156) described repetition that was required prior to hospitalization. The article points out that trust between caregiver-physician, caregiver-PHC, and caregiver-healthcare worker suffered in these patients who had multiple visits prior to hospitalization.   4.4.2 Limitations This study had a small number of participants, and their responses cannot be generalized to the entire population. There is also significant selection bias, as participants were not randomly chosen, and it is likely that those who chose to respond to the invitation to participate had recent experiences, and are more likely to have had experience from the ?extremes? ? either extremely positive, or extremely negative. The data was not cross coded by a second researcher, and theme identification may be biased by the coder (LR?s) interpretation of what was said.  72   4.4.3 Conclusion This is the first study to qualitatively investigate the experience of parents whose children were admitted to hospital with RSV. The results of this study can be used to improve understanding of health care professionals, and influence policy and procedures on an individual level (e.g. compassion and interpersonal relationships between patients, families, and professionals) and hospital level (i.e. education policies, or policies around discharge and follow up).  73  Chapter 5.  Time trade-off experiment  5.1 Background  5.1.1 Purpose and advantage of TTO Utilities provide the opportunity to use a single, cardinal value to represent quality of life. This value can be used to calculate a quality adjusted life year (QALY), incorporating both quality and quantity of life, for use in economic evaluations, and are important to incorporate when quality of life is an important outcome of interest (68). The time trade-off is a direct elicitation method for deriving utilities developed in the 1970?s, and its process has been explained in Chapter 2. The TTO provides a viable option for utility derivation in cases where the standard gamble may not be feasible (e.g. due to complexity) (68).  The TTO and SG are accepted as preferred choice based elicitation methods, as health state utility value tools (e.g. HUI, SF-36) have not been validated in children less than 5 years (117). The TTO has successfully been used in the past for eliciting preferences in children, as well as for temporary health states (126), and may be preferred as parents have refused in the past to do such experiments as a standard gamble (SG). The TTO is also considered to be easier to understand by respondents than a standard gamble, as it does not require an understanding of probabilities (73,157).   74  5.1.2 Challenges in the infant health state of RSV Several challenges are pertinent to the derivation of a utility in the context of infant RSV, and have been described in detail in section 2.3 of this thesis. Respondents? aversion to any risk of death in a child (and gambling, such as would be required in a standard gamble) presents the possibility that a large proportion of respondents would trade ?irrationally?, against the principles of TTO and utility theory, and show an ?all or nothing? trend ? refusing to trade, or ?giving up everything? for their child. This could potentially render the utility estimates unreliable, despite the methodological information gained. The TTO is likely to be a more feasible utility exercise in the case of RSV, as parents have, however, been more willing to trade time off the end of their own lives to prevent a negative health state in their child (123?126).  Establishing from which point of view to frame the TTO task is important (68). The viewpoint must specify whether it is to be constructed from the point of view of, for example, the patient, the patient and family, society, the hospital, or the government. In the case of RSV, framing the question is complicated by the fact that the respondent must be a proxy for the affected individual. Therefore, is it possible to separate the corresponding utilities of parent and child, as the parent?s life is naturally affected by the child?s state of health? Is it realistic to ask respondents to imagine they have a child? Is it realistic to ask respondents to imagine they have a health condition they would have no memory of?   Respondents also require the cognitive ability to answer the elicitation task. Direct elicitation techniques such as the SG and TTO require fairly complex and detailed thought processes. In addition, respondents will sometimes value health states outside the limits of the anchored scale, 75  more often outside the lower limits (below zero - worse than death), however, respondents actually value certain temporary health states as better than perfect health (134). This is interesting as the most commonly documented is valuing health states as worse than death (outside the lower limit, less than 1). Valuing states as worse than death in the TTO is an issue because it was originally only validated in health states preferred to death (158). It is entirely possible that some of this outer use of limits is due to cognitive inability to grasp the task at hand.   The prognosis, outcomes, and measurement tool specifications should not be left vague, or this leads to individual interpretation (68). Response can be influenced by current or past life experiences, which can only be assessed qualitatively and individually. It is extremely difficult, if not entirely impossible, to accurately present a risk of mortality due to RSV that is separate from the mortality associated with another underlying condition. The potential for confounding would be very high (68). Although death is an inherent part of life, it can be argued the TTO should have no uncertainty attached ? if used for economic evaluation uncertainty pertaining to survival would be incorporated into the decision model, and does not need to be in the elicitation task. However, RSV contains a unique combination of a short-term, acute health state, as well as a certain probability of long term consequences (childhood wheeze). As of yet, there are no guidelines in place for this particular scenario.  Using a traditional TTO approach, with death as the anchor, some believe that questions involving temporary, shorter health states may be more respondent friendly, as people may be more accustomed to thinking in shorter term scenarios than chronic disease (134), however, there 76  are other relevant time issues. In temporary health states, valuing short health states on a scale that includes death may induce sensitivity to the inclusion of death in the valuation questions (134). Asking people to imagine they are in a particular, relatively benign health state for a short period of time followed by death then becomes unrealistic. Another approach, the chained time trade off which is suggested for mild or temporary health states (explanation is provided in section 2.4), increases respondent burden significantly. In the context of RSV, another issue is combining the acute phase of RSV with the potential for long-term health outcomes such as wheeze, which adds an element of complexity when choosing the anchor state.  5.1.3 Elicitation task ? possible approaches Based on the considerations mentioned in the Chapter 2, a variety of approaches could have been taken in wording the elicitation task. The primary approaches that were considered are presented below: 1. How much time from the end of [your life], in days, would you give up in order to avoid the above situation described in your child? You can trade a maximum of [x] days from the end of your life? Where x is based on age and life expectancy.(126) 2. How much time from the end of [your infants life], in days, would you give up in order to avoid the above situation described in your child? You can trade a maximum of [x] days from the end of your life. (159) 3. How much time from the end of your life would you give up in order to prevent the worry, anxiety, fretfulness, and unease you would experience associated with the hospitalization and breathing problems of your infant? (124) 77  4. Would [you or your infant] be willing to risk a treatment with an [x] % chance of avoiding the above situation described in your child task and a [1-x] % chance of death? (123)  The objective of this chapter was to derive health utilities from the societal perspective in Canada for four different severities of RSV using the time trade-off method of valuation. Two different perspectives were used ? an adult perspective and a child perspective. This would allow a methodological comparison between perspectives, and obtain infant RSV-specific utilities that would be able to be applied to economic evaluations involving any current or potential treatment in infant RSV.  5.1.4 Incorporating qualitative research into choice experiments It is important to accurately represent the health condition and health state presented in choice experiments. Using qualitative research is useful and important in ensuring attributes that matter most to those experiencing the disease are depicted, however, several considerations need to be made in this application. In this case, families of infants with RSV are who must be relied upon to provide information on what is important to infants and families who experience RSV (see Chapter 4). Using interviews provided a very rich environment for exploring in-depth questions; however, using too few participants may lead to mis-specification of attributes. The qualitative results that came from this study are still valuable, however, but attributes need to be developed from this information while considering the need for levels to be clinically useful or useful on a policy level, the need to avoid personal characteristics (such as emotions), and the need to 78  carefully present attributes in meaningful and understandable terms while still reducing the concept to relevant attribute levels. (160)  5.2 Methods  5.2.1 Methodology The TTO methodology developed for this component of the study was similar to the approach taken by Lee et al. who compared the short- and long-term time trade-off using both proxy-proxy and proxy-patient perspectives for adult, adolescent, and infant pertussis (126) By trading the proxy?s life for both, the proxy was not asked to trade years from the infant?s life. However, elicitation questions were included from two different perspectives ? an adult and infant perspective ? to facilitate comparison, and asked the respondent to trade time from the end of their life (as opposed to the child?s life) to avoid either: a) a hypothetical health condition in themselves right now, or b) a hypothetical health condition in their (hypothetical) infant right now. As it seemed illogical to ask respondents to imagine themselves in the health state ?for the rest of their lives?, the time horizon of the disease reflected the average duration of an illness with RSV (approximately two weeks), after which they would return to full health for the rest of their remaining life. This approach enabled comparison with the study by Lee et al. in that it matched well with the type of disease (respiratory), temporary time frame (weeks), and provided a framework for discounting. This study, however, described social preferences for each perspective (i.e. the adult and the infant), and derived utilities for identical health profiles in adults and infants.  79  5.2.2 Attributes Attributes were selected after careful consideration. Based on the literature review, potential attributes were explored, along with open-ended questions about RSV, in interviews and focus groups (see Chapter 4). Results from the qualitative study supported the decisions on which attributes to include, and attribute levels were discussed with clinical experts to ensure their clinical relevance. Concerted effort was made for fewer attributes and levels to enable ease of understanding for participants.  Four specific disease characteristics were included in the TTO ? location of care, breathing at worst point, supplemental oxygen, and tubes. Attributes were informed by the qualitative study described above (Chapter 4), and outcome used in RSV literature, as well as clinical relevance. Using attributes present in the literature enables immediate comparison and potential use of the results of this study to existing evaluations. The preamble contained a set level of the fifth attribute ? risk of wheeze in the next ten years. Risk of wheeze was included as it is clearly associated with RSV, is of interest to stakeholders, and allowed comparison with the best-worst scaling experiment (see Chapter 6). Attributes were presented to respondents with definitions, and are presented here in Table 5.1.   80  Table 5.1 Health profile descriptions Health State Mild ?OP? Moderate ?WARD? Severe ?BIPAP? Very severe ?ICU? Location of care Emergency room or family doctor?s office Hospital ward Intensive care unit Intensive care unit Breathing at worst point Mild breathing problems Moderate breathing problems Severe breathing problems Breathing failure Oxygen support No oxygen support Nasal prongs Mask-assisted breathing (BIPAP/CPAP) Breathing machine Tubes No tubes Intravenous Intravenous and nasogastric tube Intravenous and nasogastric tube Recovery After recovery, approximately two in five people will experience wheezing over the next 10 years, which might result in unplanned doctor visits for asthma-related problems. Your infant lives an otherwise healthy life.  5.2.3 Recruitment and ethics 1000 adult respondents were recruited to complete an online survey using IPSOS Reid Canada, a Canadian market research company. Respondents were eligible to participate if they were over the age of 18 and could communicate in written English. IPSOS provided a representative sample of the Canadian population with respect to age, gender, and geographical location. Ethical approval was obtained from the University of British Columbia Research Ethics Board and Children and Women?s Research Ethics Board.  5.2.4 Elicitation design Preamble and follow up questions An online, web-based questionnaire containing time trade-off elicitation task was designed using Sawtooth Software version 8.2 (Sawtooth Software, Inc. Orem, UT). Although online methods often have been shown to have greater use of extreme choices (0,1,-1) (161), there is a significant cost advantage to computer based administration. This approach has also been used in 81  the past by the research group. A pilot study assessed feasibility and face validity of the elicitation tool by assessing time to completion and whether respondents were trading appropriately.   Accounting for the argument that there may be respondents who consider states worse than death, and also the uncertainty surrounding how to measure utility in temporary health states, an exploratory question was inserted. If a respondent chose to trade nothing (zero), they were prompted to confirm that this was the correct answer. If they said no, they were given the option to revise their answer. Conversely, if the respondent chose to trade everything (their maximum days available) they were prompted to indicate if they would trade more than this, and if so, how much more. They were asked to specify how many days, months, and years.  ?Health profile? versus ?health state? Vignettes were reviewed by clinical experts for accuracy. As this combination of an acute health state description with a risk of long term sequelae is not a typical approach to time trade off, the overall description is perhaps better referred to as a ?health profile?, rather than a ?health state?.  Elicitation task Respondents were presented with four vignettes, and asked to complete a time trade-off for each level of severity of RSV infection, with the exception of the mildest case which is unlikely to be clinically identified. This was thought to be both unlikely to have a measurable impact on quality of life, and difficult (if not impossible) to differentiate from other common respiratory viruses such as a cold. Based on the different levels of respiratory tract infection characteristic of RSV, 82  vignettes were created for levels of severity including a visit to emergency or the family doctor, admission to a ward with minimal oxygen requirements, admission to ICU with significant oxygen intervention, and admission to ICU with respiratory failure and intubation.   Two sets of four time trade-off questions depicting the four levels of severity of RSV infection were developed from the perspective of an adult (imagine this was happening to you), and from the perspective of an infant (imagine this was happening to your infant).  Both perspectives asked the respondent to trade time from their own life. From the infant perspective, age of the infant was set at six months, as infants under this age are more at risk of RSV. The preamble provided a statement of the respondent?s perspective (whether this was happening to them or to their child), the description of the health state (five attributes at a given level), a statement about the possible long term outcomes of the disease (risk of wheeze and repeat hospital visits), and the elicitation question. Respondents were asked to imagine that the health profile described the symptoms and attributes of the worst day of the illness. A short description of the follow up was provided, including reference to recurrent wheeze, repeat medical visits, and specification that the rest of life was ?otherwise healthy? (the follow up description was the same in all severity levels). (Figure 5.1). 83  Figure 5.1 TTO elicitation task (adult)   Vignettes were designed to elicit utilities for one day in a particular health state. Theoretically, a utility could therefore be applied to each day spent in that health condition, and allows the potential to capture the unique experience of each child?s admission which often includes one or more of the levels of RSV described. For instance, a child could have a visit to an emergency room, then return multiple days later requiring admission to a ward, and then progress to respiratory failure requiring intubation and ICU admission. Providing four levels of RSV in one-day health profiles also facilitated less complicated description of the health profiles, as then these health profiles encompassed the potential experience of both infants with comorbidity and 84  those who were otherwise healthy. The utility of an illness could then be individualized on a case by case basis by providing a specific utility for each day spent with that particular health profile.  The maximum amount of time a respondent could trade, the denominator, was equal to the typical course of an RSV virus ? approximately 14 days. As respondents are known to value time differently when trading at the end of life versus in the present, the maximum number of days available to trade was a value discounted forward calculated based on the respondent?s age and gender, and national Canadian age/gender charts. It accordance with utility theory, and the assumption that the worst possible health state is death and its utility is equal to zero, the maximum days available for a respondent to choose was equivalent to the discounted time needed to trade to reach a utility of zero. This followed the assumption that if a respondent chose the maximum number of days, that they would be saying that the health state described was equivalent to a utility of zero, or death.   