Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The health of Canadian women in the workforce : a comparison between homemaker women, workforce women… Caruth, Fran 1987

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

Item Metadata

Download

Media
831-UBC_1987_A6_7 C36_9.pdf [ 10.33MB ]
Metadata
JSON: 831-1.0096872.json
JSON-LD: 831-1.0096872-ld.json
RDF/XML (Pretty): 831-1.0096872-rdf.xml
RDF/JSON: 831-1.0096872-rdf.json
Turtle: 831-1.0096872-turtle.txt
N-Triples: 831-1.0096872-rdf-ntriples.txt
Original Record: 831-1.0096872-source.json
Full Text
831-1.0096872-fulltext.txt
Citation
831-1.0096872.ris

Full Text

THE HEALTH OF CANADIAN WOMEN IN THE WORKFORCE: A Comparison Between Horaemaker Women, Workforce Women and Workforce Men Based on the 1979 Canada Health Survey by: FRANCES M. CARUTH B.Sc, University of Toronto, 1980 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Health Services Planning & Administration) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1987 @ Frances M. Caruth, 1987 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6(3/81) ABSTRACT In the past twenty-five years there has been a marked increase in the number of women in the paid labour force, especially among women with young children. Time studies have shown that when a woman has a young family plus a position in the paid labour force, she works a very long day and has l i t t l e time for recreational or leisure pursuits. This thesis therefore poses the following questions: 1. Do women who participate in the paid labour force report poorer health status than their counterparts who are homemakers? 2. Do women who participate in the paid labour force exhibit lifestyle patterns significantly different from their homemaker counterparts? 3. Do women in the paid labour force exhibit health care utilization patterns significantly different from their homemaker counterparts? and 4. Do women's lifestyles, reported health status and health care utilization patterns differ from those of their male counterparts in the paid labour force? Data from the 1978-79 Canada Health Survey (C.H.S.), which had asked a wide cross-section of Canadians about their li f e s t y l e , health status and use of the health care system, were used to explore these questions. A model was then developed for this study which linked health risk behaviours, health status and health care related behaviours, and which used the variables available in the C.H.S. data base. - i i i -Multiple Classification Analyses were carried out to determine the best predictors of women's health risk behaviours, health status and health care related behaviours. The three study groups were then standardized using the top two predictors and the rates of the various states and behaviours were compared. First, in the prediction of women's health risk behaviours, the demographic variables included in the model were not effective as only 3-4% of the variance in the scores could be explained. Secondly, in the prediction of health status scores, the composite health risk scores developed for each subject plus the demographic variables were able to explain 4 - 11% of the variation. Thirdly, in the prediction of women's health care related behaviours the composite health risk scores, the health status scores and the demographic variables were together able to explain 14 - 27% of the variance. When the standardized rates for high health risk behaviours were compared, there were significant differences between the three groups but no group was consistently better or worse than any other. The men's group however, consistently reported better health and less use of the health care system. The women's groups reported similar health states but women in the paid labour force reported a higher use of medications and fewer days in hospital. The C.H.S. was designed to address issues which affect the whole population. The questions therefore, were not always sufficiently specific to describe the special circumstances of women, especially for example in their childbearing and nurturing years. The rapidly changing social and economic circumstances of women and their families, as women enter the paid labour force, plus the need for more information on their health risk behaviours - what these behaviours are, and what predisposes women to engage in them - point to the need for more research focused specifically on this section of the population. - V -TABLE OF CONTENTS PAGE ABSTRACT i i TABLE OF CONTENTS v LIST OF TABLES v i i LIST OF FIGURES x i i ACKNOWLEDGEMENTS x i i i CHAPTER 1. INTRODUCTION 1 CHAPTER 2. BACKGROUND 6 2.1 Lifestyle and Health 6 2.2 Women and Work in Canada 15 2.3 Women and Their Health 22 2.4 Employed Women and their Health 24 CHAPTER 3. THE CANADA HEALTH SURVEY 28 3.1 Overview 28 3.2 The Sample Design 29 3.3 The Scope of the Survey 31 3.4 The Canada Health Survey Data 33 CHAPTER 4. THE MODEL 36 4.1 The Questions and the Structure 36 4.2 Demographic Factors 42 4.3 Health Risk Behaviours 48 4.4 Health Status Indicators 71 4.5 Health Care Behaviours 77 CHAPTER 5. DATA ANALYSIS 82 5.1 The Data .' 82 5.2 Multiple Classification Analysis 83 5.3 Standardization of Rates 85 5.4 Testing for Significance between Rates 87 CHAPTER 6. THE RESULTS 88 6.1 Health Risk Behaviours 89 6.2 Health Status Indicators 139 6.3 Health Care Behaviours 151 6.4 Demographic Predictors 173 - v i -CHAPTER 7. DISCUSSION 186 7.1 The Question 186 7.2 The Model 187 7.3 Health Risk Behaviours 189 7.4 Health Status 192 7.5 Health Care Behaviours 192 7.6 Limitations 193 7.7 Issues for Further Study 194 7.8 Conclusion 195 SELECTED BIBLIOGRAPHY 197 APPENDIX 203 - v i i -LIST OF TABLES PAGE 2.1 Lifestyle Risks to Health 8 2.2 Allocation of Mortality Responsibility to Four Health Areas 9 2.3 Relative Risks for a l l CHD Deaths with Three Risk Factors - Cigarette Smoking, Hyper Cholesterolemia, and Hypertension 10 2.4 Seven Health Habits 11 2.5 Risk Factor - Cause of Death Matrix for Health Hazard Appraisal System 13 2.6 Percentage of Women in the Paid Labour Force 1911-1946 16 2.7 Cohorts of Women - Canada 1911-1960 17 2.8 Percentage of Female Labour Force by Marital Status 17 2.9 Participation Rates of Married Women (Husband Present), by Age Group and Presence of Children in the Home, 1971 and 1976 18 2.10 Selected Activities From Five Studies of Housewives and Working Wives 22 3.1 Provincial Strata and Clusters 32 4.1 Percentage Distribution of Study Group by Region and Sex 43 4.2 " " " " Season of Data Collection and Sex 44 4.3 Percentage Distribution of Study Group by Marital Status and Sex 44 4.4 Percentage Distribution of Study Group by Family Size and Sex 45 4.5 Percentage Distribution of Study Group by Income and Sex 46 4.6 Percentage Distribution of Study Group by Education and Sex 46 4.7 Percentage Distribution of Study Group by Economic Responsibility and Sex 47 4.8 Percentage Distribution of Study Group by Cigarette Smoking Experience and Sex 50 4.9 Percentage Distribution of Study Group by Alcohol Consumption and Sex 51 4.10 Physical Activity Index Categories for a l l Subjects in the Canada Health Survey 54 4.11 The Fitness Canada Prescription for Physical Activity 54 4.12 Percentage Distribution of Study Group by Physical Activity Level and Sex .55 4.13 Social Readjustment Rating Scale 56 4.14 Percentage Distribution of Study Group by Life Events and Sex 57 4.15 Percentage Distribution of Study Group by Leisure Time Companionship and Sex 58 4.16 Percentage Distribution of Study Group by Hormone P i l l Consumption 60 - v i i i -4.17 Percentage Distribution of Frequency of Preventive Behaviours of Women in the Study Group 64 4.18 Percentage Distribution of Composite Preventive Behaviour Scores for Women in the Study Group 65 4.19 Percentage Distribution of Annual Automobile Travelling Distances of the Study Group 67 4.20 Percentage Distribution of Study Group by Seat Belt Use and Sex 68 4.21 Creating Composite Risk Scores 69 4.22 Percentage Distribution of Study Group by Composite Health Risk Score and Sex 70 4.23 C.H.S. Question Used to Identify Health Problems 71 4.24 Percentage Distribution of Study Group by Number of Chronic Health Problems and Sex 72 4.25 Subjects Excluded Due to Activity Limitations 72 4.26 Developing an Health Opinion Score 73 4.27 Percentage Distribution of Study Group by Health Opinion Score and Sex 74 4.28 Developing Affect Balance Scores 75 4.29 Percentage Distribution of Study Group by Affect Balance Score and Sex 76 4.30 C.H.S. Question to Identify Disability Days 77 4.31 Percentage Distribution of Study Group by Reduced Activity Days and Sex 78 4.32 C.H.S. Question to Identify Health Professional Consultations 78 4.33 Percentage Distribution of Study Group by Number of Health Professional Consultations and Sex 79 4.34 C.H.S. Question to Identify Medications Taken 79 4.35 Percentage Distribution of Study Group by Variety of Medications Consumed and Sex 80 4.36 Percentage Distribution of Study Group by Hospitalization and Sex 81 6.1 Sample Sizes After Adjustment for Those Reporting Activity Limitations 88 6.2 Predictors of Smoking by Women Subjects 89 6.3 Smoking Prevalence: Women Homeraakers Aged 25-44 Years 91 6.4 Smoking Prevalence: Women in the Paid Labour Force, Aged 25-44 Years 92 6.5 Smoking Prevalence: Men in the Paid Labour Force, Aged 25-44 Years 92 6.6 Percentage of Daily Cigarette Smokers Among 25-44 Year Olds, Reported by the 'Smoking Habits of Canada' Surveys and the C.H.S 91 6.7 Weighted and Unweighted C.H.S. Smoking Rates for Women Aged 25-44 Years 91 6.8 Predictors of Alcohol Consumption by Women Subjects 93 6.9 Alcohol Consumption: Women Horaemakers Aged 25-44 Years 95 6.10 Alcohol Consumption: Women in the Paid Labour Force, Aged 25-44 Years 96 - i x -6.11 Alcohol Consumption: Men Aged in the Paid Labour Force, Aged 25-44 Years 97 6.12 Non-Drinkers: Women Homemakers Aged 25-44 Years 98 6.13 Non-Drinkers: Women in the Paid Labour Force, Aged 25-44 Years .99 6.14 Non-Drinkers: Men in the Paid Labour Force, Aged 25-44 Years 100 6.15 Predictors of Physical Activity Levels For Women Subjects ..101 6.16 Physical Activity: Women Homemakers Aged 25-44 Years 102 6.17 Physical Activity: Women in the Paid Labour Force, Aged 25-44 Years 103 6.18 Physical Activity: Men in the Paid Labour Force, Aged 25-44 Years 104 6.19 Predictors of Companionship During Leisure Hours For Women Subjects 106 6.20 Companionship in Leisure Time: Women Homemakers Aged 25-44 Years 108 6.21 Companionship in Leisure Time: Women in the Paid Labour Force, Aged 25-44 Years 109 6.22 Companionship in Leisure Time: Men in the Paid Labour Force, Aged 25-44 Years 110 6.23 Predictors of Hormone P i l l Consumption by Women Subjects....111 6.24 Use of Hormone P i l l s : Women Homemakers Aged 25-44 Years....112 6.25 Use of Hormone P i l l s : Women in the Paid Labour Force, Aged 25-44 Years 113 6.26 Predictors of Life Change Events Among Women Subjects 114 6.27 Life Change Events: Women Homemakers Aged 25-44 Years 115 6.28 Life Change Events: Women in the Paid Labour Force, Aged 25-44 Years 116 6.29 Life Change Events: Men in the Paid Labour Force, Aged 25-44 Years 117 6.30 Predicting Motor Vehicle Travel by Women Subjects 118 6.31 Motor Vehicle Travel as a Driver or Passenger: Women Homemakers Aged 25-44 Years 120 6.32 Motor Vehicle Travel as a Driver or Passenger: Women in the Paid Labour Force, Aged 25-44 Years 121 6.33 Motor Vehicle Travel as a Driver or Passenger: Men in the Paid Labour Force, Aged 25-44 Years 122 6.34 Predictors of Motor Vehicle Seat Belt Use by Women Subjects 123 6.35 Use of Motor Vehicle Seat Belts: Women Homemakers Aged 25-44 Years 125 6.36 Use of Motor Vehicle Seat Belts: Women in the Paid Labour, Aged 25-44 Years 126 6.37 Use of Motor Vehicle Seat Belts: Men in the Paid Labour Force, Aged 25-44 Years 127 6.38 Predictors of Female Preventive Behaviours Among Women Subjects 128 6.39 Female Preventive Behaviours: Women Homemakers Aged 25-44 Years 130 6.40 Female Preventive Behaviours: Women in the Paid Labour Force, Aged 25-44 Years 131 6.41 Predictors of Composite Risk Scores for Women Subjects 132 -x-6.42 Poor Composite Risk Scores: Women Homemakers Aged 25-44 Years 133 6.43 Poor Composite Risk Scores: Women in the Paid Labour Force, Aged 25-44 Years 134 6.44 Poor Composite Risk Scores: Men in the Paid Labour Force, Aged 25-44 Years 135 6.45 Good Composite Risk Scores: Women Homemakers Aged 25-44 Years 136 6.46 Good Composite Risk Scores: Women in the Paid Labour Force, Aged 25-44 Years 137 6.47 Good Composite Risk Scores: Men in the Paid Labour Force, Aged 25-44 Years 138 6.48 Predictors of Chronic Health Problems Among Women Subjects 139 6.49 Chronic Health Problems: Women Homemakers Aged 25-44 Years 141 6.50 Chronic Health Problems: Women in the Paid Labour Force, Aged 25-44 Years 142 6.51 Chronic Health Problems: Men in the Paid Labour Force, Aged 25-44 Years 143 6.52 Predictors of Health Opinion Scores Among Women Subjects.... 144 6.53 Health Opinion Scores: Women Homemakers Aged 25-44 Years...145 6.54 Health Opinion Scores: Women in the Paid Labour Force, Aged 25-44 Years 146 6.55 Health Opinion Scores: Men in the Paid Labour Force, Aged 25-44 Years 146 6.56 Predictors of Affect Balance Scores Among Women Subjects.... 147 6.57 Affect Balance: Women Homemakers Aged 25-44 Years 149 6.58 Affect Balance: Women in the Paid Labour Force, Aged 25-44 Years 149 6.59 Affect Balance: Men in the Paid Labour Force, Aged 25-44 Years 150 6.60 Affect Balance: A l l Subjects, Aged 25-44 Years 150 6.61 Predictors of Disability Days Among Women Subjects 151 6.62 Disability Days: Women Homemakers Aged 25-44 Years 153 6.63 Disability Days: Women in the Paid Labour Force, Aged 25-44 Years 154 6.64 Disability Days: Men in the Paid Labour Force, Aged 25-44 Years 155 6.65 Predictors of Health Professional Consultations by Women Subjects 156 6.66 Health Professional Consultations: Women Homemakers Aged 25-44 158 6.67 Health Professional Consultations: Women in the Paid Labour Force, Aged 25-44 Years 159 6.68 Health Professional Consultations: Men in the Paid Labour Force, Aged 25-44 Years 160 6.69 Predictors of Medication Use by Women Subjects 161 6.70 Regular Taking of Medications: Women Homemakers Aged 25-44 Years 163 - x i -6.71 Regular Taking of Medications: Women in the Paid Labour Force, Aged 25-44 Years 164 6.72 Regular Taking of Medications: Men in the Paid Labour Force, Aged 25-44 Years 165 6.73 Predictors of Nights in Hospital by Women Subjects 166 6.74 Nights in Hospital: Women Homemakers Aged 25-44 Years 168 6.75 Nights in Hospital: Women in the Paid Labour Force, Aged 25-44 Years 169 6.76 Nights in Hospital: Men in the Paid Labour Force, Aged 25-44 Years 170 6.77 Multiple Classification Analysis of Health Risk Behaviours for a l l Female Subjects 175 6.78 Multiple Classification Analysis of Reported Health Status and Health Care Related Behaviours for a l l Female Subjects 176 6.79 Prevalence of Health Risk Behaviours: A Comparison Between Men and Women in the Paid Labour Force and Women Homemakers 177 6.80 Self-Reported Health Status: A Comparison Between Men and Women in the Paid Labour Force and Women Homemakers 178 6.81 Health Care Related Behaviours: A Comparison Between Men and Women in the Paid Labour Force and Women Homemakers 179 6.82 Selected Lifestyle and Health Scores for Women in the Study Group, by Region of Residence 180 6.83 Selected Lifestyle and Health Scores for Women in the Study Group, by Marital Status 181 6.84 Selected Lifestyle and Health Scores for Women in the Study Group, by Level of Education 182 6.85 Selected Lifestyle and Health Scores for Women in the Study Group, by Family Size 183 6.86 Selected Lifestyle and Health Scores for Women in the Study Group, According to Financial Responsibility for Economic Family Unit 184 6.87 Selected Lifestyle Scores for Women in the Study Group, by Family Income 184 6.88 Companionship Scores for Women in the Study Group, by the Season of the Survey 185 - x i i -LIST OF FIGURES PAGE 2.1 Age-Adjusted Mortality Rates by Health Habit Score 11 2.2 Participated Rates for Successive Cohorts of Women 1911-1960 16 3.1 Basic Model of Health 28 3.2 Summary of Data Items Induced in the Canada Health Survey 35 4.1 The Basic Components of the Model 38 4.2 Explaining Variations Within the Model 40 4.3 The Variables Contained Within the Model 41 - x i i i -ACKNOWLEDGEMENTS In preparing this thesis I have had the guidance and assistance of many generous people. I am deeply grateful to the members of my thesis committee, Dr. J.H. Milsum, Dr. B. Morrison, Dr. M. Meissner and Dr. T.M. Williams who were always responsive to my requests for assistance. I am indebted to Dr. Milsum without whose constant guidance and encouragement, this thesis could not have been prepared. I owe special thanks to Dr. Morrison for her stat i s t i c a l assistance and advice. To Wendy Klein who typed this thesis and Wayne Jones who helped with the computing, I owe a great many thanks. Their unfailing support of the project has been deeply appreciated. Thanks also to the many women who over the years have shared their experiences and thoughts with me. They have helped to make me more aware, and enabled me to have a broader perspective on women's issues. Finally I must thank the men in my l i f e , my husband and two sons: thank you for tolerating my absences from family l i f e and thank you for always believing that my work was important. Fran Caruth September 1987 - 1 -CHAPTER ONE INTRODUCTION There i s growing evidence to support the idea that the way we lead our l i v e s w i l l have a subsequent impact upon our health. "Investigators are finding that the s o c i a l environment and certain common behaviours such as physical a c t i v i t y , use of alcohol and tobacco as well as eating and sleeping habits are related to the major diseases of our time" (Berkman and Breslow, 1983). During the past 80 years i n Canada, the change from a predominantly rur a l to an urban and i n d u s t r i a l i z e d society has led to the development of new so c i a l and work patterns for many people; improved public health measures, s o c i a l assistance programmes and advances i n medicine have led to a s i g n i f i c a n t change i n the demographic p r o f i l e of our society. In p a r a l l e l with these changes, society appears to have created a new set of health problems - diseases associated with prosperous urban l i v i n g , environmental pollu t i o n and a longer l i f e . At the st a r t of t h i s century, most women worked only around the home, performing the tasks generally described as homemaking. In 1911, only 14% of a l l Canadian females over the age of 10 years participated in the paid labour force compared with 79% of a l l men^. This separation of the male-female work environments continued through the f i r s t half of th i s century, with fluctuations during war time and with only a gradual increase in the ove r a l l number of women i n the paid labour force. Census data f a i l e d to account for women doing piecework, housecleaning or having a boarder ( P h i l l i p s and P h i l l i p s , 1983). - 2 -Since the 1950's there has been a sharp increase in the proportion of women participating in the paid labour force. In 1979, the proportion was 46.8% and i t has been predicted by some to reach 70% by the end of this century (Swan, 1981). This accelerated and seemingly permanent change in work patterns, particularly for married women, must also have considerable impact on women's social and home lives, and consequently on the lives of their families, i f only for reasons of time. Women's participation in the paid labour force therefore, presents the potential for marked changes in the lifestyles as well as the social, work and home environments of large sections of society. To society and to the family, there are both rewards and costs associated with these changes. Men have to learn to live and work in a more egalitarian society where women have more economic independence and expect an equal part in decision making. Men may also have to deal with work place or social expectations based on the traditional patriarchal model while simultaneously having to recognize an obligation to help with housekeeping and childraising activities (Ramie, 1983). Men and women have to learn new ways of caring for a family (should they decide to have one) while both parents work. For women entering the workforce there are more and different conflicts. Many women seek employment out of economic necessity; some do i t to enhance the family's standard of living; some seek economic independence and the -3-sense of self-esteem that comes with some forms of paid employment. Whatever their reasons, most women wil l s t i l l assume primary responsibility for the homemaking and childrearing. Can women assume these multiple responsibilities without jeonardizing their health? Women in the paid labour force with young children at home will work an 80 hour week (Ramie, 1983); there will be l i t t l e recreational and personal time. Will there be negative consequences to the mental or physical health of these women? From a personal perspective, I am intensely interested in the answers to such questions. Through three generations of women, I have seen my schoolteacher grandmother, who lived an active 95 years, work only in the home after her marriage at the age of 21; my schoolteacher mother, who stopped work when she married at the age of 28, return to part-time teaching for another 20 years before finally retiring at 65; and myself married at 25, work continuously both in and out of the home, since leaving university except for brief periods when the children were born and when we moved to new cities . My lif e s t y l e , particularly the lack of recreational and personal time, is in sharp contrast to that of my mother's and grandmother's. Am I jeopardizing the possibility of a long and healthy l i f e by working such long hours being wife, mother, employee and graduate student? Much has been written about the bases of mortality and morbidity differences between the sexes - women having higher levels of morbidity but having a seven year greater l i f e expectancy at birth. Women aged 20 and 40 -4-in 1981 were expected to outlive men of the same age by seven and six years respectively (Statistics Canada, 1984). If one accepts that there are biological bases for these differences - for example that women have a constitutionally greater resistance to both infection and degenerative disease (Nathanson, 1975) then changes in women's lifestyles may not lead to marked changes in morbidity or mortality patterns. However, i f one considers the cultural and behavioural explanations of women's higher levels of morbidity (Waldron, 1983(a)) coupled with the evidence of the potential impact of environmental factors on health (Lalonde, 1974; Milsum, 1984) then there would seem to be a very reasonable basis for asking the question, "Will women's increasing participation in the paid labour force have an impact on their morbidity and mortality patterns, and will that in turn affect their use of health care services?" The Canada Health Survey (C.H.S.) in 1978-79 sought information on the lifestyles and health of Canadians as well as their use of the health care system. Unfortunately, what was designed as an ongoing study of the health of the Canadian population survived only 10 months due to government restraint. During that time, however, 31,000 people participated in at least part of the survey and those data are available to the public. Although the C.H.S. was not designed to answer the specific questions posed by this thesis, i t provided an existing data base against which many hypotheses could be tested. Information gleaned in this way, plus the associated literature review was assessed for this study as being an appropriate way of becoming familiar with the issues and setting a basis for any future study which might involve new data collection. -5-The questions posed in this thesis therefore are based on and limited by the model used in the Canada Health Survey. Chapter Two provides an overview of the issues surrounding the health of women and their participation in the labour force. Chapter Three outlines the scope of the Canada Health Survey and Chapter Four provides details of the variables as they relate to the model. Chapter Five explains the analyses used and Chapter Six reports the results as well as commenting on their relationship to other known data. The final Chapter attempts to bring the results of this study into sharper focus and address some of the issues raised but not necessarily addressed by this study. Ultimately i t i s hoped that this thesis, i f only through my own increased knowledge, can contribute to the awareness of employers, health planners, researchers and women themselves; that others will go on to more successfully describe relationships between women's lifestyles and their health. - 6 -CHAPTER TWO BACKGROUND 2.1 LIFESTYLE AND HEALTH In 1974 the preface to "A New Perspective on the Health of Canadians" (Lalonde, 1974) stated that 'ominous counter-forces have been at work to undo progress in raising the health status of Canadians. These counter-forces ... include environmental pollution, city living, habits of indolence, the abuse of alcohol, tobacco and drugs and eating patterns ...'. In addressing the issues related to optimizing the health of Canadians, this report identified four sets of elements which could affect the health of the individual Canadian:-human biology - those aspects of health which are developed within the basic biological and organic makeup of the individual; environment - those factors related to health which are external to the human body and over which the individual has l i t t l e or no control; lif e s t y l e - the aggregation of behavioural decisions by individuals which affect their health and over which they more or less have control; health care organization - the quantity, quality, arrangement, nature and relationships of people and resources for the provision of health care services to which the individual has access. In examining the influence on our health of factors over which we as individuals and as a society have some control, the Lalonde Report quotes McKeown whose conclusions, after tracing the level of health in England and Wales back to the 18th century were: 'in order of importance the major contributions to improvement in health in England and Wales were limitation of family size (a behavioural change), increase in food supplies and a healthier physical environment (environmental influences), and specific preventive and therapeutic measures.' (McKeown, 1973) and -7-' Past improvement has been due mainly to modification of behaviour and changes i n the environment and i t i s to these same influences that we must look p a r t i c u l a r l y for further advances.'(McKeown, 1972) Further evidence of the impact of l i f e s t y l e on health can be drawn from the Canadian mortality data. The 1984 Canadian mortality data ( S t a t i s t i c s Canada, 1984) shows accidents, (motor vehicle and other) to be the leading cause of death for men aged 25 - 44 years and for women 25 - 34 years. For women 35 - 44 years, the leading cause of death i s neoplasia, followed by accidents. For the most part, these accidental deaths can be attributed to human factors such as carelessness, impaired driving and despair. They represent self-imposed r i s k , namely l i f e s t y l e decisions by the i n d i v i d u a l . For many years heart disease or cardio-vascular disease has been among the top three causes of death i n men and women aged 25 - 44 years ( S t a t i s t i c s Canada, 1984; Milsum, 1984). Smoking, obesity, stress, lack of exercise and a high fat diet are a l l known to contribute s i g n i f i c a n t l y to these diseases. Again there i s a high element of self-imposed r i s k . Table 2.1 shows the l i s t of self-imposed r i s k s i d e n t i f i e d i n "A New Perspective on the Health of Canadians" (Lalonde, 1974). A more detailed assessment of the r e l a t i v e contributions of l i f e s t y l e , environment, human biology and health care organizations to some of the leading causes of death for the population of Georgia was developed by Dever i n 1976 (Dever, 1976). ' (See Table 2.2.) Dever developed the percentage al l o c a t i o n s of r e s p o n s i b i l i t y for mortality by averaging the opinions of a panel of health experts. -8-TABLE 2.1 LIFESTYLE RISKS TO HEALTH RISK BEHAVIOURS POSSIBLE CONSEQUENCES Excess Alcohol Consumption Cirrhosis of the liver, malnutrition, encephalopathy, obesity, increased risk of motor vehicle accidents Cigarette Smoking Chronic bronchitis, emphysema, cancer of the lung, aggravation of coronary artery disease Abuse of Pharmaceuticals drug dependence, drug reaction Use of Psychotrophic Drugs social withdrawal, acute anxiety attacks could lead to suicide, homicide, malnutrition and accidents Poor Diet obesity, atherosclerosis, coronary-artery disease, dental caries, malnutrition Lack of Exercise aggravation of coronary-artery disease, obesity, poor physical fitness Lack of Recreation & Relief From Pressures stress related diseases such as hyper-tension, coronary-artery disease and peptic ulcers Careless Driving & Failure to Wear Seat Belt accidents possibly leading to injury or even death Promiscuity & Carelessness leading to syphilis and gonorrhea* * In 1974 AIDS was unknown in Canada. Source: M. Lalonde. A New Perspective on the Health of Canadians. Ottawa: Information Canada, 1974. -9-TABLE 2.2 ALLOCATION OF MORTALITY RESPONSIBILITY TO FOUR HEALTH AREAS PERCENTAGE ALLOCATION OF RESPONSIBILITY FOR MORTALITY PERCENTAGE DISTRIBUTION ENVIRON- HUMAN HEALTH CARE OF DEATHS* CAUSE OF DEATH LIFESTYLE MENT BIOLOGY ORGANIZATION 34.0 Diseases of the heart 54 9 28 12 14.9 Cancer 37 24 29 10 13.4 Cerebrovascular disease 50 22 21 7 4.2 Motor-vehicle accidents 69 18 1 12 3.8 Accidents, a l l others 51 31 4 14 3.8 2.7 Influenza & pneumonia Diseases of the 23 20 39 18 2.6 respiratory system Diseases of a r t e r i e s , 40 24 24 13 veins, & c a p i l l a r i e s 49 8 26 18 2.2 Homicides 66 30 5 0 100.0 Average for a l l causes 43 19 27 11 Note: Due to rounding, a l l o c a t i o n s may not add up to 100 percent. * i n 1973; only top nine causes presented. Source: J.H. Milsum. Health Stress and I l l n e s s : A Systems Approach. New York: Praeger, 1984. As Milsum (1984) comments, i t emerges from t h i s data that i n nine conditions leading to 94% of a l l deaths, " l i f e s t y l e i s allocated the greatest r e s p o n s i b i l i t y for each cause except influenza and pneumonia"; for a l l nine causes of death, l i f e s t y l e was allocated 43% of the t o t a l r e s p o n s i b i l i t y . Many other writers have linked l i f e s t y l e to health including Badura (1984), Cooper and Melhuish (1984) and Berkman and Syme (1979). Other l i f e s t y l e factors, not shown i n Table 2.1 which have demonstrated a cor r e l a t i o n with health outcomes include stress as a res u l t of major changes in s o c i a l and work environments (Holmes and Rahe, 1967), use of ora l contraceptives (Gibbs, 1979), hours of sleep (Hammond, 1964) and s o c i a l support (Berkman and Syme, 1979). -10-One very important aspect of the research done on lif e s t y l e risks for health includes the combined effect of risk factors. In 1974 Stamler and Epstein reported on the individual and cumulative effects of high cholesterol, high blood pressure and cigarette smoking on coronary heart disease. The cutting points for declaring the risk factors 'present' were cholesterol > 250 mg/dl, diastolic blood pressure > 90 ram/Hg and any use of cigarettes at the time of the study. The writers acknowledged that the setting of c l i n i c a l markers for cholesterol and blood pressure distorts the reality that the risk factors and the associated risks are on a continuum, but defended the approach for i t s practical application in the c l i n i c a l setting. Table 2.3 shows the multiplicative effect of increasing numbers of risk factors on the risk of heart disease. TABLE 2.3 RELATIVE RISKS FOR ALL CHD DEATHS WITH THREE RISK FACTORS -CIGARETTE SMOKING, HYPERCHOLESTEROLEMIA, AND HYPERTENSION Number of High-Risk Variables Prevalence Ten-Death -Year Age-Adjusted Rates per 1,000 Men Relative Risk 0 .17 13 1.0 1 .45 23 1.8 2 .30 44 3.4 3 .08 82 6.3 Note: Data are for U.S. white males, age 30 to 59 at entry, and age-adjusted. Source: J. Stamler and F.H. Epstein, "Coronary Heart Disease: Risk Factors as Guides to Preventive Action," Preventive Medicine 1 (1972): 27-48. cited in: J.H. Milsum. Health, Stress and Illness: A Systems Approach. New York: Praeger, 1984. Belloc and Breslow (1972) showed that seven health habits were positively correlated with physical health. These factors were dichotomized at the levels shown (See Table 2.4) and because (a) the relative risks were in approximately the same range and (b) the factors were not highly inter-related, an accumulation of the number of these seven health habits reported could be used to form a health practice score ranging from zero to seven. Using nine and a half years of data, Breslow and Enstrom (1980) plotted average health habit scores against age adjusted mortality rates as shown in Figure 2.1. Mortality rates increased approximately two fold for women and four fold for men. As scores moved from seven health habits to three or less the mortality rates generally held across age groups, income levels and different health states. TABLE 2.4 SEVEN HEALTH HABITS 1. Smoking None 2. Alcohol No more than 1-2 Drinks at a time 3. Physical Activity Frequence 4. Weight Men (-10) - (+20) of optimum Women less than (+10) of optimum 5. Sleep 7 - 8 hours per night 6. Eating Habit - Breakfast regularly 7. Eating Habit - Very limited snacking Source: Based on the work of Belloc & Breslow (1972), cited in: J.H. Milsum. Health Stress and Illness: A Systems Approach. New York: Praeger, 1984. FIGURE 2.1 AGE - ADJUSTED MORTALITY RATES BY HEALTH HABIT SCORE 0.20 -o I & 0.00 I < 3 4 5 6 7 No. of Health Habits * For f i r s t nine and a half (9 1/2) years of the study. Source: Data from L. Breslow and J.E. Enstrom, "Persistence of Health Habits and Their Relationship to Mortality," Preventive  Medicine 9 (1980): 469-83. Cited in: J.H. Milsum. Health, Stress  and Illness: A Systems Approach. New York: Praeger, 1984. -12-Supported by the research cited here and many other studies, Health Hazard Appraisal systems have been developed. These systems provide individuals with an estimate of th e i r r i s k exposure as well as (and more importantly) information on the behaviours required to achieve or maintain an acceptable l e v e l of health r i s k . Table 2.5 shows the variables used i n the Health Hazard Appraisal system currently sponsored by the Department of National Health and Welfare through the University of B r i t i s h Columbia^. This system has been used widely i n Canada and in the U.S.A. using a similar underlying configuration. The p r a c t i c a l r e a l i t y of l i f e s t y l e generated r i s k i s that the offending behaviours are often associated with p a r t i c u l a r l i v i n g or working circumstances that are not ea s i l y and simply changed. For example, smoking and drinking may be used as stress r e l i e v e r s and therefore the whole issue of stress management would be involved i n behaviour changes. If one's spouse, work mates or s o c i a l companions lead to a sedentary existence involving drinking and smoking, then again changes w i l l require more than a change i n a single facet of behaviour. As the work by Meissner et a l . (1975) showed i n Women and Work, when a woman has a family and a fu l l - t i m e position i n the paid labour force, increasing l e i s u r e time and active l e i s u r e time which are essential for physical and mental well-being would probably require a behavioural change by her partner so that there was greater sharing of homeraaker/childcare a c t i v i t i e s . Department of Health Care and Epidemiology, Division of Preventive Medicine & Health Promotion TABLE 2 . 