INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. ProQuest Information and Learning 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 800-521-0600 UMI” EFFECTS OF THE BRITISH COLUMBIA PUBLIC HEALTH OFFICER'S HEALTH DETERMINANTS ON THE HEALTH UTILITY INDEX AND THE RICHARDSON-ZUMBO HEALTH PROFILE by Michael A. Leisinger M.B.A., McMaster University, 1982 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in COMMUNITY HEALTH © Michael A. Leisinger, 2001 THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA March 09,2001 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without the permission o f the author. 1 ^ 1 National Libraiy of Canada Bü)liolhèque nationale du Canada Acquisitions and Bibliographie Sen/ices Acquitôions et sendees bibliographiques aSSWtünglonStrMt 38S,rueWaWngton CMu m ON K1A0N4 Cwud> OttnraON K1A0N4 Canada YaurI»» VaetréHm ie t O u r m N a m tm itn e » The author has granted a non­ exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell copies of this thesis in microform, paper or electronic formats. L’auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/film, de reproduction sur papier ou sur format électronique. The author retains ownership of the copyright in this thesis. Neidier the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author’s permission. L’auteur conserve la propriété du droit d’auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation. 0-612-62482-X CanadS APPROVAL Name: Michael A Leisinger Degree: Master of Science Thesis Title: EFFECTS OF THE BRITISH COLUMBIA PUBUC HEALTH OFFICER’S HEALTH DETERMINANTS ON THE HEALTH UTILITY INDEX AND THE RICHARDSON-ZUMBO HEALTH PROFILE Examining Committee: ^ Chair: Dr. Gordon Martel, Asst. (Graduate Studies) to the Vice-President (Academic) Professor and Chair, History Program UNBC G X Supervisor: Dr. Alex Michalos Professor, Political Science Program UNBC Committee Member: Dr. Bruno Zumbo Adjunct Professor, Psychology Program UNBC Committee MetAer: Sylvia Barton, MSc. Assistant Professor, Nursing Program UNBC ExternallEExxam amin iner^ er(^^ laZ la Zimmer, i BSN, PhD (c) Assistant Professor, Nursing Program UNBC Date Approved: "WW. 9, 200 ! ABSTRACT Richardson and Zumbo questioned the ability o f any single score to measure the health status o f a population, suggesting a better approach would be to use a multi­ dimensional health profile instead. The purpose o f this thesis was to examine what effects, if any, determinants of health, identified by the British Columbia Provincial Health Officer and a literature search, would have on the Richardson-Zumbo Health Profile and the Health Utility Index (HUI) in terms of the ability o f those measures to be sensitive to underlying changes in the population’s health status. Would a profile yield more usefhl information than a summary score? Data came from the 1994/95 National Population Health Survey (NPHS) over­ sample for the population of Prince George, B.C., consisting of 838 randomly selected individuals (436 female, 402 male). The key health determinants included income level, educational attainment, employment status, single parenthood, tobacco use, alcohol consumption, gender and age. Indicators closely paralleling these determinants were selected from the 1994/95 NPHS. The five Richardson and Zumbo factors (physical impairment factor, mental illhealth factor, mental well-being factor, general health impairment factor, and social well­ being factor) were combined into a Composite Score to create an additional dependent variable. A bivariate analysis was done between the six non-dichotomous NPHS indicators and the seven dependent variables. Correlations and tests for significance were perfonned. 11 A multivariate analysis was then done. Beta values and the Pratt Index for those model predictors identified through stepwise regression were calculated. The multivariate analysis was re-run with the addition of the age and gender NPHS indicators. Beta values and the Pratt Index for those model predictors identified through stepwise regression were calculated. The results of the analysis yielded no surprises. As expected, being employed, greater income adequacy, and less tobacco consumption were all associated with a higher state of health. Gender was not a significant health determinant except on the social well­ being factor where there was an apparent advantage to being female. Overall, the of the multivariate analyses were disappointingly low ranging from .017 to .212 on the five Richardson and Zumbo Factors, .097 for the Composite Score a n d . 123 for the HUI. The gain in R^ when age and gender were added was also minimal ranging from nil to +.058. The net result of these regressions seems to be that there is a poor fit between the determinants of health and population health status. The HUI was hardly describing population health while the five Richardson-Zumbo scores and the Composite Score fared little better. This analysis, once again, demonstrates the difficulty o f capturing the complex interplay o f the myriad o f variables that form the construct o f health. The challenge to future researchers is to continue to explore profiles that accurately capture the status of the population’s health and that are also sensitive to underlying changes as they occur. Ill Although the Richardson-Zumbo Health Profile did not have strong explanatory power it was, never the less, able to yield more information than the HUI or the Composite Score. IV TABLE OF CONTENTS Abstract Page ii Table of Contents Page V List of Tables Page vi List of Figures Page vii Acknowledgments Page viii Page I Specific Determinants Page 11 National Population Health Survey Page 23 Health Utility Index Page 25 Richardson-Zumbo Health Profile Page 29 Chapter 2 Methods Page 34 Chapter 3 Results Page 44 A. Bivariate Analysis Page 44 B.l Multivariate Analysis Without Age and Page 51 Page 55 Page 60 Page 65 Page 70 Page 83 Page 84 Chapter 1 Health, Health Status and the Determinants of Health Gender B.2 Multivariate Analysis Including Age and Gender Chapter 4 Conclusion References Appendix A 17 National Population Health Survey Variables Selected for Exploratory Factor Analysis Appendix B The 5 Richardson-Zumbo Factors with Corresponding Indicator Loadings Appendix C Correlations Between Dependent Variables LIST OF TABLES Table 1. Determinant of Health Categories and Indicators Page 7 Table 2. Health Determinant and Health Determinant Category Page 11 Table 3. Richardson-Zumbo Factors, Indicators and Loading Page 30 Table 4. Correlation ofHealth Utility Index with Richardson-Zumbo Page 32 Factors Table 5. National Population Health Survey Indicators and PHO Page 34 Health Determinants Table 6. Reporting n, Means, and Standard Deviations for the Page 46 Dependent Variables Table 7. Bivariate Correlations Page 50 Table 8. Multivariate Analysis Where Only the Model Predictors Page 54 Page 59 Identified Through Stepwise Regression are Indicated Table 9. Multivariate Analysis Where Only the Model Predictors Identified Through Stepwise Regression are Indicated (Plus Age and Gender) VI LIST OF FIGURES Figure 1. Population Health Determinants Flow Diagram Page 10 Vll ACKNOWLEDGEMENTS This work is dedicated to the memory o f Dr. David Fish (1929-2000) Former Dean of the Faculty ofHealth and Human Sciences University of Northern British Columbia Special thanks go to the following: My wife, Barbara, and two sons, Matthew and Erik, for their patience, understanding, and support over the past seven years despite the family time sacrificed to the program. The Terrace and Area Health Council and the Northern Interior Regional Health Board for their encouragement and financial support. My parents. Max and Gisela Leisinger, for emphasizing the value of education and the importance of life-long learning. Margaret (Dediluke) Warcup for talking me into registering. The Terrace cohort for the fun times. Alex Michalos and Bruno Zumbo for their help and guidance. The many unnamed others who helped along the way. Vlll CHAPTER 1 The Health Utility Index endeavours to capture the state o f a population’s health through a single, summary, numeric measure analogous to the Gross National Product’s (CNF) description of the national economy. The health of a population is influenced by many factors commonly referred to as determinants of health. The British Columbia Provincial Health Officer has grouped determinants of health into five broad categories and has suggested indicators that measure those determinants. Richardson and Zumbo proposed a multi-dimensioned health profile to describe the health status of a population. Using 1994/95 National Population Health Survey over-sample data for the population of Prince George, British Columbia, this thesis will investigate the sensitivity of the Health Utility Index and the Richardson-Zumbo Health Profile to a set o f key health determinants considered important by the British Columbia Provincial Health Officer. This first chapter will examine the literature with respect to the major issues discussed in the thesis. Health, Health Status and the Determinants ofHealth In the 1940’s, in Canada, data were not available that described or quantified the health status o f the population. It was not until the Sickness Survey o f 1950/51, conducted by the Dominion Bureau of Statistics, that statistical evidence based on a survey o f 40,000 households conflrmed the generally held impression that Canadians were not healthy. Self-reported levels o f illness (complaint, disability, bed rest or care) were considered unacceptably high. The poorest population cohort bore twice the disability burden of the wealthiest population cohort, raising questions o f whether this was linked with the observation that the wealthy spent 2.8 times more on health expenditures than the poor (Taylor, 1987). At the time of the 1950/51 Sickness Survey the indicator considered the most sensitive by the Dominion Bureau of Statistics was the infant mortality rate' (Taylor, 1987), not just as a measure of child health but also of the social well-being of society (Michalos, 1980^; British Columbia Provincial Health Officer, 1998). Health was largely considered the absence of disease following the biomedical disease model\ Indeed, in the past, infectious disease'* was the greatest cause of illness and death thus making the linkage between health and the absence of sickness an obvious and natural one (Epp, 1986). Evans and Stoddart (1994) observe the existence of the huge healthcare (sickness care) industry is evidence of a society which equates the use o f healthcare with health. Bergner (1985) states that until the early 1960’s mortality rates were the most relevant and sensitive measures of the health o f a population. By the mid-1960's, however, death rates no longer seemed to be sensitive to the changes that were taking place in health and healthcare. The emphasis had, instead, started to shift to measures of morbidity and quality of life to assess health status (Bergner, 1987). A significant departure from the biomedical disease model came in 1974 with the publication o f “A New Perspective on the Health o f Canadians” by the Hon. Marc ‘ In 1901 there were 134 deaths for every 1000 population under the age of one compared with 5.5 deaths/1000 in 1997 (Statistics Canada, 2000). Japan’s rate in 1996 was 3.8 deaths/1000 (Federal, Provincial and Territorial Committee on Population Health, 1999). ^ The infant mortality rate “reflects trends in general mortality, public health, sanitation, housing, and of economic development, as well as practices of infant feeding and care because infants more than any other group depend so completely on environmental conditions and the attention of others for their survival” (p. 131). ^ Cause (etiology) is followed by lesion (pathology) which is followed by symptom (outcome) (Taylor, 1987). ' NiiuToruk (1991) offers a fascinating historical oveniew of epidemics and plagues. Lalonde, the then federal Minister ofHealth. The Lalonde Report, as it is now known, took a broader view of health than just the diagnosis, treatment and cure of disease. Lalonde (1974) put forward the Health Field Concept. It divided health into four areas for analysis and evaluation: human biology, environment, life-style, and healthcare organization. Using the Health Field Concept “any health problem [could] be traced to one, or a combination of the four elements^” (p. 33). The Lalonde Report in Canada was followed by similar reports in other countries (Taylor, 1987; Evans and Stoddart, 1994). The common theme among these reports was the idea that health could be improved without continued massive cash infusions into the traditional healthcare system (Marmor, Barer and Evans, 1994)^. The Lalonde Report is widely recognized as having had an important effect on the way health is viewed, both in Canada and internationally, by focusing on the health status o f the population rather than the healthcare system alone^ (Taylor, 1987; Evans and Stoddart, 1994; Sutherland and Fulton, 1994; Federal, Provincial and Territorial Committee on Population Health, 1994; Health Canada, 1996, 1998; Institute ofHealth Promotion Research, University of British Columbia, 1999). The World Health Organization proposed a similar multidimensional approach to health in 1977 and unanimously adopted its “Health For All by the Year 2000" strategy in 1 9 8 l\ recognizing that health does not exist in isolation but is influenced by ^ An example is given by Lalonde (1974) whereby lifestyle, environment and healthcare organization were said to contribute to trahie deaths in the proportions of something like 7S%, 20% and S% respectively ( p. 33). * Poland, Cobum, Robertson and Eakin (1998) express a concern that the notion of additional healthcare expenditures producing modest improvements in health status, as compared to efforts made in other health determinant areas “provides convenient cover for those who wish to dismantle the welfare state in the name of deficit reduction” (p. 786). ^ Critics point out, however, that the report was misinterpreted by many resulting in government adopting a narrow lifestyle focus (e.g., Participaction) while largely ignoring other determinants of health (Poland et al., 1998). ' TTie “Health For All by the Year 2000” strategy was adopted by the World Health Assembly and endorsed by the United Nations General Assembly that same year. environmental, social and economic influences which interact with each other in a complex fashion (World Health Organization, 1981; Taylor, 1987; World Health Organization, Health for All Web-Site). Health is deflned in the preamble to the constitution o f the World Health Organization as “a state of complete physical, mental, and social well-being and not merely the absence o f disease or inflrmity" (Siddiqui, 1995, p. 226). Originally adopted in 1948, the World Health Organization definition continues to be widely accepted by most healthcare planners, authorities and policy makers (e.g., British Columbia Provincial Health Officer, 1994), and is cited in most of the literature when health is deflned. This deflnition o f health was reaffirmed by the World Health Organization in 1978 at Alma Ata and again in its "Global Strategy for Health for All by the Year 2000" (Koivusalo and Ollila, 1997; World Health Organization, 1998). The deflnition is also posted prominently on the World Health Organization web-site. Beigbeder (1998) points out the World Health Organization constitution raises the "enjoyment o f the highest attainable standard of health... to the level of one of the fundamental human rights” and that since "the health of all peoples is fundamental to the attainment o f peace, global health is therefore one o f the requirements for world peace” (p. 13). Hansluwka (1985) reviews the international debate over the World Health Organization deflnition of health and whether the deflnition can ever be achieved or does it, rather, serve as a lofly ideal to be striven for but