At a discount rate of 0.05, the number of days traded to obtain a utility of zero based on life expectancy was between 20 and 350 days. This was obtained using the following equation (63): PV = F/(1+r)n where PV is present value, F is future value, r is discount rate, and n is the number of years under consideration. For example, in a 19 year old today, with a life expectancy of approximately 80 years, 14 days discounted 61 years would be approximately 350 days. Or, in a 40 year old today, who has 40 years of life left based on life tables, 14 days today discounted 40 years would be equivalent to approximately 100 days. This meant that the older a participant was, the less time 85  they had to trade - after discounting, the maximum that a respondent might have to trade was 350 days, and the minimum was 25.  An interactive electronic slider was provided to indicate how much time was traded. By clicking on the slider, respondents could move a slider back and forth along the bar, setting it at the point at which they found indifference. The slider moved in increments of one day, from a minimum of zero days (not less), to a pre-determined maximum number of days. The slider defaults to zero days, and if respondents do no move the slider, they are asked a follow up question to ensure this was intentional.  5.2.5 Pilot study After internally piloting within the research group for readability and comprehension, the study was piloted in 100 respondents recruited from the market research panel. Results of the pilot showed that respondents traded in an appropriate gradient (traded more time for a more severe health state), indicating they understood the elicitation task, and gave face validity to the task and survey format. Time taken to respond and dropout rate was monitored. Responses from both groups supported the structure and design of the survey and no major changes were made.  5.3 Analysis  5.3.1 Calculation of disutility and utility Utilities were calculated by dividing the time traded (numerator) by the time available (denominator) to obtain a value between 0 and 1. Time was measured in units of days. The 86  calculation of days available was described previously as discounting forward the days available based on a respondent?s life expectancy, therefore, time was not discounted during the analysis.  5.3.2 Data exclusion criteria Data was excluded if it met at least one of four exclusionary criteria. 1. Answered in an inappropriate gradient (traded less for more severe health states) 2. Answered within an unreasonable time (<45 seconds to complete the questions)  3. Traded everything for all health states 4. Traded the same for all health states  5.4 Results  5.4.1 Sample characteristics Of 1933 people who began the survey, 1048 respondents (54%) completed either the infant or the adult perspective of the survey. 25% dropped out prior to starting the questions, and 21% dropped out before finishing either the TTO (19%) or the BWS exercise (2.9%) (the BWS exercise is described in chapter 6).551 respondents were excluded from the initial analysis based on the exclusion criteria described in section 5.3.2. The final analyses included 261 respondents from the adult perspective and 236 respondents from the infant perspective. Reasons for exclusion are indicated in Figure 5.2.  Characteristics and calculation of disutility in major subgroups are below. 80 of the respondents in the adult perspective were excluded on the criterion that they ?traded all? for each health state, 87  and 125 for trading in an inappropriate gradient. Age was not a significant in the decision to trade all, implying that the length of time available to trade (as older adults had less time to trade from, and the population is heavily weighted over the age of 50) did not explain why respondents were choosing to give up all of the time available. From the child perspective, a slightly higher percentage of women traded everything (52.7) than in the overall group (49.1), potentially indicating that women may have been more likely to give up everything for a hypothetical child, particularly in the infant perspective TTO. Figure 5.2 Reasons for exclusion from final analysis   Respondent characteristics from both perspectives are reported in Table 5.2. 51% of those who completed the survey were female, and 67% had children. Characteristics were similar between perspectives. Similarities are also between those who were included and excluded from the analysis, however, there a perceivable difference in the group who was excluded ? more were Total completed 1048 Adult perspective 506 Excluded based on: Inappropriate gradient: 125  Time: 30 Traded all for all profiles: 80 Traded same for all profiiles:  10 Included in adult analysis 261 Infant perspective 542 Excluded based on: Inappropriate gradient:  94  Time: 22 Traded all for all profiles: 174 Traded same for all profiiles:  16 Included in infant analysis 236 88  excluded from the infant perspective (306 versus 236), slightly more had children (70.3% versus 65.9%), and included more males (47% versus 43%). Table 5.2 Canada time trade-off: respondent characteristics    Adult perspective Infant perspective   All Included Excluded Included Excluded N  1048 261 245 236 306 Gender (N)       Female Male 539 (51.4) 509 (48.5) 148 (56.7) 113 (43.3) 125 (51.0) 120 (48.9) 123(52.1) 113 (47.8) 142 (46.7) 163 (53.2) Has children (N)       Yes No 700 (66.7) 348 (33.2) 172 (65.9) 89 (34.1) 164 (66.9) 81 (33.0) 166 (70.3) 70 (29.6) 198 (64.7) 108 (35.2) Age group (N)       18-24 69 (6.5) 14 (5.3) 18 (7.3) 13 (5.5) 24 (7.8) (25-29 98 (9.3) 29 (11.1) 24 (9.8) 18 (7.6) 27 (8.8) 30-39 175 (16.7) 45 (17.2) 43 (17.5) 37 (15.6) 50 (16.3) 40-44 48 (4.5) 12 (4.6) 10 (4.0) 12 (5.0) 14 (4.5) 5 (45-54) 298 (28.4) 72 (27.5) 67 (27.3) 73 (30.9) 86 (28.1) 6 (55-59) 116 (11.0) 28 (10.7) 23 (9.3) 29 (12.2) 36 (11.7) 7 (60-64) 74 (7.0) 16 (6.1) 21 (8.5) 13 (5.5) 24 (7.8) 8 (65-79) 157 (14.9) 42 (16.0) 36 (14.6) 39 (16.5) 40 (13.0) 9 (80+) 13 (1.2) 3 (1.1) 3 (1.2) 2 (0.8) 5 (1.6) Children at home (mean) 0.5 - - - - Children with RSV (N)       Yes 47 (4.4) 9(3.4) 15 (6.1) 9 (3.8) 14 (4.5)  No 655 (62.5) 163 (62.4) 146 (59.9) 158 (66.9) 188 (61.4)  Don`t know 79 (7.5) 15 (5.7) 22 (8.9) 17 (7.2) 25 (8.1)  Don`t have children 267 (25.4) 74 (28.3) 63 (25.3) 52 (22.0) 79 (25.8) Lives in the (N):       City 679 (64.7) 163 (62.4) 160 (65.3) 158 (66.9) 198 (64.7)  Town 229 (21.8) 58 (22.2) 60 (24.4) 46 (19.4) 65 (21.2)  Country 140 (13.3) 40 (15.3) 25 (102) 32 (13.5) 43 (14.0) Income (N)       Less than $20,000 87 (8.3) 12 (4.6) 33 (13.4) 12 (5.0) 30 (9.8)  $20,000-$40,000 230 (21.9) 50 (19.1) 57 (23.2) 57 (24.1) 66 (21.5)  $40,000-$80,000 344 (32.8) 84 (32.1) 74 (30.2) 84 (35.5) 102 (33.3)  $80,000 + 387 (36.9) 115 (44.0) 81 (33.0) 83 (35.1) 108 (35.2) Relationship status (N)       Married 652 (62.2) 172 (65.9) 151 (61.3) 138 (58.4) 191 (62.4)  Common-law 94 (8.97) 18 (6.9) 23 (9.3) 24 (10.1) 29 (9.4)  In a relationship (not married, not common-law) 46 (4.3) 9 (3.4) 12 (4.9) 6 (2.5) 19 (6.2)  Single 220 (20.9) 53 (20.3) 49 (20) 58 (24.5) 60 (19.6)  Widow/widower 36 (3.4) 9 (3.4) 10 (4.0) 10 (4.2) 7 (2.2)   89     Adult perspective Infant perspective   All Included Excluded Included Excluded N  1048 261 245 236 306 Education completed (N)       None 46 (4.3) 7 (2.6) 14 (5.7) 9 (3.8) 16 (5.2)         High school diploma or equivalent 280 (26.7) 56 (21.4) 72 (29.3) 58 (24.5) 94 (30.7)  College or certification 334 (31.8) 71 (27.2) 85 (34.6) 71 (30.0) 107 (34.9)  Undergraduate degree 268 (25.5) 83 (31.8) 50 (20.4) 68 (28.8) 67 (21.9)  Graduate degree 120 (11.4) 44 (16.8) 24 (9.8) 30 (12.7) 22 (7.1) Employment status (N)       Not working, looking 92 (8.7) 16 (6.1) 34 (13.8) 15 (6.3) 27 (8.8)  Not working, not looking 44 (4.2) 13 (4.9) 10 (4.0) 8 (3.3) 13 (4.2)  Full-time 479 (45.7) 116 (44.4) 104 (42.4) 116 (49.1) 143 (46.7)  Part-time 133 (12.6) 47 (18.0) 22 (8.9) 27 (11.4) 37 (12.0)  On disability 35 (3.3) 4 (1.5) 8 (3.2) 12 (5.8) 11 (3.5)  Student 38 (3.6) 10 (3.8) 13 (5.3) 5 (2.1) 10 (3.2)  Retired 212 (20.2) 49 (18.7) 52 (21.2) 50 (21.1) 61 (19.9)  Unable to work 15 (1.4) 6 (2.3) 2 (0.8) 3 (1.2) 4 (1.3)  All TTO results are presented in terms of the ?disutility? associated with a particular health profile; therefore the utility of any given health profile equals 1-disutilty. The following tables provide the mean disutility and standard error for each of the four health profiles valued by overall or subgroup (gender, whether respondents have children, and age) ? admission to ICU and intubation, admission to ICU and use of assistive breathing machines (BIPAP), admission to a ward with minimal oxygen intervention, and a visit to the emergency department or family physician. Each health profile represents a different severity of RSV. The mean disutilities trend consistently, for both perspectives, in an appropriate gradient (disutility increases as the health profile becomes more severe), indicating that most respondents understood the task, and gives face validity to the time trade-off question.  90  5.4.2 Overall disutilities Overall disutilities are presented in Table 5.3. Disutilities were consistently higher in the infant perspective than the adult perspective (Wilcoxon rank sum, p<0.05). Disutility ranged from 0.09 to 0.5 in the adult perspective, and from 0.16 to 0.62 in the infant perspective, for mildest and most severe health states, respectively. Table 5.3 Canada time trade-off: Mean disutility for adult and infant perspective   ICU/INTUBATED ICU/BIPAP WARD OUTPATIENT  N Mean SE Mean SE Mean SE Mean SE Adult perspective 261 0.5* 0.03 0.43* 0.03 0.25* 0.02 0.09* 0.01 Infant perspective 236 0.62 0.03 0.57 0.03 0.41 0.03 0.16 0.02 *significantly different than infant perspective (Wilcoxon rank sum test, p<0.01)  5.4.3 Subgroup analyses The subgroups included in this analysis were selected based on a priori hypotheses that there may be different values among males and females and those who have children and those who do not. Age was examined to rule out an age-effect as a result of respondents of older age groups being less willing to trade since they had less time available to trade (see section 5.2.4 on denominator).   Disutility from adult and infant perspectives according to gender is shown in Table 5.4. From the adult perspective, the disutility was slightly higher in than in females, although this was not statistically significant. From the infant perspective, females reported statistically more disutility than males for the three most severe health profiles (Wilcoxon rank sum, p<0.05). The disutility for an infant intubated and in the ICU was 0.7 (versus 0.54), for an infant on BIPAP and in the ICU was 0.65 (versus 0.49), and for an infant admitted to the ward was 0.47 (versus 0.35). 91  Table 5.4 Canada time trade-off: Mean disutility for adult and infant perspective by gender    ICU/intubated ICU/BIPAP WARD OUTPATIENT   N Mean SE Mean SE Mean SE Mean SE Adult perspective            Female 148 0.49 0.04 0.4 0.04 0.24 0.03 0.08 0.02  Male 113 0.51 0.04 0.48 0.04 0.27 0.4 0.1 0.02 Infant perspective            Female 123 0.7* 0.04 0.65* 0.04 0.47* 0.04 0.018 0.03  Male 113 0.54 0.04 0.49 0.04 0.35 0.04 0.13 0.02 *significantly different than males (Wilcoxon ranks sum test, p<0.05)  Disutility from adult and infant perspectives according to whether respondents had children or did not have children is shown in Table 5.5. Respondents who had children were more likely to trade more in the infant perspective than those who do not have children. For an intubated child in the ICU, disutility was 0.65 (females with children) versus 0.53 (females without children), and 0.58 (males with children) versus 0.44 (males without children). Responses were more similar in the adult perspective, although none of the differences were statistically significant (Wilcoxon rank sum, p<0.05). Table 5.5 Canada time trade-off: Mean disutility for adult and infant perspective by whether respondents have children    ICU/ INTUBATED ICU/ BIPAP WARD OUTPATIENT   N Mean SE Mean SE Mean SE Mean SE Adult perspective           Has children 172 0.53 0.03 0.45 0.03 0.27 0.03 0.08 0.02  Does not have children 89 0.44 0.05 0.4 0.05 0.22 0.04 0.11 0.03 Infant perspective           Has children 166 0.64 0.03 0.59 0.03 0.43 0.03 0.17 0.02  Does not have children 70 0.58 0.05 0.54 0.05 0.38 0.05 0.13 0.03  Disutility for adult and infant perspectives by age group is shown in Tables 5.6 and 5.7, respectively. Generally speaking, older respondents reported higher disutility across all health 92  profiles. From the adult perspective, those aged between 60-64 had the highest disutility ? the disutility for an infant intubated and in the ICU was 0.72 versus 0.4 in the group aged 18-24 (p<0.05), and for infants admitted to ICU on BIPAP, disutility was 0.61. In the two milder health profiles, however, the age group (40-44) had the highest disutility (0.35 and 0.22). This group also had the fewest respondents (n=12). Several of these differences were statistically significant, and are indicated in Table 5.6.  From the infant perspective, disutility was more similar between age groups, with the 18-24 age group having the highest disutility for the two most severe health profiles (Table 5.7). Disutility was 0.75 for infants in the ICU and intubated, and 0.72 for infants on BIPAP and in the ICU. The highest disutility for an infant admitted to the ward and for outpatient visit was in those aged 25-29 (disutility = 0.5 and 0.22, respectively). These differences were not statistically different. Table 5.6 Canada time trade-off: Mean disutility by age for adult perspective   ICU/intubated ICU/BIPAP WARD OUTPATIENT* Age N Mean SE Mean SE Mean SE Mean SE 18-24 14 0.4* 0.12 0.36 0.12 0.15 0.08 0.06^ 0.03 25-29 29 0.35* 0.08 0.32 0.08 0.26 0.07 0.07^ 0.03 30-39 45 0.38* 0.07 0.3* 0.06 0.14 0.04 0.05^ 0.02 40-44 12 0.56 0.14 0.49 0.13 0.35 0.13 0.22 0.11 45-54 72 0.51 0.05 0.44 0.05 0.28 0.05 0.07^ 0.02 55-59 28 0.58 0.09 0.47 0.09 0.29 0.08 0.09 0.05 60-64 16 0.72 0.09 0.61 0.09 0.24 0.09 0.13 0.07 65+ 45 0.6 0.07 0.56 0.07 0.31 0.06 0.11 0.03 *Statistically different than age group 60-64 (Duncan?s multiple comparison test) ^ statistically different than age group 40-44 (Duncan?s multiple comparison test)    93  Table 5.7 Canada time trade-off: Mean disutility by age for infant perspective   ICU/INTUBATED ICU/BIPAP WARD OUTPATIENT Age N Mean SE Mean SE Mean SE Mean SE 18-24 13 0.75 0.1 0.72 0.1 0.48 0.11 0.18 0.08 25-29 18 0.71 0.1 0.67 0.1 0.5 0.1 0.22 0.08 30-39 37 0.55 0.07 0.49 0.07 0.36 0.07 0.12 0.03 40-44 12 0.47 0.13 0.47 0.13 0.31 0.13 0.04 0.03 45-54 73 0.59 0.05 0.57 0.05 0.47 0.05 0.18 0.03 55-59 29 0.63 0.08 0.56 0.08 0.39 0.08 0.19 0.06 60-64 13 0.74 0.11 0.66 0.11 0.45 0.12 0.15 0.06 65+ 41 0.66 0.07 0.59 0.07 0.38 0.06 0.12 0.04   5.5 Discussion  This was the first study to derive health utilities specifically for infants with RSV. It is important to understand the value of disutility for this health condition so priorities can be set for treatments for RSV. The methodological exploration into infant utilities is also crucial to enabling effective and appropriate decisions for this pediatric population.  5.5.1 Major findings This study found that there is a significant disutility associated with RSV despite it being a temporary, short-term health condition with almost no risk of mortality in otherwise normal, healthy children. The disutility associated with minor symptoms and mild illness was as low as 0.09, and the disutility associated with intubation, admission to intensive care, and respiratory failure was as high as 0.62. These findings demonstrate that the impact of even relatively minor, short term health states on quality of life is significant despite essentially no risk of mortality. 94  This study showed contrasting subgroup trading behaviours. Disutility was lower when the elicitation task described an infant in the health state as compared to an adult in the health state, and reinforced the importance of establishing clear and intentional ?perspective? when designing elicitation questions. We also found that female respondents and respondents with children reported greater disutility than their counterparts.   The explanation for the difference in utility between perspectives and subgroups is still unclear. However, interpreting the results of this study must take into account the possible reasons behind these differences.  Respondents with children may report lower utility because of their life experience, and being able to identify easier with the challenge and worry of having a sick child. Similarly, females may provide lower utilities as well due to more often being the primary caregiver. It is possible that proxies perceive that there is truly a difference in the utility of the health state as experienced in an adult compared to that of the infant, or that adults value the health and life of infants differently than their own. If this were the case, the difference we see would likely remain apparent even with changes in methodological approaches. It is also possible that the difference seen is attributable to spillover from being in close relationship with the person experiencing the health condition, and makes understanding the reasons behind people?s choices important.  Although our study is in infants, previous studies have reported similar challenges in utility elicitation for temporary health states in children. It was not surprising that there was a difference between the child and adult perspectives, as previous studies have found differences here, too (138,150,151,153,162). However, it is valuable to see the magnitude of the difference in the 95  context of infants. The relatively low utilities that we found for the infant health condition are also consistent with a study conducted by in which infant pertussis complications were ranked as worse than adolescent or adult disease (126). Prosser et al. (138) also reports lower utilities in infants than adults. The utilities associated with RSV, however, are lower than the utilities reported in their study.   