5 RISK FACTOR - CAUSE OF DEATH MATRIX FOR HEALTH HAZARD APPRAISAL SYSTEM ^Causc of Death Risk Factor Car dio-Vascular Cancers Heart Attack Arterial Disease Stroke (Other) Hyper-tension Diabetes Lung Breast Cervical Intestinal 4 Rectal Chronic Bronchitis/ Emphysema Pneumonia M.V.A. Suicide Liver Cirrhosis (1) A f e Smoking Alcohol PhysicaJ-Unfitrjess Overweight Distance Driven Seat-Belt Non-Usage Disabling Depressions (No) Pap Smear (No) Breast Self-Exam Age Regular Sexual Intercourse Cholesterol Hypertension Uncontrolled Diabetes (2) H. Rectal Growth H. Rectal Bleeding H. Chronic Bronchitis (2) F. H. Heart Attack F. H. Diabetes F. H. Breast Cancer F. H. Suicide X XO XO XO X XO X XO X XO X XO XO XO XO XO XO X XO XO XO I XO X X X XO XO XO XO XO XO XO (3) X X X - RF appbes to this cause of deuh; O - reduction of Ihb mk recommended when appropriate. (I) Age is a RF in combination with Mother RF. eg. smoking, cholesterol: (2) H. History of; FH - Family History of; (3) Pretence of this disease invalidates any RF estimates for that disease as cause pf death. S o u r c e : J . H . Milsum. H e a l t h , S t r e s s and I l l n e s s : A Systems Approach. New York: P r a e g e r , 1984. -14-As indicated e a r l i e r , researchers c i t e s o c i a l and behavioural as well as genetic ( b i o l o g i c a l ) reasons for the variations i n mortality and morbidity rates between the sexes. Many of the behavioural differences have i n the past stemmed from the di f f e r e n t work roles that men and women have assumed. For example, women while working i n the home are much less prone to accidents than men working i n primary industries such as logging and f i s h i n g . Even t r a v e l l i n g to and from the work place has, i n the past, exposed men to a greater r i s k of accident. Also behavioural differences such as smoking and drinking have i n the past ref l e c t e d the s o c i a l mores of our society which did not condone these behaviours by women. However, l i f e s t y l e s are changing. Men and women's labour force a c t i v i t i e s are becoming more a l i k e . Fewer men are working i n primary industries, more women are moving into the paid labour force and both groups are increasingly moving into the service industries (Naisbitt, 1982). As their occupational environments converge one might expect that some of the behavioural and therefore r i s k exposure differences between the sexes would be reduced. One of the questions addressed by th i s thesis i s 'When women move into the paid labour force, does th e i r exposure to health r i s k s become more l i k e men's?' - 1 5 -2.2 WOMEN AND WORK IN CANADA THE HISTORICAL CONTEXT At the turn of the century, Canada was only just beginning i t s move toward being an industrialized society. Fifty percent (50%) of men s t i l l 2 worked in agriculture. The industrial labour force was quite small and although only 11% of a l l females over the age of ten were in the paid labour force, they almost a l l worked in light industry, making up 34% of the industrial work force. Industrialization brought urbanization; this in turn moved the paid labour force away from the home and into factories and offices. The majority of women worked at home and were not part of the paid labour force. Those women who did work outside the home (14% in 1911), generally worked in the factories that had replaced the family based industries such as cloth and clothing manufacturing. The women worked long hours (as much as 60 hours/week) and were paid poorly, the rationale being that the girls were expected to live at home until they married, at which time they were expected to leave the paid labour force (Phillips and Phillips, 1983). Table 2.6 shows the slow increase in the proportion of Canadian women in the paid labour force during the f i r s t half of this century and the extent to which they entered or left the work force at times of war and depression. Note the low level of women workers in the early thirties during the depression and the increased levels in 1918 and 1945. In 1979 only 8.6% of the male work force worked in agriculture. -16-TABLE 2.6 PERCENTAGE OF WOMEN IN THE PAID LABOUR FORCE 1911 - 1946 1911 - 14% 1939 - 24% 1918 - 22% 1941 - 24% 1921 - 18% 1945 - 33% 1931 - 13% 1946 - 25% Source: Phillips, P. and Phillips, E. Women and Work. Toronto: Lorimer, 1983. Through the f i f t i e s , the steady increase in the proportion of women working for pay continued but i t is since the sixties that the dramatically increasing rate of participation has become a significant social phenomenon. Figure 2.2 shows the participation rates for successive cohorts of women for the years 1955, 1965 and 1975. Table 2.7 shows the ages of each cohort of women for these years. FIGURE 2.2 PARTICIPATION RATES FOR SUCCESSIVE COHORTS OF WOMEN CANADA: 1911-1960 10 0 14-24 25-34 35-44 45-54 55-64 AGE IN YEARS See Table 2.7 for birthdates of cohorts. -17-TABLE 2.7 COHORTS OF WOMEN - CANADA 1911-1960 COHORT 1 2 3 A 5 BIRTHDATE 1951-1960 1941-1950 1931-1940 1921-1930 1911-1920 1955 AGES IN 1965 1975 15-24 15-24 25-34 15-24 25-34 35-44 25-34 35-44 45-54 35-44 45-54 55-64 Source: Adopted from Department of Finance Canada. Participation Rate and Labour Force Growth in Canada, Ottawa: April, 1980. Most significant perhaps i s the fact that the majority of women in the work force are now married women. Table 2.8 shows how the percentage of married women in the labour force has increased since the 1950's. The status of 'married' generally goes hand in hand with domestic and child raising responsibilities. For married women, entering the work force means adding a work commitment outside the home to an already established work role as a homeraaker. TABLE 2.8 PERCENTAGE OF FEMALE LABOUR FORCE BY MARITAL STATUS 1951 1975 1980 Single 62.1 31.0 29.9 Married 30.0 59.6 59.9 Other 7.9 9.4 10.2 Source: Phillips, P. and Phillips, E. Women and Work. Toronto: Lorimer, 1983. Many reasons have been put forward to explain this upsurge in married women's participation in the labour force. They include:-a. to maintain family income; b. to stave off boredom and to meet the higher expectations that are associated with smaller families, higher levels of education and higher standards of living; -18-c. the enticement of higher real wages for women, that i s , income more than covering the costs incurred by their going out to work; or d. the rewards of independence and self esteem that a career can provide. None of these reasons would seem to be mutually exclusive and for many women several of them may play a part in the decision to work outside the home as well as inside the home. TABLE 2.9 PARTICIPATION RATES OF MARRIED WOMEN (HUSBANDS PRESENT), BY AGE GROUP AND PRESENCE OF CHILDREN IN THE HOME, 1971 AND 1976 Wives Wives Aged Aged 15 - 34 35 - 44 With no children present 1971 73.9 59.4 1976 77.5 65.5 Absolute increase (percentage points) 3.6 6.1 Relative increase (percent) 4.9% 10.3% With children, a l l over six 1971 46.0 44.2 1976 54.9 53.6 Absolute increase (percentage points) 8.9 9.4 Relative increase (percent) 19.3% 21.3% With children under six 1971 28.0 25.4 1976 36.9 35.8 Absolute increase (percentage points) 8.9 10.4 Relative increase (percent) 31.8% 40.9% Source: 1971: Statistics Canada, 1971 Census of Canada, Labour Force Activity - Work Experience, Female Labour  Force Participation by Schooling, Marital Status, Age,  and Presence of Children, for Canada and the Regions. Cat. 94-774, Vol. 3, Part 7 (Bulletin 3. 7-4). 1976: Statistics Canada, 1976 Census of Canada, Supplementary Bulletins; Economic Characteristics,  Female Labour Force Participation Rates by Level of  Schooling, Age, Marital Status and Presence of Children. Cat. 94-836, (Bulletin 10SE7). -19-WOMEN IN THE WORKFORCE In Canada in 1979, 48.9% of women participated in the paid labour force. Table 2.8 shows the participation rate of Canadian women, by marital status and Table 2.9 shows the participation rates of married women according to the presence and ages of children in the household. Of greatest significance, not only to women but also to their families and employers, i s the increasing number of women with children under six years of age, who have joined the paid labour force. There i s an inverse relationship between family income (minus the woman's earnings) and married women's participation in the workforce. In 1971, 47% of the women in families with an income less than $3,000 participated in the workforce, compared to 27% of women from families with incomes over $15,000. Much has been written about the segregation of women in the work place. In 1979, 34% of female workers were in clerical positions and their presence was concentrated in a very few industries. For example, 43% of a l l female workers worked in service industries and 19% worked in trade. The industries and occupations into which women workers were concentrated were also those in which monetary rewards were lowest. Even within similar occupational categories men earned more than women. In only one occupational category (retail food workers) did women earn more than their male counterparts (Statistics Canada 1979). -20-Women make up a greater proportion of the part-time workers. In 1980 23.8% of women workers were part-time compared to 6% of male workers. This is significant because of the generally lower wage rates and benefits offered to part-time workers. Finally, women workers are far less likely to be unionized. In 1982, 25% of a l l female workers were unionized, compared with 37% of men. Again, this would suggest less bargaining power regarding working conditions for women. WOMEN WORKING IN THE HOME Housework was defined by Proulx (1978) in her study of the Canadian housewife as "both the activities relating to the physical and educational care of the children and those involving housework proper". Walker (1976) defined i t as "the sum of a l l useful activities performed in the home with a view to providing the goods and services which enable the family to function as a family". Traditionally, these have been tasks that have been assigned to women. Although there may be anecdotal evidence that this tradition i s changing, there i s l i t t l e evidence in the literature of any major change. In the 1979 C.H.S., only two males in B.C. reported their principal activity over the previous year as homemaking and neither of them was married. In data from a 1975 study in the Vancouver region by Meissner et a l . (1975), the very limited contribution of employed husbands is well illustrated. In "No Exit for Wives", Meissner et a l . (1975) stated: "The domestic work week of housewives without employment is a f u l l equivalent (give or take three hours depending on whether they have a young child) to the husband's 40 hour week on the job." -21-For homemakers who also have employment outside the home the burden i n terras of work hours simply increases. Meissner found that even though housewives adapt and reduce the amount of housework they do when they also work outside the home, they s t i l l add about 18 hours to th e i r average work week when they add f u l l - t i m e paid employment to their domestic r e s p o n s i b i l i t i e s . Meissner et a l . ' s studies further showed that in households with no young children, the men increased their contribution to housework by approximately six minutes per week (for an o v e r a l l average of about 3.2 hours per week) when the wife gained employment outside the home. In households with a c h i l d under 10 years of age, the men increased their contribution to housework by about one hour per week (for a weekly average of 6.0 hours), when the wife went out to work. Table 2.10 d e t a i l s well the extra workload carried by a mother working outside the home; of pa r t i c u l a r note i s the very small amount of weekday l e i s u r e time (1.8 hours per day), s o c i a l i z a t i o n (0.6 hours per day) and active l e i s u r e (0.1 hours per day) - a t o t a l of 2.5 hours compared with a t o t a l of 7.2 hours for the unemployed housewife. These data add further support to Gove and Hughes' (1979) idea of the nurturant role leaving the woman l i t t l e time for herself. Pleck (1985) in summarising recent writings and interpretations of studies on the d i v i s i o n of family work, found that i n general, men with families have increased the time they spend with their children but that in the narrower sense of family work the increase has been small and that wives, employed or unemployed, continue to do the bulk of the family work. -22-TABLE 2.10 TIME STUDIES OF HOUSEWIVES AND WIVES IN THE LABOUR FORCE WITH/WITHOUT A CHILD UNDER 10 YEARS HOURS PER DAY SPENT ON SELECTED ACTIVITIES WORK 1 )AY OFF DAY No Child Child No Child Child Self Maintenance (sleep, personal care, eating) Lab. force wife Housewife 10.7(Hrs) 11.6 10.9(Hrs 11.2 10.8(Hrs) 11.3 11.6(Hrs) 11.2 Total Employment Related Work Lab. force wife Housewife 6.9 0.0 6.0 0.0 0.5 0.0 0.1 0.0 Total Regular Housework Lab. force wife Housewife 1.8 4.8 3.5 6.1 3.2 2.5 3.6 3.6 Active Leisure Lab. force wife Housewife 0.3 1.0 0.1 0.9 0.6 1.1 0.2 0.8 Socialization Lab. force wife Housewife 1.2 1.6 0.6 1.7 2.2 2.3 1.2 1.9 Leisure Lab. force wife Housewife 3.1 5.0 1.8 4.6 5.8 6.6 4.9 5.7 Source: Adapted from M. Meissner et a l . "No Exit for Wives: Sexual Division of Labour and the Cumulation of Household Demands." Canadian Revue in  Sociology and Anthropology 12(4) Part 1, 1975. 2.3 WOMEN AND THEIR HEALTH Many writers have tried to explain or rationalize women's high use of the health care system, despite women's mortality rates being lower than those of their male counterparts. Writers such as Waldron (1983) have examined the physiological aspects of the differences while others (for example, Verbrugge, 1983) have concentrated on the social factors. It is important to recognize that use of the health care system reflects not only health status but also perceptions and behaviour as they relate to that health status. -23-Some of the genetic factors believed to contribute to sex differences in morbidity and mortality, include:-a. the possible protective effect of endogenous female sex hormones, for example in reducing women's risk of ischaemic heart disease; b. the presence of immune factors in the X chromosome offering women greater resistance to infectious diseases; and c. the differences associated with women's more demanding and complex reproductive functions. Many social and behavioural factors have been offered as explanations for differences in health care utilization. These are most often associated with traditional sex roles and suggest that women use the health care system more because: a. women are more health conscious, being responsible for the family's health; b. for the non-employed woman, the opportunity cost of seeking medical care i s less than for the employed person; c. medical visits associated with women's reproductive systems account for a large proportion of their demand for health care services; d. disability surveys (including the C.H.S.) show women as having more days of illness and disability than men: i t has been suggested that women generally are more likely to adopt a 'sick' role - in part because i t i s more consistent with the traditional female role; e. the social acceptability of admitting illness, discussing symptoms and seeking help, may be more part of the female socialization; f. demand for health services, when economic barriers are removed, is cli n i c a l l y related to age and inversely related to socio-economic -24-condition; women, especially elderly single women and single mothers, make up a large part of the lower socio-economic group. An examination of these reasons raises the need to distinguish between reported, actual health status and health care utilization. Several of the reasons cited for women's greater use of the health care system reflect sociological and cultural motivations and not necessarily poorer health status. Many researchers (Cleary, Mechanic and Greenly, 1982; Nathanson, 1975; Marcus, Seeman and Telesky, 1982) have examined differences between male and female morbidity. In a l l instances a higher level of morbidity in women was reported but this was always via a health record or from self-report surveys. None of these studies addressed the question of whether men and women experience similar levels of morbidity (although from different causes) but that women are simply more comfortable about acknowledging a discomfort or pain while men regard any sort of morbidity as a sign of weakness. To address this issue and identify true differences in health status would require c l i n i c a l and functional assessments of male and female subjects as well as some measure of pain thresholds for accepting or attempting to avoid discomfort. No such studies were found in the literature. 2.4 EMPLOYED WOMEN AND THEIR HEALTH Nathanson and Lorenz (1982) suggest that "the implications of increased participation by women in the labour force for their mortality, illness patterns and use of the health care system have only just begun to be considered..." -25-There are several components to this issue:-a. the possible changes in women's actual health in terras of morbidity and mortality, and b. changes in their behaviour with regard to their mental and physical health. The overwhelming short-coming of most of the research to-date is that i t is cross-sectional and relies on correlational analysis. These data therefore shed l i t t l e light upon the sequence of events and f a i l to address the issue of whether employed women are self selecting, in part on the basis of health factors. Studies as early as 1946 reported less illness among employed women than among housewives. Nathanson (1980) suggested that increased self-esteem and social support may explain the employed woman's better health. This idea is supported when the reduction in illness among employed women is shown to be greatest for those groups with fewer social ties (the unmarried, the divorced) and those with lower educational achievements (those not graduated from high school). Jougla, et al.(1983) suggest that in any comparisons between housewives and employed wives, the degree of role satisfaction must be considered and that this may be more( important in determining health than the activity i t s e l f . Nathanson and Lorenz (1982) reported on a study (Ladbrook, 1977) of the Wisconsin Labour Force which showed women in professional and technical occupations experiencing higher mortality rates than men in similar occupational groups, for every age group from 16 - 64. Ladbrook is reported -26-to assert that the li f e s t y l e of professional women places them at greater risk, but no detail was provided by Nathanson and Lorenz. Ladbrook, noting that the professional males have mortality rates better than those for the total male population, suggested that reduced risk taking by the men in the professional and technical occupations must be recognized. Other studies, for example, the C.H.S. (1979), have shown an inverse relationship between men's education and their smoking. In the Framingham study (Haynes and Feinleib, 1980), among women with three or more children, those in the paid labour force were two and a half times as likely as the housewives to develop coronary heart disease. These findings however were confined to clerical workers married to blue collar workers suggesting that factors other than labour force experience may be contributing to increased risk. The possible effects of occupational stress on the incidence of heart disease among women has been explored but there are no definitive findings to-date (Nathanson and Lorenz, 1982). It may be that this lack of definitive findings is a reflection of women's level of occupational involvement and may yet be seen to increase as women strive for higher occupational achievement. Morton and Ungs in their Lane County, Oregon study (1979) found a higher cancer mortality rate among housewives but that study failed to take into account many potentially significant demographic variables. Several writers report lower levels of depression among employed women. Aneshenel et a l . (1981) found the lowest levels of depression among employed married men. For women, either having a family or being employed offered -27-protection against depression, but unlike men there was no additional benefit to the woman who was both married and employed outside the home. On the other hand, the findings of Gore and Mangione (1983) showed married, employed women to have depression levels similar to those of married employed men and lower than those of married housewives. These findings f a i l to address the question of causality; whether depressed women don't enter the paid labour force, or whether entering the paid labour force reduces depression. The concept of multiple roles and role density has been used by several writers (Verbrugge, 1983) to describe the situation of women. Verbrugge reported that both few roles (for example the housekeeper with no children) and many roles (the employed mother with young children) were associated with poorer health. The women with multiple roles also reported higher use of the health care system. This appeared to be the one form of illness behaviour most available to them. This is a multifaceted issue with few areas of consensus. The time studies discussed earlier in this Chapter demonstrated very well the differences between the lifestyles of women who work both at home and in the paid labour force and those who do not. The significance of these differences in terms of their effect on health status and health care utilization is not so clear. Socio-demographic variables such as family income and education would seem to be significant factors in health, in a woman's decision to enter the work force and in the degree of satisfaction she is likely to derive from that activity. -28-CHAPTER THREE THE CANADA HEALTH SURVEY1 3.1 OVERVIEW In accepting the philosophy of A New Perspective on the Health of  Canadians (Lalonde, 1974), the Federal Government also recognized the need for more data on risk exposure leading to future health problems, on health problems that were not treated through the health care system, on the personal cost of ill-health and on the positive aspects of health. Such data would complement already available data on the causes of mortality, morbidity leading to the use of health care services, the cost of those services and the exposure of specific populations to environmental risks. The Canada Health Survey (C.H.S.) was designed to meet these new data requirements. The basic concept underlying the model used for the development of the C.H.S. is shown in Figure 3.1. FIGURE 3.1 BASIC MODEL OF HEALTH RISK HEALTH CONSEQUENCES FACTORS STATUS Three types of health risk were measured: i . l i f e s t y l e ; i i . biomedical; and i i i . environmental. The questions regarding lifestyle risk explored past exposure to alcohol This description of the C.H.S., unless indicated otherwise, is taken from the Data User's Guide which is provided with the computer data f i l e . The major printed report on the Canada Health Survey is The Health of  Canadians: A Report on the Canada Health Survey. Statistics Canada and Health and Welfare, Catalogue 82-538E, Ottawa, 1981. -29-and tobacco but for the other factors (see Figure 3.2) asked only about current behaviours. Both the physical and emotional aspects of health status (recognizing that these could be the result of r i s k exposure and/or genetic predisposition) were measured. Each person's health problems and their consequences were recorded in order to assess the impact of i l l - h e a l t h . The survey was to provide information for planning and research related areas such as health care, health promotion and disease prevention. It was designed to provide data of in t e r e s t to planners, administrators, professionals and researchers working in business, government, health care i n s t i t u t i o n s , research and education. 3.2 THE SAMPLE DESIGN2 The C.H.S., as o r i g i n a l l y conceived, was to be a continuous monthly survey with an annual cycle. It was on t h i s basis that the sample design was developed. The Survey covered the non-institutionalized Canadian population excluding residents of the T e r r i t o r i e s , Indian Reserves and remote areas as defined by the Canadian Labour Force Survey. In t o t a l , these exclusions comprised less than 3% of the entire Canadian population. The survey f i e l d work commenced in May 1978 i n the Eastern provinces, in the Central provinces in June and the entire survey population was covered from July onwards. Data c o l l e c t i o n was halted in March 1979. Only data collected during the period July 1978 through March 1979 were made available for public use. A more detailed description of the sampling methodology can be found in Appendix II of The Health of Canadians: A Report of the Canada Health Survey. Ottawa: Ministry of Supply & Services, 1981. -30-Th e Canadian population was stratified i n i t i a l l y by province. Quebec and Ontario each contained three further strata, based on groups of provincial health regions. Each of these i n i t i a l strata was further stratified into three: major citi e s in the region, other major urban areas and the remaining, primarily rural parts of the region. Table 3.1 shows the distribution of the sampling both by province and by population density; that i s , from city, urban or rural regions. In the l e f t hand half of the table, the number of clusters per province is shown. In columns three through five a further disaggregation i s shown for the two most densely populated provinces. Following these numbers across into the right hand side of the table the patterns of sampling around the major cities can be seen. For example, in the province of Quebec, 19 clusters of households were identified; nine of these were in the Montreal region, six in the Quebec City region and four in urban and rural areas away from the main ci t i e s . Of the nine clusters in the Montreal region, five were among city households, two clusters were in urban areas and two were in rural regions. Having identified households as the sampling units and set an annual sample size of 40,000 for the Interview component of the survey, 12,000 households from 100 geographical clusters were then divided into monthly samples of ten households per cluster. The Physical Measures component was administered to seven of the ten households in half of the 100 clusters. When as a result of government wide budget cuts, the decision to terminate the study was made, the sampling per cluster was increased to ensure that close to 12,000 households would be surveyed before the data collection was discontinued. -31-3.3 THE SCOPE OF THE SURVEY The survey was made up of four parts: 1) the Household Record Card (HRC) which identified the characteristics of the dwelling as well as the persons residing in i t ; in a l l 10,577 households participated in the survey - a response rate of 86% (Broyles et a l . , 1982). 2) The Interviewer Administered Questionnaire (IAQ) collected data on health problems, use of drugs and health care services, accidents and disabilities; a l l age groups were included. Information was accepted on a proxy basis, i t being considered 'visible' and therefore readily able to be reported by others. While this questionnaire was completed on behalf of 31,688 individuals, a response rate of 79% (Broyles et a l . , 1982), fewer than 50% of those persons were actually present for the duration of the interview. 3) The Lifestyle and Health Questionnaire (LHQ) was a self-administered questionnaire covering emotional health, alcohol and tobacco use, activity patterns, driving habits and preventive measures. It was self administered because the topics were considered more personal than the previous set of questions. Further, this questionnaire was restricted to those over 15 because the behaviours in question were considered 'adult' and reading ability was required. It was completed by 23,791 individuals, a response rate of 89% (Broyles et a l . , 1982). 4) The Physical Measures Questionnaire (PMQ) was in two parts; the physical measurements of blood pressure, cardiorespiratory fitness, height, weight and skin fold in persons over two years were taken and blood samples were taken from persons over three years in order to determine immune status as well as biochemical and trace element levels. Only a small subset (approximately 30%) of the subjects were asked to participate in this part of the study. - 32 -TABLE 3.1 C.H.S. STRATA AND CLUSTERS PROVINCIAL PATTERN URBAN AND RURAL PATTERN PROVINCE NUMBER OF CLUSTERS SUB REGION GROUP ALLOCATION TO SUB REGION GROUP MAJOR CITY STRATA OTHER URBAN STRATUM ALLOCATION RURAL STRATUM ALLOCATION NAME ALLOCATION I I P I P P I P I P Newfoundland 6 3 St. John's 2 1 2 1 2 1 P.E.I. 3 2 - - - 1 1 2 1 Nova Scotia 7 3 Halifax 2 1 2 1 3 1 New Brunswick 6 3 - - 3 1 3 2 Quebec 19 9 1 9 4 Montreal 5 2 2 1 2 1 2 6 3 Quebec 2 1 1 1 2 1 3 4 2 - 2 1 2 1 Ontario 22 11 2 4 2 _ — 2 1 2 1 1 10 5 Ottawa 2 1 2 1 2 1 Toronto 4 2 3 8 4 Hamilton 2 1 4 2 2 1 Manitoba 8 4 Winnipeg Saskatchewan 7 4 - - - 3 2 4 2 Alberta 10 5 Edmonton 2 1 2 1 4 2 Calgary 2 1 B.C. 12 6 Vancouver 6 3 3 2 3 1 TOTAL 100 50 TOTAL 33 16 32 17 35 17 I: Interview clusters. P: Subsample of clusters for physical measures. Source: Appendix II. The Health of Canadians: Report of the Canada Health Survey. Ottawa: Ministry of Supply and Services, 1981. -33-3.4 THE CANADA HEALTH SURVEY DATA Questionnaire data were captured directly onto computer-readable f i l e s . Extensive editing procedures, manual and automated, were carried out in order to identify inconsistencies. Once editing was complete, a number of summary or indicator variables were calculated for each person and added to the f i l e , for example, The Physical Activity Index and the percentile for the family income. A series of five weights were provided with the data to allow for the production of population estimates. These weights each had relevance to different sections of the data f i l e and were calculated using 'a post-stratification ratio estimate plus relevant estimates of provincial populations by age and sex'. In addition the size of the sub-adjustments responding to a particular section of the survey and relevant adjustments for unknowns were considered in the calculation of the weights. (Because of the computational complications involved in using several different weights in Multiple Classification Analysis and the belief that the findings of this study would be equally valuable without being able to be generalized to the total population, the weights provided with the C.H.S. data were not used in this study.) The C.H.S. adjustment for non-response, took place on five levels: 1) at the household level, non-responding households were replaced conceptually by an 'average' household determined from a l l those households which responded within the same cluster in the same month; -34-2) persons for whom responses occurred only on the Household Record Card were excluded from the survey f i l e ; the number excluded was considered negligible. 3) where compulsory items in the LHQ, PMQ and Blood were missing, adjustment was made by means of the sampling weights. This adjustment used the assumption that respondents and non-respondents are similar with respect to health related data although a study of the IAQ data of non-respondents would suggest that they tended to be slightly less healthy. The adjustment via the sampling weights was considered the best of the methods available. 4) where there were no responses to a f u l l section within a questionnaire, for example when no questions in the alcohol section were answered, a l l items in that section were coded as unknown; 5) single data items lef t unanswered were coded as unknown. The C.H.S. data used for this study was extracted from a computer tape in the UBC Data Library. - 35 -FIGURE 3.2 A SUMMARY OF DATA ITEMS INCLUDED IN THE CANADA HEALTH SURVEY RISE FACTORS HEALTH STATUS CONSEQUENCES LIFESTYLE BIO-MEDICAL ENVIRONMENT PHYSICAL/EMOTIONAL PERCEIVED/OBSERVED POSITIVE/NEGATIVE UTILIZATION CONDITION IMPACT DATA LIFESTYLE REPORTED HEALTH UTILIZATION - alcohol use LHQ - a c t i v i t y limitations IAQ - professional - tobacco use LHQ - short-term conditions IAQ providing care IAQ - physical a c t i v i t i e s LHQ - accidents and injuries IAQ - location care - seat belt use LHQ - chronic conditions IAQ received IAQ - female preventive - impairments IAQ - reasons care behaviour LHQ - hearing, vision, dental not sought IAQ status IAQ - drug use IAQ - medical devices IAQ used BIO-MEDICAL PHYSICAL HEALTH CONDITION IMPACT Immune status BLOOD cholesterol, glucose, uric acid BLOOD family disease history LAQ ENVIRONMENT cardiorespiratory fitness PMQ blood pressure PMQ anemia BLOOD li v e r function BLOOD kidney function BLOOD EMOTIONAL HEALTH - d i s a b i l i t y days IAQ lead, cadmium, copper, zinc - psychological well-BLOOD being LHQ - alcohol-related problems LHQ HOUSEHOLD CHARACTERISTICS - area designation HRC - household membership HRC - dwelling characteristics HRC DEMOGRAPHIC CHARACTERISTICS social characteristics IAQ & LHQ economic characteristics IAQ mobility, immigration IAQ l i f e events LHQ KEY: HRC - Household Record Card IAQ - Interviewer Administered Questionnaire Interview Component LHQ - Life s t y l e and your Health Questionnaire PMQ - Physical Measures Questionnaire BLOOD - Blood sample Physical Measures Component Source: Canada Health Survey: Data Users Guide. S t a t i s t i c s Canada, Ottawa, 1982. -36-CHAPTER FOUR THE MODEL 4.1 THE QUESTIONS AND THE STRUCTURE THE QUESTIONS This study addresses four questions: 1. Do women who participate in the paid labour force report poorer health status than their counterparts who are homemakers? 2. Do women who participate in the paid labour force exhibit lifestyle patterns significantly different from their counterparts who are homemakers? 