never actualized. Difficulties in defining health arise out of the "vagueness of the concept, the value judgment of the deflner, the multidimensionality o f the phenomenon and the impossibility of meaningful operationalization” (p. 1208). As The Institute ofHealth Promotion Research, University of British Columbia (1999) claims, such definitions seem to classify all human activity as being health-related. Evans and Stoddart (1994) comment that such definitions are, “honoured in repetition, but rarely in application” (p. 28). Bergner (1985) notes that many definitions of health abound but most are variants of the World Health Organization declaration. Population health is defined by the Federal, Provincial and Territorial Committee on Population Health (1999) as follows; Population health refers to the health o f a population as measured by health status indicators and as influenced by social, economic and physical environments, personal health practices, individual capacity and coping skills, human biology, early childhood development, and health services. As an approach, population health focuses on the interrelated conditions and factors that influence the health o f populations over the life course, identifies systematic variations in their patterns of occurrence, and applies the resulting knowledge to develop and implement policies and actions to improve the health and well-being o f those populations. The goal of a population health approach is to maintain and improve the health status of the entire population and to reduce inequities in health status between groups. This requires a thorough, ongoing examination o f both health status and the factors that determine or influence health, (pp. 7-8) The Institute ofHealth Promotion Research, University o f British Columbia (1999) cites the above quoted definition and adds that population health research deals with whole communities or populations and not just individuals or risk groups; looks to greater intersectoral action beyond the health field; seeks to make populations less dependent on health services and professionals; and looks to the social world for determinants of health. Identifying determinants of health, as opposed to the determinants of illness and disease, is referred to by Catlin and Will (1992) as identifying the risk factors for chronic good health. In 1986 the then Federal Minister of National Health and Welfare, Jake Epp, released the document “Achieving Health For All: A Framework For Health Promotion” at the annual conference of the Canadian Public Health Association. Building on the Lalonde Report, the Epp Report also recognized the multidimensionality of health with an emphasis on reducing inequities, increasing prevention efforts, and enhancing people’s capacity to cope (e.g., chronic conditions, disabilities, and mental health problems). Improvements in health would be achieved through a health promotion framework consisting of three mechanisms: self-care, mutual aid, and healthy environments. Specific implementation strategies were: fostering public participation, strengthening community health services, and coordinating healthy public policy. This fhunework was not designed to replace the existing healthcare system but to work with it, suggesting the potential o f slowing the growth in healthcare costs. Each year the British Columbia Provincial Health Officer issues an annual report that has a central theme or focus for the year included along with the standard tables and statistics reported each year. The 1994 Annual Report concentrated on the determinants of health. The report promoted the view that the health o f a population is influenced by more factors than the healthcare system alone. It grouped the determinants o f health into five broad categories: social and economic environment, the physical environment, health services, biological influences, and health behaviours and skills. Within each o f the five determinants o f health categories the Provincial Health Officer listed several health indicators used to measure the state of the category. These are displayed in Table 1. Determinant of Health Indicator Infant mortality rate Income disparity ratio Income adequacy Poverty rate Low-income rate Income assistance rate Single-parent family rate Unemployment rate Highest level of education attained Child care Percent of income spent on housing Air pollution levels Physical Environment Ultra-violet radiation levels Biological Influences Birth defect rates Immunization rates Alcohol use during pregnancy Health Behaviours and Skills Tobacco use Alcohol consumption Cholesterol level Hypertension Diet Obesity Inactivity Stress Illicit drug use Accident rate Standardized mortality ratio Health spending Health Services Cancer screening rates Hospital utilization rates (British Columbia Provincial Health Officer, 1994) Social and Economic Environment There is no unanimous agreement on what the determinants o f health are or their relative ranking, but those enumerated by the British Columbia Provincial Health Officer are consistent with the literature as being central to health. There is, however, growing emphasis on self-esteem, social support networks, quality o f life issues, early child development, and the role of gender and culture (Mustard and Frank, 1994; Sutherland and Fulton, 1994; Evans and Stoddart, 1994 (see Figure 1); Decter, 1994; Federal, Provincial and Territorial Committee on Population Health, 1994,1999; Frank, 1995; Health Canada, 1996’; Institute ofHealth Promotion Research, University o f British Columbia, 1999; Denton and Walters, 1999; Edwards, 2000). There can also be many individual health indicators that contribute to a given health determinant. The indicators also interact with one another, just as the determinants are not isolated from one another. The complexity of the multiple interactions o f the indicators and the determinants combined with an illusive definition of health makes the quantification o f health status so very difficult. One o f the recommendations out o f the 1994 British Columbia Provincial Health Officer’s Annual Report was “to continue efforts to develop and collect the best possible indicators to measure the health of the population” (British Columbia Provincial Health Officer, 1994, p. 71). Sutherland and Fulton (1994) make an interesting observation with respect to individual versus population health. Healthcare spending and services directed at the individual are easy to measure and evaluate as opposed to policies put in place to help a population. A physician can deal with an individual by using the medical model o f diagnosis, treatment and cure. For a population, however, many of the contributors to health are outside the health domain, e.g., sewers, divided highways, adequate street lighting, etc., and are often harder to recognize as contributors to health even though the ’ Health Canada (1996) had a view of the health déterminant categories similar to that of the British Columbia Provincial Health Officer. The following list was used as a starting point for ftinire population health policy and research directions: income and social status, social support networks, education, employment working conditions, social environments, physical environments, biology and genetic endowment personal health and coping skills, health services, and three new areas; healthy child development gender, and culnire. impact per dollar spent may be greater than service delivered at the level of the individual. Perhaps as a result of the difficulties in understanding population health versus individual health the 1994 British Columbia Provincial Health Officer’s Annual Report also recommended the development of “information and educational materials to increase public understanding about the determinants o f health; build understanding about the determinants of health and support for the population health approach among government parmers in sectors outside health” (p. 71). In 1995, the year after the Provincial Health Officer’s report on the determinants of health, the British Columbia Ministry ofHealth directed Health Authorities to consider the determinants o f health in the preparation o f their health and management plans (British Columbia Ministry ofHealth, 1995).'° As a specific example to Health Authorities the British Columbia Ministry ofHealth in its “Guide to Health and Management Planning for Regions and Communities” suggested the socioeconomic determinants of health contribute 50% towards the health stanis of a community (e.g., income, employment, working conditions, education, housing, distribution of income and social support); illness treatment 25%, genetics 15%, and the physical environment 10% (British Columbia Ministry ofHealth, 1995). Figure 1. Population Health Determinants Flow Diagram Social Environment Individual Responses - Behaviour - Biology Physical Environment Health and Function Disease Genetic Endowment Healthcare ii Well-Being Prosperity (Evans and Stoddart, 1994, p. S3) based on work done by the Canadian Institute for Advanced Research. Note how healthcare is primarily shown as a response to disease (Frank, 1995). Poland et al. (1998) agree with those who contend the box labeled prosperity would be more properly called equity to get at the concept of income distribution as opposed to absolute levels of income. Specific Determinants This thesis will be focusing on specific health status determinants from the 1994 British Columbia Provincial Health Officer’s Annual Report. Table 2 indicates the determinants that will be examined further and to which broader determinant of health category they correspond. Table 2. Health Determinant and Hea th Determinant Category Health Determinant Health Determinant Category Income Level Social and Economic Environment Educational Attainment Social and Economic Environment Employment Status Social and Economic Environment Single Parenthood Social and Economic Environment Tobacco Use Health Behaviour Alcohol Consumption Health Behaviour Gender Not specifically linked with an individual health determinant but interacts with the other determinants a) Income Level The literature shows a definite linkage between income and health. Overall mortality and most forms of morbidity follow a gradient worldwide across socioeconomic classes such that lower income and lower social status is associated with poorer health (Mustard and Frank, 1994; Evans and Stoddart, 1994; Denton and Walters, 1999). Canadians with low incomes are more likely to suffer illnesses and die earlier than Canadians with high incomes" (Epp, 1986; British Columbia Provincial Health Officer, 1994; Federal, Provincial and Territorial Committee on Population Health, 1999). In 11 general, wealthier populations and countries are healthier than poorer ones (Sutherland and Fulton, 1994) and virtually no examples of any society, past or present, are evident where overall health status is inversely related to wealth, income or social class (Hertzman, Frank and Evans, 1994). Denton and Walters (1999) state, “poor health is not simply concentrated among those who are most deprived. Health status declines with each decline in socioeconomic status" (p. 1222). The British Columbia Provincial Health Officer (1994) went so far as to state that income level and social status seem to be the most important determinants of health. The Federal, Provincial and Territorial Committee on Population Health (1999) did not go quite so far, acknowledging that there is no consensus on which is the best measure of socioeconomic status, noting that some researchers prefer to use education level or occupation. In their report, however, income was used as a proxy for socioeconomic status in most cases. Many possible explanations are given for the link between higher income levels and increased health. For example, higher incomes allow people to purchase adequate housing'^, food, and other basic needs. Meeting or exceeding basic needs allows for greater security, more control over decision making, and improved supportive social networks which may in turn lead to a more nurturing environment for children, success in school, and so on (British Columbia Provincial Health Officer, 1994). This is not a case of people being unable to work because of illness and thus unable to earn higher incomes, but, rather, that low economic status leads to exposure to unhealthy life conditions and ' ‘ “Canadian men in the highest quarter of income distribution can expect to live 6.3 years longer and 14.3 more years free of disability than those in the lowest quartile. For women, the differences are 3.0 and 7.6 years respectively” (Federal, Provincial and Territorial Committee on Population Health, 1999, p. 26). 12 thus poorer health and earlier death (Federal, Provincial and Territorial Committee on Population Health, 1999). It is noted that “people with very low incomes are more likely to smoke, drink alcohol to overcome stress, take sleeping pills, are less likely to have regular pap smears or to know the causes of heart disease” (British Columbia Provincial Health Officer, 1994, p. 43) and are more likely to go to hospital (Statistics Canada, 2000). Privilege and increased self-image (Sutherland and Fulton, 1994) accompany higher incomes and social status. The Whitehall study o f civil service employees in Britain found that health generally increased with job rank, causing the researchers to conclude that “something related to higher income, social position and hierarchy provides a buffer or defense against disease, or that something about lower income and status undermines defenses” (Federal, Provincial and Territorial Committee on Population Health, 1994, p. 14). Not only is the level of income significant but it appears to be of even greater importance how equitably wealth is distributed amongst a population, that is, the gap between the rich and the poor (Mustard and Frank, 1994; Decter, 1994'^; Hertzman, Frank and Evans, 1994*^; Frank, 1995; British Columbia Provincial Health Officer, 1997; Judge, Mulligan and Benzeval, 1998*^; Poland et al., 1998; Federal, Provincial and Most low-income people are renters, while those in higher income brackets are home owners (British Columbia Provincial Health Officer, 1997). " The poorest 20% of the population would gain 13 additional disability free years if their socioeconomic status was the same as the top 40% of income earners. Japan, which has the highest life expectancy in the world, has the smallest relative difference between the average incomes of the richest and poorest 20% of the population of any OECD country. Some, however, point to the economic success of Japan as the reason for the rise in life expectancy, once again highlighting the difficulty in disentangling the intertwining factors which contribute to health stanis (Frank, 1995). Their paper reviews twelve studies (some of which were also reviews of multiple studies) the primary focus of which was the relationship between measures of income inequality and average levels of population health. All but two studies found evidence of an association. Even so, Judge, Mulligan and Benzeval (1998) are not convinced that a definite association exists citing flawed smdy design, choice of 13 Territorial Committee on Population Health, 1999). The greater the disparities between rich and poor, the greater the health consequences. This linkage seems to be constant over time and as the diseases that are responsible for mortality change. One disease merely replaces another and the social gradient remains intact (Frank, 1995). The Gini Index is a measure of income inequality. The larger the Gini coefficient, the greater the inequality in income distribution in a range between 0, representing equal income for everyone, and 1, representing complete concentration in a single person (Michalos, 1982). According to the Federal, Provincial and Territorial Committee on Population Health (1999), the Gini Coefficient for the income