5.5.2 Strengths The methodological exploration used in this section is one of the primary strengths of this research. The methodological considerations discussed at the beginning of this paper summarize valuable information regarding the challenges facing health state utility elicitation in infant temporary health states. The method chosen to elicit utilities provides a direct comparison with a previous study(126). We explored the use of infant perspective versus adult perspective, as well as a different approach to calculating the time available to trade by discounting based on age and length of the health state before presenting the question to the respondent. Direct comparison between adult and child perspectives using the same panel and identical health states strengthens knowledge surrounding choosing ?perspective? in infant health states. Having used identical health states, the difference between the utilities of the two perspectives may indicate the presence of interaction or spill-over, and further research should be done.  5.5.3 Clinical relevance Through the application of the utilities derived to economic evaluations, this study has significant potential to impact decisions around priorities in prevention and clinical treatment of RSV. The utilities can be used for any current or future programs or treatments for the disease. In 96  addition, the qualitative understanding of the strength of preference for these infant health profiles can help health professionals to understand the significant impact on quality of life of an acute respiratory infection, particularly involving hospitalization, for families and children.  5.5.4 Limitations There are significant limitations in the methodology chosen for this valuation exercise, many of which have been discussed in detail in Chapter 2, as well as in the beginning of this chapter. Temporary health states are very different than chronic conditions, and the application of traditional valuation methods to this condition is not ideal, particularly as assumptions (such as constant proportional time trade-off) are known to be violated in some temporary health conditions. Direct valuation in pediatric and temporary health states is not well researched to date, and it is likely that there will be many changes before an accepted method of eliciting utilities in this population will be established. Therefore, these disutilities may not represent the most accurate disutility for RSV. On the other hand, they are the only value specific to RSV thus far.  The inclusion criteria were quite restrictive for this analysis, excluding a large portion of the respondents, although it is not uncommon to have large numbers excluded from the analysis for similar reasons (137,163). The inclusion criteria were created on the assumptions of rational trading behaviour; however, given the nature of the temporary health state and infant perspective, it is possible that the assumptions surrounding traditional TTO may not be appropriate exclusion criteria in this case. There is a possibility that ?those who traded all? (utility of zero) and ?those who traded nothing? (utility of one) truly felt that they would be 97  willing to give up everything (although irrational), or that they might really trade nothing since the health state is relatively benign (plausible). There is opportunity here for additional examination of the characteristics of those who were excluded, and the magnitude of the influence on disutilities, perhaps illuminating areas for future research and modified elicitation tasks.  The provision of a parameter (maximum days available) and the location of the slider at the start of the question could influence the respondent?s response towards the extremes. The respondents were provided with a maximum amount of days that they could trade. It is possible that respondents unconsciously associate the full time with death ? extremely negative, and therefore were prompted towards the ?Trade nothing? to avoid the association with death (assuming that a non-fatal health state was not so bad). The slider was also positioned at zero, potentially influencing the respondent towards zero. It is unlikely that either of these played a large role, however, as the largest group excluded was those who ?traded all?, not those who ?traded nothing?.  Being a Canadian study, and only offered in English, despite an effort to obtain a representative sample of the population, a large portion of the population may have been excluded, particularly from Quebec and New Brunswick. In addition, Canada has a large community of immigrants who may also have been excluded based on language barriers.  98  5.5.5 Conclusion Given the value of children to society, the willingness to pay for health care in this population, and yet the challenge conducting trials in this population, we spend significant resources in this area without optimal evidence. As a special population - unlike adults requiring proxy, or children who can partially communicate ? research in proxy elicitation methods in young pediatrics could dramatically shape policy decisions and health costs.  This is the first study of its kind in Respiratory Syncytial Virus. The results of this study are of significant interest to decision makers, pharmaceutical industry, researchers, and health care practitioners alike. The impact of RSV on an infant?s quality of life is valuable information, and its role in social decision making and cost-utility evaluations is likely to be quickly incorporated, and hopefully will continue the movement towards better pediatric specific evaluation tools, and ultimately, improved quality of life and pediatric health outcomes.    99  Chapter 6 Best-worst scaling experiment  6.1 Background In the fourth chapter of this thesis, parents identified aspects of their hospital experience that mattered most to them ? even with similar infant health states (for example, severe illness requiring admission to ICU), there were different parts of that experience that influenced each individual the most. In the previous chapter, we derived an overall utility for entire health states (e.g. severe illness requiring admission to ICU) using a TTO. In this method, the respondent makes a choice based on an entire set of attributes as a whole, and it is impossible to tell which of the attributes presented have the most influence on the utility obtained. In other words, it is not possible to conclude which part of that health state was the most important to respondents on average, nor individually. For example, there is no way to know if Attribute A influenced the respondent?s choice the most, or whether it was Attribute B. It is also impossible to tell to what degree the respondent prefers Attribute A to Attribute B.   This objective of this chapter was to conduct a best-worst scaling experiment exploring the relative preferences of attributes of RSV.  6.1.1 Origin, purpose, and advantage of BWS There has been increasing research in the use of DCE and ordinal preference measures (82,147,164). Best-worst Scaling (BWS) has become an increasingly popular way to provide a simpler method of assessing preferences than the more time consuming traditional approaches 100  like the time trade-off. The statistical and measurement properties for BWS experiments recently been proven (165).  BWS is a modification of the traditional discrete choice experiment (DCE), which asks respondents to choose between two health states containing two sets of attribute levels. By varying the attribute levels across multiple questions, it is possible to determine which attributes influence choices. In contrast, the BWS asks respondents to look at lists of attributes, but instead of picking one set or the other, respondents select the best and worst attribute from each set. This is a more efficient way of determining preferences as it limits the pairwise choices to a single, more comprehensible, more reliable, form ? best, or worst extremes. (90)  BWS allows identification of relative preferences across all attribute levels, making utilities between attributes directly comparable. Using one attribute level as reference, BWS allows an estimate of the relative utility of each attribute level relative to the reference level. The utility of the reference level cannot be determined. These utilities provide information surrounding which attributes have the most importance, specifically the magnitude of this importance relative to other attribute levels. Health researchers gain useful insight into attitudes of health program consumers and can use this information to target programs and treatments accordingly.  BWS has also been used in combination with DCE to allow individual level estimation. (166)  There are several advantages to the BWS over traditional valuation methods or traditional DCE. In addition to increased efficiency and relative utilities, ordinal preference measures are thought to be easier to understand than traditional cardinal methods such as TTO and SG, and have 101  shown similar results (167?169). BWS is often preferred by respondents to SG and TTO, and avoids comprehension issues common in the TTO surrounding identification of a point of indifference, and correctly considering ?certainty? of the outcome being measured (170). It is also less time consuming. Lastly, BWS is not afflicted by time preference or probability rating issues.  6.1.2 Design theory (OMEP and BIBD designs) BWS surveys present respondents with a series of questions containing a specific group of attribute levels called a choice set. Each choice set is based on a pre-determined design matrix. An optimal design ensures that each respondent sees each attribute level the same number of times throughout the questionnaire, and each attribute level pair the same number of times. Respondents are asked to choose which of the attribute levels presented is ?best? and which is ?worst?. BWS makes the statistical assumption that the cognitive process the respondent goes through weighs each possible pair in a choice set against each other, implicitly calculates the difference in utility in every pair, and then chooses the pair that represents the largest difference in utility ? one is at the extreme ?most preferred? end, and one is at the extreme ?least preferred? end. This is where BWS gets its name ?maximum difference scaling?. (90)  The design for BWS requires significant methodological consideration. Best Worst Scaling uses combinatorial design theory. This theory suggests that optimal design should ensure that all attribute levels are presented to each respondent with the same frequency, and that all possible attribute pairs are presented with the same frequency. In other words, every respondent will see each attribute level and each attribute pair the same number of times as each other attribute level 102  and attribute pair. This reduces risk that the respondent gives more weight to an attribute which they observe more often, or less weight to an attribute that is observed less often.  However, it is impossible to require respondents to make a choice from each possible alternative or pair. A fractional factorial design is employed, which allows each possible alternative to be indirectly or directly answered through only a small subset of questions (e.g. 16 choice tasks is a common number given to respondents). (171,172)  Optimal design includes use of either a Balanced Incomplete Block Design (BIBD), or an Orthogonal Main Effects Plan (OMEP) (90). Both BIBD and OMEP designs are available from data libraries freely accessible online. BIBD designs are created perfectly balanced, in which each attribute pair and each attribute are presented an equal number of times throughout the survey. OMEP designs are created where BIBD designs are not possible, and are designed based on each attribute and each attribute pair being available to be chosen at least once (90). If the attribute level combinations do not fit within a BIBD or OMEP design plan, a sub-optimal design can be generated using appropriate software. An appropriate design can be selected by generating multiple designs and choosing the most precise coefficient estimates (i.e. smallest standard errors). One of the primary limitations with not using BIBD or OMEP design is that level frequency and pair-wise frequency is not perfectly balanced.   Best-worst scaling experiments can be analyzed in different ways (165,173?175). Paired models have been reported as the how-to method (90), and use best-worst pairs to estimate utility. Best-worst pairing looks at all possible best-worst pairs, but this impacts the ability to assess 103  individual characteristics (such as age), as respondent characteristics do not vary for potential best-worst pairs (90). Marginal models aggregate choice frequencies over attribute levels, but results in few observations and consequently may lead to larger standard errors (90). Unbalanced designs should require the addition of a variable to adjust for the probability that an attribute level was shown, as not all attribute levels and pairs might be shown equally (90). Programs such as Latent Gold Choice (LGC) are designed for the analysis of individual choice data containing full or partial information. LGC uses a conditional logit model, and also has capability of latent class analysis (LCA) (176). Best-worst ?scoring? has also been fairly widely used as a way to report best-worst scaling results, and reports results in terms of choice frequencies (175). BW scoring takes the difference between the frequency that each respondent chose each attribute level as best and the frequency it was chosen as worst. These frequencies can then be collectively examined across individuals using a mean, and subgroups identified. This has been suggested as an exploratory way to look at data, as well as a more intuitive way to report findings to the lay reader. As several approaches to reporting BWS results have been used, comparison of the performance of different models is important (170), and multiple reports have been published investigating the best modeling approach to BWS experiment (89,170,173).  6.1.3 Challenges of BWS choice experiment One of the primary challenges of any choice experiments involving health care is that researchers are often trying to measure or gain information about technologies and products that do not actually exist, or for which members of the population do not often have contact or familiarity. In these scenarios, researchers are faced with lack of knowledge or experience on the part of the 104  respondent (177). Particularly in health care research, this is because programs and products are studied prior to their release or exposure to the population in order to make economic decisions.   Best-worst choice experiment also follow the same principle of ?conditional demand? as other DCEs in which it is assumed that the respondent must choose ? there is not option to ?opt out?. For this reason, the utility of the choices made are conditional on the fact that they must choose from those options. There is no information gained on the choice relative to status quo, or not choosing. (90).  As they are relatively new, there are no current guidelines on sample size needed and the appropriate number of times attribute pairs need to be available to be chosen to assess heterogeneity of individuals and choices within these choice experiments (90). BWS is rooted in Random Utility Theory (RUT), which assumes the difference in utility of two attributes is proportional to the frequency that one is chosen over the other (178). This requires that variance of respondent choice consistency - whether the respondent?s preferences are clear, making choices consistent, versus preferences which are unclear, making choices inconsistent - is constant across individuals and choices (178). Recognizing variation between respondents leads to improved point estimates (conversely, not recognizing leads to incorrect estimates). Best-worst scaling, under RUT, has instigated discussion of issues surrounding the random component of utility, the nature of the ranking exercise, and how to treat respondent heterogeneity.   105  6.2 Methods  6.2.1 Choice task design Two versions of choice tasks were developed based on two perspectives, and respondents were randomized to either the adult perspective (imagine this is happening to you), or the infant perspective (imagine this is happening to your infant). These perspectives were the same as those described in the time trade-off experiment (see Chapter 5), and respondents completed the same perspective for both experiments (all respondents completed both experiments). Vignettes were designed to present respondents with different possible outcomes associated with RSV, rather than an entire health profile, and respondents chose the best possible outcome and the worst possible outcome from the list.   6.2.2 Attributes and levels Five attributes of RSV were used in the BWS experiment ? the four used in the TTO (location of care, breathing at worst point, oxygen support, and tubes), and an additional attribute ?risk of wheeze?. Risk of wheeze was presented in the preamble (base case) of the TTO. By including risk of wheeze as an attribute, it would later allow the estimation of the impact of varying levels of risk on the disutility elicited in the health profile. The attribute levels are described in Table 6.1    106  Table 6.1 Best-worst scaling attribute levels Attribute Attribute level Location of care Emergency room or family doctor visit Hospital ward Intensive care unit Breathing at worst point Respiratory failure Severe problems Moderate problems Mild problems Oxygen support Intubation Mask-assisted breathing (CPAP/BIPAP Oxygen mask Nasal prongs None Tubes Intravenous and nasogastric tube Intravenous None Risk of wheeze 80% 60% 40% 20% 0%  6.2.3 Recruitment and ethics 1000 adult respondents were recruited to complete an online survey using IPSOS Reid Canada, a Canadian market research company. IPSOS provided a representative sample of the Canadian population in terms of age, gender, and geographical location. Ethical approval was obtained from the Research Ethics Board of British Columbia.  6.2.4 Choice task An example of the choice task is given in Figure 6.1. The choice task included a statement of the perspective of the question (establishing child perspective, or adult) which was the same as the perspective the respondent received in the TTO, a reminder of the context of the options provided (respondents were reminded that this was not a health state, but rather a list of possible situations), and the list of attribute levels. Check boxes were provided on either side of the list to check the best (left of the list), and the worst (right of the list).  