3. Do women in the paid labour force exhibit health care utilization patterns significantly different from their homemaker counterparts? and 4. Do women's lifestyles, reported health status and health care utilization patterns differ from those of their male counterparts in the paid labour force? I n i t i a l l y this study set out to answer the f i r s t question: 'Do women who participate in the paid labour force report poorer health status than their counterparts who are homemakers?' One of the reasons for believing that workforce women might experience poorer health was that so often women have added their employee status to an already busy l i f e as wife, parent and homemaker. This phenomenon of multiple roles is probably most common among women aged 25 - 45 but the consequences (good or bad) may not be evident until later in their lives. The cross-sectional nature of the C.H.S. data and the lack of information regarding past work histories meant that, to search for any relationship between multiple roles and health status, this -37-study had to use the 25 - 44 year age group. This then limited the study to the question of immediate rather than long term health consequences. In an attempt to address the question of long terra health consequences, the second question was added to the study, namely, 'Do women who partici p a t e i n the paid labour force exhibit l i f e s t y l e patterns s i g n i f i c a n t l y d i f f e r e n t from th e i r counterparts who are homemakers?' In other words, i f the women i n the paid labour force do not currently appear to be i n any poorer health, i s there anything i n the i r current health r i s k exposure, as indicated by the i r l i f e s t y l e patterns, that could lead one to believe that i n the future, they may experience poorer health than t h e i r homemaker counterparts? The t h i r d question posed by th i s study, 'Do women in the paid labour force exhibit health care u t i l i z a t i o n patterns s i g n i f i c a n t l y d i f f e r e n t from their homemaker counterparts?' completes the question i n terms of planning for the health care consequences of so many women entering the paid labour force. The fourth and f i n a l question 'How do the women's l i f e s t y l e s , reported health status and health care u t i l i z a t i o n patterns compare with those of thei r male counterparts in the paid labour force?' enables the findings concerning the women's experience to be placed in the context of "the rest" of that age group. It enables one to assess the p o s s i b i l i t y that as women move into the paid labour force, their health experiences w i l l become more similar to those of men. The study group was therefore men and women aged 25 - 44 years -38-and who participated in the C.H.S.; who either participated in the paid labour force or were homemakers; who had no activity limitations which would limit or affect their options for either work or leisure. THE MODEL The basic components of the C.H.S. have been taken as a starting point for the model in this study (see Figure 4.1). FIGURE 4.1 THE BASIC COMPONENTS OF THE C.H.S. MODEL RISK FACTORS HEALTH STATUS CONSEQUENCES To this model, demographic variables were added because of their potential capacity to explain variations in the other three components. Female preventive behaviours were separated out because they are not truly risk factors in that they don't represent behaviours that actually cause disease and their practice requires the services of health care professionals (for Pap smears and professional breast examinations). Figure 4.2 shows these additional components. The figure also shows how the model could be used to ask whether: . demographic variables could explain variations in health risk behaviours(l), . demographic variables and health risk behaviours could explain variations in health status (2a & 2b), and i f . demographic variables, health risk behaviours, health status and preventive behaviours were associated with particular health care consequences (3a,b and c). -39-For each of these components of the model, the actual variables to be used were chosen from those collected in the C.H.S. Figure 3.2 shows the variables used in the C.H.S. and Figure 4.3 shows the individual variables from the C.H.S. fitted into the model for this study. A major limitation of this study is i t s inability to address the issues of chronology and causality; i t can only attempt to demonstrate associations. The model for this study shows those associations to be unidirectional. This constitutes a major simplification of the issues. In reality the reported health status and the level of health service utilization must each drive the other and perceptions of poor health probably have a negative effect on li f e s t y l e . For example, the fact that a person takes medications for high blood pressure may reinforce the recognition of this as a chronic condition, making such persons feel more depressed and not leading a normally active l i f e , thereby compounding the i n i t i a l problem. So, while the implied causality of model for this study moves in a single direction, i t is recognized that the forces within the system move both ways among a l l components and that a study at any greater level of detail would need to address the very complex question of causality. The remainder of this chapter outlines in more detail each of the variables used in the study. In the case of the health risk factors, some rationale for their inclusion also i s offered. HI HEALTH RISK BEHAVIOURS DEMO VAR I GRAPHIC ABLES FIGURE 4 , 2 THE MODEL COMPOSITE HEALTH RISK SCORE SELF-REPORTED HEALTH STATUS FEMALE SCREENING & PREVENTIVE BEHAVIOURS HEALTH CARE RELATED BEHAVIOURS © i o i D E M O G R A P H I C V A R I A B L E S • R E G I O N • S E A S O N OF S U R V E Y • M A R I T A L S T A T U S • F A M I L Y S I Z E • F I N A N C I A L R E S P O N S I B I L I T Y • I N C O M E • E D U C A T I O N ^ • SMOKING • ALCOHOL • EXERCISE § COMPANIONSHIP • HORMONES t M.V.TRAVEL t SEATBELTS • LIFE EVENTS • FEMALE PREVENTIVE FIGURE 4.3 THE VARIABLES WITHIN THE MODEL C O M P O S I T E 81SK S C O R E S . • SMOKING • ALCOHOL F E M A L E S C R E E N I N G ft P R t v b N I I V t t EXERCISE § PAP SMEARS • COMPANIONSHIP • P . B . E . § HORMONES • HORMONES • M.V.TRAVEL • LIFE EVENTS H E A L T H S T A T U S • CHRONIC HEALTH PROBLEMS t AFFECT BALANCE • HEALTH OPINION H E A L T H C A R E C U H S E U U b N C E S • DISABILITY DAYS t PROFESSIONAL CONSULTATIONS • MEDICATIONS • HOSPITALIZATION I -42-4.2 THE DEMOGRAPHIC VARIABLES The demographic variables (from the C.H.S.) included in this study are shown in Figure 4.3. They were included i f the literature indicated that the factor could influence health behaviours or outcome. The rest of this section discusses each of the individual demographic variables included in the model. REGION The C.H.S. identified subjects as coming from one of five regions: i . The Atlantic Provinces, i i . Quebec, i i i . Ontario, iv. The Prairie Provinces and v. British Columbia While the provision of health services in Canada is to some degree protected by the Canada Health Act, the actual provision of services is a provincial responsibility. The type and level of services therefore varies not only across the regions identified by the C.H.S. but also within these regions because: . some of the regions include more than one province, . there are other variations in supply such as those found between rural and urban areas, and . individual provinces develop different priorities for service provision. For example, city women have easier access to mammography because the technology is centered in the cit i e s and, further, British Columbia has a longer established and more aggressive pap smear cytology programme than the -43-other provinces. Therefore, Ontario and Quebec with large city populations and British Columbia with more highly promoted programmes in women's health are a l l more likely to report higher rates of female preventive behaviours than the other provinces. The disaggregation of the data into just five regions can therefore only be expected to identify major differences. Table 4.1 shows the distribution of the study group across the five regions. TABLE 4.1 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY REGION AND SEX REGION WOMEN MEN (n:3760) (n:3562) Atlantic Regions 23.3 21.3 Quebec 21.4 21.8 Ontario 21.4 21.8 The Prairies 21.5 22.3 British Columbia 12.4 12.8 TOTAL 100.0 100.0 SEASON OF DATA COLLECTION The C.H.S. data that were placed in the public realm were from interviews carried out between July 1978 and March 1979. This was the only period during which there was concurrent sampling from a l l parts of the country. Many of the questions in the survey asked participants to report on activities of the previous two weeks. However i t seems likely that activities such as exercise, socializing and alcohol consumption could vary according to the season. For example, people may socialize more in December but exercise less during the colder months. The season of the data collection was therefore included and the percentage distribution of the study group is shown in Table 4.2. -44-TABLE 4.2 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY SEASON OF DATA COLLECTION AND SEX SEASON OF DATA COLLECTION WOMEN (n:3760) MEN (n:3562) July to September 1978 23.5 24.0 October to December 1978 39.7 39.5 January to March 1979 37.1 36.5 TOTAL 100.0 100.0 MARITAL STATUS In the original data, marital status was recorded as: i . single (never married) i i . married (including common law) i i i . widowed iv. separated/divorced. The literature generally indicates that marriage is associated with better health; for example, studies have reported married women experiencing less depression and fewer psychiatric problems (Aneshenel, Frerichs & Clark, 1981; Gove & Mangione, 1983). It should be noted however that Goldman and Ravid (1980) in reporting on the better mental health of both married men and women caution the reader against assuming a causal relationship and suggest that i t may be that less depressed people are more likely to marry. Because of the relatively small numbers in each of the 'not married' categories, subjects were simply identified as 'married' or 'not married'(for the purposes of this study). Table 4.3 shows the distribution of the sample. TABLE 4.3 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY MARITAL STATUS AND SEX MARITAL STATUS WOMEN (n:3760) MEN (n:3562) Not Married 15.2 15.6 Married 84.7 83.8 Unknown 0.1 0.6 TOTAL 100.0 100.0 -45-FAMILY SIZE Family size has been included for i t s potential ability to indicate domestic responsibilities. It would seem reasonable to expect that the women living in larger families would spend more time on domestic activities whether they also participated in the paid labour force or not. However, there also could be women living in a large family who are in fact dependents and for whom such assumptions should not be made. Table 4.4 shows the percentage distribution of family sizes and their distribution within the subset of the C.H.S. used for this study. TABLE 4.4 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY FAMILY SIZE AND SEX FAMILY SIZE WOMEN MEN (n:3760) (n:3562) One Person 5.2 7.4 2 - 3 Persons 31.2 35.0 4 - 6 Persons 58.5 53.8 7 or more Persons 5.1 3.8 TOTAL 100.0 100.0 FAMILY INCOME Income per se, or as an indicator of socio-economic class has been shown to be associated with behaviours such as cigarette smoking and alcohol consumption. In the C.H.S. data, family incomes were reported in quintiles which were developed from the reported family incomes after those incomes had been adjusted to reflect the size of the family and the municipality of residence. For the 25 - 45 year age group included in this study the income distribution by quintile is shown in Table 4.5. -46-TABLE 4.5 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY INCOME AND SEX INCOME QUINTILE WOMEN MEN (n:3760) (n:3562) First (lowest) 14.8 9.1 Second 18.4 17.8 Third 20.3 20.3 Fourth 19.9 20.6 Fifth (highest) 21.3 26.7 Unknown 5.4 5.5 TOTAL 100.0 100.0 EDUCATION Education, another indicator of socio-economic status, also may be significant because awareness of health risks necessarily depends upon learning s k i l l s in terms of registering and assimilating new information. Also, the current growth of the 'fitness' industry has been asserted by some to be a purely middle class phenomenon. The grouping of the subjects according to the educational levels identified in the C.H.S. is shown in Table 4.6. TABLE 4.6 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY EDUCATION AND SEX EDUCATIONAL LEVELS WOMEN (n:3760) MEN (n:3562) Some Secondary 68.1 63.9 Some Post-Secondary 8.0 8.9 Post-Secondary Diploma 14.2 10.7 University Degree 9.2 15.9 Unknown .6 .7 TOTAL 100.0 100.0 ECONOMIC RESPONSIBILITY The increasing number of women supporting a family represents a further expansion of the roles women may have to play and which may or may not cause -47-sorae of their behaviours to become more like those of men who have traditionally been responsible for the economic welfare of the family (Armstrong & Armstrong, 1978). In the C.H.S. subjects were asked whether they were the principal income earner for the family. Table 4.7 shows the percentage of principal income earners in this study group. TABLE 4.7 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY ECONOMIC RESPONSIBILITY AND SEX ECONOMIC RESPONSIBILITY WOMEN MEN (n:3760) (n:3562) Principal Earner 17.8 91.5 Not Principal Earner 82.2 8.8 TOTAL 100.0 100.0 Other demographic variables available in the C.H.S. but not used in this study, include: i . type and size of dwelling i i . mobility of family/changes in residence i i i . place and language of birth iv. industry/occupation and hours of work v. individual income from employment. While size of dwelling could be used as an indication of domestic responsibilities, size of economic family was considered a better indicator because although a woman might have help with housecleaning, she was much less likely to have help with the care and nurturing of family members. The more detailed information on the subject's role in the workforce would enable questions to be asked regarding the impact on health of different types and conditions of work. While these are equally important questions, they are beyond the scope of this thesis. -48-In this model, the demographic variables have always been considered separately. Obviously however, there are many relationships among them. For example, education and income are both indicators of socio-economic status and could perhaps be merged to create one single indicator. This was not attempted. Similarly, marital status, family size and economic responsibilities could possibly be linked to create an indicator of roles and responsibilities. However such an indicator would be open to considerable debate without additional information on the ages and levels of dependence of the family members and this information was not available in the C.H.S.''' 4.3 HEALTH RISK BEHAVIOURS Figure 3.2 shows the data items in the C.H.S. included for their potential to influence health status. For this study, one indicator for each of the lif e s t y l e factors was chosen. Three additional lifestyle factors found within the C.H.S. data (motor vehicle use, l i f e events and social contacts) were also included because of the evidence supporting their potential ability to influence health. Factors such as hours of sleep and the regular consumption of breakfast which have been shown in other studies to have an impact on health status could not be included in this study because they did not appear in the original C.H.S. data set. Factors such as these, and others which may increase the ability of the model to predict health outcomes, will be discussed in the final chapter. The C.H.S. data available to the public, in order to provide the highest possible level of anonymity to the subjects, have been organized so i t is not possible to infer the actual composition of each woman's household. -49-SMOKING The health hazards of smoking have been well documented. A 1980 report by the U.S. Department of Health and Human Services provides a good overview of the major disease consequences of tobacco use by women (Gritz, 1984). Smoking accounts for 25% of a l l cancers; i t has been causally linked with cancer of the lung, larynx, oral cavity and oesophagus; i t has been strongly associated with the development of bladder, kidney and pancreatic cancer. The risk of developing coronary heart disease is increased at least by a factor of two in women who smoke. This risk becomes ten fold among women who smoke and use oral contraceptives (Gritz, 1984). Smoking also is associated with chronic obstructive lung disease. Of particular significance to women is the potential effect of smoking on reproduction: reduced birth weights, spontaneous abortions, placental abnormalities and foetal death are some of the possible outcomes related to smoking during pregnancy. The questions in the C.H.S. covered past and present smoking habits, numbers and brands of cigarettes smoked and attempts to stop smoking. Subjects answering these questions were placed in one of the following categories: i . current regular smoker, i i . current occasional smoker, i i i . past regular smoker, iv. past occasional smoker or v. never smoked. -50-For the purposes of this study, these categories were regrouped as shown in Table 4.8. TABLE 4.8 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY CIGARETTE SMOKING EXPERIENCE AND SEX CIGARETTE SMOKING EXPERIENCE WOMEN MEN (n:3760) (n:3562) Current Smoker (Regular or Occasional) 39.3 39.9 Past Smoker (Regular or Occasional) 19.0 22.9 Never Smoked 30.5 18.0 Unknown 11.3 19.2 TOTAL 100.0 100.0 ALCOHOL Alcohol is a drug and people do become addicted. Heavy drinking is associated not only with illnesses such as hypertension, cirrhosis, breast cancer and disorders of reproduction but also with cognitive dysfunction which can lead to motor vehicle accidents, other types of accidents, belligerence, domestic turmoil and ultimately interference with a person's ability to be gainfully employed. On the other hand moderate alcohol consumption has been reported as having a protective effect against chronic heart disease and may be valuable in assisting with stress reduction (Milsum, 1984). Within the literature, there does not appear to be great consistency in the definition of moderate versus heavy drinking, and in some studies no specific definitions are offered (Wilsnack, Wilsnack & Klassen, 1984; Celentano & McQueen, 1984; Johnson, 1982). Johnson's report on Sex Differences and Alcohol (1982) used a formula of ounces of pure alcohol per day converted to percentage of alcohol in the -51-blood stream. The definitions derived from this formula are such that one beer per day would classify a 135 pound woman as a moderate drinker; a 165 pound man would be a moderate drinker having two beer per day and a heavy drinker i f he drank three beer a day. Johnson suggests that problem drinking cannot be defined simply in terms of volume of alcohol consumed (unless i t ' s very high), but that the behavioural and social impairment derived from the alcohol must also be considered. The C.H.S. asked subjects to record on a daily basis, the number of alcoholic beverages they consumed over a seven-day period. A summary number of 'drinks per week' was then generated. Other questions in the survey asked about the type of alcohol usually consumed, alcohol-related problems and recent changes in drinking patterns. This study used the recorded number of drinks consumed in a week as an indicator of alcohol consumption. Table 4.9 shows the number of drinks assigned to each category and the percentage distribution of the subjects in this study across those categories. TABLE 4.9 PERCENTAGE DISTRIBUTION OF THE STUDY GROUP BY ALCOHOL CONSUMPTION AND SEX LEVELS OF CONSUMPTION WOMEN (n:3760) MEN (n:3562) No Drinks in Previous Week 9.5 5.3 One to Seven Drinks 34.3 27.3 Eight or More Drinks 11.7 31.8 Unknown 44.5 35.6 TOTAL 100.0 100.0 -52-PHYSICAL ACTIVITY In the past ten years there has been much discussion on the subject of physical fitness. The Canadian Government has sponsored the Participaction Program and employers are being encouraged to set up employee fitness centres. Community centres as well as numerous private businesses are now offering 'keep f i t ' programmes. Some of the benefits identified with regular physical activity are decreased risk of: . chronic heart disease and sudden death, . osteoporosis, . hypertension and . Type II diabetes. It has also been suggested that exercise can be used to: . alleviate the symptoms of mild to moderate depression and . reduce symptoms of general anxiety (U.S. Department of Health & Human Services, 1985). It i s not known whether these health benefits from exercise are through improvements in physical fitness or through other pathways such as improved serum lipoprotein profiles, fibrinolytic activity, decreased platelet adherence or other metabolic changes. For example, some health benefits seem to be achieved through activities such as yoga which do not improve cardiorespiratory endurance (U.S. Department of Health & Human Services, 1985). Obesity, a factor associated with hypertension, and one of the seven health risk factors identified by Berkman and Breslow (1983), results primarily from an imbalance of calories consumed and calories expended. -53-Physical activity can play a major part in remedying such imbalances. It i s uncertain just which factors affect peoples' level of physical activity, but some of those mentioned in the literature include: . previous experience in sports, . family and peer support, . self-motivational characteristics, . positive feeling resulting from the activity. Less certain are: . accessibility of f a c i l i t i e s . time constraints and . climatic conditions. The C.H.S. asked subjects to report their activity during the previous two weeks. This included activity during discretionary time, in exercise, sport, physical recreation and household chores. An index was then created which was the summation of the frequency of each activity multiplied by the average duration in minutes of each activity and by the average metabolic 2 cost of that activity. The range of scores generated in this way was zero to 5500. Table 4.10 shows the categories derived from these scores and Table 4.11 shows the level of activity that would generate a raid-range score of approximately 3000. The frequency, duration and average intensity were a l l weighted equally in the development of this Physical Activity Index. The researchers for the C.H.S. suggest that 'while there i s no compelling evidence to apply differential weights, neither i s there strong evidence to weight them equally'. A more serious limitation i s that the 'index does not require any particular mix of frequency, duration and intensity'. (Health & Welfare Canada & Statistics Canada, 1981) -54-TABLE 4.10 PHYSICAL ACTIVITY INDEX CATEGORIES FOR ALL SUBJECTS IN THE CANADA HEALTH SURVEY CATEGORY SCORE PERCENTAGE DISTRIBUTION Sedentary 0 - 749 16.0 Moderately Sedentary 750 - 1749 18.0 Moderate 1750 - 2999 16.0 Moderately Active 3000 - 5499 18.0 Very Active 5500 + 18.0 Unknown 14.0 TOTAL 100.0 For the purposes of this study, these categories were reduced to three: 1. Sedentary and moderately sedentary 2. Moderate and 3. Moderately active and very active. TABLE 4.11 THE FITNESS CANADA PRESCRIPTION FOR PHYSICAL ACTIVITY THE IDEAL MINIMUM LEVEL OF PHYSICAL ACTIVITY. 1. MOVE: walk, climb, ride a bike ... every day, as often as possible. 2. STRETCH & DEEP BREATH: take a fitness break & relax ... every day, as needed when tense. 3. PUSH, BEND, TWIST AND SWING: ... at least three times each week 4. RUN, SWIM, CYCLE, SKI: 15-20 minutes of continuous aerobic activity vigorous enough to increase your heart rate and make you breathe deeply. ... at least three times each week. 5. ENJOY LIFE: spend time at sports, hobbies or outdoor activities. ... a two hour period at least once a week. This level of activity translates into a minimum score of 3,000 on the Physical Activity Index  -55-Of the women included in this study, 32% reported being active or moderately active in the previous two weeks; 20% reported moderate activity and 31% reported being moderately inactive or sedentary. A further 17% did not provide sufficient information for a score to be generated. Table 4.12 shows the activity levels of the men and women included in this study. TABLE 4.12 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY PHYSICAL ACTIVITY LEVEL AND SEX ACTIVITY LEVELS WOMEN MEN (n:3760) (n:3652) Active or moderately active 32.0 33.5 Moderate 20.2 14.0 Inactive or moderately inactive 30.8 27.3 Unknown 17.0 25.2 TOTAL 100.0 100.0 LIFE EVENTS In 1967 Holmes and Rahe (1967) noted that clusters of l i f e events (changes) tended to characterize the living patterns experienced by patients immediately prior to their admission to hospital. This was interpreted in terras of the increased levels of stress experienced by the patient leading them to, or making them susceptible to illness. Selye (1974) has written extensively on the presence of stress in every living organism and suggests that a certain level of stress is highly desirable. However, each organism has a varying threshold beyond which i t s capacity to maintain homeostasis in the presence of a stressor is compromised. At that point the health of the organism becomes vulnerable. Holmes and Rahe found that people largely agreed about the perceived stressfulness of frequently encountered l i f e events and from their work, -56-developed a table of 42 items, each scored for their relative stressfulness (see Table 4.13). Implicit in this scale i s the concept that rather than the emotional nature of the event (happy, sad, good, bad), i t is the extent of the change involved that determines the ranking of the event. Other researchers feel that the quality of the event and other characteristics such as the subject's perception of or reaction to that event, are more important than the change i t s e l f (Haney, 1980). TABLE 4.13 SOCIAL READJUSTMENT RATING SCALE RANK LIFE EVENT MEAN VALUE ( l i f e change units) 1 Death of spouse 100 2 Divorce 73 3 Separation 65 4 J a i l term 63 5 Death of close family member 63 7 Marriage 50 8 Fired at work 47 10 Retirement 45 12 Pregnancy 40 13 Sex d i f f i c u l t i e s 39 19 Change in number of arguments with spouse 35 22 Change in responsibilities at work 29 31-36 Change in work, residence, outside 20-18 activities, etc. 38-40 Change in home habits 16-15 41 Vacation 13 42 Christmas 12 Source: J.H. Milsum, Health, Stress and Illness: A Systems  Approach. New York: Praeger, 1984. Adapted from T.H. Holmes & R.H. Rahe, "The Social Readjustment Rating Scale." Journal of  Psychosomatic Research 11 (1967): 213-218. The C.H.S. asked subjects to indicate which of the following had happened to them in the previous 12 months: . 'you stopped full-time school . lost a job or were unemployed . got married -57-. someone moved in with you . you had financial problems . you and your spouse separated . the arrival of a baby in the home . you had a serious illness . someone dear had a serious illness . you quit or retired from full-time work . you started working or changed jobs . the death of someone dear or . none of these.1 In scoring these, the C.H.S. assigned equal weight to these events and simply reported the number of events experienced by each subject. The experiences of the men and women included in this study are shown in Table 4.14. TABLE 4.14 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY NUMBER OF LIFE EVENTS AND SEX NUMBER OF LIFE EVENTS IN PREVIOUS YEAR WOMEN (n:3760) MEN (n:3562) One or no l i f e events 68.8 62.0 Two or three l i f e events 18.9 16.0 Four or more l i f e events 2.8 3.6 Unknown 9.5 18.3 TOTAL 100.0 100.0 COMPANIONSHIP - SOCIAL SUPPORT Much has been written about the relationship between social integration and levels of morbidity and mortality (Wethington & Kessler, 1986). The exact way in which social integration benefits the individual is not really understood but i t i s discussed in terms of having additional resources available during times of stress (stress being defined by one writer (Gottlieb, 1985) as that time when resources do not meet demands/needs). Jacobson (1986) outlines some of the many issues to be resolved, including more consistency in the definitions of illness, stress and measurements of social support i f the nature of the benefits is ever to be fully understood. Other writers (Wethington & Kessler, 1986; Kelner, 1985) suggest that the number of social contacts in and of i t s e l f i s not sufficient information and other details that should be used in assessing social support systems include: the degree of intimacy in the relationship, the type of need met by the relationship (for example, nurturing, attachment, feedback or mastery), the interconnections between the members of a person's social network and the number of spheres in which that support is offered (for example, whether i t is family centred, work centred, leisure centred or any combination of these). Friendships that overlap several spheres of a person's daily l i f e are deemed to be more valuable. The only indication of social relationships provided by the C.H.S. data was in response to the question: 'Which of the following best describes how you spent your leisure time during the last two weeks?' almost a l l of i t by myself, a lot of i t by myself, about half of i t by myself and half of i t with others a lot of i t with others almost a l l of i t with others. For the subset of subjects used in this study, the answers were combined to form three groups. Table A.15 shows the groups and the percentage distribution of the subjects across those groups. TABLE A.15 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY LEISURE TIME COMPANIONSHIP AND SEX LEVEL OF COMPANIONSHIP WOMEN MEN (n:3760) (n:3562) A lot or most of leisure with others 53.1 53.8 Half alone and half with others 27.9 21.0 A lot or most of leisure time alone 10.1 6.9 Unknown 8.9 18.3 TOTAL 100.0 100.0 -59-HORMONE PILLS (Women Only). Oral contraceptives have been used by women for about 25 years. The principal health issue surrounding their use has been their association with cardiovascular disease and neoplasia (Diczfalusy, 1986). In the 1960's, oral contraceptives, (primarily oestrogen), were associated with hypertension, venous thromboembolism, stroke, myocardial infarction and cancer of the endometrium. In the 1970's the introduction of progresterone as an oral contraceptive failed to reduce many of these associated risks (Diczfalusy, 1986). With the gradual reduction in the amount of steroid used and the refinement of the options such as fixed dose combinations of progesterone and oestrogen and triphasic products which attempt to imitate the cyclic profile of ovarian steroid secretion, the relative risks associated with oral contraceptives appear to have been reduced but not completely eliminated. A 1985 study by Porter et al.