distribution (after taxes) of families in Canada fell from 0.316 in 1970 to 0.300 in 1995. In an interesting and controversial study done by Ross, Wolfson, Dunn, Berthelot, Kaplan and Lynch (2000) a review of fourteen articles supports the association between income distribution and health status. They specifically compared the relation between mortality and income inequality using census and vital statistics data o f 53 Canadian cities in ten provinces and, 282 American cities in 50 states. In the United States there was greater mortality in those areas of greater unequal income distribution while in Canada no such similar association was found. The authors theorize that there may not be an automatic association between income inequality and mortality in jurisdictions, such as Canada, were social policy, such as public funding and universal access, replaces ability to pay. Statistics Canada (2000) notes that the findings of this study run counter to research to date in the United States and internationally. The distribution and income measures and questionable data manipulation. The authors do not discount the existence of a possible association and so produced their own study with a view of not replicating the errors of their predecessors. The study’s results caused the authors to conclude that income inequality is not a significant 14 concentration of wealth, two concepts closely linked to income but different from each other, was recommended by the Federal, Provincial and Territorial Committee on Population Health (1999) as an area of further research. Self-rated health status has been shown to be a reliable predictor o f health problems, healthcare utilization, and longevity. There also exists a gradient in self-rated health strongly linked to income. Canadians from the lowest income households were four times more likely to report fair or poor health than those who lived in the highest income households (Federal, Provincial and Territorial Committee on Population Health, 1999). Segovia, Bartlett and Edwards (1989) concluded that self-rated health status is a good summary indicator of health status. In Canada there continue to be income-related disparities in both infant mortality'^ and low birthweight, and a strong relationship between perceived health status and socioeconomic status (Statistics Canada, 2000). Sutherland and Fulton (1994) and Mustard and Frank (1994) write that equal access to healthcare does not improve health, citing the fifty years o f experience with the National Health Service in Great Britain which, although it provides even more comprehensive coverage than Canada’s plan, has seen a widening in the health status gap between the rich and the poor. Health Canada (1996) acknowledges universal access to health services in Canada has also not managed to eliminate or even reduce health disparities. but only modest determinant of population health in rich industrialized countries for which good income distribution data are available. According to Statistics Canada (2000) income related difierences in infant mortality in 1996 were nearly three times as large as regional differences. 15 b) Educational Attainment Higher levels of educational attainment relate directly to greater health in terms of higher self-rated health status, greater positive health behaviours, decreased activity limitation, increased opportunities for income and job security, and generally a greater sense o f well-being (British Columbia Provincial Health Officer, 1994; Health Canada, 1996; Federal, Provincial and Territorial Committee on Population Health, 1999). People with less than a secondary school education are more likely to go to hospital than people with higher levels of educational attainment (Statistics Canada, 2000). Having less than a grade nine education is considered a proxy measure for illiteracy and limited education (British Columbia Provincial Health Officer, 1994). In the 1996/97 National Population Health Survey, only 19% of respondents with less than a high school education rated their health as "excellent" compared with 30% o f university graduates. People with low literacy skills often feel alienated, have difficulty finding and accessing health information and services, and have reduced employment opportunities. As a result, they suffer poorer health than those who have higher literacy skills. Among Canadians with the lowest levels of prose literacy, 47% lived in low-income households, compared with 8% o f Canadians with the highest levels of prose literacy (Breen, 1998; Federal, Provincial and Territorial Cotnmittee on Population Health, 1999). Generally, income levels rise with greater educational attainment. People with limited educational attainment have higher unemployment rates and lower employment participation rates than those with higher levels of education. University graduates in Canada experienced one-third the unemployment o f people with less than a high school education, and had over three times the level o f employment participation rate (Federal, 16 Provincial and Territorial Committee on Population Health, 1999). Chappell (1998) writes, “education is emerging as a key tool for grassroots empowerment, as individuals must have the capacity to seek and acquire information, and they must possess the analytic skills to ferret through that information in order to make informed choices” (p. 90). Educational attainment is linked to employment, which is linked to income level; all three are important determinants of health. It is difficult to disassociate one health determinant from another, c) Employment Status Mustard and Frank (1994) and Avison (1998) report on studies done in the United States, Denmark, and by the World Health Organization that all conclude mortality (including suicide and death by accidents), and morbidity (mental and physical illhealth), increase with unemployment. Specifically, unemployed persons exhibit greater psychological distress, anxiety, depressive symptoms, panic, substance abuse, disability days, health problems, and hospitalizations than those who are employed. Employment provides not only money, but also a sense o f identity and purpose, social contacts and opportunities for personal growth. When unemployed, the effects on health go beyond the person who is unemployed but also extend to the family unit and the community in general. Those negative impacts are not immediately reversed upon reemployment (Hunt, McEwen and McKenna, 1986; Mustard and Frank, 1994; Sutherland and Fulton, 1994; British Columbia Provincial Health Officer, 1994'^, 1997; The British Columbia Provincial Health Officer (1994) also comments that the northern regions of the province generally have the highest unemployment rates in British Columbia. 17 Health Canada, 1996‘*; Avison, 1998; Denton and Walters, 1999*®; Federal, Provincial and Territorial Committee on Population Health, 1999). d) Single Parenthood The proportion of families headed by lone (single) parents is considered an indicator of socioeconomic conditions. "The living conditions o f single-parent families have been associated with a number o f problems, including poor housing conditions, behavioral problems in children, overload o f parental responsibilities^**, loneliness, dissatisfaction with social situation^', and health problems” (British Columbia Provincial Health Officer, 1994, p. 28). In 1995, almost 50% of single-parent mother-led families were in low-income situations (Federal, Provincial and Territorial Committee on Population Health, 1999). Statistics Canada calculates poverty or straitened circumstances, also called the lowincome cutoff, as being when a family spends more than 56% o f its income on food, shelter, and clothing. In 1991, and again in 1996, almost one child in five under six years o f age lived below the low-income cutoff (British Columbia Provincial Health Officer, 1994, 1998), 43% of whom lived in female lone-parent families (British Columbia Provincial Health Officer, 1994); 59% lived in lone-parent families headed by either gender (British Columbia Provincial Health Officer, 1998). Health Canada (1996) grouped underemployment and stressful work in the same category as unemployment. ” The literature explores more than the employed / unemployed dichotomy also examining degree of job security, full-time versus part-time employment, type of shifts worked, decision making latitude, psychological demands. Denton and Walters (1999) comment on the stress arising from women’s unpaid work in the home especially when coupW with participation in the paid workforce. Denton and Walters (1999) refer to the association between health and social support today being as compelling as the association was between health and tobacco use in the I960’s. 18 e) Tobacco Use According to the British Columbia Provincial Health Officer (1994) smoking is the leading preventable cause of death in the province, accounting for one-fifth of all deaths in the province. As a cause of early death, smoking far outweighs suicide, motor vehicle crashes, AIDS and murder combined (Federal, Provincial and Territorial Committee on Population Health, 1999). Smoking is a known risk factor for heart disease, stroke, cancer, chronic obstructive pulmonary disease, diabetes, and birth defects (Federal, Provincial and Territorial Committee on Population Health, 1994). Smoking among women is linked to lower fertility, cancer of the cervix, osteoporosis, and menstrual and menopausal problems; exposure to second-hand smoke is linked to breast cancer (Health Canada Web-Site, Women’s Health Bureau, 2000). Denton and Walters (1999) cite the argument that the most disadvantaged women smoke due to the tensions in their lives caused by their disadvantaged state. Smoking is a means by which they cope with their day to day stress. People with very low incomes are more likely to smoke (British Columbia Provincial Health Officer, 1994) and smokers are more likely to be hospitalized than nonsmokers (Statistics Canada, 2000). Overall, men are still more likely than women to smoke and to smoke heavily (Federal, Provincial and Territorial Committee on Population Health, 1999). Evans and Stoddart (1994) issue a warning, however, to be aware o f so-called lifestyle determinants o f health. The implication that tobacco use is an individual choice may lead to victim blaming and obscure the observation that the use o f tobacco is a 19 strongly socially conditioned practice. Smoking was once a sign o f status, an activity engaged in by the rich and famous, whereas there is now a strong negative correlation contributing to the social gradient now observed with respect to mortality and income level. Marmor, Barer and Evans (1994) also remind us that individual choice when applied “to the consumption of a toxic substance that is also addictive, and to which people typically become addicted during early adolescence” (p. 223) is particularly inappropriate, f) Alcohol Consumption Excessive alcohol consumption can lead to a range of health and social problems. Drinking alcohol during pregnancy has been linked to lower birth weights and other negative outcomes (Federal, Provincial and Territorial Committee on Population Health, 1994). Alcohol consumption increases with income; people in higher income brackets tend to be heavier drinkers. Lower income earners are less likely than upper income earners to consume any alcohol at all. However, among lower income earners who do drink alcohol, their rate o f heavy drinking tended to slightly exceed that o f higher income earners (Federal, Provincial and Territorial Committee on Population Health, 1999). The British Columbia Health Officer (1994) states that people with very low incomes are more likely to drink to overcome stress. 20 g) Gender^ The most basic health indicator difference between men and women is life expectancy. According to the Federal, Provincial and Territorial Committee on Population Health (1999) a male Canadian child bom in 1996 could expect to live to age 75.7 years; 81.4 years for a female child. Men are far more likely than women to die before age 70, mainly because of gender differences in deaths due to heart disease, cancer, suicide and unintentional injuries. Rates of potential years of life lost are almost twice as high for men than women. Suicide rates among young men are high in Canada, compared to other countries. Boys and young men tend to experience more unintentional injuries and more severe injuries than girls and young women. Although living longer, women are more likely to suffer from long-term activity limitations and chronic conditions such as osteoporosis, arthritis and migraine headaches. Young women are particularly likely to feel depressed. The British Columbia Health Officer’s Annual Report for 1995 (1996) included a special report on women’s health. Specific observations included: • Women are poorer than men. Women earn 70% o f what men earn (true worldwide not just in British Columbia). Women earn less than men in all occupational categories. Women’s earnings are lower than men’s whatever their educational qualifications. • Lone-parent families headed by women have the lowest incomes of all family types. ^ Health Canada (1996) defines gender as “a social construct rooted more in human culture than biological differences between the sexes. Gender refers to the array of society-determined roles, personality traits, attitudes, behaviours, values, relative power and influence that society ascribes to the two sexes on a differential basis” (Appendix D). 21 • More than one woman in five aged 65 years and older is living below the lowincome cut-off. • Women who work outside the home continue to carry primary responsibilities for household duties. • Most unpaid informal caregivers (73%) are women. Despite these negative statistics, the health status of women, as measured by the indicators of life expectancy, is substantially better than that o f men. This is referred to as the women’s health paradox (British Columbia Provincial Health Officer, 1996). It may be partially explained by the fact that women tend to smoke and drink less, are less likely to be employed in risky occupations, and are protected from heart disease by naturally occurring estrogens. It is suspected that women benefit from better social support networks, better communication skills, a greater willingness to seek assistance, and a greater aptitude for caregiving (British Columbia Health Officer, 1996). 22 National Population Heaith Survey Based on a recommendation from the National Health Information Council in 1991, the National Population Health Survey was conducted in four data gathering periods between June 1994 and March 1995. The survey was conducted by telephone and obtained data from 26,430 households in every province and territory^^ with a final response rate of 88%. The survey was to be conducted every two years over the course of two decades in order to obtain longitudinal data. Eight Hundred Fifty households in Prince George were part of the 1994/95 survey. This was a one-time inclusion with no longitudinal follow-up planned (Statistics Canada, 1995; Tambay and Catlin, 1995). The stated objectives of the National Population Health Survey were to: • aid in the development of public policy by providing measures of the level, trend and distribution of the health status o f the population; • provide data for analytic studies that will assist in understanding the determinants of health; • collect data on the economic, social, demographic, occupational and environmental correlates of health; • increase the understanding of the relationship between health status and healthcare utilization, including alternative as well as traditional services; • provide information on a panel of people who will be followed over time to reflect the dynamic process of health and illness; • provide the provinces and territories and other clients with a health survey capacity that will permit supplementation o f content or sample; ^ The NPHS target population excluded persons living on Indian reserves, on Canadian Forces Bases and in some remote areas (Tambay and Catlin, 1995). 