107  Figure 6.1 BWS sample choice task  6.2.5 Sawtooth Software Sawtooth Software Inc. (SSI) (MaxDiff/Webv6.0) was used for survey design. SSI constructs near optimal plans (179) using an input system where the user inputs the total number of attribute levels, the number of attributes to show in each choice set, and the number of choice sets the respondent is to see in the questionnaire. SSI then generates a design based on four items ? frequency balance, orthogonality, connectivity, and position balance - in that order. Design by SSI is near optimal. It does not ensure that all levels are presented with exact equal frequency, nor does it ensure that all pairs are presented at the same frequency. However, across multiple versions, near balance is obtained. One of the primary concerns for this survey was respondent burden given that this experiment was to be combined with a time trade-off experiment. Using 12 BWS questions per respondent, one way and two way frequency were best balanced across 12 versions. The design statistics can be found in Appendix D.  108  6.2.6 Pilot The study was piloted within the research group, and then to 100 respondents recruited by the market research panel. Piloting assessed time to completion, feasibility, and understanding of the task. Respondents made choices in an appropriate gradient (ranked more severe attribute levels worse than less severe levels) indicating they understood the choice task. Time taken to respond and dropout rate was monitored. Responses from both groups supported the original structure and design of the survey and no major changes were made.  6.2.7 Analysis  BWS data were coded for analysis in Latent Gold Choice version 4.5.0. Latent class analysis can distinguish groups of individuals with similar preferences, and effectively reveals heterogeneity of responders (180). The primary objective for this thesis was to assess overall preferences, and compare preferences for both adult and child versions. Therefore, analysis was done using one class, and a conditional logit model. The reference level was chosen after running the model multiple times, and choosing the attribute level that appeared to have the greatest impact (the furthest from ?0? ? no preference), and there is no estimate for the reference level. The parameter estimate for each attribute level is interpreted as a relative utility (to the reference level), and p-values indicate the significance of the estimates at the 0.05 level.      109  6.3 Results  6.3.1 Sample characteristics  Of 1105 respondents that began the BWS survey, 1051 completed the entire survey. Dropout rate was lower during the BWS section (<3 %). 509 respondents completed the adult perspective, and 542 respondents completed the infant perspective BWS experiment. All respondents were included in the analysis. From the adult perspective, 53.8% were female, and 66.6% had children. From the infant perspective, 49.1% were female, and 67.2% had children. In general, characteristics between groups were similar. (Table 6.2)   110  Table 6.2 Characteristics of BWS respondents   All Adult perspective Infant perspective N  1051 509 542 Gender (N)     Female Male 540 (51.4) 511 (48.6) 274 (53.8) 235 (46.2) 266 (49.1) 276 (50.9) Has children (N)     Yes No 703 (66.9) 348 (33.1) 339 (66.6) 170 (33.4) 364 (67.2) 178(32.8) Age group (N)     1 (18-24) 69 (6.6) 32 (6.3) 37 (6.8) 2 (25-29) 99 (9.4) 54 (10.6) 45 (8.3) 3 (30-39) 175 (16.7) 88 (17.3) 87 (16.1) 4 (40-44) 48 (4.6) 22 (4.3) 26 (4.8) 5 (45-54) 298 (28.4) 139 (27.3) 159 (29.3) 6 (55-59) 116 (11.0) 51 (10.0) 65 (12.0) 7 (60-64) 74 (7.0) 37 (7.3) 37 (6.8) 8 (65-79) 159 (15.1) 80 (15.7) 79 (14.6) 9 (80+) 9 (0.8) 6 (1.2) 7 (1.3) Children at home (mean) 0.5 0.5 0.5 (0,9) Children with RSV (N)     Yes 48 (4.6) 25 (4.9) 23 (4.2)  No 657 (62.5) 311 (61.1) 346 (63.8)  Don`t know 79 (7.5) 37 (7.3) 42(7.7)  Don`t have children 267 (25.4) 136 (26.7) 131(24.2) Lives in the:     City 682 (64.9) 326 (64.0) 356 (65.7)  Town 229 (21.8) 118 (23.2) 111 (20.5)  Country 140 (13.3) 65 (12.8) 75 (13.8) Income     Less than $20,000 87 (8.3) 45 (8.8) 42 (7.7)  $20,000-$40,000 231 (8.3) 108 (21.2) 123 (22.7)  $40,000-$80,000 345 (32.8) 159 (31.2) 186 (34.3)  $80,000 + 388 (36.9) 197 (38.7) 191 (35.2) Relationship status     Married 655 (62.3) 326 (64.0) 329 (60.7)  Common-law 94 (8.9) 41 (8.1) 53 (9.8)  In a relationship (not married, not common-law) 46 (4.4) 21 (4.1) 25 (4.6)  Single 220 (20.9) 102 (21.0) 118 (21.8)  Widow/widower 36 (3.4) 19 (3.7) 17 (3.1) Education completed     None 46 (4.4) 21 (4.1) 25 (4.6)  High school diploma or equivalent 280 (26.7) 128 (25.1) 152 (28.0)  College or certification 336 (32.0) 158 (31.0) 178 (32.9)  Undergraduate degree 269 (25.6) 134 (26.3) 135 (24.9)  Graduate degree 120 (11.4) 68 (13.4) 52 (9.6) Employment status     Not working, looking 92 (8.7) 50 (9.8) 42 (7.7)  Not working, not looking 44 (4.1) 23 (4.5) 21 (3.9)  Full-time 480 (45.7) 221 (43.4) 259 (47.8)  Part-time 133 (12.7) 69 (13.6) 64 (11.8)  On disability 35 (3.3) 12 (2.4) 23 (4.2)  Student 38 (3.6) 23 (4.5) 15 (2.8)  Retired 214 (20.4) 103 (20.2) 111 (20.5)  Unable to work 15 (1.4) 8 (1.6) 7 (1.3)  111  6.3.2 Relative preferences Coefficients were estimated for 19 attribute levels relative to the 20th attribute level. In this analysis, preferences were relative to the long term outcome ?risk of wheeze: 0%?. The coefficient of each attribute level can be interpreted as the ?utility? of that attribute level in comparison to the common reference level. The estimated utilities for each attribute level are reported in Table 6.3. The average of the coefficients of each attribute level in an attribute is presented as the relative importance for that attribute (90,181). 112  Table 6.3 Relative utilities (conditional logit model)   All  Adult perspective  Child perspective    Class1 SE p-value Class1 SE p-value Class1 SE p-value Location of care -2.21   -2.13   -2.28    ICU -3.56 0.0527 <0.01 -3.42 0.075 <0.01 -3.71 0.0741 <0.01  Ward -2.15 0.0517 <0.01 -2.11 0.0738 <0.01 -2.20 0.0723 <0.01  ED/FD -0.90 0.049 <0.01 -0.87 0.0702 <0.01 -0.93 0.0685 <0.01 Breathing at worst point -2.64   -2.54   -2.74    Respiratory failure -4.73 0.056 <0.01 -4.52 0.0787 <0.01 -4.94 0.0799 <0.01  Severe -3.65 0.0527 <0.01 -3.49 0.0749 <0.01 -3.82 0.0741 <0.01  Moderate -1.52 0.0499 <0.01 -1.51 0.0716 <0.01 -1.53 0.0697 <0.01  Mild -0.67 0.0479 <0.01 -0.65 0.0686 <0.01 -0.68 0.067 <0.01 Oxygen support -2.15   -2.14   -2.15    Mechanical ventilation -3.64 0.0526 <0.01 -3.59 0.075 <0.01 -3.68 0.0738 <0.01  Mask machine (BIPAP/CPAP) -2.56 0.0521 <0.01 -2.56 0.0745 <0.01 -2.57 0.0729 <0.01  Mask -1.94 0.0509 <0.01 -1.94 0.0728 <0.01 -1.93 0.0712 <0.01  Nasal prongs -1.74 0.0506 <0.01 -1.75 0.0724 <0.01 -1.73 0.0708 <0.01  None -0.88 0.0485 <0.01 -0.90 0.0696 <0.01 -0.87 0.0676 <0.01 Tubes -2.11   -2.07   -2.14    Intravenous and nasogastric tube -3.32 0.0522 <0.01 -3.36 0.0747 <0.01 -3.28 0.073 <0.01  Intravenous -2.27 0.0513 <0.01 -2.18 0.0731 <0.01 -2.36 0.0721 <0.01  None -0.73 0.0482 <0.01 -0.68 0.069 <0.01 -0.78 0.0675 <0.01 Risk of wheeze -   -   -    80% -2.79 0.0518 <0.01 -2.66 0.0739 <0.01 -2.91 0.0727 <0.01  60% -2.21 0.0513 <0.01 -2.14 0.0732 <0.01 -2.28 0.0721 <0.01  40% -1.76 0.0506 <0.01 -1.65 0.0721 <0.01 -1.88 0.071 <0.01  20% -1.32 0.05 <0.01 -1.32 0.0716 <0.01 -1.32 0.0698 <0.01  0% (reference) -   -   -    113    All  Adult perspective  Child perspective    Class1 SE p-value Class1 SE p-value Class1 SE p-value R2  0.02   0.02   0.02   R2 (0) 0.22   0.21   0.22   Number of respondents 1051   509   542   Number of observations 25215   12208   13007   Log-likelihood -28302.4   -13925.6   -14347.2   BIC (LL) 56736.9   27969.7   28814.1   Prediction error 0.8   0.8   0.8    114  The parameter estimates within each attribute followed a gradient from less preferred to more preferred corresponding with disease severity and a priori hypotheses, given face validity to the instrument. For example, less severe disease has ?mild? breathing problems, and is more preferred than ?severe? breathing problems.  Interestingly, the attribute level ?oxygen support: mask? and ?oxygen support: nasal prongs? are very close (utility = -1.94 and -1.74). Clinically, an oxygen mask is indicative of increased oxygen demands (a sicker child), and one would expect that the disutility associated with it might be greater.  The least preferred attribute level was respiratory failure (utility estimate -4.73). Similarly, admission to the ICU, severe breathing problems, mechanical ventilation, and intravenous and nasogastric tubes showed low preference, with utility estimates of -3.56, -3.65, -3.64, and -3.32, respectively.   The most preferred attribute level was 0% risk of wheeze. The most preferred characteristics of the disease slightly less preferred with care in the emergency department (utility = -0.90), mild problems breathing (utility = -0.67), no oxygen support required (utility = -0.88), and no invasive tubes (utility = -0.73).   Preference for risk of wheeze consistently decreased as risk increased. 80% risk of wheeze ranked 6/19 for ?least preferred? attribute levels, and 20% risk of wheeze ranked 5/19 for ?most preferred? attribute levels, suggesting that long term health consequences are importance to respondents.  115  The relative importance of each overall attribute can be was interpreted as the average of all attribute levels within an attribute. ?Breathing at worst point? (weighted average utility = -2.64) is the least preferred attribute, with ?risk of wheeze? as the reference. Sequentially preferred attributes are the presence of tubes such as intravenous and nasogastric (weighted average utility = -2.11), requirement of oxygen support (weighted average utility = -2.14), and the location of care (weighted average utility = -2.21). As there is an attribute level parameter estimate missing from the attribute ?risk of wheeze?, it is not possible to determine an estimate for the relative importance of this attribute.  The preferences for both adult and child perspectives were similar to the overall preferences, with slightly stronger preferences from the child perspective  (shown by utility estimates slightly lower (less preferred) for the child perspective. This is consistent through all attributes except oxygen support, where the utility estimates for each attribute level are essentially the same between child and adult perspectives.  6.4 Discussion This is the first preference study to examine preferences of a representative sample of the Canadian population for the health consequences associated with Respiratory Syncytial Virus from two perspectives (as they would experience in themselves, and as they would experience it in their hypothetical infant). The results indicate a willingness to accept potentially severe, short term outcomes in order to avoid the risk of long term consequences. The implication for this is significant regarding decisions and priorities for future treatments and programs targeting RSV in Canada.  116   6.4.1 Major findings The most preferred attribute was the absence of risk for long term wheeze, suggesting that people may be willing to accept immediate, short term, potentially clinically significant consequences to avoid long term risk of wheeze. In comparison, respiratory failure, ICU admission, and mechanical ventilation were the least preferred. At the opposite end of the spectrum, this may suggest willingness to accept milder health states, including some risk of wheeze, to avoid these particular severe short term health outcomes.  Respondents also indicated a strong negative preference for intravenous and nasogastric tube (utility = -3.32), over a mask breathing machine (BIPAP/CPAP) (utility = -2.56). Intravenous alone (utility = -2.27) was also less preferred than admission to a hospital ward (utility = -2.15). This gives strong indication of the patient (or parent) experience of these medical interventions, despite the potential perspective of medical practitioners of these interventions as routine. This was also seen in the qualitative interviews (Chapter 3), in which the presence of tubes and wires was commonly remembered with a negative memory. Our results support this finding on a population level.  Attribute utilities as a whole indicate that breathing at worst point is the least preferred attribute, followed by location of care, oxygen support, and tubes. This could potentially direct the focus of future treatments and interventions. Additionally, this utility could be used to predict the impact each attribute has on the utility of the health state.  117  High risk of childhood wheeze was less preferred than many significant immediate clinical outcomes such as admission to hospital, moderate breathing problems, and need for a mask breathing machine. The reason for this is not established, however, there is a possibility that this is due to negative preferences for the health state, negative time-related preferences (where the preference is influenced by the time spent with the health problem as opposed to the health state itself), the use of probabilities (humans are generally risk averse), or a mix of the three.   The differences observed between adult and child perspectives is, overall, consistent with utilities derived in the previous chapter ? utility is lower in the child perspective. However, there are more inconsistencies in these results, as in the oxygen support and tubes attribute, adult and child preferences are nearly the same and actually reversed slightly for a few attribute levels (mask, nasal prongs, none, IV/NG). If you were to follow an apriori expectation that respondents would trade more for a child than for themselves when the outcome is perceived as unpleasant (for whatever reason ? this is not explored here), this may suggest that the lower levels of the oxygen support are not as important to respondents. This would not explain the IV/NG result, however.  6.4.2 Clinical relevance The use of best-worst scaling provides results in a form that is more easily interpreted by those readers who are not in the choice research field. These results can be easily explained and interpreted by health care practitioners and policy makers involved with infant care, respiratory care, and RSV. There is an implication that it is important to avoid long term consequences, and 118  a reminder that certain procedures which are considered very routine in acute care settings, such as intravenous and oxygen mask use, can have a significant impact on families? experiences.   6.4.3 Limitations This study is the first preference study of RSV in infants, and gives a view of preferences from two perspectives. However, there are limitations that originate from the nature of the disease state in question - conceptual and methodological issues surrounding proxy response for the child perspective are echoed in this chapter as well as the last. Questions surround the interpretation of the ?true? impact of the disease on infants when using a proxy respondent. The reason explaining the differences seen between the responses for adult (?imagine you had??) and child (?imagine your child had..?) is unclear. However, the fact that there is a difference has been shown in child and adult proxy research previously, and its application to infant health states may be important to future development of preference measurement tools.  One of the advantages of the statistical program used is the ability to potentially identify latent classes within these groups. The current analysis does not dissect the potential explanatory impact of individual level characteristics such as age or gender. The addition of covariates to a model such as this would lend additional information about the differences between groups, facilitating understanding of the features behind the choices seen. Further subclass and latent class analyses should be completed to maximize the information gained from this experiment, and tease out potential new methodological approaches to infant preference elicitation.  119  Like other best-worst studies, the choice of attributes and attribute levels can limit the ability to include coverage of all attributes that might affect a patient or family member?s experience with this health state. Minimizing the number of levels is partly because of the need for efficiency in the questionnaire, and also to reduce complexity of the choice tasks. The use of focus groups and expert opinion to inform the levels supports the expectation that the levels excluded from this survey would not have a significant impact on respondent?s choices and the resulting preferences.  This study uses a Canadian sample of the population, and as such, should not be directly generalized to all other geographical areas with RSV. The use of an internet survey may also have introduced some bias, however, internet surveys tend to be more efficient, cost-effective, and are highly accessible today. There is a potential bias if non-respondents represent a different class of individuals based on underlying values and preferences.  6.5 Conclusion Identifying strength of preferences for long term consequences of this short term health state is important to understanding and applying priorities in health care. The results of this study can be used for any future research and decisions requiring an understanding of population preferences surrounding RSV in infants. The use of preference information to inform policy and research will ensure that treatments and programs are targeted appropriately, and maximize the potential for improved patient outcomes.  120  Chapter 7 Discussion  7.1 Important findings and implications As technology and research advances, a plethora of medications and programs are being developed that are, or will soon be, available in Canada. The availability of funding for all possible treatments is quickly shrinking due to budget constraints and competing demands on the health care system. The ability to defend the choice to fund a health care intervention now requires that the intervention in question demonstrate the magnitude of impact on quality of life of recipients. The accepted form of ?quality of life impact? is the quality adjusted life year (QALY), whose calculation requires a utility for the given health state in question.   This thesis provides important exploration into the derivation of utilities and preferences in a methodologically challenging population ? infants. Utilities and utility theory have been around for years, however, by large it has been developed and applied in an adult, self-reporting population. Given the rise in available interventions for pediatric care, and a cultural value for, and large focus on, pediatric health, there is significant need for quality of life measures in both pediatric and infant age groups. Until recently, due to the complexity of measuring quality of life in children, who both experience health differently than adults and cannot always answer for themselves, few pediatric interventions had the ability to use pediatric specific utilities in economic evaluations. Instead, adult utilities were applied to pediatric health states. The challenge of pediatric utility derivation has been more highly studied in recent years, and no firm conclusion has been met as to the best method for deriving utility.   121  The second chapter of this thesis gives prime examples of how likely misleading information are guiding policy decisions. This literature review describes a picture of the disease state in question (RSV), as well as previous economic evaluations of palivizumab (the primary drug available for the treatment of RSV), and the methodological approaches used in deriving utility in temporary pediatric health states. This chapter highlighted the recurring problem of the absence of age group specific or disease specific utility measures in economic evaluations. It is clearly outlined that previous cost-utility analyses in RSV have used utilities representing the health state of a 5 year old child living with CLD, with a history of RSV as an infant. The utility was not specific for the RSV illness itself during infancy. Other studies used utilities from another disease (adult heart failure, and childhood asthma) and applied this to RSV. The results of these evaluations have been used to make policy level decisions.  