(1985), found a positive association between current oral contraceptive use and stroke and M.I. Diczfalusy (1985) reported varying risks of cancer with oral contraceptive use: . ovarian and endometrial cancer reduced 50%, . no overall increase in the risk of breast cancer but certain subgroups may be at greater risk when they use oral contraceptives at specific stages of their reproductive l i f e , . i t i s possible that prolonged use (over five years) produces a slightly increased risk of cervical cancer but assessment of this risk is highly complex because of the confounding effect of differences in sexual practice, . actually reduced risks of benign breast disease, ovarian retention cysts, pelvic inflammatory disease, iron deficient anaemia, ectopic pregnancy and possibly rheumatoid arthritis. -60-The C.H.S. asked subjects, 'Are you taking either, . birth control p i l l s (for contraception, to regulate the menstrual cycle or for some other reason), . female hormone p i l l s (to control the symptoms of menopause or for some other reason), . or neither of these.' The results were tabulated simply as "Yes" or "No" and for the women included in this study, the frequencies are shown in Table 4.16. TABLE 4.16 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY HORMONE PILL CONSUMPTION TAKING HORMONE PILLS WOMEN (n:3760) Yes 15.5 No 76.0 Unknown 8.5 TOTAL 100.0 FEMALE PREVENTIVE BEHAVIOURS This i s a derived indicator which includes the subject's score on Professional Breast Examination, Breast Self Examination and Papanicolaou Smear test. Each behaviour represents a screening process aimed at identifying cancer at as early a stage as possible because better survival rates are.associated with intervention at an early stage of the disease. BREAST SCREENING Breast cancer causes more deaths than any other cancer in women and is the leading cause of death in women aged 45 - 54 years (Ellerton & Smillie, 1986). Several studies (Ellerton & Smillie, 1986; Hislop, Coldman & Skippen, 1984; Foster, 1978) reported that among women with tumours, a high percentage -61-(86-90%) report having detected their own tumour. In a study on 335 patients with breast cancer reported by Foster et a l . , the benefits of regular breast examination were apparent in terms of the size of the tumour at the time of it s discovery and the lower incidence of axillary node metastases. Women routinely practicing breast self examination (about 25%) were found to have an age adjusted maximum tumour size of 1.97 cm _+ 0.22 while women who never performed breast self examination (about 50%) were found to have an age adjusted maximum tumour size of 3.59 cm +^  0.15. On the other hand Greenwald (1978) reported a greater percentage of cases (53.8%) in Clinical Stage I when detection is by professional examination rather than by self examination or accidental (27.0%). Greenwald also reported that of a l l the women reportedly practicing breast self-examination at the time their tumour was discovered, only 69.9% actually discovered the tumour during a self examination. A possible explanation for the great variations in study findings may come from the work of Hislop, Coldman and Skippen (1984). They reported that in a group of 416 women, 72% reported that they regularly performed breast self examination but on further inquiry only 10% could be said to practice a l l components of the examination adequately; that i s , monthly visual inspection plus thorough palpation of the breasts and axillae. Professional breast examination i s normally associated with routine medical examinations unless the woman presents a specific concern regarding breast disease. Hislop et a l . , found that among women with breast cancer, the annual medical examination was associated with smaller tumours but that this was sta t i s t i c a l l y significant only for those women who did not examine their -62-own breasts. Morrison (1986), reporting on the Health Insurance Plan of Greater New York's randomized controlled t r i a l of breast cancer screening, comments on the lack of evidence about the efficacy of the physical examination separate from the associated mammography. The Canadian National Study of Breast Cancer Screening currently underway wil l report on the results from mammography and physical examination versus physical examination only. Greenwald et a l . , estimated that breast self examination and routine professional examination could result respectively in an 18.8% to 24.4% reduction in mortality from breast cancer. CERVICAL CANCER SCREENING The 1982 Canadian Task Force on Cervical Cancer Screening Programs (Department of National Health & Welfare, 1982) stated that squamous c e l l carcinoma of the cervix can be controlled by means of a cytological screening programme for the following reasons: . invasive squamous c e l l carcinoma of the cervix is preceded by a spectrum of disease, extending over many years that may be recognized at the stages of dysplasia and carcinoma in situ; . in a significant portion of patients with severe dysplasia or carcinoma in situ the disease, i f untreated, will develop into invasive squamous c e l l carcinoma; . cytological evidence of dysplasia and carcinoma in situ can be easily, -63-safely and economically obtained by the preparation and examination of smears; . once dysplasia or carcinoma in situ has been identified further progress of the disease can be prevented by simple therapeutic procedures and continuing surveillance. The Canadian rates for malignant neoplasms of the uterus have dropped significantly between 1952 and 1980. For women aged 35 - 64 the rate per thousand has dropped from around 30 to approximately 10. British Columbia has had a more extensive cytology programme operating for longer than any other area in the world. Kinlen and Doll (1973) compared the mortality rates for B.C. against those of Ontario and the rest of Canada and found that although the mortality from cervical cancer had declined materially for women under 45, there was l i t t l e difference between the rates for British Columbia and those for the rest of Canada. For women aged 45 -64, British Columbia has experienced a significantly greater decline in mortality than the rest of Canada. This may in part be a reflection of the fact that the progress of the disease is slow and that i t may develop into i t s l i f e threatening stage over a period as great as twenty years (Boyes, 1987). The frequency of screening recommended by the Task Force was annual for women under the age of 35, and for those over 35, every five years unless there are reasons to believe that the woman is in a higher risk group. The C.H.S. asked women, 'When did you last have a Pap smear test: -64-less than 12 months ago? between one and two years ago? more than two years ago? never? don11 know?' The same question was asked about a breast examination by a doctor or a nurse. Subjects were also asked, 'How often do you examine your own breasts for tumours or cysts: at least monthly? once every two to three months? less often? never? don't know how to do i t ? ' Table 4.17 shows the responses of the subjects included in this study, as they were categorized in the C.H.S. results. TABLE 4.17 PERCENTAGE DISTRIBUTION OF FREQUENCY OF PREVENTIVE BEHAVIOURS OF WOMEN IN THE STUDY GROUP FREQUENCY PAP SMEAR PROF. BREAST (n:3760) EXAM (n:3760) Less than two years ago 71.5 66.7 More than two years 12.4 13.4 Never 5.1 10.7 Not sure 2.9 1.2 Missing 8.1 8.0 TOTAL 100.0 100.0 BREAST SELF EXAM More frequently than every three months 47.2 Less often 19.8 Never 19.1 Don't know how 5.5 Missing 8.4 TOTAL 100.0 -65-For developing a single indicator of women's preventive behaviours for this study, Professional Breast Examination and Pap Smear within two years were scored '1', less recently than the past two years '2' and never or not sure responses '3'. For Breast Self Examination, those reporting the examination at least every three months were scored '1' , those practicing less frequently '2' and those who never practiced self examination or didn't know how '3'. This provided a possible score range of 3 - 9 and the distribution of the scores i s shown in Table 4.18. TABLE 4.18 PERCENTAGE DISTRIBUTION OF COMPOSITE PREVENTIVE BEHAVIOURS SCORES FOR WOMEN IN THE STUDY GROUP COMPOSITE SCORE WOMEN (n:3760) 3 35.7 4 17.5 5 20.3 6 5.8 7 6.8 8 2.1 9 2.8 Missing 8.9 TOTAL 100.0 MOTOR VEHICLE TRAVEL Motor vehicle accidents account for 4% of a l l deaths (Statistics Canada, 1986). Among young adults they are the leading cause of death. While mechanical failure may account for a small fraction of these deaths, by far the most significant factor is driver behaviour. Milsum quotes mileage driven as having a risk factor between 0.2 and 3.0 (1984). The advent of the two car family and the chauffeuring mother taking children to their extracurricular activities plus women driving to and from their work place could contribute to women being on the road more and -66-therefore being at greater risk of being involved in a motor vehicle accident. It i s recognized however, that simply being on the road represents probably the smallest component of the overall risk from motor vehicle travel and that factors such as driving experience, speed, risk taking and the associated use of alcohol a l l contribute more to the level of risk. The C.H.S. asked subjects 'During the last two weeks, about how many miles/kilometers have you travelled as a passenger: in automobiles? in trucks or vans? on motorcycles? or was not a passenger in the past two weeks.' The same questions were asked regarding travel as a driver. From the travelling distances reported for the two week period the C.H.S. developed 3 estimates of distances travelled annually. For the purposes of this study, the mileages as a driver and passenger were combined and the subjects' travelling distances used to place them in one of three categories. The boundaries for the categories were derived by dividing the female subjects into approximately three equal groups as shown in Table A.19. Each two-week report of kilometres driven was multiplied by 6.5 to represent the winter, summer or autumn travelling distance according to the season in which the data were actually collected. To each of these estimates was added the average seasonal estimate for respondents in the same age-sex-community size category for each of the other three seasons. As there were no data collected in the spring, autumn data were used in place of spring travelling estimates. -67-TABLE 4.19 PERCENTAGE DISTRIBUTION OF ANNUAL AUTOMOBILE TRAVELLING DISTANCES OF THE STUDY GROUP DISTANCE TRAVELLED BY AUTOMOBILE WOMEN (n:3760) MEN (n:3562) less than 4,000 km 23.5 6.3 4,001 km to 11,000 km 22.1 9.0 more than 11,001 km 22.5 21.0 missing 31.9 63.7 TOTAL 100.0 100.0 SEAT BELT USE At the time that the C.H.S. data were collected, seat belts were mandatory for 76% of the Canadian population: for residents of Quebec, Ontario, Saskatchewan and British Columbia (Stephen, 1985). There does not appear to be any doubt that regular wearing of a lap and sash seat belt while travelling in a motor vehicle reduces the risk of death or serious injury. Numerous studies have been able to demonstrate not only the benefits of wearing a seat belt but also the effectiveness of seat belt legislation. The C.H.S. results for a l l subjects showed a compliance rate of about 60% in provinces where seat belt wearing was mandatory and only about 16% in regions where i t was voluntary. Figures from Britain show changes from 30% to 80-100% in the rate of compliance following legislation (Allen, Barne & Bodiwala, 1985; Dreghorne, 1985). The C.H.S. asked subjects whether they fastened their seat belts as (a) a passenger and (b) as a driver: always? most of the time? rarely? never? -68-For the purposes of this study a composite score on seat belt use was developed so that subjects were classified as: always wearing seat belt as passenger or driver, sometimes wearing seat belt as a passenger, or rarely or never wearing seat belt as passenger or driver. Table 4.20 shows the distribution of scores across the study group. TABLE 4.20 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY SEAT BELT USE AND SEX FREQUENCY OF WEARING SEAT BELT WOMEN (n:3760) MEN (n:3562) always 24.7 19.9 sometimes 2.7 2.8 never 20.3 13.4 missing 52.3 69.6 TOTAL 100.0 100.0 COMPOSITE HEALTH RISK SCORE To reduce the number of predictor variables being used to address the second study question, namely , 'To what extent can demographic characteristics and health risk, behaviours explain variations in reported health status?', a composite health risk score was developed from seven of the health risk factors just discussed. The eighth factor 'Seat Belt Use' was omitted because in so many cases i t was a behaviour affected more by the law (an external agent) than by the individual's internalized drive to minimize their risk exposure. Female scores for screening behaviours were also not included because they represent a behaviour that could affect the severity of the outcome of disease but not the i n i t i a l development of the disease i t s e l f . For women, the seven factors as shown in Table 4.21 have each been scored '1' through '3', '1' representing the lowest risk exposure and '3' the highest. In keeping with the method used by Belloc and Breslow (1972), no -69-weighting for these factors was used. The scores were simply summed, providing a possible range of scores from '7' to '21'. Men's scores were treated s i m i l a r l y but the i r possible range of scores was '6' to '18' because the use of hormone p i l l s was not applicable. It i s acknowledged that the assignment of 'high' r i s k scores of three to some factors represents a more categorical statement of r i s k than the l i t e r a t u r e quoted e a r l i e r might support. For example, alcohol consumption and hormone p i l l s are two variables which might have been c l a s s i f i e d d i f f e r e n t l y given a greater l e v e l of d e t a i l . TABLE 4.21 CREATING COMPOSITE HEALTH RISK SCORES VARIABLE SCORED '1' SCORED '2' SCORED '3' Smoking Never Smoked Past Smoker, Regular or Occasional Current Smoker Regular or Occasional Alcohol None i n Previous Week 1-7 Drinks i n Previous Week More than 7 drinks i n the Previous Week Physical A c t i v i t y Active, 3000+ on A c t i v i t y Index Moderate, 1750-2999 on A c t i v i t y Index Sedentary, 0-1749 on A c t i v i t y Index L i f e Change Events None or one event Two or three Events Four or More Events Companionship for Leisure A c t i v i t y Mostly with Company Half and Half Mostly Alone Hormone P i l l s Not Taken - Taken Annual Motor Vehicle Travel Less than 4000 km. 4001 to 11,000 km. More than 11,000 km. -70-Based on this scoring system, the resulting distribution of scores i s shown in Table 4.22. The large number of missing scores arises from the fact that any subject for whom there was not a score on each of the seven (six for men) items was deemed missing. Missing scores on automobile travel (as passenger or driver) account for over half the missing composite risk scores. TABLE 4.22 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY COMPOSITE RISK SCORE AND SEX WOMEN MEN (n:3760) (n:3562) Score 7 0.2 Score 6 0.0 8 0.8 7 0.1 9 1.9 8 0.5 10 4.5 9 1.9 11 7.2 10 4.5 12 8.4 11 4.7 13 6.6 12 6.1 14 5.4 13 4.5 15 2.7 14 2.3 16 1.9 15 0.9 17 0.6 16 0.4 18 0.2 17 0.1 19 0.0 18 0.0 20 0.0 21 0.0 Missing 59.6 Missing 74.0 TOTAL 100.0 TOTAL 100.0 -71-4.4 HEALTH STATUS INDICATORS CHRONIC HEALTH PROBLEMS The Interviewer Administered Questionnaire asked subjects the questions outlined in Table 4.23. TABLE 4.23 C.H.S. QUESTION USED TO IDENTIFY HEALTH PROBLEMS 'Do you or anyone in your family presently have: Anaemia? Skin allergies? Hayfever or other allergies? Asthma? Arthritis or rheumatism? Cancer? Cerebral palsy? Diabetes? Emphysema or chronic bronchitis? Mental retardation? Any emotional disorder (excluding mental retardation)? Epilepsy? High blood pressure? Heart disease? Stomach ulcer? Thyroid trouble or goitre? Recurring migraine headaches? Missing arms or legs? Missing fingers or toes? Paralysis of any kind? Excluding any health problem mentioned earlier such as arthritis or paralysis, does anyone in the family have: Serious trouble with their back or spine? Serious trouble with their legs or hips? Serious trouble with their arms or shoulders? Serious trouble with any other bones or joints?'  The interviewer then recorded which person in the family had the problem. They also asked whether anyone in the family had any other long term illness or impairment. Of a l l subjects included in the C.H.S., approximately 50% reported some chronic condition. For the group included in this study, the distribution of -72-chronic problems is shown in Table 4.24. Subjects were also asked whether they were limited from their normal daily activities by any health problem. For the present study, subjects that indicated that they did have activity limitations (7.8%) were excluded because i t was fel t that such subjects may self-select out of the workforce and could be expected to have different l i f e s t y l e behaviours (for example, reduced physical activity). Table 4.25 shows the distribution of subjects excluded on the grounds of activity limitations. TABLE 4.24 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY NUMBER OF CHRONIC HEALTH PROBLEMS AND SEX NUMBER OF CHRONIC PROBLEMS WOMEN (n:3760) MEN (n:3526) None 41.8 54.6 One 30.6 29.7 Two 16.1 10.5 Three 7.3 3.3 Four 2.7 1.3 More than four 1.5 0.6 TOTAL 100.0 100.0 TABLE 4.25 SUBJECTS EXCLUDED DUE TO ACTIVITY LIMITATIONS SUBJECTS REPORTING ACTIVITY LIMITATIONS DUE TO HEALTH PROBLEMS Women Homemakers 10.5% Women in the paid labour force 7.2% Men in the paid labour force 6.4% These subjects were excluded from the study. HEALTH OPINION SCORE A psychological screening test developed by Allister MacMillan was used to measure the reported frequency of psycho-physiological symptoms of anxiety and depression. The test i s made up of 16 questions (see Table 4.26) which generate a possible range of scores of 16 (experiencing a l l of the symptoms -73-frequently) to 48 (never experiencing any of the symptoms). TABLE 4.26 DEVELOPING A HEALTH OPINION SCORE HEALTH OPINION SCORE The following sixteen questions were scored '1' for 'Often', '2' for 'Sometimes' and '3' for 'Never' providing a possible score of 16 - 48. 1. 2. 3. 4. 6. 7. 8. 10. 11. 12. 13. 14. 15. 16. Have you ever been bothered by your heart beating? How often are you bothered by an upset stomach? Do your hands ever tremble enough to bother you? Are you ever troubled by your hands or feet sweating so that they feel damp and clammy? Have you ever been bothered by shortness of breath while not exerting yourself? Do you ever have spells of dizziness? Do you feel weak a l l over much of the time? Do you feel healthy enough to carry out the things you would like to?* Do you feel you are bothered by a l l sorts (different kinds) of ailments in different parts of your body? Do you ever have loss of appetite? Do you have any trouble in getting asleep and staying asleep? Has i l l health affected the amount of work you do? Have you ever felt that you were going to have a nervous breakdown? Are you ever bothered by nightmares? Do you tend to lose weight when important things are bothering you? Do you tend to feel tired in the mornings? * Question 8 is scored in reverse. Tousignant et a l . (Tousignant, Denis & Lachapelle, 1974) question the capacity of this instrument to serve as a population screen for mental disorders - the original purpose of the test. However they do suggest that the test may be well suited for showing variations in the subject's physical health. For the purposes of this study, comparisons can s t i l l usefully be drawn between groups assuming that the biases suggested by Tousignant, namely physical health and pressure to provide socially desirable answers, equally -74-a f f e c t a l l groups. (This i s probably a safer assumption when comparing the two women's groups than when comparing men's and women's groups.) These Health Opinion questions were administered via the self-administered component of the C.H.S.. In compiling the Health Opinion Scores for the C.H.S. data, i f 13 of the 16 component scores were known, the unknown scores were imputed from the average of the known scores. If more than three items were missing, the Health Opinion Score was deemed missing. Further, t h i s section of the health survey was added af t e r the survey began thereby increasing the number of subjects without scores. Table 4.27 shows the d i s t r i b u t i o n of scores. For the purposes of making comparisons between the groups, the scores were grouped ( i ) 22 - 40, ( i i ) 41 - 44 and ( i i i ) 45 - 48. The group boundaries were based on the d i s t r i b u t i o n of the scores (see Table 4.27). TABLE 4.27 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY HEALTH OPINION SCORE AND SEX HEALTH OPINION SCORES WOMEN (n:3760) MEN (n:3562) 22 - 40 22.5 11.7 41 - 44 27.0 23.4 45 - 48 15.9 23.6 Missing 34.6 41.3 TOTAL 100.0 100.0 -75-AFFECT BALANCE The C.H.S. used Bradburn's measure of psychological well-being (1969) to ask subjects directly about subjective states - both pleasant and unpleasant, in the recent past. Table A. 28 details the questions asked via the self-administered segment of the survey. This measure is based on the concept that psychological well-being is the result of two almost completely unrelated dimensions of affect which have been labelled positive and negative. To reflect the theory that well-being is the resultant of positive and negative feelings (which may vary independently of one another) an overall Affect Balance Score is normally expressed as the difference between the positive and negative affect scores. The resultant scores can then be expressed as positive, negative or neutral (or balanced). TABLE A.28 DEVELOPING AFFECT BALANCE SCORES Five of the questions listed below were used to develop a Positive Affect Score and the other five were used to develop a Negative Affect Score. Subjects were told, 'Here i s a l i s t that describes some of the ways people feel at different times. During the past few weeks how often have you f e l t . . . 1. On top of the world? 2. Very lonely or remote from other people? 3. Particularly excited or interested in something? A. Depressed or very unhappy? 5. Pleased about accomplishing something? 6. Bored? 7. Proud because someone complimented you on something you had done? 8. So restless you couldn't s i t long in a chair? 9. That things were going your way? 10. Upset because someone criticized you?' Subjects were asked to choose between 'Often', 'Sometimes' or 'Never'. -76-For the C.H.S. data, scores were reported i f four out of each set of five items had been answered, the one unknown score being imputed as an average of the known. If more than one item was unanswered, the Affect Balance Score was shown as missing. Table 4.29 shows the Affect Balance Scores for the subjects included in this study. TABLE 4.29 PERCENTAGE DISTRIBUTION OF THE STUDY GROUP BY AFFECT BALANCE SCORE AND SEX AFFECT BALANCE WOMEN MEN (n:3760) (n:3562) Positive Balance 44.7 42.2 Neutral 39.6 35.2 Negative Balance 3.7 2.0 Missing 12.0 20.6 TOTAL 100.0 100.0 -77-4.5 HEALTH CARE CONSEQUENCES DISABILITY DAYS In the interviewer administered segment of the C.H.S., a representative of the household was asked about members of the household experiencing reduced activity days during the previous two weeks. Table 4.30 shows the questions from which a summary score was developed. This summary score equalled the number of days on which a person experienced any reduction in activity as a result of their health. TABLE 4.30 C.H.S. QUESTION TO IDENTIFY DISABILITY DAYS REDUCED ACTIVITY DUE TO POOR HEALTH 1. (a) During the past two weeks did you* stay in bed because of your health? (b) How many days did you stay in bed for a l l or most of the day? (Includes nights spent in hospital as a patient.) 2. (a) During the past two weeks did your health keep you from work/housework for a l l or most of the day? (b) How many days did illness keep you from work/housework for a l l or most of the day? (c) On how many of those days lost from work/housework did you stay in bed? 3. (a) Not counting any days mentioned earlier, were there any days during those two weeks that you cut down on things that you usually do, because of your health? * 'You' has been used for simplicity in reporting the question. Any responsible adult was deemed able to answer on behalf of other members of the household.  Table 4.31 shows the distribution of reduced activity days amongst this study group. -78-TABLE 4.31 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY REDUCED ACTIVITY DAYS AND SEX NUMBER OF DISABILITY DAYS IN THE PREVIOUS TWO WEEKS WOMEN (n:3760) MEN (n:3562) None 88.5 92.8 One 3.4 2.7 Two 3.1 1.9 More than two 5.0 2.6 TOTAL 100.0 100.0 HEALTH PROFESSIONAL CONSULTATIONS Through the interviewer administered section of the survey, subjects or their proxy were asked about contact with health professionals during the previous year. Table 4.32 shows the questions as they were asked and Table 4.33 summarizes the number of health professional v i s i t s reported by this study group. TABLE 4.32 C.H.S. QUESTION TO IDENTIFY HEALTH PROFESSIONAL CONSULTATIONS 'During the past two weeks did anyone in the family see or talk to any of the following health professionals about their health? 1. a medical doctor? 2. a dentist? 3. a nurse? 4. a pharmacist or druggist for advice? (exclude prescriptions) 5. an optometrist or optician? 6. a chiropractor? 7. a psychologist, social worker or other counsellor? 8. any other health professional?' -79-TABLE 4.33 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY NUMBER OF HEALTH PROFESSIONAL CONSULTATIONS AND SEX NUMBER OF HEALTH CONSULTATIONS IN THE PREVIOUS YEAR WOMEN (n:3760) MEN (n:3562) None 7.6 15.3 One or two 27.5 37.7 Three to five 28.2 26.4 Six to ten 20.6 13.9 Eleven to twenty 11.5 4.2 More than twenty 3.2 1.0 Missing 1.4 1.5 TOTAL 100.0 100.0 Subjects were also asked which health professional they had seen most recently, where the contact had taken place (for example in a c l i n i c , a hospital or an office) and what health problem led to that consultation. This additional detail was not used in this study. MEDICATIONS Information on medications was gathered via the interviewer administered part of the survey so that any responsible adult in the family was considered able to answer on behalf of the other family members. Table 4.34 shows the questions that were asked. TABLE 4.34 C.H.S. QUESTION TO IDENTIFY MEDICATIONS TAKEN 'These questions refer to the use of medicines, p i l l s or ointments in the last two days. Yesterday or the day before, did you or any one in the family take or use any of the following: 1. Pain relievers such as aspirin? 2. Tranquilizers, medicines for the nerves or medicines to help you sleep? 3. Medicines for heart or blood pressure? 4. Antibiotics? 5. Stomach remedies or medicines? 6. Laxatives? 7. Cough or cold remedies? 8. Skin ointments or salves? 9. Vitamins or minerals? 10. Any other medication?' -80-Subjects were also asked whether the medication was taken on the advice of a medical doctor and whether the medicine was taken at least once a week over the previous month. These additional pieces of information were not used in this study. Table 4.35 shows the variety of medications consumed over a two-day period by the subjects included in this study. TABLE 4.35 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY VARIETY OF MEDICATIONS CONSUMED AND SEX NUMBER OF DIFFERENT MEDICATIONS CONSUMED OVER A TWO-DAY PERIOD WOMEN (n:3760) MEN (n:3562) None 48.8 67.3 One 32.6 24.4 Two 12.6 6.4 Three 4.9 1.6 Four or more 1.1 0.3 TOTAL 100.0 100.0 HOSPITALIZATION The interviewer asked whether, in the previous 12 months, anyone in the family had been a patient in a hospital, a nursing home or a convalescent home. If so they were asked how many times and for how many nights the person had been a patient. There is no indication in the survey how day hospital v i s i t s may have been treated. Table 4.36 shows the percentage distribution of subjects who had been in hospital and the number of nights they had been a patient in the previous 12 months. -81-TABLE 4.36 PERCENTAGE DISTRIBUTION OF STUDY GROUP BY HOSPITALIZATION AND SEX NUMBER OF NIGHTS IN HOSPITAL IN THE PREVIOUS 12 MONTHS WOMEN (n:3760) MEN (n:3562) None 84.8 95.0 One to five 8.1 3.1 More than five 7.1 1.9 TOTAL 100.0 100.0 Al l this information on health risk exposure, self reported health status and use of health care services, for men and women aged 25 - 44 years who were not experiencing any activity limitations due to health and were in the paid labour force or keeping house, was then analysed, as outlined in Chapter Five. -82-CHAPTER FIVE DATA ANALYSIS 5.1 THE DATA The Canadian Health Survey data collected between July 1978 and March 1979 was committed to tape and placed in the public domain in 1982. A copy of that tape and the code book (Statistics Canada, 1982) is available through the University of British Columbia Data Library at the Computing Centre. Using the Statistical Package for the Social Sciences (SPSS-X), on the U.B.C. Amdahl computer, the data items listed in Appendix A, and described in Chapter Four, were extracted for a l l men and women aged between 25 and 44. Subjects reporting any activity limitations were excluded as were those who indicated that they were neither going out to work nor keeping house. (That is , they were 'at school', 'retired', or 'other'). The remaining subjects were grouped according to their reported major activity for the previous year: women who were keeping house, women who were in the labour force, men who were in the labour force. Using SPSS-X some variables were recoded (see Appendix A for details): to reduce the number of categories within a variable, or to make scoring of one variable consistent with that of another so that the two could be combined to create one summary variable. For example, information about the wearing of a seat belt as a passenger or a driver was combined to reflect the subject's overall use of seat belts. Similarly new summary variables were created for driving distance, for women's preventive health behaviours that depended on a v i s i t to a health professional and for each subject's lifestyle-related health risk. The -83-rationale for, and greater d e t a i l on each of these variables i s provided i n Chapter Four. In order to address the four questions posed i n t h i s study, namely: 1. Do women i n the paid labour force exhibit l i f e s t y l e patterns s i g n i f i c a n t l y d i f f e r e n t from th e i r counterparts who are homemakers? 2. Do women i n the paid labour force report poorer health status than th e i r counterparts who are homemakers? 3. Do women i n the paid labour force exhibit health care u t i l i z a t i o n patterns s i g n i f i c a n t l y d i f f e r e n t from their homemaker counterparts? and 4. Do women's l i f e s t y l e s , reported health status and health care u t i l i z a t i o n patterns d i f f e r from those of t h e i r male counterparts i n the paid labour force? It was recognized that some standardization of the rates for the various behaviours would need to be made i f the comparisons were to be at a l l meaningful. Multiple C l a s s i f i c a t i o n Analysis (University of Michigan 1981) was chosen as a suitable instrument for determining the variables that were best able to explain variations i n scores, these then being the same variables that could be used to reduce population v a r i a b i l i t y across a l l three groups. 5.2 MULTIPLE CLASSIFICATION ANALYSIS Multiple C l a s s i f i c a t i o n Analysis (MCA) examines the relationship between several categorical predictor variables and a single dependent variable and determines the e f f e c t of each predictor before and after adjustment for i t s i n t e r c o r r e l a t i o n with other predictor variables i n the analysis. MCA i s s i m i l a r to regression analysis but has the advantage of being able to use predictor variables i n nominal, ordinal or i n t e r v a l scales. MCA assumes that the e f f e c t s of the predictors are additive and that there i s no i n t e r a c t i o n . - 84 -MCA assumes a model of the form: in which 'Yijk' i s the k t h value of the dependent variable 'Y' and 'a/ and 'gj' are respectively the i t ' 1 and j 1"* 1 categories of predictor variables 'a' and ' g'; eijk i s the within group or sampling error. Through an iterative process, the grand unadjusted mean of the dependent variable 'Y' i s adjusted to reflect the effect of successive values of the predictor variables 'a' through 'g'. In this study for example, smoking could be the dependent variable, 'Y' and the predictor variables ('a' through ' g') would be the seven demographic variables. The MCA produces an unadjusted mean score for smoking for each category of each predictor variable (for example each of four levels of education) as well as an adjusted mean score for each of those same categories when the effect of the other predictor variables is held constant. For each dependent variable (Y) the following statistics were generated: grand mean standard deviation sum of Y sum of Y squared explained sum of squares residual sum of squares the number of cases used in the analysis. and for each category of each predictor variable: the number of cases with valid data the mean of the dependent variables for the category deviation of the category mean from the grand mean (unadjusted) deviation of the category mean from the grand mean adjusted for the effect of the other predictors adjusted class mean standard deviation of the dependent variable for the category - 85 -The unadjusted category deviations can be used to assess the relationship of the predictor variable with the dependent variable (be i t positive, negative, curilinear or any other form). The adjusted deviation or co-efficients show this relationship after the effects of the other predictors have been partialled out. The following Analysis Summary Statistics were also generated: The proportion of variance explained by the model (the unadjusted R-square) a co-efficient indicating the amount of adjustment for degree of freedom used in f i t t i n g the model the adjusted proportion of variance explained by the model (the multiple R-square, adjusted) the adjusted multiple correlation co-efficient (the multiple R, adjusted) the eta squared which represents the proportion of variance in the dependent variable explained by the unadjusted deviations of the predictor and the beta squared co-efficients which indicate the relative importance of the various predictors The two women's groups were combined and MCA was run based on the predictor and dependent variables outlined in the model. The predictors best able to explain variation in the dependent variable scores were then used to standardize the rates of the dependent variables. 5.3 STANDARDIZATION OF RATES When comparing crude rates among several populations for an event or population characteristic, the difference between the two rates is made up of: the difference within the two populations and the difference between the frequency of the event or characteristic being compared. (Fleiss, 1981) Since i t is the latter that is of major interest, the degree to which the differences between the populations can be eliminated will determine the - 86 -accuracy with which a comparison of rates reflects a true difference in frequency of the event or behaviour of interest. It could be expected that there would be significant population differences between the women horaemaker group and the workforce women group. For example, unmarried women are most likely to be in the workforce with very few of them being homemakers whereas married women with large families may well be very under represented in the workforce group. For some dependent variables, these characteristics may be strongly associated with that variable's absence or presence. Using the results of the MCA, the two predictor variables shown to be most able to explain variation in each dependent variable were used to standardize the populations and therefore reduce the amount of variability that could be attributed to population differences. It had originally been planned that the rates for workforce men and women would be standardized against the total female group rate. However, within the total female group there were strata that were made up almost exclusively of workforce women, so i t was decided that the homemaker group should be used as the standard population. The rates for the workforce men and women groups have therefore been standardized against the women homemaker group. As pointed out by Fleiss, although comparing single standardized rates provides a convenient single measure of comparison i t will not indicate the differences that exist across the various strata within the population. The standardized tables for each variable have therefore been included in the results section so that readers may examine for themselves the variations across strata. - 87 -5.4 TESTING FOR SIGNIFICANCE BETWEEN RATES On the premise that a comparison of the rates constituted a comparison of population means, a standard t-test was used to test for significance between the rates. The weighted difference in proportions of individuals classified as 'positive' between corrpsponding cells of the two populations was used as the numerator for the t-test and a pooled variance for the denominator. The frequencies for the i 1 " * 1 c e l l of populations A and B as outlined below are used to illustrate the formulae. (Armitage, 1971) A B A & B Positive r.. r n . r. Ai Bi l Negative n. . - r.. n D. - r B . n. - r. ° Ai Ai Bi Bi i i Total n. . n n. n. A i B i l Proportion Positive p.. = r../n.. p D. = r^./n-n. p . = r./n. r A i Ai Ai *BI Bi Bi * o i l l q . = 1 - p . M o i *-oi The difference between the proportions (d) was calculated: d = 2 n . ( p A i - p B i ) / l n and the pooled variance (d) d = P o i V ( n A i + n B i ) / n A i n B i so that , oL Tests of significance were done on a l l the dependent variables. Tables 6.78-80 summarize the rates and the significance level of their differences. -88-CHAPTER SIX RESULTS The extraction of the data points shown in Appendix I for men and women aged 25 to 44 years produced a sample of 7,939 subjects. Based on an assumption that subjects with activity limitations would not have the same lif e s t y l e options as functionally unimpaired subjects, 617 of this group were then omitted from this study because they indicated that they experienced some level of activity limitation due to health problems. Table 6.1 shows the distribution of these subjects across the three groups. The 7.8% overall incidence of activity limitations for this subject group was much lower than for the 45 to 64 year olds in the C.H.S. (17%). The higher incidence of health related activity limitations among women at home should not be surprising and could be used to support the argument that women in the paid labour force are to some extent, self-selecting. TABLE 6.1 SAMPLE SIZES AFTER ADJUSTMENT FOR THOSE REPORTING ACTIVITY LIMITATIONS GROUPS AGED 25 - 44 YEARS NO. IN C.H.S. N0.