23 • allow the possibility o f linking survey data to routinely collected administrative data such as vital statistics, environmental measures, community variables, and health services utilization. (Statistics Canada, 1995, p. 6) Survey content was selected according to the following criteria: 1) Information should relate to, and help monitor, the health goals and objectives of the provinces and territories. Where health goals have not been established, for example, at the national level, policy and programs could be considered in the selection of survey content. 2) The information should not duplicate data available from other sources. 3) With a view to increasing the understanding o f health and its determinants, information collected should provide new knowledge in areas that have not been adequately studied. 4) The survey should focus on behaviours or conditions amenable to prevention, treatment, or intervention. 5) The survey should collect information about conditions that impose the greatest burden, in terms of suffering or cost, on affected individuals, the general population, or the healthcare system. 6) The survey should collect information on factors related to good health, not just those related to illness. (Statistics Canada, 1995, p. 7) Wolfson (1994) sees population health surveys such as the National Population Health Survey as an important beginning towards gathering self-reported data on the 24 health-related problems o f the individual and how those problems might relate to socioeconomic or cultural variations. Such new information and knowledge will be made even more powerful if it can be linked back to the existing traditional databases which contain vast detail on the individuals’ utilization of the healthcare system. Health Utility Index The literature of the mid-l980’s (Hansluwka, 1985; Bergner, 1985) commented that there was a shift away from individual health indicators towards the creation of health profiles and o f single aggregated indices for the measurement of health status. Such profiles and indices would be useful for the comparison o f groups across time. In order for them to be meaningful, however, the critical components of health would need to be identified and included, which assumes that health can be measured on a single continuum. Bergner (1985) calls this the single-continuum dilemma. Hansluwka (1985) was more pessimistic about the success of such an approach, stating that while specific views differ, the majority are inclined to agree that it is not possible to construct a single index of health capable of "summarizing the various aspects o f health in a way similar to the Gross Domestic Product concept” (p. 1208). Richardson and Zumbo (2000) note that interpretation o f a summary statistic would be problematic since improvement or worsening of individual components of the index would be hidden (see also Bergner, 1987). Wolfson (1994), however, points out that despite the flaws and imperfections o f the Gross National Product as a measure of the economy, no one suggests we would be better off without the index and concludes, “the best should not be the enemy o f the good” (p. 291) and as such the pursuit of a comprehensive health index continues. 25 One such aggregate index is the Health Utility Index. Richardson (1999) provides a concise history of the development of the Health Utility Index: The first index in the series, the HUI-Mark I, was designed to evaluate outcomes associated with neonatal intensive care of very low birthweight infants. Health status was classified using the following four attributes: physical function, role function, socio-emotional function and health problems, each with four to eight levels of functioning. The second index in the series, the HUI-Mark II, was developed for use in a cost-utility analysis of childhood cancer treatments (comparing aggressive with intent to cure versus palliative treatments). It described health status using the following set of attributes: sensation, mobility, emotion, cognition, self-care, pain, and fertility .... The third and most recent index, the HUI-Mark III, has been labeled a measure of functional health status for the general population. The eight attributes selected to describe health status for the HUI-Mark III are vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain^**.... Despite being extensively used to measure population health status for over five years, a multi-attribute utility function specific to the HUI-Mark III has yet to be developed. Instead, each o f the HUI-Mark III attribute scores have been translated into corresponding scores on the HUI-Mark II attribute system and the HUI-Mark II utility function then applied to generate a HUI-Mark III score. While the translation process was based on the best estimate of the McMaster research team, the differences between the two attribute classification systems 26 resulted in a number of compromises. First, the three HUI-Mark III sensation variables (vision, hearing and speech) had to be combined into a single sensation variable. Second, the closest counterpart to the HUI-Mark III dexterity attribute was the self-care attribute. Lastly, the HUI-Mark III does not contain an attribute comparable to fertility.... While the National Population Health Survey contains a detailed assessment of the psycho-social components o f health (e.g., social support, self­ esteem, perceived stress), the HUI-Mark III relies on an extremely limited selection o f health status indicators. O f the 31 questions used to derive the HUIMark III attribute scores, only three questions directly assess mental functioning and none appear to examine social functioning. The apparent paucity of items assessing mental and social well-being raise doubts concerning the ability of the HUI-Mark III to provide a valid assessment of health status for use in the National Population Health Survey. More specifically, it appears that the HUI-Mark III is primarily a measure of physical functioning incapable o f adequately assessing the mental and social dimensions of health. Given the dimensionality o f the World Health Organization definition of health and recent literature citing stress, self-esteem and social support as the most important factors in explaining today’s health gradients, it seems appropriate to incorporate indicators o f mental and social well-being into the summary measure o f health status used in the National Population Health Survey. The failure to include such indicators would not only limit the ability o f the summary Each of the eight functional attributes has six levels of classification (except speech, emotion and pain with only 5) ranging from no impairment to complete impairment. A full table is reproduced in Boyle, 27 measure to provide a comprehensive description o f health status, but more importantly, would substantially limit its ability to provide information on a large proportion of health determinants thought to achieve their effects through changes in mental and/or social functioning, (pp. 8-12) Contrast the description of the Health Utility Index provided by Richardson (1999) with the description provided by Statistics Canada (1995); The Health Utility Index is a generic health status index that is able to synthesize both quantitative and qualitative aspects o f health. The system developed at McMaster University's Centre for Health Economics and Policy Analysis, the Comprehensive Health Status Measurement System (CHSMS), provides a description of an individual's overall functional health, based on eight attributes: vision, hearing, speech, mobility (ability to get around), dexterity (use of hands and fingers), cognition (memory and thinking), emotion (feelings), and pain and discomfort. In addition to describing fimctional health status levels, the CHSMS is the basis for a provisional Health Utility Index (HI). The HI is a single numerical value for any possible combination of levels of these eight self-reported health attributes. The HI maps any one of the vectors of eight health attribute levels into a summary health value between 0 and I. For instance, an individual who is near-sighted, yet fully healthy on the other seven attributes, receives a score of 0.95 or 95% of full health. The Health Utility Index value also embodies the views of society concerning health status. These views are termed societal preferences, since Furlong, Feeny, Torrance and Hatcher (1995) and Richardson (1999). 28 preferences about various health states are elicited from a representative sample of individuals. The specific HI calculated here is provisional. The societal preferences were derived from the small-scale Childhood Cancer Study using a precursor of the CHSMS and were adapted for use with the Ontario Health Survey. Some adjustments were also made to the health attributes reported in the Ontario Health Survey. Consequently, the HI results are preliminary and approximate. This version of the CHSMS, however, was tested for consistency’^ and was deemed to provide a realistic appraisal of individual health status. (Statistics Canada, 1995, p. 28) This difference in outlook is what prompted Richardson’s (1999) review and statistical examination of the Health Utility Index as a summary measure o f health status for use in the National Population Health Survey. Richardson-Zumbo Health Profile Richardson (1999) followed by Richardson and Zumbo (2000) studied how well the Health Utility Index was able to describe the health status o f the population as a single summary (GNP-like) measure. Data were taken from the 1994.^95 National Population Health Survey over-sample of 838 residents of the Prince George / Northern Interior region of British Columbia. Seventeen variables from the 1994/95 National Population Health Survey were selected for exploratory factor analysis in order to see if, ^ Boyle et al. (1995) in a test-retest study of the reliability of the Health Utility Index-Mark HI state the reliability was substantial for the attributes of vision, ambulation and emotion; moderate for hearing, cognition and pain; speech and dexterity had the lowest estimates of reliability. Bergner (1987) in an examination of the McMaster Health Index (Questionnaire stated it could be expected to show a skewness of scores because of being designed to assess a dysfunctional population. Hunt, McEwen and McKenna (1986) commenting on the McMaster Health Index Questionnaire, precursor to the Health Utility Index, 29 and how well, they would identify the broader dimensions of health i.e., physical health, mental health, social and role functioning, and general perceptions o f well-being. The variables are^®: 1 - Health Status; Vision Attribute 2 - Health Status: Hearing Attribute 3 - Health Status: Speech 4 - Health Status: Mobility Attribute 5 - Health Status: Dexterity Attribute 6 - Health Status: Emotion Attribute 7 - Health Status: Cognition Attribute 8 - Health Status: Pain and Discomfort Attribute 9 - Adjusted Specific Chronic Stress Index 10 - Work Stress Index 11 - Self-esteem index 12 - Mastery index 13 - Sense of Coherence scale 14 - Distress score 15 - Perceived social support index 16 - Average frequency of contact index 17 - Derived health description index Various statistical measures supported the use o f factor analysis. Following further statistical examination and manipulation it was found that the 17 variables loaded on to five factors as displayed in Table 3. state that for physical function items the validity is robust but considerably weaker for social and emotional items. Complete descnptions of the 17 indicators, including the survey questions, exceipied from Statistics Canada (1995) are included in Appendix A. 30 Table 3. Richardson-Zumbo Factors. Indicators and Loading Indicator Loading Factor + Dexterity 1 • Physical Impairment Speech + Emotion + Chronic Stress + Mastery - Coherence - Distress + Emotion - Work Stress - Self-Esteem + Mastery + Vision + Hearing + Mobility + Cognition + Pain + Health Description - Emotion - Social Support + Frequency o f Contact + 2 - Mental Ill-Health 3 - Mental Well-Being 4 - General Health Impairment 5 - Social Well-Being (Richardson and Zumbo, 2000, p. 183) For example, the physical impairment factor was primarily comprised of dexterity and speech health status attributes. Both indicators were positively correlated with the factor i.e., as speech and dexterity problems increase in severity so does physical The panem matrix with the S factors and the respective indicator loading is provided in Appendix B. 31 impairment. In the mental ill-health factor, mastery is negatively correlated, i.e., as one’s sense of control over life’s situations worsens, mental ill-health increases.^^ The next phase was to run a multiple regression of the Health Utility Index scores on to the five Richardson and Zumbo Factors to determine the relative proportion of variation in the Health Utility Index accounted for by each factor. A relative Pratt index was also generated to determine the relative contribution of each factor to the regression. The results are displayed in Table 4. Beta-Weight Physical Impairment Correlation with Health Utility Index -.301 -.083 Relative Pratt Score 3.6% Mental Ill-Health -.403 -.008 0.5% Mental Well-Being .379 .031 1.7% General Health Impairment -.775 -.655 72.2% Social Well-Being .518 301 22.2% Factor (Richardson and Zumbo, 20C10, p. 186) Kûte» R was 70.2% Based on these findings it would seem that the only factors contributing to any significant degree to scores on the Health Utility Index are general health impairment at 72.2% and social well-being at 22.2% o f explained variation. Richardson and Zumbo (2000) concluded that the Health Utility Index used by the National Population Health Survey fails to capture the multi-dimensionality o f health. Most o f the explained variation comes from states o f ill-health and is unable to differentiate among various levels of well-being. It was hypothesized that this should not It is important to review and understand the definitions of the indicators in order to understand the correlations, many of which are double negatives. 32 be too surprising since the Health Utility Index was initially developed to measure the health status of a paediatric oncology population whose state o f ill-health would be much higher than the general population. The general population tended to rate its health at or near the highest health level states almost all of the time. Commenting on health assessment measures designed to examine the absence of ill-health, Bergner (1987) writes that even if they include some measures of good health, the measures do not, as an overall measure, assess positive health or its gradations. Richardson and Zumbo (2000) suggested as future research directions the integration of Health Utility Index scores and additional health indicators into a health profile in order to provide a better summary description of the construct o f health status and how the various health determinants interact with one another. See Appendix C for a Table showing correlations between all the dependent variables. 