Although the use of an appropriate utility in the evaluation of RSV is essential, perhaps more importantly, this thesis used RSV to explore the methodological approaches to infant utility derivation. This information can advance the understanding surrounding utility derivation in any pediatric or infant health state. The second half of chapter two summarized the challenges of pediatric and temporary health state utilities, identifying the complexity of proxy elicitation and diversity of approaches used in temporary health states. It is important to recall that many papers reviewed did not report utilities, but rather highlighted the difficulty in doing so. This emphasized the need to find novel, reliable and theoretically sound methods.  The first and fourth chapters identified attributes for development of a utility elicitation task for RSV. Focus groups and interviews provided insight into the experience of parents with infants 122  and children admitted to hospital with RSV. The emotions and challenges expressed provide important perspective for health care practitioners? that could improve future interactions with parents and impact use of health care services. The parents interviewed voiced several key memories of their experience, including significant emotional response (fear, anxiety, worry), and reported needs surrounding timely diagnosis, support and education on assessment of child status, and support in the follow up period. These interviews also provided backing for the selection of attributes used in the following chapters. This is the first qualitative study that has been done in RSV, and opens the door to an understanding of the impact of the disease on the infant?s experience, and the experience between the parent and the hospital or health care system. The integration of qualitative research into otherwise quantitative methods is an important step in ensuring the voice of the people is heard ? one of the limitations when choosing attributes for a choice task is that you can only measure what is included, so it is imperative to ensure the most important attributes are included.   Next, utilities specific to the infant health state of RSV were derived from a Canadian social perspective, applicable across different provinces, providing the first utility study specific to RSV. By directly comparing two perspectives, findings show a systematic difference between child and adult perspectives using a modified time trade off method. This confirms the presence of either proxy/patient utility interaction or a true difference in experience of the health state in adults and children. The side by side comparison allows direct comparison of the effect of perspective on utility. The extra question ?would you trade more? instigates further questioning about the possibility of negative utilities and states worse than death in the context of this modified time trade off approach. Significant discussion was also had over the resulting number 123  of people who did not meet the inclusion criteria for the final utility mean. These observations may question the use of traditional utility methods and theory with temporary health states.   The fifth chapter conducted a best-worst scaling experiment, also comparing both child and adult perspectives. The resulting choices show high preference in both perspectives for long term health (e.g. risk of wheeze: 0%), indicating potential to trade off potentially clinically significant short term outcomes (e.g. mechanical ventilation, ICU admission) in order to avoid negative long term consequences. The implication of this finding can have high impact on future directions of research as well as treatment choices. As a time trade off does not separate out which attributes contribute most to the utility of the health state, the information gained from this choice experiment also sheds important light on the potential weight of individual attributes within the overall health state.  7.2 Limitations There are several limitations to the research conducted, and these are briefly outlined in this section. Limitations of research are a reality, however, if identified and considered during study design and when interpreting results, they can strengthen the overall knowledge gain by clearly recognizing parallels, applications, and real-life complications.  The literature review conducted for this study, although comprehensive, was not a systematic review, and should not be seen as inclusive of all related papers. The search strategy is reported in the appendix. The review was conducted by a single person (the author of this thesis), and is considered in depth and extensive for the purposes of this research. Only major methodological 124  issues and considerations were discussed as deemed pertinent by the author and research committee.  The small number of qualitative interviews completed makes it difficult to ensure a clear picture of the RSV experience for parents across all parents or subgroups of infants admitted (e.g. otherwise health, with lung disease, with heart disease, etc.). Paired with the literature surrounding attributes of the illness, the use of the most acutely ill (those hospitalized ? this included infants with significant comorbidities) patients was thought to be sufficient in identifying the entire spectrum of disease attributes.   The methodology used in this thesis is chosen from multiple previously developed approaches for measuring temporary health states, and from examples of proxy elicitation questions, all of which have been reported as problematic in some way. Although these methods may be negatively described as ?problematic at best?, the challenge lay in developing an approach to an otherwise unexplored health state such that respondents would be willing to trade, that they would understand the task, that the resulting utility was as theoretically sound as possible given the current state of the literature.  The TTO and BWS both had adequate sample size to detect difference in utilities and preferences, although we needed to split the projected sample size in half (from 1000 to 500) in order to reduce respondent burden. The BWS design is sub-optimal, resulting in some efficiency lost in this experiment. Given the nature of the attributes, little could be done to change the levels 125  without making the tool inapplicable to real life situations. This thesis also reports basic utilities and preference values, and has not explored further modelling of latent class and covariates.   7.3 Future directions for research The results of this research have significant potential for impact on future research in pediatric temporary health conditions, and further research in utility derivation is essential in pediatric and infant populations. The potential impact on health outcomes, cost, and policy changes are major, as this population does not have an accepted approach to utilities, a measurement which is required at the government level. The development of this technology will be applicable across any infant health state, and potentially for other pediatric age groups who cannot respond on their own. Each utility gained is pertinent and applicable to all current treatments, as well as any future treatments or programs that come down the pipeline.  The qualitative piece of this project also has potential for future research. Qualitative research can easily be overlooked or put aside due to the lack of numbers and attraction of numerical comparability of results between (or within) studies. Directly based in the population of interest, these open framed research approaches could provide insight into different targetable areas for improvement in health care at the clinical and patient level. For example, wider research into the areas that were felt to need additional interventions could identify more concise idea of how to effectively reach and communicate to parents. In addition, as RSV is extremely common, the distribution of this kind of research can largely impact other members of the population by inducing compassion and awareness.  126  Perhaps the most significant area for future research is in theoretical investigation of temporary health states and best-worst scaling. How do temporary health states differ in the context of assumptions of the traditional time trade off? Is there a better, different way to measure utility? How do you separate the utility of the parent and the utility of the child, and do you need to? What are the impacts of framing bias in this experiment ? for example, the position of the slider and order of health states. The chained gamble or chained trade off may be worth exploring in comparison to the methods used here, as well.  This research will give opportunity for comparison with other studies that used similar methodologies, providing new areas for discussion. The difference in disutility between the child and adult perspectives should be explored more closely in relation to previous work in ?spill-over? and measuring proxy-patient utility interaction. There is also need for examination of the theory and rational for varying the denominator (time available to trade) in health states, chronic or temporary, and how the denominator is interpreted by the respondent.   The use of BWS in research has to be considered. Is it well-grounded theoretically? What exactly does it contribute to utility knowledge? How should it best be applied? Can these results be linked to time trade off experiments? What is the most useful reported outcome (BW scoring, parameter estimates), and how what is the gold standard for analysis? The community at large may benefit from additional latent class and covariate analysis of the BWS and TTO, and the information this may contribute to understanding subgroup differences.  127  7.4 Conclusion Given the value of children to society, the willingness to pay for health care in this population, and yet the challenge conducting trials in this population, we spend significant resources in this area without optimal evidence. As a special population - unlike adults requiring proxy, or children who can partially communicate ? research in proxy elicitation methods in young pediatrics could dramatically shape policy decisions and health costs, and   By considering and studying patient- and disease-specific preferences, researchers and governments improve the likelihood of having target populations take up a service or treatment, results in happier consumers, and better health and cost outcomes. The implications of this research should be attractive to multiple dimensions ? public, government, clinical professionals, and parents.    128  References 1.  Ch?vez-Bueno S, Mej?as A, Jafri HS, Ramilo O. Respiratory syncytial virus: old challenges and new approaches. Pediatr Ann. 2005 Jan;34(1):62?8.  2.  Cane P, editor. Respiratory Syncytial Virus. Elsevier; 2007.  3.  STENSBALLE LG, DEVASUNDARAM JK, SIMOES EA. Respiratory syncytial virus epidemics: the ups and downs of a seasonal virus. Pediatr Infect Dis J. 2003;22(2):S21.  4.  Nair H, Nokes DJ, Gessner BD, Dherani M, Madhi SA, Singleton RJ, et al. Global burden of acute lower respiratory infections due to respiratory syncytial virus in young children: a systematic review and meta-analysis. The Lancet. 2010;375(9725):1545?55.  5.  Respiratory Syncytial Virus Infection (RSV): Infection and Incidence [Internet]. Centers Dis. Control Prev. 2010 [cited 2013 Aug 9]. Available from: http://www.cdc.gov/rsv/about/infection.html 6.  Chidgey SM, Broadley KJ. Respiratory syncytial virus infections: characteristics and treatment. J Pharm Pharmacol. 2005 Nov;57(11):1371?81.  7.  Henderson FW, Clyde Jr WA, Collier AM, Denny FW, Senior RJ, Sheaffer CI, et al. The etiologic and epidemiologic spectrum of bronchiolitis in pediatric practice. J Pediatr. 1979;95(2):35?9.  8.  Reed G, Jewett PH, Thompson J, Tollefson S, Wright PF. Epidemiology and clinical impact of parainfluenza virus infections in otherwise healthy infants and young children < 5 years old. J Infect Dis. 1997 Apr;175(4):807?13.  9.  World Health Organization. Acute Respiratory Infections (update September 2009): Respiratory syncytial virus and parainfluenza viruses [Internet]. WHO. 2009 [cited 2013 Aug 9]. Available from: http://www.who.int/vaccine_research/diseases/ari/en/index2.html 10.  Thwaites R, Piercy J. Nosocomial respiratory syncytial virus infection in neonatal units in the United Kingdom. Acta Paediatr. 2007 Jan 2;93:23?5.  11.  Henrickson KJ, Hoover S, Kehl KS, Hua W. National disease burden of respiratory viruses detected in children by polymerase chain reaction. Pediatr Infect Dis J. 2004 Jan;23(1 Suppl):S11?18.  12.  Holberg CJ, Wright AL, Martinez FD, Ray CG, Taussig LM, Lebowitz MD. Risk factors for respiratory syncytial virus-associated lower respiratory illnesses in the first year of life. Am J Epidemiol. 1991 Jun 1;133(11):1135?51.  13.  Glezen WP, Denny FW. Epidemiology of Acute Lower Respiratory Disease in Children. N Engl J Med. 1973 Mar 8;288(10):498?505.  129  14.  Domachowske JB, Rosenberg HF. Respiratory syncytial virus infection: immune response, immunopathogenesis, and treatment. Clin Microbiol Rev. 1999;12(2):298?309.  15.  Ch?vez-Bueno S, Mej?as A, Merryman RA, Ahmad N, Jafri HS, Ramilo O. Intravenous Palivizumab and Ribavirin Combination for Respiratory Syncytial Virus Disease in High-Risk Pediatric Patients. Pediatr Infect Dis J. 2007 Dec;26(12):1089?93.  16.  Wang EE, Law BJ, Stephens D. Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) prospective study of risk factors and outcomes in patients hospitalized with respiratory syncytial viral lower respiratory tract infection. J Pediatr. 1995;126(2):212?9.  17.  Law BJ, Wang EE, MacDonald N, McDonald J, Dobson S, Boucher F, et al. Does ribavirin impact on the hospital course of children with respiratory syncytial virus (RSV) infection? An analysis using the pediatric investigators collaborative network on infections in Canada (PICNIC) RSV database. Pediatrics. 1997 Mar;99(3):E7.  18.  Langley JM, LeBlanc JC, Wang EEL, Law BJ, MacDonald NE, Mitchell I, et al. Nosocomial Respiratory Syncytial Virus Infection in Canadian Pediatric Hospitals: A Pediatric Investigators Collaborative Network on Infections in Canada Study. PEDIATRICS. 1997 Dec 1;100(6):943?6.  19.  Wang EEL, Law BJ, Robinson JL, Dobson S, al Jumaah S, Stephens D, et al. PICNIC (Pediatric Investigators Collaborative Network on Infections in Canada) Study of the Role of Age and Respiratory Syncytial Virus Neutralizing Antibody on Respiratory Syncytial Virus Illness in Patients With Underlying Heart or Lung Disease. PEDIATRICS. 1997 Mar 1;99(3):e9?e9.  20.  Banerji A, Greenberg D, White LF, Macdonald WA, Saxton A, Thomas E, et al. Risk Factors and Viruses Associated With Hospitalization Due to Lower Respiratory Tract Infections in Canadian Inuit Children. Pediatr Infect Dis J. 2009 Aug;28(8):697?701.  21.  Stensballe LG, Ravn H, Kristensen K, Agerskov K, Meakins T, Aaby P, et al. Respiratory syncytial virus neutralizing antibodies in cord blood, respiratory syncytial virus hospitalization, and recurrent wheeze. J Allergy Clin Immunol. 2009 Feb;123(2):398?403.  22.  Bont L, Aalderen WMC va., Kimpen JLL. Long-term consequences of respiratory syncytial virus (RSV) bronchiolitis. Paediatr Respir Rev. 2000 Sep;1(3):221?7.  23.  Bont L, Steijn M, van Aalderen WMC, Kimpen JLL. Impact of Wheezing After Respiratory Syncytial Virus Infection on Health-Related Quality of Life. Pediatr Infect Dis J. 2004 May;23(5):414?7.  24.  Stein RT, Sherrill D, Morgan WJ, Holberg CJ, Halonen M, Taussig LM, et al. Respiratory syncytial virus in early life and risk of wheeze and allergy by age 13 years. Lancet. 1999 Aug 14;354(9178):541?5.  130  25.  Poulsen A, Stensballe LG, Nielsen J, Benn CS, Balde A, Roth A, et al. Long-term consequences of respiratory syncytial virus acute lower respiratory tract infection in early childhood in Guinea-bissau. Pediatr Infect Dis J. 2006;25(11):1025?31.  26.  The IMpact-RSV Study Group. Palivizumab, a Humanized Respiratory Syncytial Virus Monoclonal Antibody, Reduces Hospitalization From Respiratory Syncytial Virus Infection in High-risk Infants. PEDIATRICS. 1998 Sep 1;102(3):531?7.  27.  Groothius.1993.rsv ig.pdf.  28.  Rodriguez WJ, Gruber WC, Groothuis JR, Simoes EAF, Rosas AJ, Lepow M, et al. Respiratory Syncytial Virus Immune Globulin Treatment of RSV Lower Respiratory Tract Infection in Previously Healthy Children. PEDIATRICS. 1997 Dec 1;100(6):937?42.  29.  Groothuis JR, Woodin KA, Katz R, Robertson AD, McBride JT, Hall CB, et al. Early ribavirin treatment of respiratory syncytial viral infection in high-risk children. J Pediatr. 1990;117(5):792?8.  30.  Feltes TF, Cabalka AK, Meissner HC, Piazza FM, Carlin DA, Top FH, et al. Palivizumab prophylaxis reduces hospitalization due to respiratory syncytial virus in young children with hemodynamically significant congenital heart disease. J Pediatr. 2003 Oct;143(4):532?40.  31.  Law BJ, Langley JM, Allen U, Paes B, Lee DSC, Mitchell I, et al. The Pediatric Investigators Collaborative Network on Infections in Canada Study of Predictors of Hospitalization for Respiratory Syncytial Virus Infection for Infants Born at 33 Through 35 Completed Weeks of Gestation. Pediatr Infect Dis J. 2004 Sep;23(9):806?14.  32.  Nachman SA, Navaie-Waliser M, Qureshi MZ. Rehospitalization With Respiratory Syncytial Virus After Neonatal Intensive Care Unit Discharge: A 3-Year Follow-up. PEDIATRICS. 1997 Dec 1;100(6):e8?e8.  33.  The PREVENT Study Group. Reduction of Respiratory Syncytial Virus Hospitalization Among Premature Infants and Infants With Bronchopulmonary Dysplasia Using Respiratory Syncytial Virus Immune Globulin Prophylaxis. PEDIATRICS. 1997 Jan 1;99(1):93?9.  34.  Groothuis JR, Gutierrez KM, Lauer BA. Respiratory syncytial virus infection in children with bronchopulmonary dysplasia. Pediatrics. 1988 Aug;82(2):199?203.  35.  Thomas M, Bedford-Russell A, Sharland M. Hospitalisation for RSV infection in ex-preterm infants?implications for use of RSV immune globulin. Arch Dis Child. 2000;83(2):122?7.  36.  Moler FW, Khan AS, Meliones JN, Custer JR, Palmisano J, Shope TC. Respiratory syncytial virus morbiity and mortality estimates in congential heart disease patients: a recent experience. Crit Care Med. 20(10):1406?13.  131  37.  Navas L, Wang E, de Carvalho V, Robinson J, the Pediatric Investigators Collaborative Network on infections in Canada. Improved outcome of respiratory syncytial virus infection in a high-hospitalized population of Canadian children. J Pediatr. 121(3):348?54.  38.  Stein RT. Long-term airway morbidity following viral LRTI in early infancy: recurrent wheezing or asthma? Paediatr Respir Rev. 2009;10:29?31.  39.  Shefali-Patel D, Paris MA, Watson F, Peacock JL, Campbell M, Greenough A. RSV hospitalisation and healthcare utilisation in moderately prematurely born infants. Eur J Pediatr. 2012 Jul;171(7):1055?61.  40.  Sampalis JS. Morbidity and mortality after RSV-associated hospitalizations among premature Canadian infants. J Pediatr. 