(%) WITH ACTIVITY LIMITATIONS NO. IN THIS STUDY Homemaker Women 2302 241 (10.5%) 2061 Workforce Women 1830 131 ( 7.2%) 1699 Workforce Men 3807 245 ( 6.4%) 3562 TOTAL 7939 617 ( 7.8%) 7322 The rest of this chapter reports on each of the variables outlining the Multiple Classification findings and discussing the standardized rates in terras of the findings of other studies. Tables 6.76-80 at the end of the chapter, provide a summary of a l l the -89-multiple classification analyses plus an overview of the comparison of rates for health risk behaviours, for reported health status and for health consequences across a l l three groups. Also at the end of this chapter, the overall impact of the various predictor variables are tabulated (Tables 6.81-87). 6.1 HEALTH RISK BEHAVIOURS SMOKING The multiple classification analysis using the women's smoking scores as the dependent variables and the predictor variables as outlined in the model (Fig. A.3), could explain only 3.9% of the variation in scores. Of those in the model, the three strongest predictors were education, marital status and region of residence. Given that smoking was scored one to three, (three indicating current smoking either regular or occasional and one indicating a non-smoker), one can seen from Table 6.2 that higher education and marriage are associated with lower group mean scores. For the regions, the smoking rate simply varies from province to province, Quebec being the highest and Ontario the lowest. The beta score provided with each predictor variable indicates i t s relative strength as a predictor. TABLE 6.2 PREDICTORS OF SMOKING BY WOMEN SUBJECTS EDUCATION GROUP MARITAL GROUP REGION GROUP MEAN STATUS MEAN MEAN (beta:0.16) (beta:0.11) (beta:0.07) Secondary 2.19 Not married 2.28 Quebec 2.16 Post. Sec. 2.01 Married 2.06 Atlantic 2.13 Diploma 1.96 Prairies 2.12 Degree 1.79 B.C. 2.11 Ontario 1.99 grand mean 2. .0 - percentage variance explained 3.9 -90-From among a l l the subsets created by the M.C.A., women with a degree participating in the paid labour force (group mean 1.73) and women in the paid labour force living in large families (group mean 1.88) reported the lowest number of smokers. Single women (group mean 2.28) and women living alone (group mean 2.20) reported the highest levels of smoking. Using the top two strongest predictors within this model, that is education and marital status, the smoking rates for the two workforce groups were standardized against the women homemaker group. The workforce men had the highest smoking rate (52.4%) and workforce women, the lowest (44.8%). Tables 6.3-5 show the variation across the different strata of the subject groups as well as the summary rates. The differences between the two workforce groups and the homemaker group were not significant at the 0.05 level, (t values of 0.745 and 1.71 for workforce women and men respectively). The smoking rates reported here are high compared with those generally reported. The Report on the C.H.S. (Health and Welfare Canada/Statistics Canada, 1982) (see Table 6.6) shows the smoking rates for a l l women aged 25 -44 as 38.9% (and 36% for the 1977 and 1979 'Smoking Habits of Canadians' surveys). Table 6.7 shows the unweighted smoking rates for the sample used in this study and the rates for the same sample when the weights provided with the C.H.S. (to match the sample population with the general population on the basis of age and sex) are applied. This information shows that smokers must have been slightly over represented in this sample. When this, plus the fact that occasional as well as regular smokers have been included under the heading of 'smokers' in this study, the rates do not appear too disparate with other findings. -91-TABLE 6.6 PERCENTAGE OF DAILY CIGARETTE SMOKERS AMONG 25-44 YEAR OLDS, REPORTED BY THE 'SMOKING HABITS OF CANADA' SURVEYS AND THE C.H.S. SMOKING HABITS OF CANADA HEALTH CANADA SURVEY 1977 1979 1978-79 Women 25-44 years 36.6 36.0 38.9 Men 25-44 years 47.0 44.0 46.7 Source: The Health of Canadians. TABLE 6.7 WEIGHTED AND UNWEIGHTED C.H.S. SMOKING RATES FOR WOMEN 25-44 YEARS UNWEIGHTED RATES WEIGHTED RATES Home Gp. W'force Gp. Home Gp. W'force Gp Regular Smokers 41.2 39.2 Occasional Smokers 3.5 4.5 TOTALS 44.7 43.7 38.7 36.5 3.9 4.8 42.6 41.3 TABLE 6:3 SMOKING PREVALENCE WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting Current, Regular or Occasional Cigarette Smoking EDUCATION LEVEL 2 EDUCATION LEVEL 3 EDUCATION LEVEL 4 EDUCATION LEVEL 5 CRUDE RATE ADJUSTED RATE NOT MARRIED 63.7 (91) 50.0 (8) 50.0 (6) 66.7 (3) 62.0 (108) 61.2 MARRIED 48.3 (1291) 37.5 (112) 30.5 (233) 26.6 (94) 44.1 (1730) 44.2 CRUDE RATE 49.3 (1382) 38.3 (120) 31.0 (239) 27.8 (97) 45.1 (1838) ADJUSTED RATE 49.3 38.2 31.6 29.0 45.2 -92-TABLE 6.4 SMOKING PREVALENCE WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting Current, Regular or Occasional Cigarette Smoking^ EDUCATION LEVEL 2 EDUCATION LEVEL 3 EDUCATION LEVEL 4 EDUCATION LEVEL 5 CRUDE RATE ADJUSTED RATE NOT MARRIED 63.0 (184) 50.0 (58) 44.9 (78) 26.0 (77) 50.4 (397) 57.9 MARRIED 47.9 (685) 31.5 (92) 36.8 (163) 20.4 (142) 41.2 (1082) 43.9 CRUDE RATE 55.4 (869) 38.7 (150) 39.4 (241) 22.4 (219) 43.7 (1479) ADJUSTED RATE 48.8 32.6 37.3 20.8 44.8 TABLE 6.5 SMOKING PREVALENCE MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting Current, Regular or Occasional Cigarette Smoking EDUCATION LEVEL 2 EDUCATION LEVEL 3 EDUCATION LEVEL 4 EDUCATION LEVEL 5 CRUDE RATE ADJUSTED RATE NOT MARRIED 56.5 (1555) 46.0 (198) 40.2 (271) 29.1 (402) 49.3 (2426) 53.6 MARRIED 56.9 (253) 42.6 (54) 47.9 (48) 34.5 (87) 49.8 (442) 52.3 CRUDE RATE • 56.6 (1808) 45.2 (252) 41.4 (319) 30.1 (489) 49.4 (2868) ADJUSTED RATE 56.7 45.8 40.7 29.4 52.4 Rates standardized against the "Women Homemakers" group. -93-ALCOHOL Using the women's alcohol consumption scores as the dependent variable and the predictor variables listed in the model (Figure 4.3), the multiple classification analysis was able to explain only 4.7% of the variation in the scores. Subjects were allocated a score of three i f they reported more than seven alcoholic beverages in the week preceding the survey, a score of two i f they had between one and seven drinks and a score of one i f they had not consumed any alcohol at a l l in the preceding week. The top three predictor variables were income, region of residence and marital status. Table 6.8 shows the sub group mean scores for each category of the top three predictors as well as their beta scores which indicate their relative strengths as predictors. TABLE 6.8 PREDICTORS OF ALCOHOL CONSUMPTION BY WOMEN SUBJECTS INCOME GROUP QUINTILE MEAN (beta:0.12) REGION GROUP MEAN (beta:0.09) MARITAL GROUP STATUS MEAN (beta:0.08) First (lowest) 1.86 Second 2.01 Third 1.98 Fourth 2.10 Fifth (highest) 2.13 Atlantic 1.91 Quebec 2.01 Ontario 2.10 Prairies 2.08 B.C. 2.11 Not married 2.19 Married 2.01 grand mean 2.04 - percentage variance explained 4.7 Across a l l the subgroups, those reporting the least alcohol consumption were those with the lowest incomes, living in the Atlantic provinces or living in a large family; those reporting the highest consumption were single women and women in the highest income group. The rates for workforce men and women consuming more than seven drinks a week were standardized on the basis of income and region against the -94-women homemaker's group. Overall the women reported considerably less alcohol consumption than the men; women abstained more and reported more moderate drinking (see Table 4.9). When the populations were standardized the rate for men having seven or more drinks (49.7%) was twice that of the homemaker and workforce women's groups (19.6% and 21.0% respectively, p<.002). The rate for men abstaining from alcohol (8.5%) was half that of the rate for the women's group (19.9% and 16.2%, p<.002 respectively for the homemaker and workforce women). (See Tables 6.9-14) The tables also show some interesting differences related to income. For example, in B .C. among workforce women, the percentage reporting seven or more drinks decreases with increasing income yet for the rest of the group the relationship to income seems to be the reverse. Of interest is the fact that more women (44.54%) than men (35.6%) did not complete the question on the previous week's alcohol consumption. One can only speculate as to whether a higher percentage of the non-reporters were heavier drinkers and to what extent social acceptance and perceived social values have influenced responses as well as non-responses. Other studies (Wilsnack, Wilsnack and Klassen, 1981) have found no evidence of heavier drinking among employed wives but found the heaviest drinking among single women aged 21 to 34, and among women whose spouse or companion was a heavy drinker. The rates for moderate to heavy drinkers (averaging more than 0.22 ounces of alcohol per day) shown in Wilsnack, Wilsnack and Klassen's study would suggest, i f anything, that the rates in this study might be low. -95-TABLE 6.9 ALCOHOL CONSUMPTION WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting more than Seven Drinks in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 7.9 (63) 9.0 (67) 8.9 . (45) 25.0 (24) 23.1 (13) 11.3 (212) 13.7 QUEBEC 10.4 (48) 6.3 (64) 17.2 (58) 13.8 (29) 22.2 (18) 12.4 (217) 13.2 ONTARIO 27.5 (40) 25.0 (48) 20.0 (55) 36.5 (52) 26.3 (19) 27.1 (214) 26.7 PRAIRIES 21.1 (19) 15.3 (59) 19.7 (61) 31.9 (72) 17.7 (51) 21.8 (262) 21.0 B.C. 21.4 (14) 32.4 (34) 15.2 (46) 23.7 (38) 29.0 (31) 23.9 (163) 24.1 CRUDE RATE 15.2 (184) 15.4 (272) 16.6 (265) 28.4 (215) 22.7 (132) 19.2 (1068) ADJUSTED RATE 17.6 16.7 16.4 26.5 23.1 19.6 Celanto and McQueen (1984) in a study of Baltimore women found only weak support for the hypothesis that women drink more as they entered the male world and that attitudes and social roles were more important than occupational status. -96-TABLE 6.10 ALCOHOL CONSUMPTION WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting more than Seven Drinks in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 6.3 (16) 23.1 (13) 16.7 (30) 21.6 (51) 22.0 (59) 18.9 (169) 18.2 QUEBEC 18.2 (ID 8.3 (12) 33.3 (24) 19.2 (47) 23.3 (103) 22.3 (197) 20.3 ONTARIO 27.3 (11) 16.7 (12) 15.4 (39) 21.0 (62) 27.2 (103) 22.9 (227) 20.3 PRAIRIES 12.5 (8) 20.0 (10) 15.2 (33) 25.0 (48) 29.0 (100) 24.6 (199) 19.6 B.C. 0 (2) 42.1 (19) 31.3 (16) 33.3 (21) 28.6 (84) 31.0 (142) 28.7 CRUDE RATE 14.6 (48) 24.2 (66) 20.4 (142) 22.7 (229) 26.3 (449) 23.8 (934) ADJUSTED RATE 13.5 21.0 21.7 23.6 26.0 21.0 Rates standardized against the "Women Homemakers" group -97-TABLE 6.11 ALCOHOL CONSUMPTION MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting more than Seven Drinks in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 38.6 (70) 53.3 (92) 46.3 (82) 47.9 (94) 50.0 (80) 47.6 (418) 47.5 QUEBEC 35.3 (51) 39.8 (113) 45.1 (113) 48.7 (117) 44.5 (128) 43.7 (522) 42.7 ONTARIO 68.4 (19) 49.1 (55) 47.2 (106) 51.6 (126) 48.7 (158) 50.0 (464) 52.4 PRAIRIES 38.1 (21) 53.4 (73) 50.0 (98) 58.5 (106) 52.8 (176) 53.0 (474) 50.9 B.C. 75.0 (4) 51.4 (35) 50.7 (71) 48.5 (66) 64.8 (145) 57.0 (321) 56.4 CRUDE RATE 41.8 (165) 48.4 (368) 47.7 (470) 51.3 (509) 52.6 (687) 49.7 (2199) ADJUSTED RATE 49.3 49.5 47.8 51.5 51.6 49.7 Rates standardized against the "Women Homemakers" group. -98-TABLE 6.12 NON-DRINKERS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting NO Alcohol Consumption in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 41.3 (63) 20.9 (67) 17.8 (45) 29.2 (24) 15.4 (13) 26.9 (212) 24.6 QUEBEC 29.2 (48) 14.1 (64) 20.7 (58) 13.8 (29) 16.7 (18) 19.4 (217) 18.6 ONTARIO 20.0 (40) 14.6 (48) 25.5 (55) 17.3 (52) 47.4 (19) 22.0 (214) 22.8 PRAIRIES 21.1 (19) 18.6 (59) 18.0 (61) 11.1 (72) 17.7 (51) 16.4 (262) 15.3 B.C. 42.9 (14) 11.8 (34) 23.9 (46) 10.5 (38) 3.2 (31) 15.9 (163) 18.8 CRUDE RATE 31.5 (184) 16.5 (272) 21.1 (265) 14.9 (215) 18.2 (132) 20.1 (1068) ADJUSTED RATE 29.9 16.3 20.9 14.0 20.7 19.9 -99-TABLE 6.13 NON-DRINKERS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting NO Alcohol Consumption in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 12.5 (16) 30.8 (13) 20.0 (30) 17.7 (51) 20.3 (59) 19.5 (169) 21.0 QUEBEC 18.2 (ID 8.3 (12) 8.3 (24) 21.3 (47) 10.7 (103) 13.2 (197) 12.9 ONTARIO 9.1 (ID 8.3 (12) 20.5 (39) 12.9 (62) 1.0 (103) 8.4 (227) 11.5 PRAIRIES 37.5 (8) 0 (10) 18.2 (33) 10.4 (48) 10.0 (100) 12.1 (199) 14.3 B.C. 50.0 (2) 26.3 (19) 12.5 (16) 14.3 (21) 15.5 (84) 16.9 (142) 23.2 CRUDE RATE 18.8 (48) 16.7 (66) 16.9 (142) 15.3 (229) 10.5 (449) 13.5 (934) ADJUSTED RATE 24.8 13.5 16.1 15.2 11.2 16.2 Rates standardized against the "Women Homemakers" group. -100-TABLE 6.14 NON-DRINKERS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting NO Alcohol Consumption in the Previous Week REGION INCOME QUINT. 1 INCOME QUINT. 2 INCOME QUINT. 3 INCOME QUINT. 4 INCOME QUINT. 5 CRUDE RATE ADJUSTED RATE ATLANTIC REGIONS 17.1 (70) 13.0 (92) 13.4 (82) 17.0 (94) 8.8 (80) 13.9 (418) 14.1 QUEBEC 17.7 (51) 6.2 (113) 4.4 (113) 6.8 (117) 4.7 (128) 6.7 (522) 7.7 ONTARIO 15.8 (19) 10.9 (55) 6.6 (106) 8.7 (126) 11.4 (158) 9.7 (464) 10.3 PRAIRIES 0 (21) 12.3 (73) 3.1 (98) 6.6 (106) 7.4 (176) 6.8 (474) 6.2 B.C. 0 (4) 0 (35) 7.0 (71) 7.6 (66) 2.8 (145) 4.4 (321) 3.6 CRUDE RATE 14.6 (165) 9.2 (368) 6.6 (470) 9.2 (509) 7.0 (687) 8.4 (2199) ADJUSTED RATE 10.2 9.1 6.7 9.3 7.2 8.5 Rates standardized against the "Women Homemakers" group. -101-PHYSICAL ACTIVITY The top predictors of the amount of physical activity women reported were the season the survey was taken, the region of residence and the level of education. Activity levels were scored according to the physical activity index (see Chapter Four) with a score of three being assigned to the lowest reported levels of activity and a score of one being assigned to the highest. The grand mean score for a l l female subjects was 1.99. Table 6.15 shows the beta values of the top three predictors, season of survey, region and education, and the group means for each of their categories. TABLE 6:15 PREDICTORS OF PHYSICAL ACTIVITY LEVELS FOR WOMEN SUBJECTS SEASON OF GROUP SURVEY MEAN (betarO.ll) REGION GROUP MEAN (betarO.ll) EDUCATION GROUP MEAN (beta:0.06) Fall 1978 1.83 Winter 1978 2.06 Spring 1979 2.01 Atlantic 2.13 Quebec 2.00 Ontario 1.94 Prairies 1.93 B.C. 1.85 Secondary 2.01 Post Sec. 2.01 Diploma 1.89 Degree 1.93 grand mean 1.99 - percentage variance explained 3.1 Among a l l subgroups, those surveyed in July - September 1978 and those living in B.C., reported the most activity. Given the association between warmer weather and higher levels of activity, i t should not be surprising that B.C., with the mildest year-round climate is also associated with higher activity levels. People living in the Atlantic provinces and those living alone had the lowest average reported levels of activity. When the percentage of subjects reporting the lowest levels of activity were standardized on the basis of season and region against the homemaker's group, workforce women were shown to have the greatest number of sedentary subjects -102-TABLE 6.16 PHYSICAL ACTIVITY WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Scoring Less than 1750 on the Physical Activity Index SEASON OF SURVEY ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE FALL 32.0 (94) 29.0 (76) 23.3 (73) 18.3 (82) 24.1 (54) 25.6 (379) 25.6 WINTER 48.3 (149) 43.6 (165) 37.7 (122) 35.2 (165) 31.7 (79) 40.2 (680) 40.3 SPRING 43.1 (153) 39.1 (161) 30.4 (125) 37.8 (135) 23.0 (74) 36.7 (648) 36.6 CRUDE RATE 43.2 (396) 39.1 (402) 31.6 (320) 32.5 (382) 26.6 (207) 35.6 (1707) ADJUSTED RATE 43.5 38.7 31.7 32.4 26.7 35.6 -103-TABLE 6.17 PHYSICAL ACTIVITY WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Scoring Less than 1750 on the Physical Activity Index* SEASON OF SURVEY ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE FALL 48.7 (78) 33.3 (69) 30.6 (85) 26.6 (79) 17.8 (45) 32.6 (356) 33.0 WINTER 51.6 (91) 44.4 (126) 40.0 (125) 37.7 (HA) 38.8 (80) 42.4 (536) 43.1 SPRING 46.9 (98) 38.3 (115) 40.2 (112) 40.2 (117) 38.5 (65) 38.6 (507) 41.1 CRUDE RATE 49.1 (267) 39.7 (310) 37.6 (322) 35.8 (310) 33.7 (190) 39.3 (1399) ADJUSTED RATE 49.2 39.6 38.0 36.2 34.0 40.1 Rates standardized against the "Women Homemakers" group. -1 OA-TABLE 6.18 PHYSICAL ACTIVITY MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Scoring Less than 1750 on the Physical Activity Index SEASON OF SURVEY ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE FALL 41.9 (136) 38.7 (142) 31.9 (138) 33.6 (146) 22.1 (86) 34.6 (648) 35.0 WINTER 41.6 (185) 49.6 (270) 29.0 (221) 34.1 (223) 28.8 (146) 37.6 (1045) 37.9 SPRING 50.5 (192) 29.8 (235) 32.8 (186) 34.7 (239) 34.8 (112) 36.3 (964) 38.1 CRUDE RATE 45.0 (513) 40.0 (647) 31.0 (545) 34.2 (608) 29.1 (344) 36.4 (2657) ADJUSTED RATE 45.1 43.5 30.4 34.0 27.4 37.4 Rates standardized against the "Women Homemakers" group. (40.1%, p<.01), compared with the homemaker group (35.6%) and the workforce men (37.4%). (See Table 6.16-18) These results should probably not be considered surprising when we know that women tend to work in service and clerical type jobs (Phillips and Phillips, 1983) and that for workforce women with families there i s very l i t t l e time for leisure activities (Meissner, 1975). The Canada Fitness Survey (Stephens, Craig and Ferris, 1986) found that only 25% of the 20 - 60 year-old population were active enough to provide -105-cardiovascular health benefit. This i s lower than the 32-33% found in the 25 - 44 year-olds included in this study, but age alone could account for the difference. The Fitness Survey found that activity was related to the type of occupation (a question not addressed in this study) and that while professionals and managers had the highest proportion of adequate activity, blue collar workers had the lowest. This would support the results of this analysis which showed that education was positively correlated with activity. The Fitness Survey also identified an increasing mean level of energy expenditure as the study moved from East to West. One U.S. study (U.S. Department of Health and Human Services, 1985) reported physical activity to be positively associated with being male and having higher socioeconomic status but associated inversely with age. More men (25.2%) than women (17.0%) failed to complete the section on physical activity which summed reported domestic, work-related and leisure time activity to create the Physical Activity Index. The Health of Canadians (Health and Welfare/Statistics Canada, 1982), reporting on the C.H.S. reported an overall 86% response rate and that the proportion of responses decreased with age. It i s unclear how we should interpret the fact that a greater percentage of males failed to answer sufficient questions to generate a score. -106-C0MPANI0NSHIP Multiple Classification Analysis showed that the season of the survey, the size of the family and marital status were the best indicators, in this model, of the amount of leisure time with company subjects would report. However, the model was able to explain only 3.2% of the variation in scores (Table 6.19). The subgroups with the highest mean scores, indicating that they reported spending most or a l l of their leisure time alone, were those not married, living alone and surveyed between October and December 1978. The groups with the lowest mean scores, indicating that they spent most or a l l of their leisure time with others, were from large families, surveyed in the late summer of 1978 or January to March 1979 and those living in British Columbia. Table 6.19 shows the sub group means for each category of the top three predictors. TABLE 6.19 PREDICTORS OF COMPANIONSHIP DURING LEISURE HOURS FOR WOMEN SUBJECTS SEASON OF GROUP SURVEY MEAN (beta:0.10) FAMILY GROUP SIZE MEAN (beta:0.08) MARITAL GROUP STATUS MEAN (beta:0.08) July to Sept. 1978 1.44 Oct. to Dec. 1978 2.06 Jan. to Mar. 1978 2.01 One 1.66 Two or Three 1.58 Four to Six 1.50 Seven or More 1.41 Not Married 1.65 Married 1.51 grand mean 1.53 - percent variance exp! ained 3.2 Rates for subjects reporting that they spent a l l or most of their leisure time alone were standardized on 'season of survey' and 'family size' using the homemaker populations as the standard (see Tables 6.20-22). A o significantly greater proportion of women homemakers reported spending leisure time alone (12.2%) than was the case for men (7.3%, p<.002) and workforce women (8.1%, p<.01). -107-Much has been written about the relative isolation of the housewife in our society; about the higher levels of psychological distress experienced by housewives (Gore and Mangione, 1983) and the positive effects of marriage (Morgan, 1980). Nathanson (1980) stated that "employment for women, much like marriage for men, has been found to be socially integrating and consequently protective of health". The results of this study would seem to support at least the notion of employment being associated with greater levels of social interaction. The discussion on demographic variables (see Tables 6.82 and 6.84) at the end of this chapter further emphasize the apparent protective effect of marriage and family. More men (18.3%) than women (8.9%) failed to answer the question about companionship and more workforce women (10.4%) than homemakers (7.7%) were non-respondents. The direct relationship between non-response rates and higher rates of social interaction should be noted. In theory, i f a l l the non-respondents were socially isolated, the results would look very different. However, there seems no a priori reason to assume that the non-respondents might come from one group than the other. -108-TABLE 6.20 COMPANIONSHIP IN LEISURE TIME WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reported Spending a l l or most of their Leisure Time Al SEASON OF SURVEY FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE FALL 0 (0) 10.1 (89) 6.4 (299) 22.6 (31) 7.6 (419) 8.2 WINTER 50.0 (4) 22.2 (162) 14.9 (544) 15.9 (44) 16.7 (754) 16.6 SPRING 0 (0) 14.6 (171) 9.0 (511) 6.4 (47) 10.2 (729) 10.0 CRUDE RATE 50.0 (4) 16.6 (422) 10.8 (1354) 11.5 (122) 12.2 (1902) ADJUSTED RATE 19.8 16.6 10.7 13.7 12.2 -109-TABLE 6.21 COMPANIONSHIP IN LEISURE TIME WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reported Spending a l l or most of their Leisure Time Alone SEASON OF SURVEY FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE FALL 20.8 (53) 9.6 (146) 5.9 (169) 0 (17) 9.1 (385) 6.4 WINTER 15.2 (66) 12.6 (254) 9.5 (253) 7.7 (13) 11.4 (586) 10.1 SPRING 12.9 (62) 8.3 (252) 6.7 (223) 6.7 (15) 8.2 (552) 7.1 CRUDE RATE 16.0 (181) 10.3 (652) 7.6 (645) 4.4 (45) 9.7 (1523) ADJUSTED RATE 15.5 10.3 7.6 5.6 8.1 Rates standardized against the "Women Homemakers" group. i -110-TABLE 6.22 COMPANIONSHIP IN LEISURE TIME MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reported Spending most or a l l of their Leisure Time Alone SEASON OF SURVEY FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE FALL 22.8 (57) 6.3 (255) 6.1 (363) 0 (26) 7.3 (701 6.2 WINTER 25.9 (85) 10.0 (412) 6.6 (611) 10.3 (29) 9.3 (1137) 7.6 SPRING 19.3 (88) 7.4 (393) 7.1 (560) 15.6 (32) 8.2 (1073) 7.8 CRUDE RATE 22.6 (230) 8.1 (1060) 6.7 (1534) 5.8 (87) 8.4 (2911) ADJUSTED RATE 22.7 8.2 6.7 11.7 7.3 Rates standardized against the "Women Homemakers" group. HORMONE PILLS In this model, family size, region of residence and economic responsibility were the best predictors of hormone p i l l taking. Table 6.23 shows the group means for each subgroup of these variables as well as the beta scores and percentage of variance explained. The higher the group mean score, the greater number of subjects engaging in the behaviour. The subgroups reporting the highest average p i l l consumption were those who were a family of one (group mean 1.74), those who described themselves as the principal breadwinner of the family (1.57) and women living in Quebec (1.45). Those reporting the lowest hormone p i l l consumption rates were those from large families (group mean 1.18) and those living in the Atlantic Provinces (1.28). The inverse relationship between family size and hormone p i l l taking -111-in this age group should be no surprise. The higher consumption rates in Quebec, a predominantly Roman Catholic population (88%, Statistics Canada, 1981), leaves much room for musing. Eight and a half percent (8 1/2%) of a l l female subjects did not answer the question on hormone p i l l s , there being slightly more non-respondents among the workforce women. TABLE 6.23 PREDICTORS OF HORMONE PILL CONSUMPTION BY WOMEN SUBJECTS FAMILY SIZE GROUP MEAN (beta:0.10) • REGION GROUP MEAN (beta:0.08) ECONOMIC GROUP RESPONSIBILITY MEAN (beta:0.07) One 1.74 Two or three 1.43 Four to six 1.27 Seven or more 1.18 Atlantic 1.28 Quebec 1.45 Ontario 1.32 Prairies 1.35 B.C. 1.32 Principal Breadwinner 1.57 Not principal Breadwinner 1.29 grand mean 1.34 - percentage variance explained 4.2 When the workforce women's rates were standardized on the basis of family size and region against the homemaker population, a significantly higher proportion of workforce women (16.8%, p<.002) reported taking hormone p i l l s . (See Tables 6.24-25) This difference can perhaps be in part explained by the fact that the homemaker group necessarily includes women who have temporarily le f t the workforce in order to have a family. Information required to identify such factors i s not readily available in a cross-sectional study. -112-TABLE 6.24 USE OF HORMONE PILLS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Taking Hormones for Birth Control or other Reasons FAMILY SIZE ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE FAMILY OF 1 0 (1) 0 (2) 0 (1) 0 (1) 0 (0) 0 (5) 0 FAMILY OF 2-3 14.6 (89) 22.8 (92) 12.2 (98) 17.6 (91) 12.3 (65) 16.1 (435) 16.4 FAMILY OF 4-6 8.5 (340) 18.0 (316) 12.3 (244) 15.3 (301) 13.7 (146) 12.0 (1347) 14.4 FAMILY OF 7+ 5.2 (58) 12.5 (16) 8.7 (23) 5.6 (18) 0 (5) 6.7 (120) 7.0 CRUDE RATE 9.2 (488) 14.1 (426) 12.0 (366) 15.3 (411) 13.0 (216) 13.6 (1907) ADJUSTED RATE 9.7 18.7 12.0 15.2 12.5 13.6 -113-TABLE 6.25 USE OF HORMONE PILLS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Taking Hormones for Birth Control or other Reasons FAMILY SIZE ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE FAMILY OF 1 4.0 (21) 44.4 (45) 39.0 (41) 40.4 (47) 32.0 (25) 38.0 (179) 30.8 FAMILY OF 2-3 25.2 (119) 25.0 (148) 23.6 (140) 25.4 (138) 22.6 (102) 24.4 (647) 24.6 FAMILY OF 4-6 19.1 (152) 22.5 (120) 9.4 (159) 7.6 (145) 6.1 (87) 13.4 (663) 14.0 FAMILY OF 7+ 11.5 (26) 22.2 (9) 33.3 (3) 20.0 (5) 0 (2) 15.6 (45) 18.6 CRUDE RATE 21.1 (318) 26.7 (322) 19.0 (343) 19.7 (335) 17.6 (216) 21.0 (1534) ADJUSTED RATE 20.0 23.1 14.3 12.5 9.5 16.8 Rates standardized against the "Women Homemakers" group. LIFE CHANGE EVENTS Family size, income and economic responsibility were shown to be the best predictors of l i f e change events. A score of '3' was assigned i f subjects reported more than one l i f e change event, a score of '2' i f they reported one l i f e change event and a score of '1' i f they reported no l i f e change events. The subgroups reporting the highest mean scores and therefore the highest average number of events were people in a family of one (mean score 1.39), those who were the principal income earners for their family (1.34) and those -114-in the lowest income families (1.31). The subgroups reporting the lowest average number of events were people from families of four or more (group means 1.20 to 1.22) and women living in Quebec (1.23). Table 6.26 shows the subgroup mean scores for each of the top three predictors, their beta scores and the percentage of variance in the scores explained by the model. TABLE 6.26 PREDICTORS OF LIFE CHANGE EVENTS AMONG WOMEN SUBJECTS FAMILY SIZE GROUP MEAN (beta:0.12) INCOME GROUP QUINTILE MEAN (betarO.lO) ECONOMIC GROUP RESPONSIBILITY MEAN (beta:0.07) One 1.39 Two to three 1.36 Four to six 1.22 Seven or more 1.20 Lowest 1.31 Second 1.26 Third 1.26 Fourth 1.24 Highest 1.29 Principal Breadwinner 1.34 Not principal Breadwinner 1.25 group mean ] ..27 - percentage variance explained 3.6 Again, men had the highest percentage of non-respondents (19.6), then workforce women (11.3) and with the lowest percentage, homemakers (8.1). When the workforce men's and women's populations were standardized against the homemakers group on the basis of family size and income, there was a significantly higher percentage (29.0%, p<.002) of workforce women than homemakers (17.6%) reporting more than one l i f e event in the previous year. Men also reported a higher rate (23.3%, p<.05). Tables 6.27-29 show the crude and standardized rates for each of the groups. One might speculate that workforce women experience more l i f e changes because they are exposed to an additional sphere of interpersonal/employer related events. On the other hand this may be countered by the family centered social contacts experienced by the homemaker. It should be noted that 25% of the questions in the survey were 'exclusively' workforce related -115-TABLE 6.27 LIFE CHANGE EVENTS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting More than One Major Life Change Event INCOME QUINTILES FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE LOWEST QUINTILE 50.0 (4) 27.0 (89) 1.5 (267) 20.5 (44) 9.7 (404) 8.5 SECOND QUINTILE 100.0 (1) 30.8 (65) 19.5 (389) 19.4 (36) 21.2 (491) 22.3 THIRD QUINTILE 0 (0) 19.-7 (76) 18.2 (362) 10.0 (20) 18.1 (458) 18.0 FOURTH QUINTILE 0 (0) 33.3 (105) 14.7 (218) 0 (5) 20.4 (328) 18.0 HIGHEST QUINTILE 0 (0) 24.8 (83) 25.0 (96) 0 (4) 27.3 (183) 23.4 CRUDE RATE 60.0 (5) 28.7 (418) 15.2 (1332) 16.5 (109) 18.4 (1864) ADJUSTED RATE 37.3 27.1 15.0 12.0 17.6 -116-TABLE 6.28 LIFE CHANGE EVENTS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting More than One Major Life Change Event INCOME QUINTILES FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE LOWEST QUINTILE 33.3 (9) 58.8 (34) 36.4 (44) 14.3 (7) 42.6 (94) 40.1 SECOND QUINTILE 25.0 (8) 46.5 (43) 22.8 (79) 21.4 (14) 29.9 (144) 28.0 THIRD QUINTILE 91.7 (12) 38.7 (62) 24.8 (153) 27.3 (11) 31.9 (238) 28.3 FOURTH QUINTILE 41.0 (39) 30.4 (135) 15.1 (166) 33.3 (6) 24.3 (346) 19.6 HIGHEST QUINTILE 25.2 (111) 26.4 (329) 20.2 (124) 100.0 (1) 25.0 (565) 26.3 CRUDE RATE 33.5 (179) 31.8 (603) 21.6 (566) 25.6 (39) 27.7 (1387) ADJUSTED RATE 46.0 42.5 24.6 31.1 29.0 Rates standardized against the "Women Homemakers" group. -117-TABLE 6.29 LIFE CHANGE EVENTS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting More than One Major Life Change Event INCOME QUINTILES FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7 + CRUDE RATE ADJUSTED RATE LOWEST QUINTILE 100.0 (5) 43.2 (44) 30.9 (181) 25.0 (24) 33.9 (254) 33.5 SECOND QUINTILE 33.3 (3) 26.5 (48) 24.4 (390) 16.1 (31) 24.3 (522) 24.4 THIRD QUINTILE 66.7 (12) 20.8 (149) 20.6 (417) 18.2 (11) 21.6 (589) 20.6 FOURTH QUINTILE 53.1 (32) 29.8 (272) 13.5 (311) 12.5 (8) 22.6 (623) 17.2 HIGHEST QUINTILE 29.8 (181) 25.5 (448) 13.3 (173) 0 (4) 23.7 (805) 15.3 CRUDE RATE 36.5 (233) 26.8 (1011) 20.5 (1472) 18.0 (78) 24.1 (2794) ADJUSTED RATE 59.1 29.2 21.9 16.3 23.3 Rates standardized against the "Women Homemakers" group. while there were no equivalent events 'exclusive' to the homemaker population. Even the arrival of a baby in the household was presented in such a way as to imply equal impact on a l l members of the family. Goldberg and Comstock (1980) reporting on a two site study involving 2780 subjects point out that different groups in society are more likely to experience different types of events and that people at different phases of -118-the l i f e cycle (for example, early career and later career) are also prone to different types of l i f e events. Skinner and Lei (1980) suggest that there are categories of l i f e events such as social and personal, work or school related, domestic and family centered and that any questionnaire asking about l i f e events needs to reflect a l l of these. It should also be noted that some l i f e events seem more likely than others to be associated with a second event. For example, in the C.H.S. l i s t of possible l i f e events, 'getting married' and 'having someone move in with you' are specified. It seems likely that the f i r s t might be strongly associated with the second; two other events, 'finishing school' and 'getting a job' also are likely to be associated. These many facets of evaluating l i f e events do not appear to have been considered in the C.H.S.'s abbreviated l i s t of l i f e events and therefore caution i s advised in the evaluation of the summary rates. MOTOR VEHICLE TRAVEL The top three predictors of travelling distances (as a driver or passenger) reported by subjects were income, economic responsibility and region of residence. Table 6.30 shows the mean scores of each of the subgroups of these three variables, the beta scores and the percentage of variance in the scores explained by this model. TABLE 6.30 PREDICTING MOTOR VEHICLE TRAVEL BY WOMEN SUBJECTS INCOME GROUP QUINTILE MEAN (beta:0.18) ECONOMIC GROUP RESPONSIBILITY MEAN (beta:0.08) REGION GROUP MEAN (beta:0.07) Lowest 1.74 Second 1.86 Third 1.98 Fourth 2.03 Highest 2.17 Principal Breadwinner 2.13 Not principal Breadwinner 1.95 Atlantic 1.99 Quebec 1.89 Ontario 1.96 Prairies 2.07 B.C. 2.02 grand mean '. .98 - percentage variance explained 4.0 -119-The direct association between income and distances travelled should be noted. Overall, the subgroups reporting the highest travelling distances were the high income group (group mean 2.17), women from large families (2.13) and women who were the principal income earners for their families (2.13). The subgroups reporting the least motor vehicle travel were those in the lowest income group (1.74) and people making up a family of one (1.88). Given that a l l but this last group are likely to have children who require a certain amount of 'chauffering', difference between travelling distances of women from large versus small families, should not be surprising. In this study respondents had to provide both passenger and driver distances i f they were to be included. The number of missing responses in this section of the survey was very high; 32% of female subjects and 63% of male. In the report on the C.H.S. (Health and Welfare/Statistics Canada, 1982) i t was suggested that subjects may have failed to answer the questions because they had difficulty recalling the number of miles or kilometres they had driven or ridden in the previous two weeks. It could also be that people who did not travel as a passenger during that time, simply left the question unanswered and this could account for the markedly higher number of absent scores among the male subjects. When the workforce groups were standardized on the basis of family income and economic responsibility against the homemaker group (see Tables 6.31-6.33), significantly