33 CHAPTER 2 - METHODS The objective of this thesis is to investigate the sensitivity of the Health Utility Index and the Richardson-Zumbo Health Profile to a set of key health determinants considered important by the Provincial Health Officer o f British Columbia (1994). Six indicators were selected from the 1994/95 National Population Health Survey that correspond to significant determinants o f health described in the British Columbia Provincial Health Officer’s 1994 Annual Report. Based on the literature review, age and gender were also selected. Table 5 displays the National Population Health Survey indicators chosen and the corresponding Provincial Health Officer’s health determinant. No indicators were selected from the domains o f physical environment, biological influences or health services. Table 5. National Population Health Survey Indicators and PHO Health Determinants NPHS Indicator PHO Health Determinant 1- Single Parenthood Social and Economic Environment 2- Derived Variable for Working Status Social and Economic Environment 3 - Derived Highest Education Level Attained 4 - Derived Income Adequacy Social and Economic Environment 5 - Type o f Smoker Health Behaviours and Skills 6 - Derived Type of Drinker Health Behaviours and Skills 7 - Age Nil 8 - Gender Nil Social and Economic Environment The National Population Health Survey variables (Statistics Canada, 1995) are explained as follows: 34 1 - Single Parenthood is a dichotomous indicator we created which segregates Derived Type o f Household into two categories: 0 OTHER 1 SINGLE PARENT Where, 0 OTHER is comprised of: 1 Couple With Children < 25 defined as a married or common-law couple with at least one partner being the parent of the dependent child. No other relationships are allowed. 2 Couple With Children>25 With or Without Other child(ren) defined as a married or common-law couple with no dependent< 25 years old. Any other relationships are allowed. 3 Single defined as an unattached individual living alone. Household size=l. 4 Single With Others defined as unattached individuals living together. There cannot be a marital/common-law or parental relationship but other relationships such as siblings are allowed. 5 Couple With Dependent Child(ren)<25 And Other Relatives defined as at least one partner must be the parent of one child < 25 years old in the household. Other relationships are allowed. 6 Couple Alone defined as married or common-law couple alone. No other relationships are permitted. Household size=2. 9 Other Household Types defined as all other household types. And, 1 SINGLE PARENT is comprised of: 35 7 Single parent With Dependent Child(ren) < 25 where one child must be less than 25 years old. No other relationships are permitted. 8 Other Single-parent Households where one child must be less than 25 years old. Other relationships are permitted. 2 - Derived Variable for Working Status is a derived indicator based on the respondent’s recent employment history where: 1 CURRENTLY WORKING 2 NOT CURRENTLY WORKING-BUT HAD A JOB 3 DID NOT WORK DURING LAST 12 MONTHS 5 NOT APPLICABLE 9 NOT STATED A higher score indicates greater unemployment. 3 - Derived Highest Education Level Attained is a derived variable based on the responses to questions EDUC-Ql to EDUC-Q4. EDUC-Ql Excluding kindergarten, how many years of elementary and high school have/has... successfully completed? EDUC-Q2 Have/has... graduated from high school? EDUC-Q3 Have/has... ever attended any other kind of school such as university, community college, business school, trade or vocational school, CEGEP or other post-secondary institution? EDUC-Q4 What is the highest level of education th at... have/has attained? Some trade, technical, vocational school or business college Some community college, CEGEP or nursing school 36 Some university Diploma or certificate from trade, technical or vocational school, or business college Diploma or certificate from community college, CEGEP, or nursing school) Bachelor's or undergraduate degree or teacher's college (e.g., B.A., B.Sc., LL.B.) Masters (e.g., M.A., M. Sc., M.Ed.) Degree in medicine, dentistry, veterinary medicine or optometry (M.D., D.D.S., M.D., D.V.M., O.D.) Earned doctorate (e.g., Ph D., D.Sc., D.Ed.) Other (Specify ) The responses were assigned to one o f the following categories: 1 NO SCHOOLING 2 ELEMENTARY SCHOOL 3 SOME SECONDARY SCHOOL 4 SECONDARY SCHOOL GRADUATION 5 OTHER BEYOND HIGH SCHOOL 6 SOME TRADE SCHOOL ETC 7 SOME COMMUNITY COLLEGE 8 SOME UNIVERSITY 9 DIPLOMA/CERTIFICATE TRADE SCHOOL 10 DIPLOMA/CERTIFICATE COM. COL.CEGEP 37 11 BACHELOR DEGREE (INCLUDES LLB) 12 MASTER/DEGREE IN MEDICINE/DOCTORATE 96 NOT APPLICABLE 99 NOT STATED A higher score indicates more schooling. 4 - Derived Income Adequacy is based on household income and the size of the household where: 1 2 3 4 5 9 Lowest income Less than $10,000 1 to 4 persons Less than $15,000 5 or more persons Lower middle income $10,000 to $14,999 1 or 2 persons $10,000 to $19,999 3 or 4 persons $15,000 to $29,999 5 or more persons $15,000 to $29,999 1 or 2 persons $20,000 to $39,999 3 or 4 persons $30,000 to $59,999 5 or more persons Middle income Upper middle income $30,000 to $59,999 1 or 2 persons $40,000 to $79,999 3 or 4 persons $60,000 to $79,999 5 or more persons $60,000 or more 1 or 2 persons $80,000 or more 3 persons or more Not stated Not applicable Highest Income Unknown A higher score indicates greater income adequacy. 38 5 - Type of Smoker is based on questions SM0K-Q2, Q4a, Q5: SM0K-Q2 At the present time do/does... smoke cigarettes daily, occasionally or not at all? SM0K-Q4a Have/has you/he/she ever smoked cigarettes at all? SMOK-QS Have/has you/he/she ever smoked cigarettes daily? The responses were assigned to one o f the following categories: 1 Daily smoker SM0K-Q2 = 1 2 Occasional smoker but SMOK-Q2=2 AND former daily smoker SM0K-Q5=1 Always an occasional smoker SMOK-Q2=2 AND SMOK-Q5=2 Former daily smoker SMOK-Q2=3 AND SM0K-Q4A=1AND SMOKQ5=l Former occasional smoker SMOK-Q2=3 AND SMOKQ4A=1AND SMOK-Q5=2 Never smoked SMOK-Q2=3 AND SMOK-Q4A=2 Not stated Not stated A lower score indicates greater smoking frequency. 6 - Derived Tvpe of Drinker is based on questions ALC0-Q2 and ALC0-Q5B: ALC0-Q2 During the past 12 months, how often did you/he/she drink alcoholic beverages? 39 Every day 4-6 times a week 2-3 times a week Once a week 2-3 times a month Once a month Less than once a month ALC0-Q5B Did you/he/she ever have a drink? Yes No The responses were assigned to one of the following categories: 1 Regular drinker: a drink at ALCO-Q2<7 least once a month 2 Occasional drinker: less than ALCO-Q2=7 one drink a month 3 Don't drink now: did not have ALC0-Q5B=1 a drink in the last 12 months 4 Abstinent (never drank) ALCO-Q5B=2 9 Not stated Not stated A lower score indicates greater frequency of alcohol consumption. 7 - Age created grouped age cohorts: 1 12 TO 14 YEARS 2 15 TO 19 YEARS 40 3 20 TO 24 YEARS 4 25 TO 29 YEARS 5 30 TO 34 YEARS 6 35 TO 39 YEARS 7 40 TO 44 YEARS 8 45 TO 49 YEARS 9 50 TO 54 YEARS 10 55 TO 59 YEARS 11 60 TO 64 YEARS 12 65 TO 69 YEARS 13 70 TO 74 YEARS 14 75 TO 79 YEARS 15 80 YEARS OR OLDER A higher score indicates greater age. 8 - Gender is a dichotomous variable. 1 MALE 2 FEMALE Data for analysis were extracted from the full 1994/95 National Population Health Survey data set specific to the Prince George over-sample. The five Richardson and Zumbo factors (physical impairment factor, mental ill-health factor, mental well-being factor, general health impairment factor, and social well-being factor) were also combined into a Composite Score to create an additional dependent variable such that; 41 Composite of 5 Factors = sum (social well-being, mental well-being) - sum (physical impairment, mental ill-health, general health impairment). The negative scales (physical impairment, mental ill-health, general health impairment) were reverse-coded so that the Composite of 5 Factors would measure health in a manner that a large positive number would mean healthier, as in a positive aspect of health. The two dichotomous indicators. Single Parent and Gender were examined against the seven dependent variables (Health Utility Index, Composite o f 5 factors, physical impairment factor, mental ill-health factor, mental well-being factor, general health impairment factor, and social well-being factor). The sample size, means and standard deviations were calculated. T-tests were performed to determine if the differences in the means were statistically significant. A bivariate analysis was done between the six non-dichotomous National Population Health Survey indicators (working status, highest level o f education attained, income adequacy, type of smoker, type of drinker, and age) and the seven dependent variables (Health Utility Index, Composite of 5 factors, physical impairment factor, mental ill-health factor, mental well-being factor, general health impairment factor, and social well-being factor). Correlations and tests for significance were performed resulting in a 6 X 7 matrix. A Multivariate analysis was then done on the 6 x 7 matrix. Beta values^” and the Pratt Index^' for those model predictors identified through stepwise regression were “When all variables are standardized to have means of zero and standard deviations of one, the standardized regression coefficients (Betas) measure the percent of movement in the dependent variable when a predictor variable moves one full unit and every other predictor in the set is held constant” (Michalos, 1996, p. 55). “The Pratt Index quantifies the relative contribution each explanatory variable makes to the overall regression equation by partitioning the model into that proportion attributable to each explanatory variable. The scores are additive and will therefore sum to 1.0” (Richardson, 1999, p. 32 ). 42 calculated. The Multivariate analysis was re-run with the addition of the age and gender National Population Health Survey indicators. Beta values and the Pratt Index for those model predictors identified through stepwise regression were calculated for this 8 x 7 matrix. Pratt Index = Beta • corr ^ 100 43 CHAPTER 3 -RESULTS A. Bivariate Analysis The two dichotomous indicators, Single Parent and Gender, were examined against the seven dependent variables: Health Utility Index, Composite o f 5 Factors, physical impairment factor, mental ill-health factor, mental well-being factor, general health impairment factor, and social well-being factor. The sample size, means and standard deviations were calculated. T-tests were performed to determine if the differences in the means were statistically significant. See Table 6. SiPfijg Paient The sample size for Health Utility Index was 833, and 838 for the other six dependent variables (Composite of 5 Factors, physical impairment factor, mental illhealth factor, mental well-being factor, general health impairment factor and social well­ being factor). Single-parent households accounted for 11.2% (93/833) and 11.1% (93/838) of the samples respectively. A statistically significant difference in the means as determined by t-test was found in four of the dependent variables: Composite of S Factors B = .002, mental-ill health g < 001, mental well-being g = .006, and social well-being g = .017. Single parents were found to have a lower Composite Score (mean = -.863 versus mean = . 108), greater mental ill-health (mean = .496 versus mean = -.062), less mental well-being (mean = .878 versus mean = .029), and less social well-being (mean = -.168 versus mean .021) than the not single-parent cohort. There were no statistically significant differences found for the dependent variables: Health Utility Index, physical impairment, or general health impairment. 44 Gender The sample size for Health Utility Index was 833, and 838 for the other six dependent variables (Composite of 5 Factors, physical impairment factor, mental illhealth factor, mental well-being factor, general health impairment factor and social well­ being factor). Females accounted for 52.2% (435/833) and 52.0% (436/838) of the samples respectively. A statistically significant difference in the means as determined by t-test was found in two of the dependent variables; mental-ill health p = .017, and social well-being p = .048. Females were found to have greater mental ill-health (mean =-.077 versus mean = .071) and greater social well-being (mean = -.051 versus mean = .047) than the males in the sample. There were no statistically significant differences found for the dependent variables: Health Utility Index, Composite of 5 Factors, physical impairment, mental well-being, or general health impairment. 45 Table 6. Reporting n« Means, and Standard Deviations for the Dependent Variables Dependent Variable Health Utility Index Composite of 5 Factors 1 - Physical Impairment Factor 2 - Mental IllHealth Factor 3 - Mental Well-Being Factor 4 - General Health impairment Factor 5 - Social Well-Being Factor Single Parent Single Parent n = 93 Mean = .903 Std. Dev. = .115 Single Parent 0 = 93 Mean = -.863 Std. Dev. ==2,67 Single Parent 0 = 93 Mean = -.062 Std. Dev. = 117 Single Parent 0 = 93 Mean = .496 Std. Dev. = .963 Single Parent 0 = 93 Mean = -.236 Std. Dev. = .878 Single Parent 0 = 93 Mean = .024 Std. Dev. = .666 Single Parent 0 = 93 Mean = -.168 Std. Dev. = .751 Not Single Parent n = 740 Mean = .893 Std.Dev. = .130 Not Single Parent 0 = 745 Mean = .108 Std.Dev. = 2.79 Not Single Parent 0 = 745 Mean = .008 Std Dev. = 1.5 Not Single Parent 0=745 Mean = -.062 Std.Dev. = .869 Not Single Parent 0 = 745 Mean = .029 Std.Dev. = .871 Not Single Parent 0 = 745 Mean = -.003 Std.Dev. = .771 Not Single Parent 0 = 745 Mean = .021 Std.Dev. = .715 n.s. n.s. Males 0 = 402 Mean = .014 Std. Dev = 1.199 1(836) = 5.76, D Health Utility Index Composite of 5 Factors I - Physical Impairment Factor 2 - Mental IllHealth Factor 3 - Mental Well-Being Factor n=773 Corr. = -.269 *** n = 776 n=776 Corr. = +.106 #* n=776 Corr. = -.204 *** 5 - Social Well-Being Factor Corr. = +.042 n.s. Corr. = +.101 Corr. = +.110 *** Corr. = +.182 Type of Smoker Corr. = +.152 *** Corr. = +.222 *** Corr. = -.017 n.s. Corr. = -.199 *** Corr. = +.179 *** Corr. = -.177 *** Corr. = +.188 *** Derived Type of Drinker Corr. = -.095 ** Corr. = -.085 Corr = +.099 ** Corr. = -.013 n.s. Corr. = -.073 * Corr. = +.147 *** Corr. = +.036 n.s. Age Cohort Corr. = -.250 *** Corr. = -.028 n.s. Corr. = +.070 • Corr. = -.187 *** Corr. = +.072 * Corr. = +.349 Corr. = +.040 n.s. ** ** *** Note. * denotes p < .05, ** denotes p < .01, ♦** denotes p < .001, n.s. denotes that the test o f the correlation was not statistically significant. o B.1 Multivariate Analysis Without Age and Gender A Multivariate analysis was then done between the six National Population Health Survey indicators (single-parenthood, derived variable for working status, derived highest level of education attained, derived income adequacy, type o f smoker, and derived type of drinker) and the seven dependent variables (Health Utility Index, Composite of 5 Factors, physical impairment factor, mental ill-health factor, mental well­ being factor, general health impairment factor and social well-being factor). Beta values and the Pratt Index scores for those model predictors identified through stepwise regression were calculated. See Table 8. Health Utilitv Index Stepwise regression identified the derived variable for working status and type of smoker as significant model predictors for the dependent variable Health Utility Index. That is, in the presence of the six predictors taken together, only two influenced the dependent variable. Greater unemployment and more tobacco consumption are related to a lower Health Utility Index Score. Employment status is responsible for 75.9% (Beta = -.268) of the R-squared value (R^ = .095) and smoking 23.8% (Beta = +.149). Composite of 5 Factors Stepwise regression identified the derived variable for working status, derived income adequacy, and type o f smoker as significant model predictors for the dependent variable Composite of 5 Factors. That is, in the presence o f the six predictors taken together, only three influenced the dependent variable. Greater unemployment, less income adequacy and more tobacco consumption are related to a lower Composite of 5 Factors score. Employment status is responsible for 34.7% (Beta = -.165) o f the R- 51 squared value (R^ = .097), income adequacy 18.0% (Beta = +.096) and smoking 47.8% (Beta = +.209). Physical Impairment Factor Stepwise regression identified the derived variable for working status and derived type o f drinker as significant model predictors for the dependent variable physical impairment factor. That is, in the presence of