2003 Nov;143(5 Suppl):S150?156.  41.  WELLIVER RC. Respiratory syncytial virus and other respiratory viruses. Pediatr Infect Dis J. 2003;22(2):S6?S12.  42.  Welliver RC, Checchia PA, Bauman JH, Fernandes AW, Mahadevia PJ, Hall CB. Fatality rates in published reports of RSV hospitalizations among high-risk and otherwise healthy children. Curr Med Res Opin. 2010 Sep;26(9):2175?81.  43.  Deshpande SA, Northern V. The clinical and health economic burden of respiratory syncytial virus disease among children under 2 years of age in a defined geographical area. Arch Dis Child. 2003;88(12):1065?9.  44.  Chan PKS, Sung RYT, Fung KSC, Hui M, Chik KW, Adeyemi-Doro FA., et al. Epidemiology of Respiratory Syncytial Virus Infection among Paediatric Patients in Hong Kong: Seasonality and Disease Impact. Epidemiol Infect. 123(2):257?62.  45.  Kristensen K, Hjuler T, Ravn H, Simoes EAF, Stensballe LG. Chronic Diseases, Chromosomal Abnormalities, and Congenital Malformations as Risk Factors for Respiratory Syncytial Virus Hospitalization: A Population-Based Cohort Study. Clin Infect Dis. 2012 Jan 12;54(6):810?7.  46.  Kristensen K, Stensballe LG, Bjerre J, Roth D, Fisker N, Kongstad T, et al. Risk factors for respiratory syncytial virus hospitalisation in children with heart disease. Arch Dis Child. 2009 Jun 18;94(10):785?9.  47.  Chan P-K, Abdel-Latif M-A. Cost of hospitalization for respiratory syncytial virus chest infection and implications for passive immunization strategies in a developing nation. Acta Paediatr. 2003;92(4):481?5.  48.  Thorburn K. Pre-existing disease is associated with a significantly higher risk of death in severe respiratory syncytial virus infection. Arch Dis Child. 2008 Sep 18;94(2):99?103.  49.  Tsolia MN, Kafetzis D, Danelatou K, Astra H, Kallergi K, Spyridis P, et al. Epidemiology of Respiratory Syncytial Virus Bronchiolitis in Hospitalized Infants in Greece. Eur J Epidemiol. 18(1):55?61.  132  50.  Eriksson M, Bennet R, Rotz?n-?stlund M, Sydow M von, Wirgart BZ. Population-based rates of severe respiratory syncytial virus infection in children with and without risk factors, and outcome in a tertiary care setting. Acta Paediatr. 2002;91(5):593?8.  51.  Ghazi A. Jamjoom, Ahmad M. AlSemrani, Abdul Rahaman Al-Frayh, Fach Artz, Khalid F. Al-Mobaireek. Respiratory Syncytial Virus Infection in Young Children Hospitalized with Respiratory Illness in Riyadh. J Trop Pediatr. 1933;39:346?8.  52.  S??ez-Llorens X, Moreno MT, Ramilo O, S??nchez PJ, Top FH, Connor EM. Safety and Pharmacokinetics of Palivizumab Therapy in Children Hospitalized with Respiratory Syncytial Virus Infection. Pediatr Infect Dis J. 2004 Aug;23(8):707?12.  53.  Avendano LF, Palomino MA, Larranaga C. Surveillance for Respiratory Syncytial Virus in Infants Hospitalized for Acute Lower Respiratory Infection in Chile (1989 to 2000). J Clin Microbiol. 2003 Oct 7;41(10):4879?82.  54.  The Swiss Pediatric Surveillance Unit, Berger TM, Aebi C, Duppenthaler A, Stocker M. Prospective Population-based Study of RSV-related Intermediate Care and Intensive Care Unit Admissions in Switzerland over a 4-Year Period (2001?2005). Infection. 2008 Dec 9;37(2):109?16.  55.  Deshpande; S. RSV prevention. Arch Dis Child. 2000 Jan 1;82(1):88?88.  56.  Wang D, Cummins C, Sandercock J, Burls C. Immunoprophylaxis against respiratory syncytial virus (RSV) with palivizumab in children: a systematic review and economic evaluation. Health Technol Assess. 2008;12(36).  57.  MacDonald NE, Hall CB, Suffin SC, Alexson C, Harris PJ, Manning JA. Respiratory Syncytial Viral Infection in Infants with Congenital Heart Disease. N Engl J Med. 1982 Aug 12;307(7):397?400.  58.  Joffe S, Escobar GJ, Black SB, Armstrong MA, Lieu TA. Rehospitalization for respiratory syncytial virus among premature infants. Pediatrics. 1999;104(4):894?9.  59.  Wilkesmann A, Ammann RA, Schildgen O, Eis-H??binger AM, M??ller A, Seidenberg J, et al. Hospitalized Children With Respiratory Syncytial Virus Infection and Neuromuscular Impairment Face an Increased Risk of a Complicated Course: Pediatr Infect Dis J. 2007 Jun;26(6):485?91.  60.  Simon A, Ammann RA, Wilkesmann A, Eis-H?binger AM, Schildgen O, Weimann E, et al. Respiratory syncytial virus infection in 406 hospitalized premature infants: results from a prospective German multicentre database. Eur J Pediatr. 2007 Feb 16;166(12):1273?83.  61.  bc guidelines for rsv immunoprophylaxis 2012-13.pdf.  62.  Smart KA, Lanct?t KL, Paes BA. The cost effectiveness of palivizumab: a systematic review of the evidence. J Med Econ. 2010 Sep;13(3):453?63.  133  63.  Drummond, M.F. Methods for the economic evalution of health care programs. 3rd ed.  64.  Grauer, D.W., Lee, J., Odom, T.D., et al. Pharmacoeconomics and Outcomes: Application Patient Care. Second Edition. Kansas City: American College of Clinical Pharmacy; 2003.  65.  Loomes G. Evidence of a new violation of the independence axiom. J Risk Uncertain. 1991;4(1):91?108.  66.  Cohen BJ. Is Expected Utility Theory Normative for Medical Decision Making? Med Decis Making. 1996 Feb 1;16(1):1?6.  67.  Nease RF. Do Violations of the Axioms of Expected Utility Theory Threaten Decision Analysis? Med Decis Making. 1996 Oct 1;16(4):399?403.  68.  Torrance, George W. Measurement of health state utilities for economic appraisal: a review. J Health Econ. 1986;5:1?30.  69.  Weinstein MC, Torrance G, McGuire A. QALYs: The Basics. Value Health. 2009 Mar;12:S5?S9.  70.  Stalmeier, Peep F.M., Busschbach, Jan J.V., Lamers, Leida M., Krabbe, Paul F.M. The gap effect: discontinuities of preferences around dead. Health Econ. 2005;14:679?85.  71.  Stalmeier PF, Lamers LM, Busschbach JJ, Krabbe PF. On the assessment of preferences for health and duration: maximal endurable time and better than dead preferences. Med Care. 2007;45(9):835.  72.  Attema AE, Brouwer WBF. On the (not so) constant proportional trade-off in TTO. Qual Life Res. 2010 Feb 12;19(4):489?97.  73.  Bleichrodt H. A new explanation for the difference between time trade-off utilities and standard gamble utilities. Health Econ. 2002 Jul;11(5):447?56.  74.  Dolan P, Gudex C, Kind P, Williams A. Valuing health states: a comparison of methods. J Health Econ. 1996;15(2):209?31.  75.  Sharma R, Stano M. Implications of an economic model of health states worse than dead. J Health Econ. 2010 Jul;29(4):536?40.  76.  Devlin NJ, Tsuchiya A, Buckingham K, Tilling C. A uniform time trade off method for states better and worse than dead: feasibility study of the ?lead time? approach. Health Econ. 2011 Mar;20(3):348?61.  77.  Dolan P, Roberts J. To what extent can we explain time trade-off values from other information about respondents? Soc Sci Med. 2002;54(6):919?29.  78.  Patrick DL, Starks HE, Cain KC, Uhlmann RF, Pearlman RA. Measuring Preferences for Health States Worse than Death. Med Decis Making. 1994 Feb 1;14(1):9?18.  134  79.  Tilling C, Devlin N, Tsuchiya A, Buckingham K. Protocols for Time Tradeoff Valuations of Health States Worse than Dead: A Literature Review. Med Decis Making. 2010 Jan 12;30(5):610?9.  80.  A Different Approach to Health State Valuation. Value Health. 2009 Mar;12:S16?S17.  81.  Ungar WJ. Challenges in health state valuation in paediatric economic evaluation: are QALYs contraindicated? Pharmacoeconomics. 2011 Aug;29(8):641?52.  82.  Griebsch I. Quality-Adjusted Life-Years Lack Quality in Pediatric Care: A Critical Review of Published Cost-Utility Studies in Child Health. PEDIATRICS. 2005 May 1;115(5):e600?e614.  83.  Prosser LA. Current Challenges and Future Research in Measuring Preferences for Pediatric Health Outcomes. J Pediatr. 2009 Jul;155(1):7?9.  84.  Petrou S. Methodological issues raised by preference-based approaches to measuring the health status of children. Health Econ. 2003 Aug;12(8):697?702.  85.  Jansen SJT, Stiggelbout AM, Wakker PP, Nooij MA, Noordijk EM, Kievit J. Unstable Preferences:: A Shift in Valuation or an Effect of the Elicitation Procedure? Med Decis Making. 2000 Jan 1;20(1):62?71.  86.  Coast J, Flynn T, Sutton E, Al-Janabi H, Vosper J, Lavender S, et al. Investigating Choice Experiments for Preferences of Older People (ICEPOP): evaluative spaces in health economics. J Health Serv Res Policy. 2008 Oct 1;13(suppl 3):31?7.  87.  Bansback N, Brazier J, Tsuchiya A, Anis A. Using a discrete choice experiment to estimate health state utility values. J Health Econ. 2012 Jan;31(1):306?18.  88.  Porteous T, Ryan M, Bond CM, Hannaford P. Preferences for self-care or professional advice for minor illness: a discrete choice experiment. Br J Gen Pract. 2006;56(533):911.  89.  Potoglou D, Burge P, Flynn T, Netten A, Malley J, Forder J, et al. Best?worst scaling vs. discrete choice experiments: An empirical comparison using social care data. Soc Sci Med. 2011 May;72(10):1717?27.  90.  Flynn TN, Louviere JJ, Peters TJ, Coast J. Best?worst scaling: What it can do for health care research and how to do it. J Health Econ. 2007 Jan;26(1):171?89.  91.  Harris KC, Anis AH, Crosby MC, Cender LM, Potts JE, Human DG. Economic Evaluation of Palivizumab in Children With Congenital Heart Disease: A Canadian Perspective. Can J Cardiol. 2011 Jul;27(4):523.e11?523.e15.  92.  Abbott Laboratories, Limited (distributor), Boehringer Ingelheim (BI) Pharma KG (manufacturer). Product Monograph: Synagis (palivizumab) lyophilized powder (50 mg and 100 mg). 2009.  135  93.  Simoes EA. Respiratory syncytial virus infection. The Lancet. 1999;354(9181):847.  94.  Wang D, Bayliss S, Meads C. Palivizumab for immunoprophylaxis of respiratory syncitial virus (RSV) bronchiolitis in high-risk infants and young children: a systematic review and additional economic modelling of subgroup analyses. 2011 [cited 2013 Jul 30]; Available from: https://www.pubmedcentral.nih.gov/pubmedhealth/PMH0014983/ 95.  Lanct?t KL, Masoud ST, Paes BA, Tarride J-E, Chiu A, Hui C, et al. The cost-effectiveness of palivizumab for respiratory syncytial virus prophylaxis in premature infants with a gestational age of 32?35 weeks: a Canadian-based analysis*. Curr Med Res Opin. 2008 Nov;24(11):3223?37.  96.  Nuijten M, Lebmeier M, Wittenberg W. Cost effectiveness of palivizumab for RSV prevention in high-risk children in the Netherlands. J Med Econ. 2009;12(4):291?300.  97.  Nuijten M, Lebmeier M, Wittenberg W. Cost effectiveness of palivizumab in children with congenital heart disease in Germany. J Med Econ. 2009;12(4):301?8.  98.  Derrick Y. Tam, Anna Banerji, Bosco A. Paes, Charles Hui, Jean-Eric Tarride, Krista L. Lanctot. The cost effectiveness of palivizumab in term Inuit infants in the Eastern Canadian Arctic. J Med Econ. 2009;12(4):361?70.  99.  Nuijten MJ, Wittenberg W, Lebmeier M. Cost effectiveness of palivizumab for respiratory syncytial virus prophylaxis in high-risk children. Pharmacoeconomics. 2007;25(1):55?71.  100.  ElHassan NO, Sorbero ME, Hall CB, Stevens TP, Dick AW. Cost-effectiveness analysis of palivizumab in premature infants without chronic lung disease. Arch Pediatr Adolesc Med. 2006;160(10):1070.  101.  Bentley A, Filipovic I, Gooch K, Buesch K. P170 A cost-effectiveness analysis of respiratory syncytial virus (RSV) prophylaxis in infants in the UK. Thorax. 2011 Dec 2;66(Suppl 4):A136?A137.  102.  Neovius K, Buesch K, Sandstr?m K, Neovius M. Cost-effectiveness analysis of palivizumab as respiratory syncytial virus prophylaxis in preterm infants in Sweden: Cost-effectiveness of RSV prophylaxis. Acta Paediatr. 2011 Oct;100(10):1306?14.  103.  Chirico G, Ravasio R, Sbarigia U. Cost-utility analysis of palivizumab in Italy: results from a simulation model in the prophylaxis of respiratory syncytial virus infection (RSV) among high-risk preterm infants. Ital J Pediatr. 2009;35(1):4.  104.  Resch B, Gusenleitner W, Nuijten MJC, Lebmeier M, Wittenberg W. Cost-effectiveness of palivizumab against respiratory syncytial viral infection in high-risk children in Austria. Clin Ther. 2008 Apr;30(4):749?60.  105.  Nuijten MJ, Wittenberg W. Cost effectiveness of palivizumab in Spain: an analysis using observational data. Eur J Health Econ. 2009 Dec 5;11(1):105?15.  136  106.  Yount LE, Mahle WT. Economic Analysis of Palivizumab in Infants With Congenital Heart Disease. PEDIATRICS. 2004 Dec 1;114(6):1606?11.  107.  Weiner, LB, Polak, MJ, Masaquel, A, Mahadevia, PJ. Cost-effectiveness of respiratory syncytial virus prophylaxis with palivizumab among preterm infants covered by medicaid in the united states. Value Health. 2011;14:A118.  108.  Derek G Human, Kevin C Harris, Aslam H Anis, Marsha C Crosby, Laurie M Cender, James E Potts. Economic Evaluation of Palivizumab in Children with Congenital Heart Disease: A Canadian Perspective. Cardiology. 20(Supplement 1):30 (#3117).  109.  Bernhard Resch, Constanze Sommer, Mark JC Nuijten, Susanne Seidinger, Evelyn Walter, Vera Schoelbauer, et al. Cost-effectiveness of palivizumab for respiratory syncytial virus infection in high-risk children, based on long-term epidemiologic data from Austria. Pediatr Infect Dis J. 2012;31:e1?e8.  110.  Eduardo G. Perez-Yarza, Antonio Moreno, Pablo Lazaro, Asuncion Mejia. The association between respiratory syncytial virus infection and the develpment of childhood asthma. Pediatr Infect Dis J. 2007;26:733?9.  111.  Romero, Jose R. Palivizumab prophylaxis of respiratory syncytial virus disease from 1998 to 2002: results from four years of palivizumab usage. Pediatr Infect Dis J. 2003;22(2 (Supplement)):S46?S54.  112.  Liisa Palmer, Caroline B. Hall, Julie P. Katkin, Nianwen Shi, Anthony S. Masaquel, Kimmie K. McLaurin, et al. Respiratory outcomes, utilization and costs 12 months following a respiratory syncytial virus diagnosis among commercially insured late-preterm infants. Curr Med Res Opin. 2011;27(2):403?12.  113.  Greenough A. Health care utilisation of prematurely born, preschool children related to hospitalisation for RSV infection. Arch Dis Child. 2004 Jul 1;89(7):673?8.  114.  Fryback DG, Dasbach EJ, Klein R, Klein BEK, Dorn N, Peterson K, et al. The Beaver Dam Health Outcomes study: Initial Catalog of Health-state Quality Factors. Med Decis Making. 1993 Jun 1;13(2):89?102.  115.  Center for Evaluation and Value of Risk in Health. Cost Effectiveness Analysis Registry [Internet]. Cost Eff. Anal. Regist. [cited 2012 Jun 23]. Available from: https://research.tufts-nemc.org/cear4/ 116.  Meijboom MJ, Rozenbaum MH, Benedictus A, Luytjes W, Kneyber MCJ, Wilschut JC, et al. Cost-effectiveness of potential infant vaccination against respiratory syncytial virus infection in The Netherlands. Vaccine. 2012 Jun;30(31):4691?700.  117.  Prosser LA, Hammitt JK, Keren R. Measuring health preferences for use in cost-utility and cost-benefit analyses of interventions in children. Pharmacoeconomics. 2007;25(9):713?26.  137  118.  Callery P, Luker K. The use of qualitative methods in the study of parents? experiences of care on a children?s surgical ward. J Adv Nurs. 1996;23(2):338?45.  119.  Canadian Agency for Drugs and Technologies in Health. Guidelines for the economic evaluation of health technologies Canada. [Internet]. Ottawa, Ont.: Canadian Agency for Drugs and Technologies in Health; 2006 [cited 2013 Jul 30]. Available from: http://site.ebrary.com/id/10119368 120.  Gu NY, Wolf C, Leopold S, Manner PA, Doctor JN. A Comparison of Physician and Patient Time Trade-Offs for Postoperative Hip Outcomes. Value Health. 2009 Jun;12(4):618?20.  121.  Dolan P. Whose Preferences Count? Med Decis Making. 1999 Oct 1;19(4):482?6.  122.  De Wit, GA, BusschBach, JJV, De Charro, FTD. Sensitivity and perspective in the valuation of health status: whose values count? Health Econ. 2000;9:109?26.  123.  Lipton, G, Wittenberg, E., Nichols, J., Rich, M., Harper, M. Assessment of parental health state preferences for diagnostic evaluation and clinical outcomes in urinary tract infection in young children (Poster board). 33rd Annu Meet Soc Med Decis Mak [Internet]. 2011. Available from: http://smdm.confex.com/smdm/2011ch/webprogram/Paper6648.html 124.  Basu A, Dale W, Elstein A, Meltzer D. A Time Tradeoff Method for Eliciting Partner?s Quality of Life due to Patient?s Health States in Prostate Cancer. Med Decis Making. 2010 Mar 12;30(3):355?65.  125.  Prosser LA, Ray GT, O?Brien M, Kleinman K, Santoli J, Lieu TA. Preferences and willingness to pay for health states prevented by pneumococcal conjugate vaccine. Pediatrics. 2004;113(2):283?90.  126.  Lee GM, Salomon JA, LeBaron CW, Lieu TA. Health-state valuations for pertussis: methods for valuing short-term health states. Health Qual Life Outcomes. 2005;3(1):17.  127.  Wright DR, Wittenberg E, Swan JS, Miksad RA, Prosser LA. Methods for measuring temporary health states for cost-utility analyses. Pharmacoeconomics. 2009;27(9):713?23.  128.  Bala MV, Wood LL, Zarkin GA, Norton EC, Gafni A, O?Brien BJ. Are health states ?timeless?? The case of the standard gamble method. J Clin Epidemiol. 1999;52(11):1047?53.  129.  Locadia M. A Comparison of 3 Valuation Methods for Temporary Health States in Patients Treated with Oral Anticoagulants. Med Decis Making. 2004 Dec 1;24(6):625?33.  130.  Merlino LA, Bagchi I, Taylor TN, Utrie P, Chrischilles E, Sumner W, et al. Preferences for Fractures and Other Glucocorticoid-Associated Adverse Effects among Rheumatoid Arthritis Patients. Med Decis Making. 2001 Apr 1;21(2):122?32.  131.  Swan JS. Process Utility in Breast Biopsy. Med Decis Making. 2006 Jul 1;26(4):347?59.  138  132.  Jansen SJT, Stiggelbout AM, Wakker PP, Vliet Vlieland TPM, Leer J-WH, Nooy MA, et al. Patients? Utilities for Cancer Treatments: A Study of the Chained Procedure for the Standard Gamble and Time Tradeoff. Med Decis Making. 1998 Oct 1;18(4):391?9.  133.  Llewellyn-Thomas H, Sutherland HJ, Tibshirani R, Ciampi A, Till JE, Boyd NF. The Measurement of Patients? Values in Medicine. Med Decis Making. 1982 Jan 1;2(4):449?62.  134.  Johnston K, Brown J, Gerard K, O?Hanlon M, Morton A. Valuing temporary and chronic health states associated with breast screening. Soc Sci Med 1982. 1998 Jul;47(2):213?22.  135.  Stalmeier PFM. Discrepancies between chained and classic utilities induced by anchoring with occasional adjustments. Med Decis Mak Int J Soc Med Decis Mak. 2002 Feb;22(1):53?64.  136.  Rutten-van M?lken MP, Bakker CH, van Doorslaer EK, van der Linden S. Methodological issues of patient utility measurement. Experience from two clinical trials. Med Care. 1995 Sep;33(9):922?37.  137.  Lieu TA, Ortega-Sanchez I, Ray GT, Rusinak D, Yih WK, Choo PW, et al. Community and patient values for preventing herpes zoster. Pharmacoeconomics. 2008;26(3):235?49.  138.  Prosser LA, Payne K, Rusinak D, Shi P, Uyeki T, Messonnier M. Valuing health across the lifespan: health state preferences for seasonal influenza illnesses in patients of different ages. Value Heal J Int Soc Pharmacoeconomics Outcomes Res. 2011 Jan;14(1):135?43.  139.  Prosser LA, Bridges CB, Uyeki TM, R?go VH, Ray GT, Meltzer MI, et al. Values for preventing influenza-related morbidity and vaccine adverse events in children. Heal Qual Life Outcomes. 2005;3(1):18.  140.  Robbins JM, Tilford JM, Gillaspy SR, Shaw JL, Simpson DD, Jacobs RF, et al. Parental emotional and time costs predict compliance with respiratory syncytial virus prophylaxis. Ambul Pediatr. 