more men (40.5%, p<.02) reported travelling in excess of 11,000 kilometres in the previous year. Fewer homemakers (29.0%) than workforce women (33.7%) travelled in excess of 11,000 kilometres but this difference was not significant. -120-Using principal wage earner and income as the standardizing factors changes the men's rates much more than i t does the women's. This can be explained by the polarization of 'principal wage earners' (being mainly men) and 'not principal wage earners' (being primarily women) coupled with the inverse relationship between c e l l size and income for the homemaker group compared with the men's group. Because of the high number of missing scores and the lack of indication as to how these scores might have been distributed, these results must be interpreted with caution. TABLE 6.31 MOTOR VEHICLE TRAVEL AS A DRIVER OR PASSENGER WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Indicating that they Travelled More than 11,000 km by Motor Vehicle in the Previous Year LOWEST INCOME QUINT. SECOND INCOME QUINT. THIRD INCOME QUINT. FOURTH INCOME QUINT. HIGHEST INCOME QUINT. CRUDE RATE ADJUSTED RATE PRINCIPAL INCOME EARNER 20.5 (44) 60.0 (5) 66.7 (3) 0 (0) 25.0 (4) 26.8 (56) 39.2 NOT PRINCIPAL INCOME EARNER 20.4 (231) 24.8 (375) 31.1 (347) 33.8 (263) 38.4 (146) 28.9 (1362) 28.6 CRUDE RATE 20.4 (275) 25.3 (380) 31.4 (350) 33.8 (263) 38.0 (150) 28.8 (1418) ADJUSTED RATE 20.4 26.2 32.5 32.5 37.8 29.0 -121-TABLE 6.32 MOTOR VEHICLE TRAVEL AS A DRIVER OR PASSENGER WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Indicating that they Travelled More than 11,000 km by Motor Vehicle in the Previous Year^ LOWEST INCOME QUINT. SECOND INCOME QUINT. THIRD INCOME QUINT. FOURTH INCOME QUINT. HIGHEST INCOME QUINT. CRUDE RATE ADJUSTED RATE PRINCIPAL INCOME EARNER 25.7 (35) 41.5 (41) 40.0 . (50) 40.3 (77) 45.8 (153) 41.3 (356) 38.3 NOT PRINCIPAL INCOME EARNER 33.3 (15) 27.9 (68) 35.7 (126) 34.4 (183) 41.2 (296) 36.9 (688) 33.5 CRUDE RATE 28.0 (50) 33.0 (109) 36.9 (176) 36.2 (260) 42.8 (449) 38.4 (1044) ADJUSTED RATE 33.0 28.4 35.9 34.6 41.4 33.7 $ Rates standardized against the "Women Homemakers" group. -122-TABLE 6.33 MOTOR VEHICLE TRAVEL AS A DRIVER OR PASSENGER MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Indicating that they Travelled More than 11,000 km by Motor Vehicle in the Previous Year^ LOWEST INCOME QUINT. SECOND INCOME QUINT. THIRD INCOME QUINT. FOURTH INCOME QUINT. HIGHEST INCOME QUINT. CRUDE RATE ADJUSTED RATE PRINCIPAL INCOME EARNER 52.1 (117) 59.7 (201) 60.6 (231) 56.0 (282) 58.8 (323) 58.0 (1154) 57.7 NOT PRINCIPAL INCOME EARNER 0 (3) 46.2 (13) 42.9 (14) 57.1 (14) 59.1 (44) 52.9 (87) 39.8 CRUDE RATE 50.8 (120) 58.9 (214) 59.6 (245) 56.1 (296) 58.9 (367) 57.6 (1241) ADJUSTED RATE 2.0 46.7 43.6 57.1 59.1 40.5 £ Rates standardized against the "Women Homemakers" group. MOTOR VEHICLE SEAT BELT USE Region of residence, education and marital status were the top three predictors of seat belt use. In the entire model, this was the one dependent variable for which a significant degree of the variance in scores could be explained (30.2%) and i t was due to the strong association between region and legislation for seat belt use. In 1978-79 the Atlantic provinces, Manitoba and Alberta had no seat belt legislation. (Alberta, the last province to do so, introduced seat belt legislation in February, 1987.) Residents of the Prairies and the Atlantic provinces therefore reported the lowest levels of seat belt use. British Columbia reported the highest use of seat belts. Table 6.34 shows the subgroup mean scores for each of the top three predictor -123-variables. It can be seen from the beta scores that region is a strong predictor when compared with the other two. TABLE 6.34 PREDICTORS OF MOTOR VEHICLE SEAT BELT USE BY WOMEN SUBJECTS REGION GROUP EDUCATION GROUP MARITAL GROUP MEAN MEAN STATUS MEAN (beta:0.52) (beta:0.14) (beta:0.06) Atlantic 2.63 Secondary 1.99 Not Quebec 1.53 Post Sec. 1.98 married 1.96 Ontario 1.46 Diploma 1.79 Prairies 2.33 Degree 1.64 Married 1.90 B.C. 1.42 grand mean 1.9 . - percentage variance explained 30.2 An explanation for the high number of missing scores, 52% for women and almost 70% for men, offered in the report on the C.H.S. (Health and Welfare/Statistics Canada, 1982) was that "people may have been reluctant to report failure to wear seat belts in areas of the country where the wearing of seat belts is required by law." Such an explanation would suggest that non-respondents were non-wearers. However, for the homemakers group, the percentage of missing scores gradually decreases as the region of residence moves from east to west and for the two workforce groups the rates for missing scores are similar for a l l regions except B.C., which in a l l instances has the lowest rate of non-responses. There is therefore no pattern of missing scores being associated with regions with legislation. One might s t i l l argue however, that in a country where seat belt use has such wide acceptance as a desirable behaviour, non-users may be more inclined than users to avoid declaring their behaviour. When the rates for each of the workforce groups were standardized on the basis of region and education against the homemakers group they each showed -124-significantly higher proportions of non-users. While 41.1% of homemakers reported not using their seat belts, 48.6% (p>.002) of workforce women and 47.1% (p>.01) of workforce men reported not using vehicular seat belts. (See Tables 6.35-37) The strong association between education and seat belt use can also be scan in each of the tables. Other researchers (Christian, 1984 and Dreghorn, 1985) have reported seat belt wearing rates of approximately 30% when no legislation was in effect and rates as high as 86-88% once i t became law. In this study, the Atlantic provinces were the only region totally unaffected by legislation (in the Prairie region, Saskatchewan had legislation) and i t s reported non-iise rate was higher than the rates quoted in these other studies. The non-use rate reported in the provinces with legislation was also slightly lower than the literature would lead the reader to expect. This would be expected i f non-respondents were indeed non-users. -125-TABLE 6.35 USE OF MOTOR VEHICLE SEAT BELTS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting NO USE of a Seat Belt as a Passenger or Driver ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 88.6 (131) 22.4 (143) 13.0 (123) 65.0 (157) 18.2 (88) 45.5 (642) 44.2 SOME POST-SECONDARY 87.5 (16) 10.0 (10) 27.3 (11) 59.1 (22) 16.7 (18) 44.2 (77) 42.4 CERTIFI-CATE OR DIPLOMA 71.0 (31) 20.6 (34) 13.2 (38) 45.8 (59) 13.0 (23) 34.6 (185) 34.6 UNIVERSITY DEGREE 33.3 (15) 15.4 (13) 12.5 (16) 57.9 (19) 9.1 (11) 28.4 (74) 28.6 CRUDE RATE 81.4 (193) 21.0 (200) 19.2 (188) 61.9 (257) 16.4 (140) 41.0 (978) ADJUSTED RATE 81.0 20.5 14.1 60.3 16.4 41.1 -126-TABLE 6.36 USE OF MOTOR VEHICLE SEAT BELTS WOMEN IN THE PAID LABOUR FORCE AGED 25-44 YEARS Percentage of Women Reporting NO USE of a Seat Belt as a Passenger or Driver ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 87.0 (77) 34.4 (96) 23.7 (97) 76.7 (86) 28.8 (66) 49.3 (422) 53.0 SOME POST-SECONDARY 72.2 (18) 0 (5) 29.4 (17) 64.0 (25) 20.0 (15) 46.3 (80) 39.6 CERTIFI-CATE OR DIPLOMA 72.1 (43) 23.1 (39) 3.2 (31) 77.3 (22) 10.5 (19) 39.0 (154) 41.4 UNIVERSITY DEGREE 48.3 (29) 20.8 (24) 18.0 (39) 50.0 (30) 3.5 (29) 27.8 (151) 30.9 CRUDE RATE 74.9 (167) 28.7 (164) 19.6 (184) 70.0 (163) 19.4 (129) 43.0 (807) ADJUSTED RATE 80.1 28.5 19.9 73.8 22.7 48.1 Rates standardized against the "Women Homemakers" group. -127-TABLE 6.37 USE OF MOTOR VEHICLE SEAT BELTS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting NO USE of a Seat Belt as a Passenger or Driver ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 90.1 (151) 34.2 (120) 35.6 (104) 63.6 (143) 35.9 (78) 55.9 (596) 53.4 SOME POST-SECONDARY 84.6 (13) 30.0 (20) 14.3 (21) 57.9 (19) 21.7 (23) 37.5 (96) 43.9 CERTIFI-CATE OR DIPLOMA 57.8 (19) 21.4 (28) 6.7 (30) 68.0 (25) 15.4 (26) 31.3 (128) 37.1 UNIVERSITY DEGREE 46.2 (39) 22.2 (54) 12.3 (65) 39.7 (58) 6.8 (44) 24.6 (260) 27.4 CRUDE RATE 79.3 (222) 29.3 (222) 22.5 (220) 58.0 (245) 23.4 (171) 43.8 (1080) ADJUSTED RATE 80.2 30.5 26.7 62.2 28.7 47.6 Rates standardized against the "Women Homemakers" group. -128-FEMALE PREVENTIVE BEHAVIOURS Female preventive scores were developed from reported rates of breast self-examination, professional breast examination and pap smears. A score of '3' indicated that a l l three tests had been done within an appropriate period of time while at the other end of the spectrum, a score of '9' indicated that the woman had not been exposed to any of these screening measures. Women were not included in this section of the study unless they had answered a l l three questions. The rate for non-responses was in fact quite low (8.9%). As can be seen from Table 4.17, many women do avail themselves of these screening tests. Overall, the best three predictors of variance were region, marital status and education. The overall mean score for female preventive behaviours was 4.43. The subgroups reporting the highest mean scores and therefore the least screening were single women (group mean 4.89), women in Quebec (4.86) and women living in large families (4.71). The subgroups reporting the best scores were women living in B.C. (4.08) and Alberta (4.10) and women with post-secondary education (4.06). Table 6.38 shows the subgroup means for each of the top three predictors, their beta scores and the percentage of variance in the scores explained by the model. TABLE 6.38 PREDICTORS OF FEMALE PREVENTIVE BEHAVIOURS AMONG WOMEN SUBJECTS REGION GROUP MARITAL GROUP EDUCATION GROUP MEAN STATUS MEAN MEAN (beta:0.18) (beta:0.13) (beta:0.11) Atlantic 4.61 Not Secondary 4.57 Quebec 4.86 married 4.89 Post Sec. 4.06 Ontario 4.41 Diploma 4.12 Prairies 4.06 Married 4.35 Degree 4.22 B.C. 4.01 grand mean 4.43 - percentage variance explained 7 0 -129-Income, while not one of the top three predictors, exhibited a direct relationship with preventive behaviours. The higher the income group the better the group mean score. Of the three components of the preventive score, pap smears were the most often reported to have been done within two years (73%); professional breast examinations were done within two years for 66.7% of the women. Regular breast self-examination was reported by only 47% of the women while 25% reported that they never performed breast self-examination. When the workforce womens group was standardized against the homemaker group on the basis of region and education (see Tables 6.39-40) , the rates for never or almost never availing themselves of these screening measures (that i s scores of 7-9) were significantly higher (p<.002) for homemakers (14.3%) than for workforce women (9.8%). Part of this difference between the scores could be further reduced i f the groups were standardized also for education because education was positively associated with use of screening measures and the levels of education were slightly higher for the workforce group. However, standardizing on three variables was found to produce too many very small c e l l sizes. In other studies (Ellerton and Smillie, 1986) the reported level of breast self-examination was lower (17-18%) than in this study but similar predictors, namely marital status and education were identified. Ellerton and Smillie also found a slightly higher rate for breast self-examination among workforce women than among housewives. A telephone survey of Edraontonian women (Kurji and MacDonald, 1986) found that 73% of women -130-reported having a professional breast examination in the previous year and 66% reported a pap smear during that time. The rates for this study which recorded behaviours during the previous two years were 67% and 73% respectively. In the Edmonton study, 37% reported regular breast self-examination and 10% reported never having had a pap smear In this study 47% reported performing breast self-examinations and 6.5% never having had a pap smear. TABLE 6.39 FEMALE PREVENTIVE BEHAVIOURS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Rarely or Never Having a Pap Smear and Breast Examinations ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE NOT MARRIED 33.3 (42) 44.0 (25) 14.3 (21) 21.4 (14) 16.7 (12) 29.0 (114) 27.6 MARRIED 17.6 (443) 20.4 (397) 13.5 (334) 6.3 (397) 4.7 (212) 13.4 (1783) 13.5 CRUDE RATE 19.0 (485) 21.8 (422) 13.5 (355) 6.8 (411) 5.4 (224) 14.3 (1897) ADJUSTED RATE 18.6 21.8 13.5 7.2 5.4 14.3 -131-TABLE 6.40 FEMALE PREVENTIVE BEHAVIOURS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Rarely or Never Having a Pap Smear and Breast Examinations ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. CRUDE RATE ADJUSTED RATE NOT MARRIED 23.3 (73) 25.3 (91) 11.8 (93) 14.3 (91) 9.3 (54) 17.2 (402) 18.0 MARRIED 11.6 (241) 12.7 (228) 9.8 (255) 5.8 (241) 3.4 (148) 9.1 (1113) 9.3 CRUDE RATE 14.3 (314) 16.3 (319) 10.3 (348) 8.1 (332) 5.0 (202) 11.2 (1515) ADJUSTED RATE 12.3 13.5 9.9 6.3 3.7 9.8 Rates standardized against the "Women Homemakers" group. COMPOSITE RISK SCORE Each subject's composite r i s k score was generated by combining th e i r scores f or: i . cigarette smoking i i . alcohol consumption i i i . physical a c t i v i t y i v . companionship v. l i f e change events v i . motor vehicle t r a v e l v i i . hormone p i l l s (women only). Each of these variables had been scored '1' through '3' (hormone p i l l s was dichotomized, '1' for "No" and '3' for "Yes") and when the scores were summed, for women there was a possible range of 7 (the least r i s k score) to 21 (the maximum r i s k score), and for men, 6 to 18. (See Chapter Four for -132-raore details on the development of composite risk scores.) No composite risk was generated for subjects who had missing scores on any one of these variables. For women this excluded 60% of subjects, and for men almost 74%. Among the female subjects, the top three predictors of composite risk scores were family size, education and marital status. The model overall was able to explain 7.7% of the variation in the women's scores. Table 6.41 shows the subgroup means for each of the top three predictors. TABLE 6.41 PREDICTORS OF COMPOSITE RISK SCORES FOR WOMEN SUBJECTS FAMILY SIZE GROUP MEAN (beta:0.15) EDUCATION GROUP MEAN (beta:0.11) MARITAL GROUP STATUS MEAN (beta:0.07) One person 13.6 2 - 3 people 12.63 4 - 6 people 11.98 7 or more 11.58 Secondary 12.28 Post Sec. 12.90 Diploma 12.16 Degree 12.25 Not Married 13.21 Married 12.15 grand mean 12.30 - percentage variance explained'7.7 The two workforce groups were standardized against the homemaker group on the basis of family size and education and the rates calculated for both the highest and lowest risk groups. (See Tables 6.42-47.) In each instance, the rates for the homemaker group were better than for the workforce groups. Ten and a half percent (10.5%) of homemakers compared with 11.8% (n.s.) of workforce women and 30.5% (p>.002) of workforce men reported composite risk scores in the highest category. More homemakers (23%) than workforce women (15.9%), p>.002) or workforce men (8.6%, p>..002) reported low (good) composite risk scores. Education was inversely related to health risk exposure in workforce men with high risk scores (Table 6.44) and positively related for workforce -133-women with low r i s k scores (Table 6.46). For the other groups, the association was more mixed. In a l l three groups, the number of missing scores was p o s i t i v e l y related to education. There was nearly a 20% difference i n missing score rates between the lowest education group and the highest and workforce men had the highest rate of non-respondents o v e r a l l . Therefore caution may be needed i n interpreting the adjusted rates i f i t were not for the fact that on 4 out of 6 indi v i d u a l items, the lowest education group was shown to have s i g n i f i c a n t l y higher rates of r i s k related behaviour. TABLE 6.42 POOR COMPOSITE RISK SCORES WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women with the Highest Risk Scores (15-21) in the Summary Measure of Risk Behaviours FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 0 (1) 17.2 (116) 10.2 (391) 0 (34) 11.1 (542) 11.4 SOME POST-SECONDARY 0 (0) 26.7 (15) 18.0 (39) 0 (3) 19.3 (57) 19.1 CERTIFICATE OR DIPLOMA 0 (0) 6.8 (44) 4.5 (89) 0 (4) 5.1 (137) 4.8 UNIVERSITY DEGREE 0 (1) 11.8 (17) 6.8 (44) 0 (1) 7.9 (63) 7.6 CRUDE RATE 0 (2) 15.1 (192) 9.6 (563) 0 (42) 10.4 (799) ADJUSTED RATE 0 15.7 9.5 0 10.5 -134-TABLE 6.43 POOR COMPOSITE RISK SCORES WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women with the Highest Risk Scores (15-21) in the Summary Measure of Risk Behaviours^ FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 42.3 (26) 18.2 (165) 10.1 (159) 16.7 (12) 16.3 (362) 12.5 SOME POST-SECONDARY 57.1 (14) 27.5 (40) 14.3 (21) 0 (0) 29.3 (75) 16.9 CERTIFICATE OR DIPLOMA 24.0 (25) 14.8 (61) 4.2 (48) 0 (1) 12.6 (135) 6.6 UNIVERSITY DEGREE 27.6 (29) 15.9 (88) 12.5 (24) 0 (1) 17.6 (142) 12.7 CRUDE RATE 35.1 (94) 18.1 (354) 9.5 (252) 14.3 (14) 17.2 (714) ADJUSTED RATE 39.0 18.1 9.5 11.3 11.8 ^Rates standardized against the "Women Homemakers" group. It i s interesting to note that i t was only the rates for men with good scores that were negatively associated with family size. In a l l other instances there was some level of protection in living in a big family. -135-TABLE 6.44 POOR COMPOSITE RISK SCORES MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men with the Highest Risk Scores (13-18) in the Summary Measure of Risk Behaviours^ FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 31.3 (32) 38.9 (180) 31.2 (285) 47.41 (19) 34.5 (516) 34.0 SOME. POST-SECONDARY 55.6 (9) 46.0 (37) 30.0 (30) 0 (1) 40.3 (77) 32.3 CERTIFICATE OR DIPLOMA 42.9 (14) 23.9 (46) 21.3 (47) 0 (0) 25.2 (107) 20.9 UNIVERSITY DEGREE 45.5 (33) 20.2 (109) 21.1 (90) 0 (2) 23.9 (234) 19.9 CRUDE RATE 40.9 (88) 32.3 (372) 28.1 (452) 40.9 (22) 31.3 (934) ADJUSTED RATE 36.1 35.3 28.6 32.1 30.5 .Rates standardized against the "Women Homemakers" group. -136-TABLE 6.45 GOOD COMPOSITE RISK SCORES WOMEN HOMEMAKERS AGED 25-44 YEARS Percentage of Women with the Lowest Risk Scores (7-10) in the Summary Measure of Risk Behaviours FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 100.0 (1) 13.8 (116) 25.1 (391) 29.4 (34) 23.1 (542) 22.8 SOME POST-SECONDARY 0 (0) 13.3 (15) 18.0 (39) 33.3 (3) 17.5 (57) 17.6 CERTIFICATE OR DIPLOMA 0 (0) 11.4 (44) 25.8 (89) 75.0 (4) 22.6 (137) 24.9 UNIVERSITY DEGREE 0 (1) 23.5 (17) 27.3 (44) 0 (1) 25.4 (63) 24.9 CRUDE RATE 50.0 (2) 14.1 (192) 24.9 (563) 33.3 (42) 22.8 (799) ADJUSTED RATE 67.8 • 14.1 24.8 35.1 23.0 -137-TABLE 6.46 GOOD COMPOSITE RISK SCORES WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women with the Lowest Risk Scores (7-10) in the Summary Meas of Risk Behaviours,,. FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 7.7 (26) 9.7 (165) 14.5 (159) 33.3 (12) 12.4 (362) 14.3 SOME POST-SECONDARY 0 (14) 7.5 (40) 19.1 (21) 0 (0) 9.3 (75) 15.2 CERTIFICATE OR DIPLOMA 12.0 (25) 11.5 (61) 16.7 (48) 100.0 (1) 14.1 (135) 19.8 UNIVERSITY DEGREE 10.3 (29) 17.1 (88) 25.0 (24) 0 (1) 16.9 (142) 21.7 CRUDE RATE 8.5 (94) 11.6 (354) 16.3 (252) 35.7 (14) 13.3 (714) ADJUSTED RATE 8.1 10.4 16.0 39.7 15.9 Rates standardized against the "Women Homemakers" group. -138-TABLE 6.47 GOOD COMPOSITE RISK SCORES MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men with the Lowest Risk Scores (6-9) in the Summary Measure of Risk Behaviours^. FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE SOME SECONDARY EDUCATION 15.6 (32) 7.2 (180) 7.0 (285) 0 (19) 7.4 (516) 6.7 SOME POST-SECONDARY 0 (9) 10.8 (37) 13.3 (30) 100.0 (1) 11.7 (77) 17.3 CERTIFICATE OR DIPLOMA 0 (14) 10.9 (46) 10.6 (47) 0 (0) 9.4 (107) 10.1 UNIVERSITY DEGREE 3.0 (33) 16.5 (109) 13.3 (90) 0 (2) 13.3 (234) 13.4 CRUDE RATE 6.8 (88) 10.8 (372) 9.1 (452) 4.6 (22) 9.4 (934) ADJUSTED RATE 10.8 8.8 8.6 7.1 8.6 Rates standardized against the "Women Homemakers" group. 6.2 HEALTH STATUS INDICATORS CHRONIC HEALTH PROBLEMS Almost 50% of the study group reported at least one chronic health problem. Table 4.23 shows the distribution of chronic health problems across the study group. - 1 3 9 -M u l t i p l e c l a s s i f i c a t i o n a n a l y s i s of the female data showed the composite r i s k score , fami ly s i z e and reg ion to be the best p r e d i c t o r s of problems. Table 6.48 shows the subgroup means fo r each of these top three p r e d i c t o r s , t h e i r beta scores and the percentage of var iance i n the scores exp la ined by the model. TABLE 6.48 PREDICTORS OF CHRONIC HEALTH PROBLEMS AMONG WOMEN SUBJECTS COMPOSITE GROUP RISK SCORE MEAN (be ta :0 .12) FAMILY GROUP SIZE MEAN (be ta :0 .11) REGION GROUP MEAN (be ta :0 .10) 7-10 (good) 1.07 11-14 1.05 15-20 (poor) 1.03 One person 1.31 2 - 3 people 1.13 4 - 6 people 0.97 7 or more 0.85 A t l a n t i c 1.05 Quebec 0.88 Ontar io 1.09 P r a i r i e s 1.06 B.C. 1.21 grand mean 1.05 - percentage var iance exp la ined 4.1 Across a l l subgroups, those e x h i b i t i n g the h ighest average number of chron ic hea l th problems were from a fami ly of one (group mean 1.31), not marr ied (1.23) and l i v i n g in B r i t i s h Columbia (1 .21 ) . The groups wi th lowest average number of ch ron ic problems were sub jec ts from la rge f a m i l i e s ( 0 .85 ) , those wi th a u n i v e r s i t y degree (0.88) and women l i v i n g i n Quebec (0 .88 ) . When the workforce groups were s tandard ized aga ins t the homemaker group on the bas is of composite r i s k score and fami ly s i z e , workforce men were shown to have the s i g n i f i c a n t l y lower ra te of chron ic hea l th problems (44.8%, p<.002 for both comparisons) than e i t h e r of the women's groups fo r whom the ra tes were very s i m i l a r (57.8% and 57.2%). (See Tables 6 .49-51 . ) A negat ive a s s o c i a t i o n between hea l th r i s k scores and chron ic hea l th problems among women can be seen i n Table 6 .49. Such an a s s o c i a t i o n i s hard to exp la i n un less one be l i eves that people more l i k e l y to become hea l th consc ious ( that i s , reduce t h e i r r i s k exposure) a f t e r they have exper ienced some f a i l u r e i n - 140 -t h e i r h e a l t h . Among the workforce women (Table 6.50) changes i n hea l th r i s k scores appear to have l i t t l e e f f e c t on the number of chron ic hea l th problems repo r ted . For men (Table 6 .51 ) , there i s at l e a s t a lower ra te of chron ic hea l th problems among those repo r t i ng the l e a s t r i s k exposure. Th is might lead one to ask the q u e s t i o " , "Are the hea l th r i s k f a c t o r s used i n t h i s study r e f l e c t i v e of the types of hea l th r i s k s to which men are most commonly exposed and are there some others that would more accu ra te l y address the hea l th r i s k s faced by women?" One other study which inc luded chron ic hea l th problems (WHO, 1973) repor ted much lower ra tes fo r problems (13-18%). Th is probably r e f l e c t s only a d i f f e rence i n the d e f i n i t i o n s used. As can be seen from the l i s t i n Table 4 .23 , cond i t i ons such as sk i n a l l e r g i e s and hayfever have been inc luded i n the C . H . S . quest ion regard ing chron ic cond i t i ons whereas the World Heal th Organ iza t ion study d id not de f ine " long s tanding hea l th problems or chron ic i l l n e s s " or o f f e r any examples of what might have been expected. The context of the quest ion could have led respondents to cons ider only f u n c t i o n a l l y l i m i t i n g c o n d i t i o n s . The h igher ra te of chron ic problems among women sub jec ts l i v i n g i n B r i t i s h Columbia i s of i n t e r e s t (Table 6 .48 ) . Older people (wi th and without hea l th problems) are known to r e t i r e to the West Coast because the c l ima te i s e a s i e r but t h i s does not exp la i n a high ra te of chron ic problems among women aged 25 - 44. An examinat ion of other C . H . S . data showing the types of hea l th problems reported by these women may guide f u r t he r specu la t i on but cannot address the i s sue of c a u s a l i t y . -141-Also of i n t e r e s t i s the a s s o c i a t i o n between d o l l a r s per person spent on phys i c i an s e r v i c e s i n the prov inces and the average number of hea l th problems repor ted i n each r e g i o n . B r i t i s h Columbia and Ontar io repor ted the h ighest average number of problems (see Table 6.48) and a l so have the h ighest l e v e l of spending per person on phys ic ian s e r v i c e s (Heal th and Welfare Canada, 1987). The A t l a n t i c prov inces and Quebec spend l e s s on phys i c i an s e r v i c e s and i n t h i s study the sub jec ts from these areas a l s o reported fewer hea l th problems. Th is again r a i s e s the i ssue of c a u s a l i t y which cannot be addressed i n t h i s s tudy. TABLE 6.49 CHRONIC HEALTH PROBLEMS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Repor t ing One or More Chronic Heal th Problems FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE COMPOSITE • RISK 7-10 (LO) 0 (0) 66.7 (27) 68.8 (141) 57.1 (14) 67.6 (182) 67.0 COMPOSITE RISK 11-14 0 (2) 62.5 (136) 54.6 (370) 50.0 (28) 56.2 (536) 55.8 COMPOSITE RISK 15-21 (HI) 0 (0) 58.6 (29) 51.9 (54) 47.6 (42) 52.0 (125) 52.9 CRUDE RATE 0 (2) 62.5 (192) 57.9 (565) 50.0 (84) 58.0 (843) ADJUSTED RATE 0 62.8 57.3 51.2 57.8 -142-TABLE 6.50 CHRONIC HEALTH PROBLEMS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting One or More Chronic Health Problems FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE COMPOSITE RISK 7-10 (LO) 87.5 (8) 61.0 (41) 53.7 (41) 60.0 (5) 60.0 (95) 56.0 COMPOSITE RISK 11-14 66.0 (53) 61.0 (251) 59.0 (188) 42.9 (7) 60.5 (499) 57.9 COMPOSITE RISK 15-21 (HI) 75.8 (33) 56.3 (64) 64.0 (25) 0 (2) 62.1 (124) 55.9 CRUDE RATE 71.3 (94) 60.1 (356) 58.7 (254) 42.9 (14) 60.7 (718) ADJUSTED RATE 72.1 60.3 58.6 40.2 57.2 Rates standardized against the "Women Homemakers" group. -143-TABLE 6.51 CHRONIC HEALTH PROBLEMS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting One or More Chronic Health Problems FAMILY OF 1 FAMILY OF 2-3 FAMILY OF 4-6 FAMILY OF 7+ CRUDE RATE ADJUSTED RATE COMPOSITE RISK 7-10 (LO) 50.0 (6) 50.0 (40) 36.6 (41) 0 (1) 43.2 (88) 36.0 COMPOSITE RISK 11-14 87.1 (31) 53.3 (212) 47.5 (274) 41.7 (12) 52.0 (529) 48.3 COMPOSITE RISK 15-21 (HI) 56.8 (37) 52.3 (HI) 40.7 (128) 31.7 (41) 44.8 (317) 42.4 CRUDE RATE 68.9 (74) 52.6 (363) 44.0 (443) 33.3 (54) 48.7 (934) ADJUSTED RATE 74.6 52.4 44.1 31.2 44.8 Rates standardized against the "Women Homemakers" group. -144-HEALTH OPINION SCORES The health opinion scores (possible range 16-48, see Table 4.26) are a measure of psychophysiological symptoms of anxiety. The lower the score, the more frequently subjects reported experiencing the symptoms. (See Chapter Four for details of questions and scoring.) The mean score for a l l female subjects was 41.5. Table 4.27 shows the distribution of the scores across the male and female study groups. Multiple classification analysis of the female data showed composite risk scores, marital status and education to be the best predictors of the women's health opinion scores. Table 6.52 shows the subgroup means for each of these predictors. The percentage of variance in the scores explained by the model was 11.13%. TABLE 6.52 PREDICTORS OF HEALTH OPINION SCORES AMONG WOMEN SUBJECTS COMPOSITE GROUP RISK SCORE MEAN (beta:0.23) MARITAL GROUP STATUS MEAN (beta:0.18) EDUCATION GROUP MEAN (beta:0.15) 7-10 42.33 11 - 14 41.50 15 - 20 40.21 Not Married 40.38 Married 41.69 Secondary 40.99 Post Sec. 42.39 Diploma 42.07 Degree 42.24 grand mean 41.5 - percentage variance explained 11.13 Across a l l the female subjects, the subgroups with the poorest scores were those in the highest li f e s t y l e risk group (group mean 40.21), those in the lowest income quintile (40.29) and those not married (40.38). The subgroups reporting the lowest incidence of psychophysiological symptoms were those with post-secondary education (42.39), with the lowest li f e s t y l e risk scores (42.33) and those with a university degree (42.24). The workforce groups were standardized against the homemaker group on the basis of marital status and composite li f e s t y l e risk scores. Thirty-five -145-percent (35%) of the homemaker group reported scores in the lowest range (22-40) compared with 32.5% of workforce women and 16% (p>.002) of workforce men. Tables 6.53-55 show the rates for each of the groups. The significant difference between male and female scores for this variable again raises the issues of illness perceptions and behaviours; questions of personal awareness and willingness to report symptoms. For example, do scores on instruments such as this truly reflect variations in the presence of psychophysiological symptoms or are they indicators of the presence of some intervening variable(s). Several researcers (Gove & Hughes, 1979; Waldron, 1983; Verbrugge, 1983), have attempted to address this question. Generally these studies support the contention that there are real differences in morbidity between the sexes. However, Gove and Hughes found that when one controls for marital status, living arrangements, psychiatric symptoms and nurturant role obligations, the health differences between men and women disappear. TABLE 6.53 HEALTH OPINION SCORES WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women with Poorest Scores in the Health Opinion Questionnaires COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 75.0 (4) 53.3 (15) 66.7 (3) 59.1 (22) 59.5 MARRIED 23.5 (119) 32.7 (352) 68.8 (48) 33.9 (519) 34.0 CRUDE RATE 25.2 (123) 33.5 (367) 68.6 (51) 34.9 (541) ADJUSTED RATE 25.6 33.5 68.7 35.0 -146-TABLE 6.54 HEALTH OPINION SCORES WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women with Poorest Scores in the Health Opinion Questionnaires COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK -15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 22.2 (9) 37.3 (75) 46.0 (37) 38.8 (121) 34.7 MARRIED 38.6 (44) 29.2 (281) 40.4 (47) 31.7 (372) 32.4 CRUDE RATE 35.9 (53) 30.9 (356) 42.9 (84) 33.5 (493) ADJUSTED RATE 38.0 29.5 40.7 32.5 Rates standardized against the "Women Homemakers" group. TABLE 6.55 HEALTH OPINION SCORES MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of men with Poorest Scores in the Health Opinion Questionnaires COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 0 (9) 31.4 (51) 37.5 (32) 30.4 (92) 24.8 MARRIED 5.5 (55) 18.3 (328) 22.2 (162) 18.2 (545) 15.7 CRUDE RATE 4.7 (64) 20.1 (379) 24.7 (194) 19.9 (637) ADJUSTED RATE 5.2 18.8 22.9 16.1 Rates standardized against the "Women Homemakers" group. -147-AFFECT BALANCE Affect balance scores were derived from Bradburn's measure of pleasant and unpleasant subjective states. Subjects' scores could be classified as positive, negative or neutral. Table 4.28 l i s t s the questions used to develop subjects' affect balance scores and Table 4.29 shows the distribution of the scores for this study group. The multiple classification analysis of the women's affect balance scores showed that subject's composite risk scores, marital status and family size were the best predictors of those scores. Table 6.56 l i s t s the subgroup means for each of these variables. The closer the mean score i s to 3, the more subjects in the group reporting a negative affect balance. TABLE 6.56 PREDICTORS OF AFFECT BALANCE SCORES AMONG WOMEN SUBJECTS COMPOSITE GROUP RISK SCORE MEAN (beta:0.17) MARITAL GROUP STATUS MEAN (beta:0.12) FAMILY SIZE GROUP MEAN (beta:0.09) 7 - 1 0 (good) 1.39 11 - 14 1.51 15 - 20 (poor) 1.63 Not Married 1.62 Married 1.49 One person 1.49 2-3 people 1.52 4-6 people 1.49 7 or more 1.58 grand mean 1.51 - percentage variance explained 5.8 Among the female subjects, the subgroups with the poorest affect balance scores were unmarried women (group mean 1.62), women in the lowest income quintile (1.62) and women with the poorest lifestyle risk scores (1.63). The groups reporting most positive affect balance scores were those with the lowest li f e s t y l e risk scores (1.39), and those with post-secondary or diploma level education (1.44). One apparently contradictory situation i s the high score of the 'not married' women (1.62) and the low score of women living alone (1.49). Since unmarried (including divorced, widowed and -148-separated) women with families appear in the f i r s t group but not in the second, one has to surmise that as a group these women report many more negative affect balance scores. The workforce groups were standardized against the women's homemaker group on the basis of lifestyle risk scores and marital status. Because a very limited number of subjects had negative scores, the rates for positive scores were compared. Tables 6.57-59 show the rates for each of the three subject groups. Fewer homemaker women reported positive scores (53.5%) compared with the workforce women's (56.2%, n.s.) and the workforce men's groups (58.5%, p<.05). The tables demonstrate the constantly positive effect of marriage and the positive relationship with good lif e s t y l e risk scores. When the very small groups of subjects with negative affect balance scores were compared, controlling only for composite health risk scores, there was no significant difference between the two women's groups (homemakers 4.9%, workforce women 3.0%) but significantly fewer men (1.2%, p<.01) had negative affect scores. (See Table 6.60) These results are in keeping with other findings which suggest that the traditional female role as homemaker and the absence of a partner are associated with higher psychosocial stress (Cleary & Mechanic, 1983; Gore & Mangione, 1983). -149-TABLE 6.57 AFFECT BALANCE WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Recording a Positive Affect Balance Score COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 50.0 (6) 38.9 (18) 0 (6) 33.3 (30) 37.5 MARRIED 62.0 (171) 52.5 (497) 48.6 (72) 54.3 (740) 54.3 CRUDE RATE 61.6 (177) 52.0 (515) 44.9 (78) 53.5 (770) ADJUSTED RATE 61.5 52.0 46.7 53.7 TABLE 6.58 AFFECT BALANCE WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Recording a Positive Affect Balance Score COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 55.6 (18) 46.7 (107) 43.6 (55) 46.7 (180) 48.5 MARRIED 70.1 (77) 54.1 (379) 41.8 (67) 54.9 (523) 56.5 CRUDE RATE 67.4 (95) 52.5 (486) 42.6 (122) 52.8 (703) ADJUSTED RATE 69.6 53.8 41.9 56.2 Rates standardized against the "Women Homemakers" group. -150-TABLE 6.59 AFFECT BALANCE MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Recording a Positive Affect Balance Score COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE NOT MARRIED 45.5 (ID 59.2 (76) 35.3 (51) 49.3 (138) 53.6 MARRIED 71.1 (76) 56.0 (455) 47.9 (238) 55.0 (769) 58.7 CRUDE RATE 67.8 (87) 56.5 (531) 45.7 (289) 54.1 (907) ADJUSTED RATE 70.1 56.2 47.4 58.5 TABLE 6.60 AFFECT BALANCE ALL SUBJECTS, AGED 25-44 YEARS Percentage of A l l Subjects Recording a Negative Affect Balance Score GROUP COMPOSITE RISK 7-10 (LO) COMPOSITE RISK 11-14 COMPOSITE RISK 15-21 (HI) CRUDE RATE ADJUSTED RATE HOMEMAKERS 2.3 (177) 3.3 (515) 9.0 (78) 4.9 (770) 4.9 WORKFORCE WOMEN 3.2 (95) 2.7 (486) 4.1 (122) 3.0 (703) 3.0 WORKFORCE MEN 0 (87) 1.1 (531) 4.8 (289) 2.2 (907) 1.2 Rates standardized against the "Women Homemakers" group. -151-6.3 HEALTH CARE CONSEQUENCES DISABILITY DAYS The number of disability days reported by the subjects was the number of days during a two week period on which they experienced any level of decreased activity as a result of their health. Table 4.30 l i s t s the questions used to generate the scores and Table 4.31 shows the distribution of the scores across the subject groups. Multiple classification analysis of the female scores using the intervening variables shown in Figure 4.3 showed that number of health problems, health opinion scores and composite lifestyle risk scores were the best predictors of the number of disability days. Table 6.61 shows the subgroup means for each of these top three predictors, their beta scores and the percentage of variance in the scores explained by the model. TABLE 6.61 PREDICTORS OF DISABILITY DAYS AMONG WOMEN SUBJECTS HEALTH GROUP PROBLEMS MEAN (beta:0.32) HEALTH GROUP OPINION MEAN (beta:0.22) COMPOSITE GROUP RISK MEAN (beta:0.16) No problems 0.0 One problem 0.48 > one problem 1.06 22-40 (poorer) 0.67 41-44 0.39 45-48 (better) 0.19 7-10(better) 0.57 11-14 0.36 15-20 0.69 grand mean 0.44 - percentage variance explained 19.8 Among the female study group, the subgroups with the greatest number of disability days were those with more than one chronic health problem (group mean 1.06 disability days), those with negative affect balance scores (average 0.76 days) and those with some post-secondary education (0.72 days). The subgroups with the lowest average number of disability days were women with no chronic health problems (reporting no disability days), those with high health opinion scores (0.19 days) and those in the second quintile for family income (0.23). -152-The workforce groups were standardized against the homemaker group on the basis of health opinion scores and numbers of chronic health problems. Tables 6.62-64 show the rates for each of the groups. A higher percentage of women in the workforce (15.7% compared with 12.5% of homemakers and 12.7% of men) reported one or more disability days but the differences were not significant at the .05 level. One explanation for a higher number of disability days among the workforce women might be that the separation of the workplace from the resting place, makes a reduced activity day more identifiable for persons in the workforce. Especially with proxy reporting, i t i s possible that reduced days for homemakers could go unrecognized at least by anyone other than the homemaker. Alternatively one could argue on the basis of Gove's fixed role study (1984) that i f a fixed and well defined role has a health protective effect, then women in the workforce may be suffering the consequences of having conflicting roles (nurturant homemaker versus employee). The similarly low rates for homemakers and men in the workforce further support Gove's hypothesis. The work of Meissner (1975) and similar studies by other researchers show that the total work week for women in the workforce is much longer than for men and women homemakers. The longer work days and fewer hours of rest may lead to a higher illness rate. -153-TABLE 6.62 DISABILITY DAYS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting One or More Disability Days in the Previous Two Weeks 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 15.0 (80) 14.3 (49) 29.4 (17) 16.4 (146) 18.6 H.O.S. 41 - 44 0 (105) 10.3 (78) 23.8 (63) 9.4 (246) 9.6 POORER H.O.S. 16 - 40 0 (54) 17.7 (51) 24.7 (73) 15.2 (178) 12.1 CRUDE RATE 5.0 (239) 13.5 (178) 24.8 (153) 13.0 (570) ADJUSTED RATE 3.4 13.7 25.3 12.5 -154-TABLE 6.63 DISABILITY DAYS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting One or More Disability Days in the Previous Two Weeks^ 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 0 (55) 18.8 (32) 50.0 (22) 15.6 (109) 19.3 H.O.S. 41 - 44 0 (92) 18.4 (76) 34.1 (44) 13.7 (212) 14.9 POORER H.O.S. 16 - 40 0 (51) 14.3 (56) 37.3 (59) 18.1 (166) 14.5 CRUDE RATE 0 (198) 17.1 (164) 38.4 (125) 15.6 (487) ADJUSTED RATE 0 17.1 38.7 15.7 Rates standardized against the "Women Homemakers" group. -155-TABLE 6.64 DISABILITY DAYS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting One or More Disability Days in the Previous Two Weeks^ 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATF ADJUSTED RATE BETTER H.O.S. 45 - 48 0 (117) 14.9 (67) 15.2 (33 6.9 (217) 8.7 H.O.S. 41 - 44 0 (161) 21.1 (95) 34.2 (38) 11.2 (294) 15.7 POORER H.O.S. 16 - 40 0 (54) 18.0 (39) 22.2 (45) 12.3 (138) 11.6 CRUDE RATE 0 (332) 18.4 (201) 32.8 (116) 10.0 (649) ADJUSTED RATE 0 18.6 25.8 12.7 Rates standardized against the "Women Homemakers" group. -156 -HEALTH PROFESSIONAL CONSULTATIONS Th is score represents the number of hea l th p r o f e s s i o n a l con tac ts that the sub jec ts repor ted fo r a two week p e r i o d , converted to an annual r a t e . Table 4.33 shows the d i s t r i b u t i o n of consu l t s repor ted by t h i s group. M u l t i p l e c l a s s i f i c a t i o n a n a l y s i s of the female scores us ing the i n te r ven ing v a r i a b l e s shown i n F igure 4 . 