the six predictors taken together, only two influenced the dependent variable. Greater unemployment and less alcohol consumption are related to greater physical impairment. Employment status is responsible for 54.9% (Beta = +.088) of the R-squared value (R^ = .017) and drinking 46.0% (Beta = +.079). Mental Ill-Health Factor Stepwise regression identified single parent, the derived variable for income adequacy and type of smoker as significant model predictors for the dependent variable mental ill-health factor. That is, in the presence o f the six predictors taken together, only three influenced the dependent variable. Being a single parent, lower income adequacy and greater tobacco consumption are related to greater mental ill-health. Single­ parenthood is responsible for 35.6% (Beta = +.156) of the R-squared value (R^ = .085), income adequacy 23.7% (Beta = -.118) and smoking 40.7% (Beta = -.174). Mental Well-Being Factor Stepwise regression identified the derived variable for income adequacy, type of smoker and derived type of drinker as significant model predictors for the dependent variable mental well-being factor. That is, in the presence o f the six predictors taken together, only three influenced the dependent variable. Greater income adequacy, less tobacco consumption and greater alcohol consumption are related to greater mental well­ 52 being. Income adequacy is responsible for 34.2% (Beta = +.125) o f the R-squared value (R^ = .057), smoking 55.3% (Beta = +.176) and drinking 10.4% (Beta = -.081). General Health Impairment Factor Stepwise regression identified the derived variable for working status, type of smoker and derived type of drinker as significant model predictors for the dependent variable general health impairment factor. That is, in the presence of the six predictors taken together, only three influenced the dependent variable. Increasing levels of unemployment, greater tobacco consumption and less alcohol consumption are related to greater general health impairment. Employment status is responsible for 68.7% (Beta = +.313) of the R-squared value (R^ = .154), smoking 21.6% (Beta = -.188) and drinking 9.8% (Beta = +.313). Social Well-Being Factor Stepwise regression identified the derived variable for income adequacy and type of smoker as significant model predictors for the dependent variable social well-being factor. That is, in the presence of the six predictors taken together, only two influenced the dependent variable. Greater income adequacy and less tobacco consumption are related to greater social well-being. Income adequacy is responsible for 21.5% (Beta = +.085) o f the R-squared value (R^ = .042) and smoking 79.7% (Beta = +.178). 53 Table 8. Multivariate Analysis Where Only the Model Predictors IdentMled Through Stepwise Regression are indicated Dependent Variable -> Health Utility index Composite o f S Factors 1 - Physical impairment Factor 2 - Mental illHealth Factor 3 - Mental WellBeing Factor Model Predictors i Single Parent n = 773 n = 776 n = 776 n = 776 Beta = +.156 Pratt = 35.6% 0 = 776 Beta = -.268 Pratt = 75.9% Beta = -.165 Pratt = 34.7% Beta - +.088 Pratt = 54.9% Derived Variable for Working Status Derived Highest Level o f Education Attained Derived income Adequacy Type of Smoker Beta = +.149 Pratt = 23,8% R: Beta = -.118 Pratt = 23.7% Beta = +.125 Pratt = 34.2% Beta = +.209 Pratt = 47.8% Beta = -. 174 Pratt = 40.7% Beta = +.176 Pratt = 55.3% Beta = -. 188 Pratt = 21.6% Beta = -.081 Pratt = 10.4% Beta = +.103 Pratt = 9.8% .057 £ (3 . 772) = 15.528 .154 £ (3 .7 7 2 ) = 46.937 D<001 B <.001 Beta = +.079 Pratt = 46.0% Sie. .095 E ( 2 .770) = 40.220 D<.001 .097 £ (3 .7 7 2 ) = 27.745 D <001 .017 £ (2 .7 7 3 ) = 6.764 D= 001 5 -Social WellBeing Factor n = 776 Beta = +.313 Pratt = 68.7% Beta = +.096 Pratt = 18.0% Derived Type of Drinker £ 4 - General Health impairment Factor n = 776 .085 £ (3 .7 7 2 ) = 23.917 D<.001 Beta = +.085 Pratt = 21.5% Beta = +.178 Pratt = 79.7% .042 £ (2 .7 7 3 ) = 17.144 B<.001 B.2 Multivariate Analysis Including Age and Gender The Multivariate analysis was then repeated using the six National Population Health Survey indicators (single-parenthood, derived variable for working status, derived highest level of education attained, derived income adequacy, type of smoker, and derived type of drinker) plus age and gender, and the seven dependent variables (Health Utility Index, Composite o f 5 Factors, physical impairment factor, mental ill-health factor, mental well-being factor, general health impairment factor and social well-being factor). See Table 9. Health Utilitv Index As in the previous analysis, stepwise regression identified the derived variable for working status and type of smoker as model predictors for Health Utility Index, but this time age was also identified. That is, in the presence of the eight predictors taken together, only three influenced the dependent variable. Again, greater unemployment and more tobacco consumption are related to a lower Health Utility Index Score. Increasing age also leads to a lower Health Utility Index score. Employment status is responsible for 45.3% (Beta = -.207, previous Pratt = 75.9%) of the R-squared value (R^ = .123 versus .095 previously), smoking 18.2% (Beta = +.147, previous Pratt = 23.8%), plus age 36.2% (Beta = -.178). Composite of 5 Factors As in the previous analysis, stepwise regression identified the derived variable for working status, derived income adequacy and derived type o f smoker as model predictors for the Composite o f 5 Factors. That is, in the presence o f the eight predictors taken together, only three influenced the dependent variable. Age and gender had no significant 55 influence. Greater unemployment, less income adequacy and greater tobacco consumption are related to lower a Composite score. Employment status is responsible for 34.7% (Beta = -.165) of the R-squared value (R^ = .097), income adequacy 18.0% (Beta = +.096), and smoking 47.8% (Beta = +.209), all unchanged from the previous analysis. Phvsical Impairment Factor As in the previous analysis, stepwise regression identified the derived variable for working status and derived type of drinker as model predictors for the physical impainnent factor. That is, in the presence of the eight predictors taken together, only two influenced the dependent variable. Age and gender had no significant influence. Greater unemployment and less alcohol consumption are related to greater physical impairment. Employment status is responsible for 54.9% (Beta = +.088) o f the R-squared value (R^ = .017) and drinking 46.0% (Beta = +.079), both unchanged from the previous analysis. Mental Ill-Health Factor As in the previous analysis, stepwise regression identified single parent, the derived variable for income adequacy, and type o f smoker as significant model predictors for the mental ill-health factor, but this time age was also identified. That is, in the presence o f the eight predictors taken together, only four influenced the dependent variable. Again, being a single parent, lower income adequacy and greater tobacco consumption are related to greater mental ill-health. Lower age, however, is also related to increased mental ill-health. Gender had no significant influence. Single parenthood is responsible for 19.6% (Beta = +.115, previous Pratt = 35.6%) o f the R-squared value (R^ = 114 versus .085 previously), income adequacy 20.9% (Beta = -.139, previous Pratt = 56 23.7%), and smoking 30.9% (Beta = -.177, previous Pratt = 40.7%), plus age 28.5% (Beta =-.174). Mental Well-Being Factor As in the previous analysis, stepwise regression identified the derived variable for income adequacy, type of smoker and derived type of drinker as significant model predictors for the mental well-being factor, but this time the derived variables for working status and age were also identified. That is, in the presence of the eight predictors taken together, five influenced the dependent variable. Again, greater income adequacy, less tobacco consumption and greater alcohol consumption are related to greater mental well-being. Increased employment and age are also related to greater mental well-being. Gender had no significant influence. Income adequacy is responsible for 19.1% (Beta = +.092, previous Pratt = 34.2%) of the R-squared value (R^ = .075 versus .057 previously), smoking 42.7% (Beta = +.179, previous Pratt = 55.3%), drinking 7.5% (Beta = -.077, previous Pratt = 10.4%), plus employment 18.2% (Beta = -.112) and age 12.3% (Beta = +.128). General Health Impairment Factor As in the previous analysis, stepwise regression identified the derived variable for working status, type of smoker and derived type o f drinker as significant model predictors for the general health impairment factor but this time age was also identified. That is, in the presence of the eight predictors taken together, four influenced the dependent variable. Again, increasing levels of unemployment, greater tobacco consumption and lower levels of alcohol consumption are related to greater general health impairment. Greater age is also related to greater general health impairment. 57 Gender bad no significant influence. Employment status is responsible for 36.4% (Beta = +.228, previous Pratt = 68.7%) of the R-squared value (R^ = .212 versus .154 previously), smoking 15.4% (Beta = -.184, previous Pratt = 21.6%), drinking 6.2% (Beta = +.089, previous Pratt = 9.8%, plus age 42.3% (Beta = +.257). Social Well-Being Factor As in the previous analysis, stepwise regression identified the derived variable for income adequacy and type of smoker as significant model predictors for the social well­ being factor but this time gender was also identified. That is, in the presence of the eight predictors taken together, only three influenced the dependent variable. Again, greater income adequacy and less tobacco consumption are related to greater social well-being. Being female is also related to greater social well-being. Age had no significant influence. Income adequacy is responsible for 20.3% (Beta = +.094, previous Pratt = 21.5%) of the R-squared value (R^ = .049 versus .042 previously), smoking 68.7% (Beta = +.179, previous Pratt = 79.7%), plus gender 11.2% (Beta = +.082). 58 T able 9. M ultivariate A nalysis W here O nly the M odel Predictors id entified T hrough Stepw ise Regression are indicated (Plus A ge and G ender) Dependent Composite o f 5 1 - Physical 2 - Mental IllHealth Utility 3 - Mental Well4 - General Health 5 - Social Well Factors impairment index Health Factor Being Factor Variable -> Impairment Factor Being Factor Factor Model n=776 n=776 n=776 n = 776 n =776 0=776 n=773 Predictors 4 Beta = +.115 Single Parent Pratt = 19.6% Beta = .112 Beta = +.088 Beta = +.228 Beta = -.165 Beta = -.207 Derived Pratt = 54.9% Pratt = 18.2% Pratt = 36.4% Pratt = 45.3% Pratt = 34,7% Variable for Working Status Derived Highest Level o f Education Attained Beta = +.092 Beta = -.139 Beta = +.094 Beta = +.096 Derived Pratt = 20.9% Pratt = 19.1% Pratt = 20.3% Pratt = 18.0% income Adequacy Beta = +.179 Bela = -.184 Beta = +. 179 Beta = -.177 Beta = +.209 Type of Beta = +.147 Pratt = 30.9% Pratt = 42.7% Pratt = 15.4% Pratt = 47.8% Pratt = 68.7% Pratt = 18.2% Smoker Beta = +.079 Bela = -.077 Beta = +.089 Derived Type Pratt = 46.0% Pratt = 6.2% Pratt = 7.5% o f Drinker Age Cohort Beta = -.174 Pratt = 28.5% Beta = -.178 Pratt = 36.2% Beta = +.128 Pratt = 12.3% Beta = +.257 Pratt = 42.3% Bela = +.082 Pratt = 11.2% Gender D=.001 .114 E (4.771) = 24.722 n<001 .075 E (5.770) = 12.477 n<-001 Nil +.029 + 018 .097 E (3,772) = 27.745 D<.OOI .017 E (2.773) = 6.764 Sig. 123 E (3,769) = 35.815 D<.001 Gain +.028 Nil R* E n<.ooi .049 E (3.772) = 13.269 D<. 00i +.058 +.007 .212 E (4.771) = 51.967 1 CHAPTER 4 - CONCLUSION Richardson and Zumbo (2000) examined the Health Utility Index as a measure of health status for use in the 1994/95 National Population Health Survey. Their results demonstrated “that the use of the [Health Utility Index] as the sole summary measure of health status ... [was] problematic ... [since it did] not appear to discriminate between the many different levels o f positive health experienced by the vast majority o f the general population ... [and it was] more or less insensitive to variation in key indicators o f mental well-being" (p. 188). Richardson and Zumbo questioned the ability of any single score to measure the health status of a population suggesting a better approach would be to use a multi-dimensional health profile instead. Utilizing the Richardson-Zumbo Health Profile and the Health Utility Index, we attempted to examine the effect determinants o f health identified by the British Columbia Provincial Health Officer and the literature search would have on those measures in terms of their ability to be sensitive to underlying changes in the population’s health status. Would a profile yield more useful information than a summary score? We began by examining the literature and the 1994 Annual Report o f the British Columbia Provincial Health Officer for significant determinants o f health. The 1994 British Columbia Provincial Health Officer Annual Report was a focus for two reasons. First, the report’s theme in 1994 was Determinants of Health and, second, 1994/95 was the year o f the National Population Health Survey for which the Prince George data set was available. The results of the literature review and the Provincial Health Officer’s Annual Report were similar and yielded the following eight health determinants for analysis: single parenthood, age, gender, employment status, education, income 60 adequacy, tobacco use, and alcohol consumption. Indicators closely paralleling these determinants were selected from the National Population Health Survey. Since single parenthood and gender are dichotomous variables they were examined against the six dependent variables from the work o f Richardson and Zumbo namely, the Health Utility Index, and the five factors of the Richardson-Zumbo Health Profile: physical impairment, mental ill-health, mental well-being, general health impairment, and social well-being. A Composite Score, being the summation o f the five Richardson and Zumbo Factors, was also constructed and examined. Sample size, means and standard deviations were calculated and t-tests were performed to determine if the differences in the means were statistically significant. A bivariate analysis was then done between the six non-dichotomous National Population Health Survey indicators (derived variable for working status, derived highest level o f education attained, derived income adequacy, type o f smoker, derived type of drinker, and age cohort) and the seven dependent variables (Health Utility Index, Composite of S Factors, physical impairment factor, mental ill-health factor, mental well­ being factor, general health impairment factor and social well-being factor). Correlations and tests for significance were calculated. A multivariate analysis was then done, first including the six health determinants identified by the Provincial Health Officer and then again with eight health determinants through the inclusion o f age and gender. The model predictors identified through stepwise regression were identified. Beta values, Pratt scores, R-squared values, and tests for significance were calculated as was the difference in R-squared values between the two multivariate analyses. 