2002;2(6):444?8.  141.  Leidy NK. The Impact of Severe Respiratory Syncytial Virus on the Child, Caregiver, and Family During Hospitalization and Recovery. PEDIATRICS. 2005 Jun 1;115(6):1536?46.  142.  Bracht M, Basevitz D, Cranis M, Paulley R. Impact of Respiratory Syncytial Virus. Drugs R D. 2011;11(3):215?26.  143.  Miles MS, Funk SG, Carlson J. Parental Stressor Scale: neonatal intensive care unit. Nurs Res. 1993 Jun;42(3):148?52.  144.  Youngblut JM, Jay SS. Emergent admission to the pediatric intensive care unit: parental concerns. AACN Clin Issues Crit Care Nurs. 1991 May;2(2):329?37.  139  145.  Leader S, Jacobson P, Marcin J, Vardis R, Sorrentino M, Murray D. A method for identifying the financial burden of hospitalized infants on families. Value Health. 2002;5(1):55?9.  146.  Francis NA, Crocker JC, Gamper A, Brookes-Howell L, Powell C, Butler CC. Missed opportunities for earlier treatment? A qualitative interview study with parents of children admitted to hospital with serious respiratory tract infections. Arch Dis Child. 2010 Nov 2;96(2):154?9.  147.  Ungar WJ, Santos MT. Quality appraisal of pediatric health economic evaluations. Int J Technol Assess Health Care. 2005;21(2):203?10.  148.  Meyer EC, Garcia Coll CT, Seifer R, Ramos A, Kilis E, Oh W. Psychological distress in mothers of preterm infants. J Dev Behav Pediatr JDBP. 1995 Dec;16(6):412?7.  149.  Tomlinson PS, Harbaugh BL, Kotchevar J, Swanson L. Caregiver mental health and family health outcomes following critical hospitalization of a child. Issues Ment Health Nurs. 1995;16(6):533?45.  150.  Eiser C, Morse R. Can parents rate their child?s health-related quality of life? Results of a systematic review. Qual Life Res. 2001;10(4):347?57.  151.  Baca CB, Vickrey BG, Hays RD, Vassar SD, Berg AT. Differences in Child versus Parent Reports of the Child?s Health-Related Quality of Life in Children with Epilepsy and Healthy Siblings. Value Health. 2010 Sep;13(6):778?86.  152.  Sherifali D, Pinelli J. Parent as Proxy Reporting: Implications and Recommendations for Quality of Life Research. J Fam Nurs. 2007 Feb 1;13(1):83?98.  153.  Gundy CM, Aaronson NK. The influence of proxy perspective on patient-proxy agreement in the evaluation of health-related quality of life: an empirical study. Med Care. 2008;46(2):209?16.  154.  Grange A, Bekker H, Noyes J, Langley P. Adequacy of health-related quality of life measures in children under 5?years old: systematic review. J Adv Nurs. 2007 Aug;59(3):197?220.  155.  Dixon-Woods M, Anwar Z, Young B, Brooke A. Lay evaluation of services for childhood asthma. Health Soc Care Community. 2002;10(6):503?11.  156.  de Oliveira, BRG, Collet, N, de Mello, DF, de Lima, RAG. The therapeutic journey of families of children with respiratory diseases in the public health service (abstract). Rev Lat-Am Enferm. 2012;20(3):453?61.  157.  Wakker P, Stiggelbout A. Explaining Distortions in Utility Elicitation through the Rank-dependent Model for Risky Choices. Med Decis Making. 1995 Jun 1;15(2):180?6.  140  158.  Torrance, GW. Measurement of health state utilities for economic appraisal: a review. J Health Econ. 1986;5:1?30.  159.  Carroll AE, Downs SM. Improving Decision Analyses: Parent Preferences (Utility Values) for Pediatric Health Outcomes. J Pediatr. 2009 Jul;155(1):21?25.e5.  160.  Coast J, Al-Janabi H, Sutton EJ, Horrocks SA, Vosper AJ, Swancutt DR, et al. Using qualitative methods for attribute development for discrete choice experiments: issues and recommendations: ATTRIBUTE DEVELOPMENT FOR DCEs. Health Econ. 2012 Jun;21(6):730?41.  161.  Norman R, King MT, Clarke D, Viney R, Cronin P, Street D. Does mode of administration matter? Comparison of online and face-to-face administration of a time trade-off task. Qual Life Res. 2010 Feb 22;19(4):499?508.  162.  Carod-Artal FJ, Coral LF, Trizotto DS, Moreira CM. Self- and Proxy-Report Agreement on the Stroke Impact Scale. Stroke. 2009 Aug 6;40(10):3308?14.  163.  Arnesen T, Trommald M. Are QALYs based on time trade-off comparable? - A systematic review of TTO methodologies. Health Econ. 2005 Jan;14(1):39?53.  164.  Brazier J, Rowen D, Yang Y, Tsuchiya A. Using rank and discrete choice data to estimate health state utility values on the QALY scale. 2009 [cited 2013 Jun 13]; Available from: http://eprints.whiterose.ac.uk/10882/ 165.  Marley AAJ, Louviere JJ. Some probabilistic models of best, worst, and best?worst choices. J Math Psychol. 2005 Dec;49(6):464?80.  166.  Lancsar E, Louviere JJ. Estimating individual level discrete choice models and welfare measures using best-worst choice experiments and sequential best-worst MNL. Univ Technol Cent Study Choice Censoc. 2008;08?004.  167.  Ratcliffe J, Brazier J, Tsuchiya A, Symonds T, Brown M. Using DCE and ranking data to estimate cardinal values for health states for deriving a preference-based single index from the sexual quality of life questionnaire. Health Econ. 2009 Nov;18(11):1261?76.  168.  Salomon JA. Reconsidering the use of rankings in the valuation of health states: a model for estimating cardinal values from ordinal data. Popul Heal Metr. 2003;1(1):12.  169.  McCabe C, Stevens K, Roberts J, Brazier J. Health state values for the HUI 2 descriptive system: results from a UK survey. Health Econ. 2005 Mar;14(3):231?44.  170.  Ratcliffe J, Couzner L, Flynn T, Sawyer M, Stevens K, Brazier J, et al. Valuing Child Health Utility 9D health states with a young adolescent sample: a feasibility study to compare best-worst scaling discrete-choice experiment, standard gamble and time trade-off methods. Appl Health Econ Health Policy. 2011;9(1):15?27.  141  171.  Lancsar E, Louviere J. Conducting discrete choice experiments to inform healthcare decision making. Pharmacoeconomics. 2008;26(8):661?77.  172.  Louviere JJ, Islam T, Wasi N, Street D, Burgess L. Designing Discrete Choice Experiments: Do Optimal Designs Come at a Price? J Consum Res. 2008 Aug;35(2):360?75.  173.  Flynn TN, Louviere JJ, Peters TJ, Coast J. Estimating preferences for a dermatology consultation using Best-Worst Scaling: Comparison of various methods of analysis. BMC Med Res Methodol. 2008;8(1):76.  174.  Flynn TN. Valuing citizen and patient preferences in health: recent developments in three types of best?worst scaling. Expert Rev Pharmacoecon Outcomes Res. 2010 Jun;10(3):259?67.  175.  Marti J. A best?worst scaling survey of adolescents? level of concern for health and non-health consequences of smoking. Soc Sci Med. 2012 Jul;75(1):87?97.  176.  Vermunt JK, Magidson J. Latent GOLD? Choice 4.0 User?s Manual. Stat Innov Inc Belmont MA [Internet]. 2005 [cited 2013 Aug 19]; Available from: http://lamp2.meganet.net/~stmdlwk/technicalsupport/LGCusersguide.pdf 177.  Louviere JJ. What You Don?t Know Might Hurt You: Some Unresolved Issues in the Design and Analysis of Discrete Choice Experiments. Environ Resour Econ. 2006 May;34(1):173?88.  178.  Flynn TN. Using Conjoint Analysis and Choice Experiments to Estimate QALY Values. Pharmacoeconomics. 2010;28(9):711?22.  179.  Sawtooth Software Inc. Sawtooth Software technical paper series: the MaxDiff/Webv6.0 Technical Paper. Sawtooth Software; 2007.  180.  Greene WH, Hensher DA. A latent class model for discrete choice analysis: contrasts with mixed logit. Transp Res Part B Methodol. 2003 Sep;37(8):681?98.  181.  Najafzadeh M, Lynd LD, Davis JC, Bryan S, Anis A, Marra M, et al. Barriers to integrating personalized medicine into clinical practice: a best?worst scaling choice experiment. Genet Med. 2012 Jan 26;14(5):520?6.     142  Appendix A: Medline searches for literature review Cost-Utility Analyses  Detailed medline search Updated: April 29, 2012 Database: Ovid MEDLINE(R) 1946 to Present with Daily Update Search Strategy: -------------------------------------------------------------------------------- 1     exp Respiratory Syncytial Viruses/ (6021) 2     RSV.mp. (6634) 3     1 or 2 (9604) 4     exp "Costs and Cost Analysis"/ (163881) 5     exp Cost-Benefit Analysis/ (53675) 6     exp Cost-Benefit Analysis/ (53675) 7     exp Economics, Pharmaceutical/ (2318) 8     cost.mp. (261061) 9     cost effective.mp. (39042) 10     exp Budgets/ (11384) 11     econom.mp. (8) 12     or/4-11 (293586) 13     3 and 12 (261)      Utilities in infants using a Proxy  Research objective: ?Explore the methodological areas pertaining to the derivation of elicitations for health outcomes experienced by infants with RSV? (Bansback proposal 2010) Whose health is being traded off, and for whose benefit How to best elicit an preference from a proxy (how is it done, how to do it) How to approach the measurement of a temporary health state  Research question: In those requiring a proxy (substitute decision maker, representation, someone to speak for them), how are (the state of the evidence regarding) preferences/ health utilities elicited? [*limited to children]  Search strategy:  PROXY: (MESH)  proxy,  (key words) proxy, surrogate, respondent, substitute decision maker, legal guardian  143  PREFERENCE/HEALTH UTILITY: (MESH) quality of life, health status, health surveys, outcome assessment, quality-adjusted life years, cost-benefit analysis, patient preference   (key words) health utility, cost-utility, preferences, utilities, risk attitude  CHILDREN: (MESH) child, infant newborn, infant, adolescent?not in search  The primary search was run in Medline, then in Embase and Cinahl  Chosen because they are accepted as thorough databases, relevant to health care research and quality of life.  Grey literature and other databases were not searched at this time Search terms were expanded after examining a few of the first relevant articles for key words used in the literature Mesh terms were exploded  Reference lists of key articles were reviewed for inclusion of relevant articles that may not have shown up in the online search of these few databases.  Inclusion Criteria: Study must use a proxy, any publishing date, English language, primary outcome ? preference (utility), use of direct elicitation (TTO, standard gamble) OR use of indirect derivation (*multi-attribute health status system EQ5D, SF36, HUI 2 or 3, **QWB scale)  Include only those in children (<18) ? rationale: the use of a proxy in someone who has BEEN able to communicate is not going to be parallel to those who have not learned/previously been able to. Also the input from someone close to the end of life/someone who is at the same stage in life is not going to be parallel **** big assumptions!   Exclusion criteria: Any language other than English, non-human subjects, elderly/adult  Secondary inclusion criteria: HRQoL outcome, uses any HRQoL tool, uses a proxy  Detailed medline search Database: Ovid MEDLINE(R) 1948 to Present with Daily Update Search Strategy: ------------------------------------------------------------------------------ 1     exp "Quality of Life"/ (95278) 2     exp Health Status/ (85692) 3     exp Health Surveys/ (343393) 4     exp "Outcome Assessment (Health Care)"/ (551741) 5     exp Quality-Adjusted Life Years/ (5241) 6     exp Cost-Benefit Analysis/ (52295) 7     exp Patient Preference/ (965) 8     cost-utility.mp. (1737) 9     utility.mp. (85822) 10     risk attitude.mp. (71) 11     time trade-off.mp. (559) 144  12     standard gamble.mp. (567) 13     or/1-12 (1087002) 14     exp Child/ (1410463) 15     exp Infant/ (859160) 16     exp Child, Preschool/ (686076) 17     exp Infant, Newborn/ (457772) 18     or/14-17 (1841799) 19     13 and 18 (176477) 20     19 (176477) 21     limit 20 to (english language and humans) (154053) 22     exp Proxy/ (805) 23     proxy.mp. (7271) 24     respondent.mp. (3867) 25     exp Legal Guardians/ (2574) 26     or/22-25 (12711) 27     21 and 26 (769)       Qualitative studies  Research objective: to develop a focused list of quality of life impacts/outcomes/factors that have been discussed in a research setting, aggregate attributes regarding critical aspects of patients/parents/people's experience with RSV (essentially things that may contribute to QOL)  Research question: what aspects of RSV most impact the quality of life of the affected individual and their family/parents?  Inclusion/exclusion criteria: include infant or pediatric studies; looking for outcomes that are measured - specify which age groups was measured each effect; [include quality of life studies, experiential studies, phenomenological studies]  Search strategy: RESPIRATORY SYNCYTIAL VIRUS + QUALITATIVE METHODOLOGY  Search terms: respiratory syncytial virus: respiratory syncytial virus AND phenomenology or focus group or case study or grounded theory or ethnography or biography or phsychology or quality of life  Databases searched: Proquest; CINAHL; PsychInfo, Medline  Detailed medline search: Database: Ovid MEDLINE(R) 1948 to Present with Daily Update Search Strategy: 145  -------------------------------------------------------------------------------- 1     exp Respiratory Syncytial Viruses/ (5995) 2     exp Qualitative Research/ (11703) 3     exp Focus Groups/ (12453) 4     phenomenology.mp. (3885) 5     grounded theory.mp. (3991) 6     exp "Quality of Life"/ (95513) 7     exp Anthropology, Cultural/ (98580) 8     exp Case Reports/ (1560227) 9     or/2-8 (1771450) 10     1 and 9 (158) 11     10 (158) 12     limit 11 to (english language and humans) (117)        146  Appendix B. Focus group guidelines Focus Group: Parents? Experience of RSV in their Children  Focus Group Guidelines & Script  SECTION I: INTRODUCTION  Before the focus group session begins, the researcher will have spoken with each participant individually regarding the signed consent forms received. The researcher will then begin the focus group by reviewing the consent and related information with the participants as a large group. The researcher will ensure the group is aware that the session will be audio recorded, and transcribed. Instruction will be given to the participants to use first names only throughout the session. Participants will be informed that when transcribed, all references to identifying names of person, places, locations, or specific events will be coded and identifiers omitted. At the end of the required REB retainment period, all documents will be destroyed.  ?Good afternoon/evening. Thank you for taking your time and choosing to come to this focus group. My name is Lilla, and I will be leading this session.  The main things I need to remind you about the consent are these: your participation is voluntary (you can leave at any time), everything will be kept strictly confidential, and we will be tape recording the session. The focus group will not last any longer than an hour and a half, so, depending how fast we are going, I?ll warn you that although I want to hear everything you have to say, I may have to interrupt you for the sake of speeding things up and getting us out of here on time.  [Freedom to leave] I would like to start by reminding you that participation is entirely optional. You may choose to leave at any time, and you do not need to provide a reason for doing so.   [Confidentiality] In order to protect confidentiality, I will ask us to use first names only today. I cannot control what you do with the information discussed in this session, but will ask you to keep it to yourself out of respect for each other. Any information shared with me, and consequently with the other researchers in my group (their names are on the information I gave you with the consent), will be kept strictly confidential. All names will be removed from the transcript when the discussion is transcribed.   147  [Information] This focus group will last no more than 90 minutes. There is coffee /tea, and a few snacks, on the table for your enjoyment. Please help yourself to some refreshments.   You may have noticed a tape recorder on the table beside me. This session will be audio taped, and then transcribed afterwards, removing all names and any other information that could identify you.  During this session, I will explain the purpose of the study we are doing, a bit about RSV, and then spend most of the time discussing a few questions. You have all been asked to come here because you have a child who has had RSV sometime in the last five years. The questions I have will be about your experiences during the time that your child had RSV.  [Icebreaker]To start, let?s go around the circle and introduce ourselves by first name, Perhaps tell us how many children you have and how old they are.  My name is [Lilla], and I do not have any children, however, I grew up in a big family back east. [a nurse, and also a masters student at UBC][why chose RSV?]  Briefly describe the purpose of the study The researcher will start the focus group session by giving a brief overview of RSV.  ?Before I open the discussion, let?s make sure we are on the same page regarding RSV. RSV (Human respiratory syncytial virus) is a respiratory disease which can severely affect ability to breathe.  It is the most common reason for hospitalization in infants. By the age of 1 year, 70% of infants have been infected with RSV, and almost all have been infected by the age of 2. RSV is an upper respiratory tract infection, which can spread to the lower lungs and cause severe breathing problems for however long it takes for the virus to be fought off  (approx 7-10 days). For some children, breathing may become so difficult that the children may have a tube put down their throat and have a machine to breathe for them. They may also be admitted to the Intensive Care Unit. There is no vaccine to entirely prevent this virus, and there is no treatment besides trying to help them feel better (symptomatic treatment) while waiting it out?   The researcher will then provide information to explain the purpose of the study:  148  ?We are doing a study to determine the impact that RSV has on an infant?s quality of life. We want to know how much it affects families, and what it means to families and to society, to have RSV in their infant. This measure of quality of life is required by health care decision makers in order to make informed decisions about the priorities, treatment, and costs related to RSV. There is currently no information on the impact that RSV has on a child?s quality of life. Part of this is because we cannot ask a child how it affects them, the same way that we can ask you how a condition affects you. Therefore, we need to rely on you, as parents, to tell us about the impact that RSV has had on your child and your lives.  The purpose of this focus group is to identify the parts of your experience with RSV that were most important to you and your child. How did this affect the quality of your life? How it might have impacted you, emotionally, or otherwise, and to determine what was most important to you while your child was unwell?  SECTION II: OPEN-ENDED QUESTIONS ABOUT RSV  Could I ask you to introduce yourself by your first name, tell me what age your child was when they had RSV, and whether they had any repeat visits for the same reason?  After describing the focus of the study, the researcher will pose open-ended questions to guide discussion:  Parent?s perspective 1.Open question: During the time that your child had RSV, how did you feel emotionally?  