3 , i d e n t i f i e d number of hea l th problems, hea l th op in ion scores and the composite l i f e s t y l e scores as the best p r e d i c t o r s of p r o f e s s i o n a l hea l th c o n s u l t s . Table 6.65 shows the subgroup means fo r these p r e d i c t o r s , t h e i r beta scores and the percentage of var iance i n the scores exp la ined by the model. I t i s of i n t e r e s t , but not s u r p r i s i n g , that women's scores f o r p r o f e s s i o n a l s e r v i c e s (female consu l t s ) was the fou r th best p r e d i c t o r . (See Table 6.78) TABLE 6.65 PREDICTORS OF HEALTH PROFESSIONAL CONSULTATIONS BY WOMEN SUBJECTS HEALTH GROUP PROBLEMS MEAN (beta : 0.29) HEALTH GROUP OPINION MEAN (be ta :0 .19) COMPOSITE GROUP RISK MEAN (be ta :0 .17) No problems 3.95 One problem 6.38 > 1 problem 7.59 21-40 (poorer) 5.93 41-44 5.63 45-48 (be t te r ) 5.01 7-10 (be t te r ) 5.22 11-14 5.75 15-21 (poorer) 6.11 grand mean 5.7 - percentage var iance exp la ined 21.8 The subgroup repor t i ng more than one chron ic hea l th problem reported the h ighest number of hea l th c ons u l t a t i ons ( 7 . 6 ) . Other high groups were those from the A t l a n t i c P rov inces , those wi th more than one d i s a b i l i t y day, wi th a u n i v e r s i t y degree or from a fami ly of two or three ( a l l r epo r t i ng an average of 6.4 c o n s u l t a t i o n s fo r a yea r ) . The subgroups repo r t i ng low numbers of hea l th p r o f e s s i o n a l contac ts were people from la rge f a m i l i e s (2 .84 ) , people wi th no chron ic hea l th problems (3.95) and women from the lowest income q u i n t i l e (4.24 v i s i t s ) . -157-The two workforce groups were standardized against the homemaker group on the basis of health problems and health opinion. Tables 6.66-68 show the rates for the three groups of subjects reporting more than five health professional contacts in a year. The rate for workforce women was greater than for the homemaker group (39.3% compared with 35.4%), but the difference was not significant at the .05 level. The men's rate (26.9%, p<.002) for five or more consults was considerably lower than the homemakers' rate (35.4%). In keeping with their higher rate of health consultations, workforce women had a higher rate for female health related professional services (preventive as well as prescriptive for hormones), had more disability days and as can be seen from the next section, also reported a greater use of medications. This might lead one to believe that they are less healthy, however homemaker and workforce women reported similar numbers of chronic health problems. An examination of the relationship between number of health problems and consultations (see Tables 6.66-68) shows that i t is workforce women with more than one chronic problem who use professional health services the most. There is no apparent explanation for this. Medication taking, although higher for workforce women than for either of the other two groups does not rise disproportionately for women with more than one chronic health problem. The C.H.S. data on reasons for health professional consultations (not included in this study) might provide some explanations. The inverse relationship between health opinion and consultations among homemaker women may be a reflection of visits associated with childbearing. For both workforce groups, a poorer health opinion i s associated with a higher number of consultations. -158-TABLE 6.66 HEALTH PROFESSIONAL CONSULTATIONS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting More than Five Health Professional Consultations in the Previous Year 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 26.5 (67) 40.9 (44) 61.1 (18) 36.4 (129) 40.5 H.O.S. 41 - 44 27.6 (105) 36.6 (71) 41.7 (60) 33.9 (236) 34.3 POORER H.O.S. 16 - 40 16.0 (50) 44.2 (52) 45.1 (71) 36.4 (173) 32.8 CRUDE RATE 24.8 (222) 40.1 (167) 45.7 (149) 35.3 (538) ADJUSTED RATE 23.6 40.1 47.5 35.4 -159-TABLE 6.67 HEALTH PROFESSIONAL CONSULTATIONS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting More than Five Health Professional Consultations in the Previous Year^ 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 14.8 (55) 31.3 (32) 57.9 (19) 27.4 (106) 31.8 H.O.S. 41 - 44 28.6 (91) 40.2 (82) 65.2 (46) 40.6 (219) 42.3 POORER H.O.S. 16 - 40 23.5 (51) 50.9 (55) 54.4 (57) 43.6 (163) 40.6 CRUDE RATE 23.4 (197) 42.0 (169) 59.0 (122) 38.8 (488) ADJUSTED RATE 23.7 41.6 60.0 39.3 Rates standardized against the "Women Homemakers" group. -160-TABLE 6.68 HEALTH PROFESSIONAL CONSULTATIONS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting More than Five Health Professional Consultations in the Previous Year^ 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 14.5 (117) 26.9 (67) 12.5 (32) 18.1 (216) 17.8 H.O.S. 41 - 44 10.0 (160) 29.4 (92) 42.1 (38) 20.3 (290) 24.9 POORER H.O.S. 16 - 40 20.4 (54) 53.9 (39) 40.0 (35) 35.9 (128) 36.2 CRUDE RATE 13.3 (331) 33.3 (198) 32.4 (105) 22.7 (634) ADJUSTED RATE 14.4 36.7 34.4 26.9 ^ Rates standardized against the "Women Homemakers" group. -161-MEDICATIONS Subjects were asked to report the variety of medications they had consumed in the previous two days. As can be seen from Table 4.34, subjects were to include every type of medication from vitamins and over-the-counter drugs, to prescribed medications. In this model, chronic health problems, health opinion scores and composite health risk scores were the best predictors of medication consumption. Table 6.69 shows the subgroup means for each of these top three predictors, their beta scores and the percentage of variance in the scores explained by the model. The percentage of variance explained by the model was higher for medication use (27.5%) than for any other variable except seat belt use (30.0%). TABLE 6.69 PREDICTORS OF MEDICATION USE BY WOMEN SUBJECTS HEALTH GROUP PROBLEMS MEAN (beta:0.34) HEALTH GROUP OPINION MEAN (beta:0.20) RISK GROUP SCORE MEAN (beta:0.18) No problems 0.52 One problem 0.88 > 1 problem 1.40 45-48 (better) 0.66 41-44 0.85 21-40 (poorer) 1.04 7-10 (better) 0.78 11-14 0.82 15-21 (poorer) 1.06 grand mean 0.85 - percentage variance explained 27.5 Subjects reporting more than one health problem and those reporting more than one disability day also reported the highest average consumption of medications (1.4). Those with no health problems and those with a good health opinion score reported the lowest levels of medication use (0.52 and 0.66 respectively). The positive correlation of health risk scores with medication consumption is interesting to note although this model is unable to provide any indication of causality. The rates of consumption of more than one medication were compared across a l l three groups. The two workforce populations were standardized against -162-the homemaker group on the basis of health problems and health opinion scores. Tables 6.70-72 show the rates for each of the study groups as well as the consistent relationships of fewer health problems and better health opinion scores with lower rates of medication consumption. It is of interest that when variations in health problems and health opinion were controlled, the men's consumption of medications was not significantly different from the homemaker women's group (16.6% compared with 17.8%). The workforce women however s t i l l had a significantly higher rate of consumption than either group (24.1%,p<.02). These findings of higher rates for women that become nearly equal when the populations are controlled for morbidity, are similar to reports in the literature (Verbrugge, 1982). The higher rate of medication consumption among workforce women compared with homemaker women could not however be validated against reports in the literature. In one study (Jennings, Mazaik and McKinlay, 1984) the use of tranquilizers was highest among the unemployed (18%) and lowest among the employed women (10%) with homemakers (14%) in the middle. The level of medication taking by workforce men is nearly the same as for homemaker women, yet their number of health professional consultations are significantly (p<,002) lower. Does this mean that men use more over-the-counter drugs than women? The inability of this study to distinguish between preventive and curative drug consumption has to be recognized in the interpretation of this study's findings. -163-TABLE 6.70 REGULAR TAKING OF MEDICATIONS WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting the Regular Taking of One or More Medication 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 2.9 (68) 13.6 (44) 33.3 (15) 10.2 (127) 14.6 H.O.S. 41 - 44 6.7 (105) 15.5 (71) 26.2 (61) 14.4 (237) 14.7 POORER H.O.S. 16 - 40 14.8 (54) 23.1 (52) 39.7 (73) 27.4 (179) 24.2 CRUDE RATE 7.5 (227) 17.4 (167) 33.6 (149) 17.7 (543) ADJUSTED RATE 8.5 17.6 32.3 17.8 -164-TABLE 6.71 REGULAR TAKING OF MEDICATIONS WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting the Regular Taking of One or More Medication 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 9.1 (55) 18.8 (32) 40.9 (22) 18.4 (109) 20.8 H.O.S. 41 - 44 9.8 (92) 25.3 (83) 47.8 (46) 23.5 (221) 25.0 POORER H.O.S. 16 - 40 17.7 (51) 25.0 (56) 37.3 (59) 27.1 (166) 25.3 CRUDE RATE 11.6 (198) 24.0 (171) 41.7 (127) 23.6 (496) ADJUSTED RATE 12.2 23.7 42.7 24.1 Rates standardized against the "Women Homemakers" group. -165-TABLE 6.72 REGULAR TAKING OF MEDICATIONS MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting the Regular Taking of One or More Medication 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE BETTER H.O.S. 45 - 48 1.7 (117) 14.9 (67) 24.2 (33) 9.2 (217) 12.0 H.O.S. 41 - 44 2.5 (161) 20.0 (95) 36.8 (38) 12.6 (294) 17.3 POORER H.O.S. 16 - 40 7.4 (54) 25.6 (39) 29.4 (34) 18.9 (127) 19.1 CRUDE RATE 3.0 (332) 19.4 (201) 30.5 (105) 12.7 (638) ADJUSTED RATE 3.9 20.7 31.4 16.6 Rates standardized against the "Women Homemakers" group. -166-HOSPITALIZATION Subjects had been asked to report the number of nights in the previous year that they had been a patient in a hospital, a nursing home or a convalescent home. Table 4.32 shows the distribution of the number of nights of hospitalization for the study group. It should not be surprising that for this childbearing age group, the women reported more hospitalization than the men. Multiple classification analysis of the women's data showed disability days, chronic health problems and health opinion scores to be the best predictors of hospitalization. Table 6.73 shows the subgroup means for each of these variables, their beta scores and the percentage of variance in the scores explained by the model. TABLE 6.73 PREDICTORS OF NIGHTS IN HOSPITAL FOR WOMEN SUBJECTS DISABILITY GROUP DAYS MEAN (beta: 0.20) HEALTH GROUP PROBLEMS MEAN (beta: 0.17) HEALTH GROUP OPINION MEAN (beta: 0.13) None 0.73 One 1.49 > 1 1.32 None 0.50 One 1.03 > 1 1.02 45-48 (better) 0.71 41-44 0.74 22-40 (poorer) 0.97 grand mean 0.81 - percentage variance explained 14.7% Among the female subjects, the subgroups reporting the highest average number of nights in hospital were those with one or more disability days (averages 1.49 and 1.32) and those in the second percentile for family income (average 1.14 nights). As seen in Table 6.73 there was a positive correlation between disability days and hospitalization. The women's groups reporting the lowest average number of nights in hospital were those that were not married (average 0.18 nights), those with a -167-negative affect score (average 0.32 nights) and those from a family of one (average 0.41 nights). These findings are in keeping with an earlier finding that negative affect balance is well correlated with not being married and with the argument that hospitalization for this age group of women is associated with their childbearing *ole (a role not usually assumed by unmarried women). When the rates of hospitalization for the two work groups were standardized against the homemaker group on the basis of disability days and chronic health problems, the homemaker group had the highest rates. Eighteen percent (18%) of the homemaker group reported one or more night in hospital compared with 11% (p<.002) of workforce women and 5% (p<.002) of workforce men. Tables 6.74-76 show the standardized rates, as well as the hospitalization trends associated with disability days and health problems. It i s interesting to note that for workforce women there was a linear association between chronic health problems and hospitalization but for the homemaker group, this association was 'U'-shaped. This observation would support the suggestion that further study should include a comparison of reasons for hospitalization among the women's groups. Higher rates of hospitalization for women regardless of occupation, is in keeping with findings of other studies (WHO, 1973; Lewis & Lewis, 1977; Statistics Canada, 1984). It is worthy of note that this method of assessing hospitalization neglects the use of hospital services on a day-care basis. As this form of service provision is rapidly expanding, particularly for gynaecological procedures, i t s use should be included in any further study of hospital service use. -168-TABLE 6.74 NIGHTS IN HOSPITAL WOMEN HOMEMAKERS, AGED 25-44 YEARS Percentage of Women Reporting Any Nights in Hospital in the Previous Year 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE NO DISABILITY DAYS 15.6 (585) 21.2 (368) 17.0 (283) 17.6 (1236) 17.6 ONE DISABILITY DAY 0 (0) 44.0 (25) 24.1 (54) 30.4 (79) 19.9 > 1 DISABILITY DAY 0 (0) 50.0 (24) 27.8 (36) 36.7 (60) 22.7 CRUDE RATE 15.6 (585) 24.2 (417) 19.0 (373) 19.1 (1375) ADJUSTED RATE 14.0 23.8 17.8 18.0 -169-TABLE 6.75 NIGHTS IN HOSPITAL WOMEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Women Reporting any Nights in Hospital in the Previous Year 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED RATE NO DISABILITY DAYS 8.4 (452) 12.3 (268) 15.5 (193) 11.1 (913) 11.5 ONE DISABILITY DAY 0 (0) 8.9 (45) 6.9 (58) 7.8 (103) 4.6 > 1 DISABILITY DAYS 0 (0) 17.1 (35) 18.9 (37) 18.1 (72) 10.4 CRUDE RATE 8.4 (452) 12.4 (348) 14.2 (288) 11.2 (1088) ADJUSTED RATE 7.6 12.3 15.2 11.1 Rates standardized against the "Women Homemakers" group. -170-TABLE 6.76 NIGHTS IN HOSPITAL MEN IN THE PAID LABOUR FORCE, AGED 25-44 YEARS Percentage of Men Reporting Any Nights in Hospital in the Previous Year 0 CHRONIC HEALTH PROBLEMS 1 CHRONIC HEALTH PROBLEM > 1 CHRONIC HEALTH PROBLEM CRUDE RATE ADJUSTED PATE NO DISABILITY DAYS 2.8 (1108) 7.0 (503) 6.1 (229) 4.4 (1840) 5.0 ONE DISABILITY DAY 0 (0) 7.8 (47) 9.8 (51) 4.1 (98) 5.0 > 1 DISABILITY DAYS 0 • (0) 13.8 (29) 26.3 (19) 18.8 (48) 11.3 CRUDE RATE 2.8 (1108) 6.0 (579) 7.7 (299) 4.7 (1986) ADJUSTED RATE 2.5 7.3 7.2 5.2 Rates standardized against the "Women Homemakers" group. -171-SUMMARY Using the model shown in Figure 4.3, M.C.A. were used to identify the variables most able to explain variations in health risk behaviour, health status and health care consequence scores. Tables 6.77-78 summarize the results which have been discussed in detail throughout the chapter. The tables reflect the ability of a l l the predictor variables provided in the model to explain variations in the dependent variable scores. In Table 6.77, using cigarette smoking as an example, the reader can read down the chart and identify that, the model was able to explain only 3.9% of the variation in smoking scores, there were 3176 female subjects who reported on their smoking behaviour, the mean score for a l l the subjects was 2.10 with a range of possible scores from '1' for non-smokers through to '3' for current regular smokers, education was the best predictor (beta score 0.16) out of the seven demographic variables shown in Figure 4.3 and economic responsibility the worst (beta score 0.02). Among the health risk behaviours, (Table 6.77) the model was most effective at explaining variations in seat belt use (30% explained) with region of residence (beta 0.52) being the best predictor. Physical activity was the health risk behaviour least able to be explained by the model (3% explained) and along the bottom row of the chart i t can be seen from the '0.00' beta scores that in some instances some of the predictor variables showed no correlation with the dependent scores. Table 6.78 shows the results of M.C.A. for health status and health care consequence scores. The model was most effective in explaining variations in disability days (19.8% explained), health professional consultations (21.8% explained) and the regular taking of medications (27.5% explained). Number -172-of chronic health problems was a strong predictor (beta 0.3 approx) for disability days, health professional consultations and medication use. From the bottom row of the chart i t can again be seen that there were some predictors with beta scores of '0.00' showing no correlation with the dependent variable scores. Tables 6.79-81 summarize the rates found in the standardized tables. For a l l the variables except Affect Balance*, the rates for poor health risk behaviour, poor health status or high use of the health care system were compared. Higher rates indicate a higher percentage of subjects engaging in the behaviour. Using smoking as an example (see Table 6.79) , the rates for regular or occasional smokers were compared across the three subject groups which had been standardized on the basis of the top two predictors, namely Education and Marital Status. For smoking there were no significant difference among the three groups. Across the various health risk behaviours, the workforce women did differ significantly from the homemaker group, as seen by the asterisks, sometimes being exposed to less risk (companionship and screening for breast and cervical cancers) but more often exposed to more risk (less exercise, more l i f e events, regular taking of hormone p i l l s and less frequent use of seat belts). In health status (Table 6.80) the two groups of women were very similar although i f there was a trend, i t was for workforce women to report slightly better health than their homemaker counterparts. In health care related behaviours (Table 6.81) except for hospitalization, workforce women reported more - more disability days, more contact with health professionals and more use of medications. For Affect Balance, because of the low number of negative affect scores, the percentages of positive affect scores were compared. Therefore, a higher score is considered desirable. -173-Generally, although men reported much higher consumption of alcohol and more motor vehicle travel (often without seat belts) they reported better health and less use of the health care system. 6.4 DEMOGRAPHIC VARIABLES In summarizing the results i t i s also interesting to separate out each of the demographic variables and identify behaviours for which i t served as a good predictor. Tables 6.82-88 show the demographic variables, the dependent variables for which they are one of the top three predictors and the subgroup means for each category of the demographic variable. For example, from Table 6.82 i t can be seen that region was one of the top three predictors for a l l health risk behaviours except companionship, l i f e change events and composite risk scores. It was also one of the top three predictors for chronic health problems and this i s included in the table. Throughout Tables 6.82-88 low subgroup means are more desirable except for health opinion scores. B.C. had the lowest and therefore the best scores for four out of the seven health risk behaviour variables for which region was a top predictor. On the other hand B.C. subjects reported the greatest number of chronic health problems. The data in this study do not provide any indication of why this strange juxtaposition of healthier behaviours with poorer health status should exist. Tables 6.83 and 6.85 show a consistently protective effect of being married and living in a large family. Table 6.84 shows only a weak trend toward better scores with higher education but those with secondary education only, consistently report higher risk behaviours. Tables 6.86-88 show other demographic predictors which feature as one of the top three predictors for one or more variable. The three variables associated with economic responsibility (Table 6.86) should not be -174-s u r p r i s i n g . Women with economic r e s p o n s i b i l i t y for the i r family are less l i k e l y to want to have more children, may do a large part of the family driving as well as driving to and from work and as women i n the workforce have more opportunity to experience l i f e change events. Table 6.88 showing that more people report spending l e i s u r e time alone between October and December i s inte r e s t i n g , especially as December tends to be a time of increased s o c i a l a c t i v i t y . The model does not provide any explanation but one could speculate that December celebrations may be more family centred and there may be less casual s o c i a l i z i n g at night school classes and clubs etc. The following chapter provides further comment on the findings reported i n Chapter Six and attempts to put them i n perspective with regard to the health of women i n the workforce. Implications for further study also are discussed. TABLE 6.77 MULTIPLE CLASSIFICATION ANALYSIS OF HEALTH RISK BEHAVIOURS FOR ALL FEMALE SUBJECTS MOTOR USE OF COMPOSITE DEPENDENT CIGARETTE ALCOHOL HORMONE PHYSICAL COMPANION- LIFE CHANGE FEMALE VEHICLE SEAT RISK VARIABLE SHORING CONSUMPTION PILLS ACTIVITY SHIP EVENTS PREVENTIVE TRAVEL BELTS SCORES PERCENTAGE 3.9 4.7 4.2 3.0 3.2 3.6 7.0 4.0 30.0 7.7 EXPLAINED 1 OF SUBJECTS 3176 2395 3273 2973 3262 3239 3261 2451 1724 1462 MEAN SCORE 2.10 2.OA 1.34 1.99 1.53 1.27 4.43 1.98 1.9 12.31 RANGE (1-3) (1-3) (1-3) (1-3) (1-3) (1-3) (3-9) (1-3) (1-3) (7-21) Beta Beta Beta Beta Beta Beta Beta Beta Beta Beta PREDICTOR Educa- Family Family Month of Family Family Family VARIABLES IN tion 0.16 Income 0.12 Size 0.10 Survey 0.11 Survey 0.10 Size 0.12 ReRion 0.18 Income 0.08 ReRlon 0.52 Size 0.15 ORDER OF Marital Family Family Marital Economic Educa- Educa-THEIR Status 0.10 ReRlon 0.09 ReRlon 0.08 ReRlon 0.11 Size 0.08 Income 0.10 Status 0.13 Respons. 0.08 tion 0.14 tion 0.11 ABILITY TO Marital Economic Educa- Marital Economic Educa- Marital Marital EXPLAIN ReRlon 0.07 Status 0.08 Respons. 0.07 tion 0.06 Status 0.08 Respons. 0.07 tion 0.11 ReRlon 0.07 Status 0.06 Status 0.07 VARIATION IN Month of Family Month of Family Family Family Family Family THE DEPENDENT Survey 0.02 Size 0.07 Survey 0.05 Size 0.05 ReRlon 0.04 Region 0.06 Income 0.06 Size 0.06 Size 0.06 Income 0.06 VARIABLE Family Month of Educa- Economic Famlly Educa- Economic Educa- Family Size 0.02 Survey 0.05 tion 0.04 Respons. 0.04 Income 0.03 tion 0.04 Respons. 0.05 tion 0.04 Income 0.05 Region 0.05 Family Educa- Faml1y Family Educa- Month of Family Ma r i t a l Month of Economic Income 0.02 tion 0.03 Income 0.03 Income 0.03 tion 0.02 Survey 0.00 Size 0.04 Status 0.02 Survey 0.05 Respons. 0.03 Economic Economic Marital Marital Economic Marital Month of Month of Economic Month of Respons. 0.02 Respons. 0.00 Status 0.00 Status 0.02 Respons. 0.00 Status 0.00 Survey 0.02 Survey 0.02 Respons. 0.03 Survey 0.03 TABLE 6.78 MULTIPLE CLASSIFICATION ANALYSIS OF REPORTED HEALTH STATUS AND HEALTH CARE RELATED BEHAVIOURS FOR ALL FEMALE SUBJECTS DEPENDENT VARIABLE CHRONIC HEALTH PROBLEMS HEALTH OPINION AFFECT BALANCE DISABILITY DAYS HEALTH PROFESSIONAL CONSULTS MEDICATIONS HOSPITALIZATION PERCENTAGE EXPLAINED 4.1 11.1 5.8 19.8 21.8 27.5 14.5 f OF SUBJECTS 1462 991 1422 968 954 965 965 MEAN SCORE RANGE 1.05 (0-10) 41.50 (16-48) 1.51 (1-3) 0.44 (0-14) 5.7 (0-50) 0.85 (0-4) 0.81 (0-22) PREDICTOR VARIABLES IN Composite Risk Score Beta 0.12 Beta Composite Risk Score 0.23 Beta Composite Risk Score 0.17 Beta Health Problems 0.32 Beta Health Problems 0.29 Beta Health Problems 0.34 Disability Days Beta 0.20 ORDER OF THEIR Family Size 0.11 Marital Status 0.18 Marital Status 0.12 Health Opinion 0.22 Health Opinion 0.19 Health Opinion 0.20 Health Problems 0.17 ABILITY TO EXPLAIN Region 0.10 Education 0.15 Family Size 0.09 Risk Score 0.16 Risk Score 0.17 Risk Score 0.18 Health Opinion 0.13 VARIATION IN THE DEPENDENT Education 0.08 Family Size 0.11 Education 0.09 Region 0.09 Female Consults 0.12 Disability Days 0.17 Family Income 0.12 VARIABLE Month of Survey 0.04 Region 0.07 Region 0.07 Education0.07 Family Size 0.10 Region 0.14 Composite Risk 0.11 Marital Status 0.03 Family Income 0.06 Family Income 0.06 Family Income 0.06 Disability Days 0.10 Female Prevent 0.09 Female Prevent 0.11 Economic Respons. 0.02 Month of Survey 0.04 Month of Survey 0.04 Family Size 0.06 Region 0.09 Family Income 0.05 Marital Status 0.11 Family Income 0.02 Economic Respons. 0.02 Economic Respons. 0.00 Affect Balance 0.06 Family Income 0.08 Family Size 0.04 Family Size 0.06 Month of Survey 0.01 Education0.07 Education0.04 Region 0.04 Marital Status 0.01 Female Prevent 0.07 Marital Status 0.02 Affect Balance 0.03 Economic Respons. 0.01 Affect Balance 0.03 Economic Respons. 0.02 Education 0.02 Month of Survey 0.00 Affect Balance 0.02 Month of Survey 0.02 Marital Status 0.00 Month of Survey 0.00 Economic Respons. 0.00 Economic Respons. 0.00 TABLE 6.79 PREVALENCE OF HEALTH RISK BEHAVIOURS A Comparison Between Men and Women i n the P a i d Labour Force and Women Homemakers RISK FACTOR DESCRIPTION PREVALENCE TOP PREDICTORS IN THE MODEL WOMEN AT HOME WOMEN IN THE PAID LABOUR FORCE MEN IN THE PAID LABOUR FORCE SMOKING Current Regular or Occasional Smokers Education M a r i t a l S t a t u s 45.2 44.8 52.4 ALCOHOL More than 7 A l c o h o l i c Drinks i n One Week Family Income Region 19.6 21.0 **** 49.7 EXERCISE Less than 1749 on the P h y s i c a l A c t i v i t y Index Time of Year Region 35.6 *** • 40.1 37.4 COMPANY Le i s u r e Time Spent Mostly Alone Time of Year Family S i z e 12.2 *** 8.1 **** 7.3 HORMONE PILLS Any Type of Hormone 'Medication' Family S i z e Region 13.6 **** 16.8 N/A LIFE EVENTS > 1 Major L i f e Change Event i n the Prev i o u s Year Family S i z e Family Income 17.6 **** 29.0 23.9* M.V. TRAVEL More than 11,000 Km i n the P r e v i o u s Year Family Income Economic Respons. 29.0 33.7 ** 40.5 SEAT BELT USE NO use of Seat B e l t on a Regular B a s i s Region Education 41.1 **** 48.1 *** 47.6 FEMALE PREVENT Infrequent or no BSE, BPE or Pap Smear Region M a r i t a l S t a t u s 14.3 **** 9.8 N/A COMPOSITE RISK Combined Score of Above Except Female Prevent & Seat B e l t Scores Family S i z e Education 10.5 11.8 •*** 30.5 * ** *** **** P<.05 ; P<.02; P<.01; P<.002 TABLE 6.80 SELF REPORTED HEALTH STATUS A Comparison between Men and Women in the Paid Labour Force and Women Homemakers HEALTH STATUS MEASURE DESCRIPTION TOP PREDICTORS PREVALENCE (%) WOMEN AT HOME WOMEN IN THE PAID LABOUR FORCE MEN IN THE PAID LABOUR FORCE HEALTH OPINION Scoring 40 or Less on a 16-48 Range1 Composite Risk Marital Status 35.0 32.5 ** 16.1 AFFECT BALANCE Reporting More Pleasant than Unpleasant Subjective Feelings Composite Risk Marital Status 53.7 56.2 * 58.5 CHRONIC HEALTH PROBLEMS More than One Chronic Health Problem Composite Risk Family Size 57.8 57.2 ** 44.8 P<.05; P<.002 .See text for f u l l details of Health Opinion Measure TABLE 6.81 HEALTH CARE RELATED BEHAVIOURS A Comparison between Men and Women in the Paid Labour Force and Women Homemakers INDICATORS DESCRIPTION TOP PREDICTORS PREVALENCE (%) WOMEN AT HOME WOMEN IN THE PAID LABOUR FORCE MEN IN THE PAID LABOUR FORCE DISABILITY DAYS Days of Reduced Activity Due to Illness Number of Health Problems Health Opinion Score 12.5 15.7 12.7 HEALTH CONSULTS More than 5 Health Professional Consults in Previous Year Number of Health Problems Health Opinion Score 35.4 39.3 ** 26.9 MEDICATIONS More than One Medication Regularly Number of Health Problems Health Opinion Score 17.8 * 24.1 16.6 HOSPITAL-IZATION Any Nights in Hospital in Previous Year Number of Disability Days Number of Health Problems 18.0 ** 11.1 ** 5.2 P<.02; P<.002 -180-TABLE 6.82 SELECTED LIFESTYLE AND HEALTH SCORES FOR WOMEN IN THE STUDY GROUP, BY REGION OF RESIDENCE SUMMARY OF REGIONAL VARIATIONS*... ...among those dependent variables for which 'region' was one of the top three Predictors FACTOR ATLANTIC REGIONS QUEBEC ONTARIO PRAIRIES B.C. ALCOHOL 1.91 2.01 2.10 2.08 2.11 CIGARETTES 2.14 2.16 .1.99 2.12 2.11 PILLS 1.31 1.45 1.31 1.35 1.29 M.V. TRAVEL 2.04 1.90 1.95 2.04 1.98 EXERCISE 2.14 1.99 1.94 1.93 1.85 SEAT BELTS 2.62 1.51 1.47 2.35 1.43 FEMALE PREVENTIVE 4.55 4.83 4.42 4.10 4.08 HEALTH PROBLEMS 1.08 0.86 1.09 1.06 1.21 The higher the score, the less desirable the behaviour or status. -181-TABLE 6.83 SELECTED LIFESTYLE AND HEALTH SCORES FOR WOMEN IN THE STUDY GROUP, BY MARITAL STATUS VARIATIONS ASSOCIATED WITH MARITAL STATUS ....among those dependent variables for which 'marital status' was one of the top three predictors. FACTOR NOT MARRIED MARRIED ALCOHOL 2.15 2.02 CIGARETTES 2.32 2.07 SEAT BELT USE 2.06 1.89 FEMALE PREVENTIVE 4.89 4.35 COMPOSITE RISK SCORE 12.67 12.25 HEALTH OPINION SCORE* 39.76 41.79 AFFECT BALANCE 1.68 1.48 COMPANY FOR LEISURE TIME 1.65 1.51 For a l l except Health Opinion Score, a lower score should be considered more desirable. -182-TABLE 6.84 SELECTED LIFESTYLE AND HEALTH SCORES FOR WOMEN IN THE STUDY GROUP, BY LEVEL OF EDUCATION VARIATIONS ASS( ...among those depend* was one of the to 5CIATED WITH EDUCATION 2nt variables for which 'education' 3 three predictors. FACTOR SECON-DARY POST SECON-DARY DIPLOMA DEGREE ALCOHOL 2.04 2.09 2.01 2.05 CIGARETTES 2.20 2.00 1.95 1.78 EXERCISE 2.02 2.02 1.89 1.92 SEAT BELT USE 2.01 1.89 1.78 1.65 FEMALE PREVENTIVE 4.54 4.12 4.15 4.33 COMPOSITE RISK SCORE 12.41 12.72 12.08 11.91 HEALTH OPINION SCORE* 41.02 42.68 41.90 42.13 For a l l except Health Opinion Scores, a lower score should be considered more desirable. -183-TABLE 6.85 SELECTED LIFESTYLE AND HEALTH SCORES FOR WOMEN IN THE STUDY GROUP, BY FAMILY SIZE VARIATIONS ASSOCIATED WITH FAMILY SIZE * ....among those dependent variables for which 'family size' was one of the top three predictors. FACTOR FAMILY OF 1 FAMILY OF 2 - 3 FAMILY OF 4 - 6 FAMILY OF 7+ HORMONE PILLS 1.57 1.40 1.30 1.20 LIFE EVENTS 1.30 1.36 1.23 1.20 COMPOSITE RISK 12.98 12.57 12.10 11.64 HEALTH PROBLEMS 1.33 1.16 0.95 0.79 AFFECT BALANCE 1.49 1.52 1.49 1.58 COMPANY FOR LEISURE TIME 1.63 1.59 1.50 1.40 * Lower scores should be considered more desirable. -184-TABLE 6.86 SELECTED LIFESTYLE SCORES FOR WOMEN IN THE STUDY GROUP ACCORDING TO FINANCIAL RESPONSIBILITY FOR ECONOMIC FAMILY UNIT VARIATIONS ASSOCIATED WITH FINANCIAL RESPONSIBILITY...