61 The results of the analysis yielded no surprises. As expected, being employed, greater income adequacy and less tobacco consumption were all associated with a higher state of health. All these were intuitively plausible and consistent with the literature. Gender was not a significant health determinant except on the social well-being factor where there was an apparent advantage to being female. Being a single parent was only significant on the mental ill-health factor where being a single parent was associated with greater mental ill-health. There was no apparent significant association between education level and any of the dependent variables. Of interest was the improvement in mental well-being and the decrease in mental ill-health with increasing age while youth, as expected, was associated with higher Health Utility Index scores and less general health impairment. One exception that was counter intuitive was the apparent health advantage gained by alcohol consumption on the physical impairment, mental well-being and general health impairment factors. Before one asserts the benefits to health o f alcohol consumption, however, the underlying indicator needs to be more closely examined. The National Population Health Survey derived variable for Derived Type of Drinker had the heaviest class o f drinker consuming one or more drinks per month. Clearly this washes out any differences which may occur between a heavy drinker and the person who has but one drink per month.^^ The Derived Type of Drinker variable could have captured higher levels of alcohol consumption by including the results from National Population Health Survey questions ALC0-Q3, Q4 and QS which capmred number of times when more than five drinks were consumed on one occasion, the greatest number of drinks on one occasion, and how many drinks the person had on each of the last seven days. Schwarz & Strack (1999) suggest asking open ended questions is better than giving the respondent a range of responses to choose from. They suggest that “respondents assume the list of response alternatives reflects the researcher’s knowledge of the distribution of the behaviour.... [and] accordingly, they use the range of the response alternatives as a frame of reference in estimating their own behavioral frequency" (p. 73). 62 Overall, the of the multivariate analyses were disappointingly low ranging from .017 to .212 on the five Richardson and Zumbo Factors, .097 for the Composite Score and .123 for the Health Utility Index. The gain in R^ when age and gender were added was also minimal ranging from nil to +.058. The net result of these regressions seems to be that there is a poor fit between the determinants of health and population health status. The determinants do not seem to be determining much which, unfortunately, was the problem we started with. The Health Utility Index was hardly describing population health while the five Richardson-Zumbo scores and the Composite Score fared little better. The general health impairment R-squared (R^ = .212) was the highest o f all dependent variables; made up of the following National Population Health Survey indicators: vision, hearing, mobility, cognition, pain, and health description.^^ This would suggest a nice, overview, summary-type variable that could on its own or in a profile be examined as a good health status indicator. Does this mean the determinants o f health are unimportant? Clearly not. What this analysis once again demonstrates is the difficulty in caphiring the complex interplay of a myriad of variables that form the construct o f health. Given the multi-faceted nature of health it seems even less useful to attempt to develop a single summative measure of health even though, as Hunt McEwen and McKenna (1986) assert, health policy makers are usually more interested in a single global number which can summarize the health status of a population into a sununary statistic akin to the way the Gross National Product is an indicator of the health of the economy. See Table 3, Appendix A, and Appendix B for the factor loadings and the survey questions underlying the general health impainnent factor. 63 The challenge to future researchers is to continue to explore profiles that accurately capture the status of the population’s health and that are also sensitive to underlying changes as they occur. For although the Richardson-Zumbo Health Profile did not have strong explanatory power it was, never the less, able to yield more information than the Health Utility Index or the Composite Score. 64 REFERENCES American Psychological Association. (1994). Publication manual of the American Psychological Association (4th éd.). Washington, DC; Author. A vison, W. R. (1998). The health consequences of unemployment, pp. 3-41 in Determinants of health: Adults and seniors. Volume 2 of Canada health action: Building on the legacy, papers commissioned bv the national forum on health. Sainte-Foy: Editions MultiMondes. Beigbeder, Y. (1998). The World Health Organization: Vol. 4. International organization and the evolution of world society. Dordrecht: Martinus Nijhoff Publishers. Bergner, M. (1985). Measurement of Health Status. Medical Care. 23 (5), 696704. Bergner, M. (1987). Health status measures: An overview and guide for selection. Annual Review of Public Health. 8. 191-210. Boyle, M. H., Furlong, W., Feeny, D., Torrance, G. W., & Hatcher, J. (1995). Reliability of the health utility index - mark III used in the 1991 cycle 6 Canadian general social survey health questionnaire. Quality of Life Research. 4 . 249-257. Breen, M. J. (1998). Promoting literacy, improving health, pp. 43-86 in Determinants o f health: Adults and seniors. Volume 2 of Canada health action: Building on the legacy, papers commissioned bv the national forum on health. Sainte-Foy: Editions MultiMondes. British Columbia Ministry of Health. (1995). Building our future: Guide to health and management planning for regions and communities. Victoria: Author. British Columbia Provincial Health Officer. (1994). A report on the health of British Columbians: Provincial Health Officer’s annual report 1994. Victoria, British Columbia: Ministry o f Health and Ministry Responsible for Seniors. British Columbia Provincial Health Officer. (1996). A report on the health of British Columbians: Provincial Health Officer’s annual report 1995. Victoria, British Columbia: Ministry of Health and Ministry Responsible for Seniors. British Columbia Provincial Health Officer. (1997). A report on the health of British Columbians: Provincial Health Officer’s annual report 1996. Victoria. British Columbia: Ministry of Health and Ministry Responsible for Seniors. British Columbia Provincial Health Officer. (1998). A report on the health of British Columbians: Provincial Health Officer's annual report 1997. Victoria, British Columbia: Ministry of Health and Ministry Responsible for Seniors. 65 Catlin, G., & Will, P. (1992). The national population health survey: Highlights of initial developments. Health Reports. 4 (3), Statistics Canada Cat. 82-003.313-9. Chappell, N. L. (1998). Maintaining and enhancing independence and well-being in old age. pp. 89-141 in Determinants of health: Adults and seniors. Volume 2 of Canada health action: Building on the legacv. papers commissioned bv the national forum on health. Sainte-Foy: Editions MultiMondes. Decter, M. B. (1994). Healing medicine: Managing health system change the Canadian way. Toronto: McGilligan Books. Denton, M., & Walters, V. (1999). Gender differences in structural and behavioural determinants of health: An analysis of the social production o f health. Social Science & Medicine. 48.1221-35. Epp, J. (1986). Achieving health for all: A framework for health promotion. Ottawa: Health and Welfare Canada Cat.H39-102/1986E. Edwards, P. (2000). Building a healthy future. Ottawa: Health Canada. Evans, R. G., & Stoddart, G. L. (1994). Producing health, consuming health care, pp. 27-64 in Wh\rare some people healthy, and others not? The determinants of health of populations!. Evans, R. G., Barer, M. L., & Marmor, T. L. (Eds.). New York: Aldine de Gruyter. Federal, Provincial and Territorial Committee on Population Health (ACPH). (1994). Strategies for population health: Investing in the health o f Canadians. Ottawa: Health Canada. Federal, Provincial and Territorial Committee on Population Health (ACPH). (1999). Toward a healthv future: Second report on the health of Canadians. Ottawa: Health Canada. Frank, J. W. (1995). Why population Health? Canadian Journal o f Public Health. M (3), 162-164. Hansluwka, H. E. (1985). Measuring the health o f populations, indicators and interpretations. Social Science and Medicine. 2 0 112!. 1207-24. Health Canada, Women’s Health Bureau - Ottawa: http://www.hcsc.gc.ca/women/english/index.htm Health Canada. (1996). Towards a common understanding: Clarifying the core concepts of population health. Ottawa: Author. 66 Health Canada. (1998). Taking action on population health: A position paper for health promotion and programs branch staff. Ottawa: Author. Hertzman, €., Frank, J., & Evans, R. G. (1994). Heterogeneities in health status and the determinants of population health, pp. 67-92 in Whv are some people healthv. and others not? The determinants of health of populations! Evans, R. G., Barer, M. L., & Marmor, T. L. (Eds.), New York: Aldine de Gruyter. Hunt, S. M., McEwen, J., & McKenna, S. P. (1986). Measuring health status. Dover, New Hampshire: Groom Helm Ltd. Institute of Health Promotion Research, University of British Columbia. (1999). Population health in Canada: A working paper. Ottawa: Health Canada. Judge, K., Mulligan, J., & Benzeval, M. (1998). Income equality and population health. Social Science and Medicine. 46 (4-5), 567-79. Koivusalo, M., & Ollila, E. (1997). Making a healthv world: Agencies, actors and policies in international health. London: Zed Books Ltd. Lalonde, M. (1974). A new perspective on the health o f Canadians: a working document. Ottawa: Government of Canada. Marmor, T. R., Barer, M. L , & Evans, R. G. (1994). The determinants of a population’s health: What can be done to improve a democratic nation’s health status? pp. 217-230 in Why are some people healthv. and others not? The determinants o f health of populations! Evans, R. G., Barer, M. L., & Marmor, T. L. (Eds.). New York: Aldine de Gruyter. Michalos, A. C. (1980). North American social report: A comparative study o f the quality of life in Canada and the USA from 1964 to 1974. Volume I: Foundations. population, and health. Dordrecht, Holland: D. Reidel Publishing Company. Michalos, A. C. (1982). North American social report: A comparative study o f the quality of life in Canada and the USA from 1964 to 1974. Volume V: Economics. religion, and morality. Dordrecht, Holland: D. Reidel Publishing Company. Michalos, A. C. (1996). Aspects o f the quality of life in Prince George: A case study in Canada. South Asian Journal o f Psychology. 1.45-70. Mustard, J. F. & Frank J. (1994). The determinants of health, pp. 7-48 in I h s determinants of health: A critical assessment (Western geographical series 29L Hayes, M. V., Foster, L. T., & Foster, H. D. (Eds.). Victoria: University o f Victoria. Nikiforuk, A. (1991). The fourth horseman: A short history of epidemics, plagues. famine and other scourges. Toronto: Viking. 67 Poland, B., Cobum, D., Robertson, A., & Eakin, J. (1998). Wealth, equity and health care: A critique of a population health perspective on the determinants of health. Social Science and Medicine. 46.785-98. Richardson, C. (1999). A statistical examination of the health utility index as a measure of health status for use in the national population health survey. Unpublished masters thesis. University of Northern British Columbia, Canada. Richardson, C., & Zumbo, B. D. (2000). A statistical examination o f the Health Utility Index - Mark ill as a summary measure o f health status for a general population health survey. Social Indicators Research. 51 (3), August, 171-191. Ross, N. A., Wolfson, M. C., Dunn, J. R., Berthelot, J., Kaplan, G. A., & Lynch, J. W. (2000). Relation between income inequality and mortality in Canada and in the United States: Cross sectional assessment using census data and vital statistics. British Medical Journal. 320.898-902. Schwarz, N., & Strack, F. (1999). Reports o f subjective well-being: Judgmental processes and their methodological implications, pp. 61-84 in Well-being: The foundations o f hedonic psychology. Kahneman, D., Diener, E., & Schwarz, N. (Eds.). New York: Russell Sage Foundation. Segovia, J., Bartlett, R. F., & Edwards, A. C. (1989). An empirical analysis of the dimensions o f health status measures. Social Science and Medicine. 29 (6), 761-768. Siddiqi, J. (1995). World health and world politics: The World Health Organization and the UN system. Columbus: University of South Carolina Press. Statistics Canada. (1995). National Population Health Survey. 1994-95. Statistics Canada. Ottawa. Catalogue number 82-567. Statistics Canada. (2000). Health reports: How healthy are Canadians? The Daily. March 31,1-4. Sutherland, R., & Fulton, J. (1994). Spending smarter and spending less: Policies and partnerships for health care in Canada. Ottawa: The Health Group. Sykes, J. B. (Ed.). (1978). The Concise Oxford Dictionary of Current English (6th ed ). Oxford: Clarendon Press. Tambay, J., & Catlin, G. (1995). Sample design of the national population health survey. Health Reports. 7 f n . 29-38. Taylor, M. G. (1987). Health insurance and Canadian public policy: The seven decisions that created the Canadian health insurance system and their outcomes (2nd Ed). Montreal: McGill-Quccn's Univcrsit>' Press. 68 Wolfson, M. C. (1994). Social proprioception: Measurement, data, and information from a population health perspective, pp. 287-316 in Why are some people healthy, and others not? The determinants o f health o f populations). Evans, R. G., Barer, M. L., & Marmor, T. L. (Eds.). New York: Aldine de Gruyter. World Health Organization - Geneva: http://www.who.int World Health Organization (Health for All) - Geneva: http://www.who.int/aboutwho/en/healthforall.htm World Health Organization. (1981). Global strategy for health for all by the year 2000. Geneva: Author. World Health Organization. (1998). The world health report 1998 - Life in the 21" centurv: A vision for all. 09921 Geneva: Author. 69 APPENDIX A 17 National Population Health Survey Variables Selected for Exploratory Factor Analysis 1 - Health Status: Vision Attribute Based on HSTAT-Ql to HSTAT-Q5. HSTAT-Ql Are/Is ... usually able to see well enough to read ordinary newsprint without glasses or contact lenses? HSTAT-Q2 Are/Is you/he/she usually able to see well enough to read ordinary newsprint with glasses or contact lenses? HSTAT-Q3 Are/Is you/he/she able to see at all? HSTAT-Q4 Are/Is you/he/she able to see well enough to recognize a friend on the other side of the street without glasses or contact lenses ? HSTAT-Q5 Are/Is you/he/she usually able to see well enough to recognize a friend on the other side o f the street with glasses or contact lenses? D W ISFG - Derived Vision Attribute 1 NO VISUAL PROBLEMS 2 PROBLEMS CORRECTED BY LENSES 3 PROBLEM SEEING DISTANCE/NOT CORRECTED 4 PROBLEM SEEING CLOSE/NOT CORRECTED 5 PROBLEM SEEING CLOSE and DISTANCE/NO SIGHT 99 NOT STATED A higher score indicates more severe problems. 2 - Health Status: Hearing Attribute Based on HSTAT-Q6 to HSTAT-Q9. HSTAT-Q6 A re/Is... usually able to hear what is said in a group conversation with at least three other people without a hearing aid? HSTAT-Q7 Are/Is you/he/she usually able to hear what is said in a group conversation with at least three other people with a hearing aid? HSTAT-Q7a Are/Is you/he/she able to hear at all? HSTAT-Q8 Are/Is you/he/she usually able to hear what is said in a conversation with one other person in a quiet room without a hearing aid ? HSTAT-Q9 Are/Is you/he/she usually able to hear what is said in a conversation with one other person in a quiet room with a hearing aid? DVHEAFG - Derived Hearing Attribute 1 NO HEARING PROBLEMS 70 2 PROBLEM HEARING/CORRECTED 3 PROBLEM HEARING/NOT CORRECTED 99 NOT STATED A higher score indicates more severe problems. 3 - Health Status: Speech Attribute Based on HSTAT-QIO to HSTAT-Ql3. HSTAT-QIO A re/Is... usually able to be understood completely when speaking with strangers in your own language? HSTAT-Ql 1 Are/Is you/he/she able to be understood partially when speaking with strangers? HSTAT-Q12 Are/Is you/he/she able to be understood completely when speaking with those who know you/him/her well? HSTAT-Ql3 Are/Is you/he/she able to be understood partially when speaking with those who know you/him/her well? DVSPEFG - Derived Speech Attribute 1 NO SPEECH PROBLEMS 2 PARTIALLY/ NOT UNDERSTOOD 9 NOT STATED A higher score indicates more severe problems. 