Physically?  2.Open question: What was most important to you during your child?s illness? (combine these two) 3.Open question: What were the most important impacts on your quality of life (?in your life?) at that time? Leading questions: did you worry about needing to take time away from work, were you anxious, did you have to take any medications, did you have childcare issues, did you have 149  trouble sleeping, were there issues with feeding/breastfeeding, separation anxieties. Concern about future implications (e.g. asthma).  4.Open question: What bothered you most to see in your child?   Child?s perspective (proxy) 1.Open question: What did you feel your child had the hardest time coping with?  Leading question: Sleeping, feeding changes, separation anxiety, strangers/uniforms, the illness itself (physical symptoms)  2. Open question: What do you feel helped relieve these things for your child?   3. How were you able to assess, or how did you asses, how your child was feeling? How did you KNOW that your child was upset or sick?   SECTION III: PILOT TTO QUESTIONS  [Introduction to TTO] As you may have read in your information letter, we will be using the information you have just given us about your experience with RSV, describing which parts were most important to you, to create a questionnaire about RSV. We have created a few possible questions that we would like you to answer. Answering these questions will help us to identify if the questions are clear, and ways to improve the question so they are easier to understand.  Please remember, the questions may not be accurate descriptions of RSV based on what is important to you ? they were created before this focus group based on information we have read.  [Explanation of TTO] One of the ways that researchers measure quality of life is to ask people to make trade-offs. It is not possible to have everything in life ? we always have to trade off something good to avoid something bad, or pay some price for some benefit. When talking about 150  health, we often ask you to trade off life (give up years of life, and die earlier) in order to prevent having to live in a certain health condition ?for example, arthritis, long-term pain, or disability such as blindness. Presumably, the worse the condition, the more the person is willing to give up (live a shorter life) in order to avoid having that health condition. This is called a time trade-off.  For example, let?s say that you are told you will have very severe arthritis for the rest of your life. The question gives you the option to live a shorter life (instead of dying at 80, let?s say you die at 70, giving up 10 years of life) in order to prevent having arthritis from now on.   Does that make sense? It is a hypothetical ? or pretend ? way of imagining a scenario, and how much you would give up to avoid it. It lets us estimate how much of an impact a health condition has on someone?s life, essentially showing how ?desirable? a condition is.  [Instructions] We will be using this time trade-off approach as another way to see how much RSV impacts people?s lives. There are two types of questions in total I will ask you to answer for me. You do not have to put your name on the paper.   First I would like you to all read through the first page I have given you. There is a detailed description of a scenario at the top. This is a description of what we thought RSV might be like for your child. This may change based on what we have talked about today.  Below the description of RSV, there is a question. Read through it slowly. I would ask that you try to answer it before asking me any questions. After we have all had some time to put down some answers, we can discuss if any parts of it were confusing to you, and if there?s anything that you think needs to be improved or changed.  Sample TTO question #1 (potentially provide as a printout of sawtooth draft?) Distribute paper copies for each participant? Have them answer it themselves first Discuss as a group   Follow up questions 151  How did you find this question?  Which parts were clear? Which parts did you find confusing? What did you think about the ping-pong approach versus the top-down approach? Were they easy to understand? Do you have any other suggestions/comments regarding how the question is presented?  Format Length       4. Do you have any other comments about the question?  Sample TTO question #2  Follow up questions How did you find this question?  Which parts were clear? Which parts did you find confusing? What did you think about the ping-pong approach versus the top-down approach? Were they easy to understand? Do you have any other suggestions/comments regarding how the question is presented?  Format Length Do you have any other comments about the question?    SECTION IV: CONCLUSION  152  Exit Question  Is there anything else you would like to tell us, or talk about, regarding your experience with RSV in your family?  Ending the Focus Group  Participants will be thanked for their time. $25 imbursement will be offered to cover parking and travel at this time.    153  Appendix C: Qualitative coding scheme RSV Focus Groups Qualitative Analysis  Coding Scheme for Parent?s Experience of Children with RSV  1. After hospitalization Parents' experiences after being discharged from the hospital.  ? Concern for future health Statements by parents expressing concerns for their child?s future health and safety. ? Going home Explanation of experiences once discharged from hospital, both emotional (e.g. fear, prolonged anxiety) and physical descriptions (e.g. changes in routine, other health changes) ? Medication Experiences and comments about medication after discharge from hospital.  2.  Parental daily life impacts Statements by parents describing how their child?s illness with RSV impacted their daily life/quality of life during that time.  ? Missing work A report of needing to take time off work, or their partner needing to take time off work. ? Childcare Statements describing the experience with arranging for childcare for other children while one child was in hospital. ? Hygiene Parents' comments about maintaining personal hygiene during the time their child had RSV. ? Support Statements describing support from family members (outside of childcare) or health team members during the child's illness with RSV. ? Food Statements surrounding the challenge of finding food for themselves. ? Sibling?s reactions Parents report of their children's reactions to the RSV experience (their parent's absence, visiting their sibling).  3. Unpleasant observations Events, symptoms or procedures that was unpleasant for parents to see happen to/see in their child.  ? Breathing Statements by parents expressing that they did not like seeing the way their child was breathing. 154  ? Xray Parents' expressions that it was hard to see the xray that they witnessed their child have. ? Monitors Statements by parents expressing dislike for the monitor alarms. ? Suction Parents expressions of dislike surrounding the process of suctioning of their child. ? Tubes Parents' expressions that it was hard to see the "tubes" that were in/on their child during their hospitalization. ? IV Parents' expressions that it was hard to see the IV and IV start.  4. Child coping challenges Aspects of illness that parents feel were challenging for their children; things they did not cope well with/did not like/had hardest time with.  ? Physical symptoms Statements that the physical symptoms themselves were among the most significant challenges for their child. ? Frequent assessments Statements that the frequency of assessments by hospital staff was a challenge for their child. ? Hospital stay Statements that simply being in the hospital presente a significant challenge for the child. This may include an explanation of how it was difficult for that child. ? Tests Any other tests completed in the hospital (during hospitalization for RSV) that was identified as challenging for their child. ? Oxygen Oxygen apparatus/maintaining oxygen apparatus was identified as challenging for their child. ? Suctioning The process of suctioning was identified as challenging for their child/disliked by their child. ? High tolerance Statements about the generally high tolerance of the child to the goings-on of the hospital (things did not bother them, really seemed to not mind).  5. Child relief measures Statement explaining what helped the child to cope/feel better during the illness.  ? Presence Statements that parents felt that their presence helped their child. ? Volunteers in hospital Statements that volunteers helped their child cope with the hospital stay. ? Entertainment Comments about how parents entertained their children and how this helped their child cope with the hospital stay. 155  ? Suction Suction was identified as being a factor in helping the child feel better. ? BIPAP BIPAP/CPAP was identified as being a factor in helping the child cope with the challenges of RSV/hospitalization. ? Nest Statements describing the "nest" built by the staff for their child, and how this helped their child cope/do better.  6. Parent assessment techniques Statements explaining how parents knew when, or what, was upsetting their child.  ? Instinct Parents' comments about instinct, gut feelings, how they ?just know? what is bothering their child, and other non-specific statements central to how they know how their child is feeling. ? Assessment Report that parents are able to tell if their child is sick or unhappy based on specific physical and behavioral signs. Statements of specific actions by their child indicate that their child is unhappy. For example, restlessness, and pulling at tubes/cords. Behavioral assessments may include comments related to mood. ? Child isn?t present Parent report of a child not being ?present? even though assessed as alert.  7. Reason for seeking medical attention Explanation for why parents sought medical attention.  8. Presence Statements about time spent with their child.  9. Parents? hospital experience  ? Relationship with other patients and families Description of the relationship with and/or perceptions of other patients and their families during the time in the hospital. ? Discomfort in ward Statements pertaining to decreased comfort level while in the ward versus the ICU, worry about child's safety and proximity of help. ? Isolation Statements relating to isolation procedures and how the parent felt about this.  10. Nasogastric tube Feelings about feedings through nasogastric tube.  13. Need for care plan Statements expressing need for care plan for future visits to hospital for RSV.  156  14. Education Statements expressing helpfulness or lack of education from hospital/medical staff about RSV.  15. Asthma Statements about asthma follow up and/or asthma investigations.  16. Opinion of hospital Statements expressing thoughts and opinions of the hospitals themselves (both BCCH and others). This may include its staff, the physical building, and atmosphere.  ? Positive ? Negative  17.  Child symptoms Statements describing the signs and symptoms their child experienced.  ? Breathing Descriptions of breathing pattern, rate, sound, or difficulty during illness with RSV. This does not include absence of breathing (apnea). ? Apnea Descriptions of apneic periods (periods of slowed or no breathing) during illness with RSV. ? Color change Descriptions of changes in coloring of their child?s skin during illness with RSV. ? Lethargy Descriptions of lethargy in their child during illness with RSV. ? Feeding changes Descriptions of feeding changes during their child's illness with RSV. ? Sleeping changes Description of sleeping pattern changes during their child's illness with RSV. ? Fever Description of fever as a sign of their child being ill.  18. Parent feelings Feelings experienced by parents during the time that their child had RSV.  a) Emotional Statements describing parent?s emotions that they experienced during hospitalization, illness, and treatment of their children.  o Fear, nervousness, worry Statements of fear, nervousness, or worry experienced during their child?s illness. o Stress Statements of stress experienced during their child?s illness. o Relief Statements of relief experienced during their child?s illness. o Guilt 157  Statements of guilt experienced during their child's illness. o Frustration Statements of frustration experienced during their child?s illness. o Sadness Statements of sadness experienced during their child?s illness. o Confusion Statements of confusion experienced during their child?s illness. o Anger Statements of anger experienced during their child?s illness. o Blurry Statements about the time their child being sick being blurry in their memory.  b) Physical Statements describing physical effects experienced by parents while their child had RSV.  o Fatigue Statements describing fatigue as a physical effect experienced during their child?s illness. o Difficulty sleeping Statements from parents that they found it difficult to sleep around the time their child had RSV. o Other health problems Report of other physical effects (increased blood pressure, swelling) during the time their child had RSV.  19. Misdiagnosis of RSV  Parents and/or doctors not recognizing illness as RSV.     158  Appendix D: BWS design statistics Number of Items (Attributes): 20 Number of Items per Set: 5 Number of Sets per Respondent: 12 Number of Versions: 12  One Way Frequencies:  Item  Times Used -----------------     1      36     2      36     3      36     4      36     5      36     6      36     7      36     8      36     9      36    10      36    11      36    12      36    13      36    14      36    15      36    16      36    17      36    18      36    19      36    20      36  Mean = 36.000000 Std Dev. = 0.000000  Two Way Frequencies:  Item|       1       2       3       4       5       6       7       8       9      10      11      12      13      14      15      16      17      18      19      20 -----+----------------------------------------------------------------------------------------------------------------------------- -----------------------------------     1|      36       7       8       8       7       9       8       8       8       8       7       7       8       7       7       7       7       8       7       8     2|       7      36       8       7       7       8       7       8       8       8      10       7       8       7       7       8       7       7       7       8     3|       8       8      36       7       7       8       8       9       7       8       8       7       7       8       7       7       7       8       7       8     4|       8       7       7      36       8       8       7       7       7       7       8       8       8       9       8       8       8       7       7       7     5|       7       7       7       8      36       7       8       8       7       8       7       7       8       8       8       8       7       8       8       8     6|       9       8       8       8       7      36       7       7       7       9       7       8       8       8       8       7       7       7       7       7     7|       8       7       8       7       8       7      36       7       7       8       7       7       7       8       8       8       8       8       8       8     8|       8       8       9       7       8       7       7      36       7       9       8       7       7       7       8       7       7       8       8       7 159      9|       8       8       7       7       7       7       7       7      36       7       8       8       7       7       7       8       8      10       8       8    10|       8       8       8       7       8       9       8       9       7      36       8       7       7       7       8       7       7       7       7       7    11|       7      10       8       8       7       7       7       8       8       8      36       8       8       7       7       7       8       7       7       7    12|       7       7       7       8       7       8       7       7       8       7       8      36       8       8       8       7       9       7       8       8    13|       8       8       7       8       8       8       7       7       7       7       8       8      36       7       8       8       8       7       8       7    14|       7       7       8       9       8       8       8       7       7       7       7       8       7      36       7       8       8       7       8       8    15|       7       7       7       8       8       8       8       8       7       8       7       8       8       7      36       7       8       8       8       7    16|       7       8       7       8       8       7       8       7       8       7       7       7       8       8       7      36       8       7       9       8    17|       7       7       7       8       7       7       8       7       8       7       8       9       8       8       8       8      36       7       7       8    18|       8       7       8       7       8       7       8       8      10       7       7       7       7       7       8       7       7      36       8       8    19|       7       7       7       7       8       7       8       8       8       7       7       8       8       8       8       9       7       8      36       7    20|       8       8       8       7       8       7       8       7       8       7       7       8       7       8       7       8       8       8       7      36  Off Diagonal Elements (not adjusted for prohibitions) Mean = 7.578947 Std Dev. = 0.616936  Positional Frequencies:           | Pos.       1       2       3       4       5 ----------+--------------------------------------------- Item     1|            7       7       7       8       7          2|            7       8       7       7       7          3|            8       7       7       7       7          4|            7       7       8       7       7          5|            7       8       6       7       8          6|            7       8       7       7       7          7|            7       7       7       7       8          8|            7       7       7       7       8          9|            7       7       8       7       7         10|            7       7       7       7       8         11|            8       7       7       7       7         12|            7       8       7       7       7         13|            7       7       7       8       7         14|            7       7       7       8       7         15|            8       8       7       6       7         16|            7       7       8       7       7         17|            7       7       7       8       7         18|            8       7       7       7       7         19|            7       7       8       7       7         20|            7       6       8       8       7 Mean = 7.200000 Std Dev. = 0.469042  160  Note: Test design tests the current design on file.  It does not reflect changes made to the MaxDiff settings dialog since the design was generated or imported.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0074259/manifest

Comment

Related Items