* ...among those dependent variables for which 'economic responsibility' was one of the top three predictors FACTOR RESPONSIBLE DEPENDENT HORMONE PILLS 1.46 1.32 M.V. TRAVEL 2.13 1.95 LIFE EVENTS 1.34 1.25 TABLE 6.87 SELECTED LIFESTYLE SCORES FOR WOMEN IN THE STUDY GROUP, BY FAMILY INCOME VARIATIONS ASSOCIATED WITH INCOME FAMILY * ....for the two dependent variables for which 'family income' was one of the top three predictors. 1ST QUINTILE (LOWEST) 2ND QUINTILE 3RD QUINTILE 4TH QUINTILE 5TH QUINTILE (HIGHEST) M.V. TRAVEL 1.73 1.86 1.99 2.04 2.17 LIFE EVENTS 1.34 1.30 1.30 1.23 1.21 * Lower scores should be considered more desirable. -185-TABLE 6.88 COMPANIONSHIP SCORES FOR WOMEN IN THE STUDY GROUP, BY THE SEASON OF THE SURVEY VARIATIONS ASSOCIATED WITH THE SEASON OF THE SURVEY * ....for the one variable for which 'time of year' was one of the top three predictors. FACTOR JULY-SEPT OCT-DEC JAN-MAR COMPANY FOR LEISURE TIME 1.45 1.61 1.49 Lower scores should be considered more desirable. -186-CHAPTER SEVEN DISCUSSION 7.1 THE QUESTION Since the 1950's there has been a dramatic increase in the number of women in the paid labour force in Canada. Table 2.9 shows that the increase has been greatest among women with children and especially among those with children under six years of age. The health of a society's mothers has an impact not only on the women themselves but also on their families because as discussed in Chapter Two, women s t i l l are responsible for the nurturing and day-to-day care of the family. Reduced time for mothers with their children can probably be compensated by 'quality time' after work hours but i f the mother i s very tired and suffering chronic aches and pains, then the after-work hours with the family may not be very satisfactory for any of it s members and the general domestic environment may be quite stressed. Such a scenario, i f i t exists, has implications for society in terms of the nature and role of the family group which in turn influences the next generation's hopes and expectations regarding marriage and family l i f e . The questions posed in this thesis: 1. Do women who participate in the paid labour force report poorer health status than their counterparts who are homemakers? 2. Do women who participate in the paid labour force exhibit lifestyle patterns significantly different from their counterparts who are homemakers? 3. Do women in the paid labour force exhibit health care utilization patterns significantly different from their homemaker counterparts? and, -187-4. Do women's lifestyles, reported health status and health care utilization patterns differ from those of their male counterparts in the paid labour force? attempt to address some of these issues. Further, the research was designed to examine the implications of women's changing roles in terms of their health care needs. 7.2 THE MODEL The C.H . S . data were chosen primarily because they provided a broad data base immediately available for analysis. Although there were limitations imposed on the study because the data base did not address a l l the issues, i t nevertheless offered an opportunity to become familiar with the issues and to do some preliminary hypothesis testing. The limitations of the C.H . S . data in terms of building the most complete model possible for this study, included: a lack of employment history information. Women tend to move in and out of the workforce more than men and the fact that they reported being in the workforce in the previous twelve months does not permit one to assume anything about work experience in the years previous to the one reported. the absence of some health behaviours identified as important by other researchers. These were diet and promiscuity (Lalonde, 1974); eating habits and sleep patterns (Belloc and Breslow, 1972). the inability to identify disability days, health professional consultations and hospitalizations associated with childbirth and reproductive functioning in general. For any study on the health of women, i t is important to be able to identify these items especially i f comparisons are being drawn against male or older populations. -188-the use of proxy reporting for health related behaviours including disability days, health professional consultations and medication consumption. This must surely reduce the accuracy of the data to varying degrees. Parents may be able to report accurately on behalf of children but is i t reasonable to believe that one adult, answering on behalf of other adults or adolescents in the house, would know about and remember visits by these others, to doctors and dentists? Would a man who is at work a l l day know whether his wife had a rest in the middle of the day because she wasn't feeling well? He probably would be better able to report her lif e s t y l e such as drinking, smoking, exercise and companionship but these were part of the survey that had to be f i l l e d in by each subject personally. missing data in the Lifestyle and Health Questionnaire. For example, a high percentage of women subjects (91.5%) answered the question about hormone p i l l consumption but a much lower percentage of the women (54.5%) reported on their alcohol consumption and only 36.3% of men answered a l l the questions concerning motor vehicle travel. This lack of answers could be attributed to: - reluctance to report what might be perceived as socially undesirable behaviour; - inability to recall the information easily; - a perception that the question does not apply to the respondent, hence failure to provide any response. The high level of missing data for some of the lifestyle questions reduced by more than 50% the number of subjects for which a composite risk score could be developed. Given these limitations, a model was developed (Figures 4.1-3) and -189-despite i t s inability to explain high proportions of variations in scores, some interesting patterns emerged. 7.3 HEALTH RISK BEHAVIOURS Workforce women generally reported health risk behaviours significantly different from their homemaker counterparts (see Table 6.78). On the positive side, they had more company during leisure time and they more consistently availed themselves of female screening behaviours. On the negative side, they did less exercise, were more likely to take hormone p i l l s , had more l i f e change events and were less likely to wear their seat belt. For two variables, workforce women also reported higher health risk behaviour than their male counterparts; they reported less exercise and more l i f e change events. The percentage of men reporting seven or more drinks in the previous week was the one aspect of their health risk behaviour that was overwhelmingly different from women's, (p<.002). Not as marked but s t i l l significant was the men's higher rate for l i f e events and motor vehicle travel. These three factors resulted in a significantly greater percentage of men (30.5%, p<.002) than homemakers (10.5%) or workforce women (11.8%) being in the highest risk group. These findings raise the question of what constitutes major health risk behaviours for women and are they different from those reported for men? The work by Breslow and Enstrom (1980) showed female mortality rates to be less affected by the non-observance of the seven health practices they had identified (see Table 2.A). Stamler and Epstein (1972) evaluated the relative risks of coronary heart disease from cigarette smoking, hypercholesterolemia -190-and hypertension, but worked only with male subjects. There would seem to be a need for more research into the possibility either that women's behaviour is less strongly related to their mortality (but possibly related to morbidity) or that there are other factors in women's lifestyles (such as role responsibilities and perceptions of autonomy) which have not yet been incorporated into the health risk model. The findings of this study, that B.C. women often reported the best scores for risk avoidance behaviours but s t i l l reported the highest proportion with one or more chronic health problems, should confirm that more research is needed to explain health risk behaviours for women in terms of morbidity as well as mortality. The model was able to explain only very small percentages of variance (around 4%) in the health risk behaviour scores, but some of the demographic variables fairly consistently appeared as the best predictors. As Table 6.81 shows, region was associated with six out of eight of the health risk behaviour scores (not l i f e events or companionship). The issues surrounding these regional differences are ones of health promotion, education and service provision. As noted earlier, although health services within Canada are broadly legislated under the Canada Health Act, within and across provinces the range of health promotion and health care services varies greatly. One important question is whether in a longitudinal study these variations in provision of services can be shown to affect morbidity and/or mortality patterns. For example, will the emphasis on cervical screening in B.C. provide that region with a mortality rate from cervical cancer that is less than the national average? Data from Kurlen and Doll (1973) showed that -191-in the ten years previous to their study the rate for 45 - 64 year olds was lower in B.C. than in the rest of Canada. More data like these are required i f service providers are to assign society's health care resource to the most effective programmes. The other demographic variables that were among the best predictors of health risk behaviour were marital status, family size and education. None of these is surprising. The protective effect of living with a family (except possibly as a single parent) has consistently been documented, (Anesheuel, Frerichs and Clark, 1981; Gore and Mangione, 1983). The significance of this finding to this study is that as mothers move into the workforce i t would seem important that the institutions of marriage and family not be altered to the extent that they become a burden more than a benefit for family members. The education effect shown by these data was not always consistent but generally, higher education was associated with lower rates of high risk behaviour. The benefits of more women seeking higher education should be, therefore, not only improved work opportunities but also healthier lifestyles. -192-7.4 HEALTH STATUS There was no significant difference between the reported health status of the two women's groups but the men's group showed significantly better rates (p<.05) on a l l three measures. The better male scores are consistent with the findings of other studies (Gove and Hughes, 1979; Waldron, 1983(b)), and this study offers no additional insight into why such a consistent difference exists. The lack of difference between the two women's groups may be a question of age. First, as i t was noted earlier, the consequences of health risk events and behaviours are not always immediate. Therefore although the groups showed different risk behaviours, the women may have been too young for the consequences to have become apparent. Second, the use of an age group as wide as 25 - 44 years may have diluted the consequences being experienced by women in the top end of the age span. One more convoluted possibility is that women in the workforce may be experiencing more stresses and strains but their very busy, involved lives may reduce their willingness to acknowledge and report chronic problems (supermom syndrome). Two such factors, i f they existed, could negate each other. 7.5 HEALTH CARE BEHAVIOURS There was no significant difference between the number of disability days or health professional consultations reported by the three groups. These findings are in keeping with those of Cleary, Mechanic and Greenley (1982) who found that when adjusted for differences in health status, health care utilization by men and women was similar. Contrary to this Verbrugge (1984) -193-reported lower health care utilization among employed women. In this study the percentage of employed women taking medications (24.1%, p<.02) was higher than for either homemaker women (17.8%) or workforce men (16.6%). Women use health care sei-vices for reasons other than illness. For example, cervical screening, professional breast examinations, hormone p i l l prescriptions and childbearing care a l l require regular health professional consultations but are not associated with poor health. Without more information regarding the nature of the consultations, i t is not possible to assess what proportion of the differences in the scores might be explained on these grounds. Another explanation for the greater use of consultations and medications by workforce women may be their need to maintain their busy work schedules. For example, for minor acute episodes such as throat infections, the workforce women may seek antibiotics rather than take days off work. This idea may appear at odds with the fact that workforce women report more disability days but as argued earlier, the higher score for workforce women may be an artifact of proxy reporting. 7.6 LIMITATIONS There were several limiting factors in this study beyond those imposed by the C.H.S. data. The age range 25 - 44 was too broad. During that age span women's responsibilities range from young very dependent families to independent children living at home; the not married status i s likely to be more common among younger members of the age group; and physical activity patterns are known to become more sedentary as people move into middle age. -194-Unfortunately, a l l age-related data were discarded when the data base was set up for this study. The high number of subjects excluded from the analyses on health status and health care consequences because they failed to provide sufficient data to generate a composite risk score raises the question of whether these people were different from those who answered a l l the questions. This study did not address this question. To know the reasons for subjects' health professional consultations and use of medications, although not required for the model, would have been very valuable for interpreting the results. These data were available on the C.H.S. tape but were not extracted when the data base was set up for this study. The study questions were very broad. While i t offered an overview of the many issues surrounding studies in women's health, i t provided no definite results. This was in part because no issue was explored in any great depth. The thesis leaves unanswered many more, and more complicated questions than i t answers. Despite this, i t allowed the student to become familiar with the issues and provided information which could help with, or even stimulate further study. 7.7 ISSUES FOR FUTURE STUDY For further study, three particular issues would seem to have greatest relevance. 1. What are the health risk behaviours for women? Are there factors such as role responsibility or perceptions of autonomy which could have a significant impact on women's health? What part does self-selection play in the apparent ability of some women to successfully enjoy multiple roles while -195-others struggle to cope with just one? Only a comprehensive longitudinal study can address these questions. 2. To what extent does health status reporting as done i n the C.H.S., r e f l e c t perceptions and attitudes toward health rather than health status i t s e l f ? Are men so c i a l i z e d to ignore aches and pains or do they deny them because they are at odds with the male image? If so, do women who are in the workforce with men learn to ignore (or take a p i l l and ignore) t h e i r minor symptoms? An a t t i t u d e - t o - i l l n e s s questionnaire i n conjunction with a health status questionnaire and an objective measure of functional status might answer some of these questions. 3. If women i n the workforce have a high need for health professional services, how are these best provided? Serious thought must be given to the type of services required and their a v a i l a b i l i t y . Could non-traditional, health professional-type c l i n i c s meet most of the women's needs? Would employer sponsored c l i n i c s be e f f e c t i v e for both the employer and the employee? To answer questions such as these, much more information i s required about the health care needs of women in the paid labour force. 7.8 CONCLUSION There are many other questions which when answered, would further our knowledge of women's health issues. This thesis has addressed just one aspect of an int e r e s t i n g , large and important topic. Men and women experience health d i f f e r e n t l y . In Chapter 2.3 some of the bi o l o g i c a l and s o c i a l reasons for t h i s difference were discussed and i t was -196-noted t h a t many of the s o c i a l r easons p r e v i o u s l y i d e n t i f i e d a r e a s s o c i a t e d w i t h women i n t r a d i t i o n a l l i f e s t y l e s . But women's o p t i o n s have expanded enormously i n the past 30 or 40 year s and many women choose t o , and are a b l e t o , e x e r c i s e those o p t i o n s . Others a r e r e q u i r e d t o , or p r e f e r t o , stay i n t r a d i t i o n a l r o l e s . T h i s j u x t a p o s i t i o n of women e x e r r . i s i n g d i f f e r e n t s o c i a l and l i f e s t y l e o p t i o n s p r o v i d e s a nation-wide o p p o r t u n i t y f o r f u r t h e r e x p l o r a t i o n of the s o c i a l f a c t o r s i n f l u e n c i n g the h e a l t h r e l a t e d e x p e r i e n c e s and behaviours of women. -197-SELECTED BIBLIOGRAPHY Allen, M.J.; Barne, M.R.; and Bodewala, G.G. 1985. " The Effect of Seat Belt Legislation on Injuries Sustained by Car Occupants." Injury 16: 471-6. Anesheuel, CP.; Frerichs, R.R.; and Clark, V.A. 1981. "Family Roles and Sex Differences in Depression." Journal of Health and Social Behaviour 22: 379-393. Arraitage, P. 1971. Statistical Methods in Medical Research. New York: John Wiley. Armstrong, P.; and Armstrong, H. 1978. The Double Ghetto. Toronto: McClelland and Stewart. Badwa, B. 1984. "Lifestyle and Health: Some Remarks on Different View Points." Social Sciences in Medicine 19(4): 341-347. Belloc, N.B.; and Breslow, L. 1972. "Relationship of Physical Health Status  and Health Practices." Preventive Medicine 1: 409-21; cited in J.H. Milsum, 1984. Health, Stress and Illness. New York: Praeger. Berkman, L.F.; and Breslow, L. 1983. Health and Ways of Living. New York: Oxford University Press. ; and Syme, S.L. 1979. "Social Networks, Host Resistance and Mortality." American Journal of Epidemiology 109(2): 186-204. Breslow, L.; and Eustrom, J.E. 1980. "Persistance of Health Habits and their Relationship to Mortality." Preventive Medicine 9: 469-83. Bradburn, N.M. 1969. The Structure of Psychological Wellbeing. Chicago: Aldine. Broyles, R.W.; Manga, P.; Binder; Angus; and Clarette. 1983. "The Use of Physician Services Under a National Health Insurance Scheme." Medical Care 21(11): 1037-54. Celento, D.D.; and McQueen, D.V. 1984. "Alcohol Consumption Patterns Among Women in Baltimore." Journal of Studies on Alcohol 45(4): 355-358. Christian, M.S. "Morbidity and Mortality of Car Occupants: A Comparitive Survey over Twenty-four Months." British Medical Journal 289: 1525-26. Cleary, P.D.; Mechanic, D.; and Greenley, J.R. 1982. "Sex Differences in Medical Care Utilization." Journal of Health and Social Behaviour 23: 106-119. Cooper, C.L.; and Melhuish, A. 1984. "Executive Stress and Health: Differences Between Men and Women." Journal of Occupational Medicine 26 (2): 99-104. Dever, G.E.A. 1976. An Epidemiological Model for Health Policy Analysis. Social Indicators Research 2; cited in J.H. Milsum. 1984. Health,  Stress and Illness. New York: Praeger. -198-Department of National Health and Welfare. 1982. Canadian Task Force on Cervical Cancer Screening Programmes. Ottawa: Ministry of Supply and Services. Diczfalusy, E. 1986. "New Developments in Oral, Injectable and Implantable Contraceptives, Vaginal Rings and Intrauterine Devices." Contraception 33(1): 7-22. Dreghorn, CR. "The Effect of Seat Belt Legislation on a District General Hospital." Injury 16: 415-8. Ellerton, M.L.; and Smillie, CL. 1986. "Demonstration of a Systematic Evaluation of a Breast Self-Examination Instruction Program Produced by a Non-Government Organization." Canadian Journal of Public Health 77: 296-300. Fleiss, J.L. 1981. Statistical Methods for Rates and Proportions. New York: J. Wiley and Son, 2nd ed. Foster, R.S.; Lang, S.P.; Costanza, M.C; Warden, J.K; Haines, C.R.; Yates, J.W. 1978. "BSE Practices and Breast Cancer Stage." New England  Journal of Medicine 299(6): 265-270. Gibbs, R. 1979. Lifestyle and Coronary Heart Disease. Melbourne: Sun Books; cited in J.H. Milsum. 1984. Health, Stress and Illness. New York: Praeger. Goldberg, E.L.; and Comstock, G.W. 1980. "Epidemiology of Life Events: Frequency in a General Population." American Journal of Epidemiology 111: 736-752. Goldman, N.; and Ravid, R. 1980. "Community Surveys: Sex Differences in Mental 111 ness." The Mental Health of Women. M. Guttentag, S. Salasin, and D. Belle, Eds., 1980. New York: Acaaderaic Press. Gore, S.; and Mangione, T.W. 1983. "Social Roles, Sex Roles and Psychological Stress." Journal of Health and Social Behaviour 24: 300-312. Gottlieb, B.H. 1985. "Social Networks and Social Support: an overview of research, practice and policy implications." Health Education  Quarterly 12(1): 5-22. Gove, W. 1984. "Gender Differences in Mental and Physical Health." Social  Sciences in Medicine 19(2): 77-91. ; and Hughes, M. 1979. "Possible Causes of the Apparent Sex Differences in Physical Health: an empirical investigation." American  Sociological Review 44: 126-146. Greenwald, P.; Nasca, P.C; Lawrence C.E.; Horton, J.; McGarrah, R.P.; Gabriele, T.; Carlton, K.; 1978. "Estimated Effect of BSE and Routine Physician Examinations on Breast Cancer Mortality." New England Journal  of Medicine 299(6): 271-273. Gritz, E.R. 1984. "Cigarette Smoking by Adolescent Females: Implications for Health and Behaviour." Women and Health 9(2&3): 103-115. -199-Hayney, CA. 1980. "Life Events as Precursors of Coronary Heart Disease." Social Science and Medicine. 14A: 119-126. Haw, M.A. 1982. "Work and Stress: A Review and Agenda for the Future." Journal of Health and Social Behaviour 23: 132-144. Health and Welfare, and Statistics Canada. 1981. The Health of Canadians: Report of the Canada Health Survey. Ottawa: Ministry of Supply and Services. Health and Welfare Canada. 1987(a). The Active Health Report -Perspectives on Canada's Health Promotion Survey. Ottawa: Health and Welfare. Hislop, G.T.; Coldman, A.J.; and Skippen, D.H. 1984. "Breast Self-Exaraination: importance of technique in early diagnosis." Canadian Medical Association Journal 131: 1349-1352. Holmes, T.H.; and Rahe, R.H. 1967. "The Social Readjustment Rating Scale." Journal of Psychosomatic Research 11: 213-18; cited in J.H. Milsum. 1984. Health, Stress and Illness. New York: Praeger. Jacobson, D.E. 1986. "Types and Timing of Social Support." Journal of  Health and Social Behaviour 27: 250-264. Jennings, S.; Mazark, C; McKinlay, S. 1984. "Women and Work: an investigation of the association between health and employment status in middle-aged women." Social Sciences in Medicine 19: 423-431. Johnson, P.B. 1982. "Sex Differences, Women's Roles and Alcohol Use: preliminary national data." Journal of Social Issues 38: 93-116. Jougla, E.; Bouvrer-Colle, M.H.; Maguin, P.; Diaz-Valdes, R.; and Minivielle, D. 1983. "Health and Employment of a Female Population in an Urban Area." International Journal of Epidemiology 12(1): 67-76. Kelner, M. 1985. "Community Support Networks: Current Issues." Canadian  Journal of Public Health 76(SuppI): 69-70. Kinlen, L.J.; and Doll, R. 1973. "Trends in Mortality from Cancer of the Uterus in Canada and in England and Wales." British Journal of  Preventive and Social Medicine 27(2): 146-149. Kurji, K.; and MacDonald, P. 1986. "Health Practices of Edmontonians." Canadian Journal of Public Health 77: 320. Ladbrooke, D.A. 1977. "Social Contexts of Premature Death in Contemporary America." Unpublished Ph.D. Dissertation, University of Wisconsin-Madison; cited in CA. Nathanson and G. Lorenz. "Women and Health: the social dimensions of biomedical data." in Women in the  Middle Years: Current Knowledge and Direction of Research and Policy. J.Z. Giele, Ed. 1982. New York: Wiley. Lalonde, M. 1974. A New Per spGctive on ths Health of Canadians* Ottawa! Information Canada. -200-Lewis, C.E.; and Lewis, M.A. 1977. "The Potential Impact of Sexual Equality on Health." New England Journal of Medicine 297(16): 863-869. McKeown, T. 1972. "An Interpretation of the Modern Rise in Population in Europe." Population Studies 27 (3); cited in M. Lalonde. 1974. A New Perspective on the Health of Canadians. Ottawa: Information Canada. McKeown, T. 1973. The Major Influences on Man's Health. Unpublished paper cited in M. Lalonde. 1974. A New Perspective on the Health of Canadians. Ottawa: Information Canada. Meissner, M.; Humphreys, E.W.; Meis, S.M.; and Scheu, W.J. 1975. "No Exit for Wives: sexual division of labour and the cumulation of household demands." Canadian Review in Sociology and Anthropology 12(4): 424-439. Milsum, J.H. 1984. Health, Stress and Illness. New York: Praeger. Morgan, M. 1980. "Marital Status, Health, Illness and Service Use." Social  Science and Medicine 14A: 633-643. Morrison, B. 1986. "The Periodic Health Examination: Breast Cancer." Canadian Medical Association Journal 134: 727-729. Nathanson, C.A. 1975. "Illness and the Feminine Role: A Theoretical Review." Social Science and Medicine 9: 57-62. . 1980. "Social Roles and Health Status Among Women; The Significance of Employment." Social Science and Medicine 14A: 463-471. ; and Lorenz, G. "Women and Health: The Social Dimensions of Biomedical Data." in Women in the Middle Years: Current Knowledge  and Direction of Research and Policy. J.Z. Giele Ed. 1982. New York: Wiley. Naisbitt, J. 1982. Megatrends. Nie, N.H.; Hull, H.C.; Jenkins, J.G.; Steinbrenner, K.; Bent, D.H. Statistical Package For the Social Sciences. New York: McGraw-Hill. Otto, R. 1979. "Negative and Positive Life Experience Among Men and Women in Selected Occupations, Sympton Awareness and Visits to the Doctor." Social Science and Medicine 13A(2): 151-164. Phillips, P.; and Phillips, E. 1983. Women and Work. Toronto: Lorimer. Pleck, J.H. 1985. Working Wives/Working Husbands. Beverly Hi l l s : Sage. Porter, J.B.; Hunter, J.R.; Jick, H.; Stergachis, A. 1985. "Oral Contraceptives and Non-Fatal Vascular Disease." Obstetrics  and Gynaecology 66: 1-4. Proulex, M. 1978. Five Million Women - A Study of the Canadian Housewife. Ottawa: Canadian Advisory Council on the Status of Women. Selye, H. 1974. Stress Without Distress. New York: Signet. -201-Social Sciences and Humanities Research Council of Canada. Women in the Canadian Labour Force. N. Hersom and D.E. Smith Eds. Ottawa: Social Sciences and Humanities Research Council of Canada. Statistics Canada. 1971. Census of Canada: Labour Force Activity - Work  Experience: Female Labour Force Participation by Schooling, Marital  Status, Age and Presence of Children for Canada and the Regions. Cat 94-774, Vol. 3, Part 7. Ottawa: Statistics Canada. . 1976. Census of Canada: Supplementary Bulletins; Economic Characteristics, Female Labour Force Participation Rates by  Level of Schooling, Age, Marital Status and Presence of Children. Cat 94-836. Ottawa: Statistics Canada. . 1979. Women in the Labour Force, Part II. Ottawa: Statistics Canada. . 1981. Religion: Census of Canada. Cat 92-912. Ottawa: Statistics Canada. . 1982. Code Book for the Canada Health Survey. Ottawa: Statistics Canada, Health Division. . 1984. Mortality. Ottawa: Statistics Canada. Stephens, T. 1985. "Healthy Lifestyles for a l l by the Year 2000: How are Canadians doing?" Health Promotion, F a l l : 2-6. ; Craig, C.L.; and Ferris, B.F. 1986. Adult Physical Activity in Canada: Findings from the Canada Fitness Survey. Canadian Journal of  Public Health 77: 285-290. Swan, C. 1981. Women in the Canadian Labour Market. Ottawa: Labour Market Development Task Force. Szalai, A. 1975. "Women's Time: Women in the Light of Contemporary Time-Budget Research." Futures 7(5): 385-399. Tousignant, M.; Guy, D.; and Lachapelle R. 1974. "Some Considerations Concerning the Validity and Use of the Health Opinion Survey." Journal of Health and Social Behaviour 15(3). Turabian, K.L. 1973. A Manual for Writers of Term Papers, Theses and Dissertations. 4th edition. Chicago: University of Chicago Press. University of Michigan. 1981. Osiris IV Users Manual, 7th Ed. Ann Arbour: Survey Research Centre. U.S. Department of Health and Human Services. 1985. "Workshop on Epidemiologic and Public Health Aspects of Physical Activity and Exercise." Morbidity and Mortality Weekly Report 34(13): 173-181. Verbrugge, L.M. 1984. "Physical Health of Clerical Workers in the U.S. Framingham and Detroit." Women and Health 9(1): 16-41. -202-Waldron, I. 1983 (a). "Sex Differences in Human Mortality: the Role of the Genetic Factor." Social Science and Medicine 17(6): 321-333. ;. 1983 (b). "Sex Differences in Illness Incidence, Prognosis and Mortality: Issues and Evidence." Social Science and Medicine 17 (16): 1107-1123. Walker, K.E.; and Woods, M.E. 1976. Time Use: A Measure of Household  Production of Family Goods and Services; cited in M. Proulx, 1978. Five Million Women - A Study of the Canadian Housewife. Ottawa: Canadian Advisory Council on the Status of Women. Wethington, E.; and Kessler, R.C. 1986. "Perceived Support, Received Support and Adjustment to Stressful Life Events." Journal of Health and  Social Behaviour 27(1): 78-79. Welsnack, R.W.; Welsnack, S.C.; and Klassen, A.D. 1984. "Women's Drinking and Drinking Problem Patterns from a 1981 National Survey." American  Journal of Public Health 74(11): 1231-8. WHO International Collaborative Study of Medical Care Utilization. 1973. Report on Basic Canadian Data. G.H. Josie Ed. Saskatoon: Dept. of Social and Preventive Medicine, University of Saskatchewan. -203-APPENDIX DATA ITEMS FROM THE C.H.S. USED IN THIS STUDY The following data for a l l men and women aged 25 - 44 were extracted. From the Interviewer Administered Questionnaire: Region of Residence Month of Survey Marital Status Size of Economic Family 10 Atlantic Provinces 20 Quebec 30 Ontario 40 The Prairies 50 British Columbia 01 July 78 02 August 78 03 September 78 04 October 78 05 November 78 06 December 78 07 January 79 08 February 79 09 March 79 01 Single (never married) 02 Married (including common law) 03 Widowed 04 Separated/divorced 01 Unattached individual 02 2 - 3 people 03 4 - 6 people 04 Separated/divorced Recoding: 01-03 - Summer 1978 (01) 04-06 - Fall 1978 (02) 07-09 - Winter 1979 (03) Recoding: Not married married (01) (02) Family Income 01 - 05 Income quintiles adjusted to reflect the size of the family and the municipality of residence. Principal Income Earner 01 02 Education 02 03 04 05 Principal earner for the economic family Not principal income earner Secondary (some or complete) Some post-secondary Post-secondary certificate or diploma University degree -204-Major Activity in past year 01 02 03 04 05 06 Working Keeping house Going to school 03 Retired or not working (health) Retired or not working (other) Baby or child 06 not included in this study. Activity Limitations 01 02 03 04 None Some Major Activity Limitations Cannot do Major Activity 01 only included in this study Disability Days in Previous Two Weeks 01 - 14 This is the sum of a l l days spent away from normal activity (whether in bed or not) and days of reduced activity due to poor health. Number of vis i t s in previous year to: Doctor Dentist Nurse Opt'rist Other 00 00 00 00 00 24+ 12+ 12+ 06+ 06+ Recoding: These numbers of visits were combined to create an 'Annual number of health professional v i s i t s ' Number of nights spent as a hospital patient in the previous year: 00 - 22+ An accident in the previous twelve months: Yes/No (Included only those accidents which resulted in injury and a limitation of normal activities.) Number of chronic health problems: 00 - 04+ Variety of medications taken at least once a week during the previous month: 00 - 04+ -205-From the self-administered Lifestyle Questionnaire: Affect Balance 01 Positive 02 Balanced (neutral) 03 Negative Health Opinion. Score 16 48 Derived from positive and negative scores on the Bradburb Items (see text) Derived from the 16 items of the McMillam Health Opinion Survey, (see text for details) Companionship for Leisure Time 01 Almost a l l by oneself 02 A lot by oneself 03 Half and half 04 A lot with others •Recoding: 01-02 - l i t t l e (03) 03 - moderate (02) 04-05 - lots (01) 05 Almost a l l with others Level of Physical Activity 01 Sedentary 02 Moderately inactive 03 Moderate 04 Moderately active 05 Very active Recoding: 01-02 - l i t t l e (03) 03 - moderate (02) 04-05 - active (01) (Based on reported leisure and work time activites) Alcohol Consumption 00 - 14+ Number of drinks in previous week Recoding: 00 - none (01) 01-07 - Av. one (02) 08-14+ - Av. two (03) Cigarette Smoking 01 Current smoker 02 Occasional 03 Former regular 04 Former occasional 05 Never Recoding: 01-02 - Current (03) 03-04 - Former (02) 05 - Never (01) Annual Distance as Passenger 00031 - 25000+ Annual Distance as Driver 00022 - 3000+ Passenger & driving distances combined & grouped: 01 Low through 4000 02 4001 through 11000 03 11001 through High Use of Seat Belt as: Passenger 01 Always 02 Rarely or never Driver 01 Always 02 Rarely or never Composite seat belt score created: 01 Always (as Driver & Passenger) 02 Variable 03 Rarely or never Life Events in Past Year 00 - 05+ -206-Data points for women subjects only. Most Recent Pap Smear 01 Less than 12 months ago 02 1-2 years ago 03 More than 2 yrs ago 04 Never 05 Not sure Recoding: 01-02 - Good (01) 03 - Moderate (02) 04-05 - Poor (03) Most Recent Professional Breast Exam 01 Less than 12 months ago 02 1-2 years ago 03 More than 2 yrs ago 04 Never 05 Not sure Recoding: 01-02 - O.K. (01) 03 - So,so (02) 04-05 - Poor (03) Frequency of Breast Self Exam 01 Once a month 02 Once every 2-3 months 03 Less often 04 Never 05 Don't know how Recoding: 01-02 - Good (01) 03 - So,so (02) 04-05 - Bad (03) Taking of Hormone/Birth Control P i l l s : 01 - 02 Yes/No Recoding: No (01) Yes (02) 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0096872/manifest

Comment

Related Items