4 - Health Status: Mobility Attribute Based on HSTAT-Q14 to HSTAT-Q20. HSTAT-Q14 A re/Is... usually able to walk around the neighbourhood without difficulty and without mechanical support such as braces, a cane or crutches? HSTAT-Ql5 Are/Is you/he/she able to walk at all? HSTAT-Q16 Do/Does you/he/she require mechanical support such as braces, a cane or crutches to be able to walk around the neighbourhood? HSTAT-Ql 7 Do/Does you/he/she require the help o f another person to be able to walk? HSTAT-Ql8 Do/Does you/he/she require a wheelchair to get around? HSTAT-Q19 How often do/does you/he/she use a wheelchair? HSTAT-Q20 Do/Does you/he/she need the help o f another person to get around in the wheelchair? DVMOBFG - Derived Mobility Attribute 1 NO MOBILITY PROBLEMS 71 2 MOBILITY PROBLEMS/NO AID 3 PROBLEMS/MECHANICAL SUPPORT 4 PROBLEMS/CANNOT WALK 99 NOT STATED A higher score indicates more severe problems. 5 - Health Status: Dexterity Attribute Based on HSTAT-Q21 to HSTAT-Q24. HSTAT-Q21 A re/Is... usually able to grasp and handle small objects such as a pencil and scissors? HSTAT-Q22 Do/Does you/he/she require the help of another person because of limitations in the use of hands or fingers? HSTAT-Q23 Do/Does you/he/she require the help of another person with: Some tasks? Most tasks? Almost all tasks? All tasks? HSTAT-Q24 Do/Does you/he/she require special equipment, for example, devices to assist in dressing because of limitations in the use o f hands or fingers? DVDEXFG - Derived Dexterity Attribute 1 NO DEXTERITY PROBLEMS 2 DEXTERITY PROBLEMS/NO HELP 3 DEXTERITY PROBLEMS/NEED HELP 99 NOT STATED A higher score indicates more severe problems. 6 - Health Status: Emotion Attribute Based on HSTAT-Q25. HSTAT-Q2S Would you describe yourself ... as being usually: Happy and interested in life? Somewhat happy? Somewhat unhappy? Unhappy with little interest in life? So uiihappy that life is not worthwhile? DVEMOF94 - Derived Emotion Attribute 1 HAPPY AND INTERESTED IN LIFE 72 2 SOMEWHAT HAPPY 3 SOMEWHAT UNHAPPY 4 UNHAPPY WITH A LITTLE INTEREST IN LIFE 5 SO UNHAPPY THAT LIFE IS NOT WORTHWHILE 9 NOT STATED A higher score indicates less perceived happiness. 7 - Health Status: Cognition Attribute Based on HSTAT-Q26 to HSTAT-Q27. HSTAT-Q26 How would you describe your/his/her usual ability to remember things? Are/Is you/he/she; Able to remember most things? Somewhat forgetful? Very forgetful? Unable to remember anything at all? HSTAT-Q27 How would you describe your/his/her usual ability to think and solve day to day problems? Are/Is you/he/she: Able to think clearly and solve problems? Having a little difficulty? Having some difficulty? Having a great deal of difficulty? Unable to think or solve problems? DVCOGFG - Derived Cognition Code 1 NO COGNITIVE PROBLEMS 2 NO MEMORY PROBLEMS 3 SOMEWHAT FORGETFUL 4 DIFFICULTY THINKING 5 VERY FORGETFUUUNABLE TO REMEMBER 99 NOT STATED A higher score indicates more severe problems. 8 - Health Status: Pain and Discomfort Attribute Based on HSTAT-Q28 and HSTAT-Q29. HSTAT-Q28 Are/Is... usually free of pain or discomfort? HSTAT-Q29 How would you describe the usual intensity o f your/his/her pain or discomfort? Mild 73 Moderate Severe DVPASF94 - Severity o f Pain Code 1 NO PAIN OR DISCOMFORT 2 MILD PAIN/DISCOMFORT 3 MODERATE PAIN/DISCOMFORT 4 SEVERE PAIN/DISCOMFORT 9 NOT STATED A higher score indicates more severe problems. 9 - Adjusted Specific Chronic Stress Index To adjust DVCSI294 according to the number of questions. DVCSI394 = (DVCSI294 * 16) / # of questions answered yes, no or don't know in DVCSI294 e.g., single with children: (DVCSI294 * 16) -î-14 In this third index, the range of scores of the second index, DVCSI294 is adjusted as if all the items were relevant to each respondent. DVCSI294 based on CSTRESS-Ql to Q4 and CSTRESS-Q12toQ18. CSTRESS-Ql You are trying to take on too many things at once. CSTRESS-Q2 There is too much pressure on you to be like other people. CSTRESS-Q3 Too much is expected of you by others. CSTRESS-Q4 You don't have enough money to buy the things you need. CSTRESS-Ql2 Your work around the home is not appreciated. CSTRESS-Ql3 Your fnends are a bad influence. CSTRESS-Q14 You would like to move but you cannot. CSTRESS-Ql5 Your neighbourhood or community is too noisy or too polluted. CSTRESS-QI6 You have a parent, a child or partner who is in very bad health and may die. CSTRESS-Ql7 Someone in your family has an alcohol or drug problem. CSTRESS-QI8 People are too critical of you or what you do. DVCSI394 - Derived adjusted specific chronic stress index 00 I 1 22 33 44 55 66 77 88 74 99 10 10 11 11 12 12 13 13 14 14 15 15 16 16 96 NOT APPLICABLE 99 NOT STATED A higher score indicates a greater number of chronic stressors. 10 - Work Stress Index Sum of all items in WSTRESS-Ql WSTRESS-Ql Now I'm going to read you a series o f statements that might describe your job situation. Please tell me if you STRONGLY AGREE, AGREE, NEITHER AGREE NOR DISAGREE, DISAGREE, or STRONGLY DISAGREE with each of the following: a) Your job requires that you learn new things b) Your job requires a high level of skill c) Your job allows you freedom to decide how you do your job d) Your job requires that you do things over and over e) Your job is very hectic f) You are free from conflicting demands that others make g) Your job security is good h) Your job requires a lot of physical effort i) You have a lot to say about what happens in your job j) You are exposed to hostility or conflict from the people you work with k) Your supervisor is helpful in getting the job done 1) The people you work with are helpful in getting the job done MIN = 0,MAX = 48 Respondents 15 and over who were currently employed were asked to evaluate their work situation. The 12-item index, based on a larger pool of items from Karasek, reflects respondents' perceptions about various dimensions o f their work including job security, social support, monotony, physical effort required and extent of participation in decision-making. DVWSI194 - Derived work stress - sum of all items 0-45 INDEX SCORE 96 NOT APPLICABLE 99 NOT STATED 75 Higher scores indicate greater work stress. 11 • Self-esteem index Sum of all items of ESTEEM-Ql ESTEEM-Ql a) You feel that you have a number of good qualities. b) You feel that you're a person o f worth at least equal to others. c) You are able to do things as well as most other people. d) You take a positive attitude toward yourself. e) On the whole you are satisfied with yourself. 0 All in all, you're inclined to feel you're a failure. MIN = 0,MAX = 24 The self-esteem index reflects the amount of positive feelings an individual holds about his/herself. Scores on the index are based on a subset of items from the self­ esteem Rosenberg scale (1969). The six items factored into one dimension in the factor analysis done by Pearlin and Schooler (1978). Respondents' answers are based on a S point scale: 0 = Strongly disagree 1 = Disagree 2 = Neither agree nor disagree 3 = Agree 4 = Strongly agree (Scores was reversed for item F.) DVESTI94 - Derived Self Esteem Scale - sum of all items 11 22 33 44 55 66 77 88 99 10 10 11 11 12 12 13 13 14 14 15 15 76 16 16 17 17 18 18 19 19 20 20 2121 22 22 23 23 24 24 99 NOT STATED Higher scores indicate greater self-esteem. 12 - Mastery index Sum of all items o f MAST-Ql MAST-Ql a) You have little control over the things that happen to you. b) There is really no way you can solve some of the problems you have. c) There is little you can do to change many o f the important things in your life. d) You often feel helpless in dealing with problems oHife. e) Sometimes you feel that you are being pushed around in life. f) What happens to you in the future mostly depends on you. g) You can do just about anything you really set your mind to. MIN = 0,MAX = 28 The index which measures sense of mastery is based on the work o f Pearlin and Schooler (1978). It measures the extent to which individuals believe that their lifechances are under their control. Respondents' answers are based on a 5 point scale: 0 = Strongly agree 1 = Agree 2 = Neither agree or disagree 3 = Disagree 4 = Strongly disagree (Scores were reversed for items F and G.) DVMASI94 - Derived Mastery Scale - sum o f all items 11 1 o 77 33 44 55 66 77 88 99 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 2 121 22 22 23 23 24 24 25 25 26 26 27 27 28 28 96 NOT APPLICABLE 99 NOT STATED Higher scores indicate superior mastery. 13 - Sense of Coherence scale Sum of SCOH-Ql to SC0H-Q13 SCOH-Ql How often do you have the feeling that you don't really care about what goes on around you? SC0H-Q2 How often in the past were you surprised by the behaviour of people whom you thought you knew well? SC0H-Q3 How often have people you counted on disappointed you? SC0H-Q4 How often do you have the feeling you're being treated unfairly? SC0H-Q5 How often do you have the feeling you are in an unfamiliar situation and don't know what to do? SC0H-Q8 Many people - even those with a strong character —sometimes feel like sad sacks (losers) in certain situations. How often have you felt this way in the past? 78 SC0H-Q9 How often do you have the feeling that there's little meaning in the things you do in your daily life? SCOH-QIO How often do you have feelings that you're not sure you can keep under control? SCOH-Ql 1 Until now your life has had no clear goals or purpose or has it had very clear goals and purpose? SC0H-Q12 When something happens, you generally find that you overestimate or underestimate its importance or you see things in the right proportion? SCOH-Ql 3 Is doing the things you do every day a source o f great pleasure and satisfaction or a source of pain and boredom? The 13-item version of the sense of coherence scale developed by Antonovsky was used in the NPHS. It denotes the extent to which individuals perceive events as comprehensible, manageable and meaningful. The concept o f manageability is addressed in questions Q3, Q4, Q8, and QIO. Items Q l, Q9, Q 11, and Q13 measure meaningfulness and items Q2, QS, Q6, Q7, Q12 are related to the comprehensibility dimension. Score was reversed for questions SCOHQl, Q2, Q3, Q8, Q13 DVSCI94 - Derived Sense of Coherence Scale 4-78 INDEX SCORE 96 NOT APPLICABLE 99 NOT STATED Higher scores indicate a stronger sense of coherence. 14 - Distress score Sum of questions MHLTH-Ql A to MHLTH-QIF MHLTH-INTa Now some questions about mental and emotional well-being. During the past month, about how often did you feel: MHLTH-Ql a ... so sad that nothing could cheer you up? All o f the time Most of the time Some of the time A little of the time None of the time M HLTH-Qlb... nervous? M HLTH-Qlc... restless or fidgety? M HLTH-Qld... hopeless? M HLTH-Qle... worthless? MHLTH-QIf During the past month, about how often did you feel that everything was an effort? 79 The items and scoring used to derive the distress score are based on the work of Kessler and Mroczek (from Michigan University). The index is based on a subset of items from the Composite International Diagnostic Interview (CIDI). The CIDI is a structure diagnostic instrument that was designed to produce diagnoses according to the definitions and criteria of both DSM-III-R and the Diagnostic Criteria for Research of the ICD-10. DVMHDS94 2 Derived Mental Health - Distress Scale 00 1 1 22 33 44 55 66 77 88 99 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 2121 22 22 23 23 24 24 99 NOT STATED Higher scores indicate more distress. 15 - Perceived social support index Sum of all true responses from questions S0CSUP-Q3 to S0CSUP-Q6 S0CSUP-Q3 Do you have someone you can confide in, or talk to about your private feelings or concerns? S0CSUP-()4 Do you have someone you can really count on to help you out in a crisis situation? 80 SOCSUP-QS Do you have someone you can really count on to give you advice when you are mahng important personal decisions? S0CSUP-Q6 Do you have someone that makes you feel loved and cared for? The perceived social support index is composed of four items which reflect whether respondents feel that they have someone they can confide in, someone they can count on, someone who can give them advice and someone who makes them feel loved. DVSS1194 • Derived Social Support Index 00 1 1 22 33 44 9 NOT STATED A higher score indicates greater perceived social support. 16 - Average frequency of contact index Based on S0CSUP-Q7A to S0CSUP-Q7H S0CSUP-Q7A How often did you have contact with your parents or parents-inlaw? S0CSUP-Q7B How often did you have contact with your grandparents? S0CSUP-Q7C How often did you have contact with your daughters or daughtersin-law? S0CSUP-Q7D How often did you have contact with your sons or sons-in-law? S0CSUP-Q7E How often did you have contact with your brothers or sisters? S0CSUP-Q7F How often did you have contact with other relatives (including in­ laws)? S0CSUP-Q7G How often did you have contact with your close fnends? S0CSUP-Q7H How often did you have contact with your neighbours? The average frequency of contact index measures the average number o f contacts in the past 12 months with family members and friends who are not part of the household and with neighbours. DVSSI394 = CONTACT /NETSIZE CONTACT is an approximate value indicating the number of contacts for all categories (S0CSUP-Q7A to S0CSUP-Q7H). NETSIŒ is a combined value indicating the existence o f possible persons to be contacted (sum of flags indicating ‘yes’ to parents, ‘yes’ to grandparents, etc.). 81 DVSSI394 - Derived average frequency of contacts 00 1 1 22 33 44 55 66 99 NOT STATED A higher number Indicates more contacts. 17 • Derived health description index Based on GENHLT-Ql. GENHLT-Ql In general, would you say ... r/’s health is: Excellent? Very good? Good? Fair? Poor? DVGHI94 - Derived health description index OPOOR IFAIR 2 GOOD 3 VERY GOOD 4 EXCELLENT A higher score indicates better health. (Statistics Canada, 1995) 82 APPENDIX B The 5 Richardson-Zumbo Factors with Corresponding Indicator Loadings 1 - Physical Impairment Factor Ind # 3 Health Status: Speech Attribute 5 Health Status: Dexterity Attribute Loading .447 1.0 2 - Mental Ill-Health Factor 6 9 12 13 14 Health Status: Emotion Attribute Adjusted Specific Chronic Stress Index Mastery Index Sense of Coherence Scale Distress Score .230 .657 -.240 -.685 .665 3 - Mental Well-Being Factor 6 10 11 12 Health Status: Emotion Attribute Work Stress Index Self-Esteem Index Mastery Index -.246 -.208 .838 .565 4 - General Health Impairment Factor 1 2 4 7 8 17 Health Status: Vision Attribute Health Status: Hearing Attribute Health Status: Mobility Attribute Health Status: Cognition Attribute Health Status: Pain and Discomfort Attribute Derived Health Description Index .257 .284 .437 .244 .451 -.527 5 - Social Well-Being Factor 6 15 16 Health Status: Emotion Attribute Perceived Social Support Index Average Frequency of Contact Index -.422 .355 .476 (Richardson and Zumbo, 2000, p. 183) 83 APPENDIX C: Correlations Between Dependent Variables Dependent Vnrisblef Heaith Utiiity index Health Utility index i.OOO n.s. 1 - Physical impairment Factor -.301 #* i.OOO n.s. 2 - Mental iiiHeaith Factor -.403 •• .019 n.s. i.OOO n.s. 3 - Mental Well-Being Factor .379 *• -.024 n.s. -.628 i.OOO n.s. -.775 »* .280 •• .312 •* -.280 ** 1.000 n.s. .518 ** -.114 *• -.566 *# .526 ## -.285 «* 4 - General Health impairment Factor 5 - Sociai Well-Being Factor 1 - Physical impairment Factor 2 - Mental illHealth Factor 3 - Mental WellBeing Factor 4 - General Health impairment Factor 5 - Sociai WellBeing Factor 1.000 n.s. Note. ** denotes correlation is significant at the p < .01 level, n.s. denotes that the test o f the correlation was not statistically significant. 2 Last page left intentionally blank. 85