Children with chronic physical illness have twice the risk of psychosocial maladjustment as healthy children. One specific aspect of mental functioning is abnormal eating attitudes, hypothesized to be an early stage of an eating disorder. To date, there has been no study of the association of a variety of physical chronic illnesses with a specific psychiatric syndrome in a diagnostically heterogeneous population, comparing chronically ill children with healthy children. This study examined the association between chronic illness and eating attitudes in a population of school children from Western Quebec. A checklist of chronic medical conditions was added to the parent questionnaire in order to determine presence of chronic illness. There were two objectives of this thesis. The first objective was to determine whether chronic illness was associated with eating attitudes in children as measured by the Eating Attitudes Test. The second study objective was to determine whether there was a meaningful group difference in othcr psychometric scores of depression, family functioning and stress between children with a chronic illness and children without a chronic illness. (Abstract shortened by UMI.)

Objectives. The purpose of this study was to analyze prostate cancer trends in Canada, determine whether the observed trends are associated with earlier detection, assess the association between prostatectomy rates and prostate cancer incidence rates and assess other possible reasons to explain the observed trends. Conclusions. Correlations between prostate cancer incidence rates and prostatectomy rates suggest that increased surgical treatment of benign prostatic disease contributed to the increase in incidence rates through increased detection of latent cancers. This hypothesis is supported by the chart review, which is the first work to show an association, other than an ecological one, between TURPs and the increased detection of prostate cancer. The increase in early stage cancers, especially incidentally discovered cancers, and the discovery of increased scrutiny of surgical specimens by histopathological staff, corroborates the ecological data. Although elevations in unestablished risk factors may have contributed to the observed increase in incidence, much of the increase can be attributed to an increase in rates of localised disease. This suggests that the increases may be due to early detection and not risk elevation. (Abstract shortened by UMI.)

PREVENT is a user-friendly, interactive computer program developed in the Netherlands which integrates the effects of demographic structure, risk factor prevalence, relative risks and mortality rates to estimate the impact of changes in risk factor prevalence on mortality. A new database for PREVENT has been developed with the following Ontario data: population structure, all-cause and disease-specific mortality, life expectancy, and birth projections. The current prevalence of smoking, obesity, alcohol and seat belt use is estimated from the Ontario Health Survey; the prevalence of hypertension and hypercholesterolemia from provincial heart health surveys. Future changes in risk factor prevalences for the simulation period are extrapolated from past trends. Sensitivity testing shows that the mortality varies directly with both the prevalence and relative risks. The shape of the mortality prevention curve is affected by two factors. In the first twenty years the latency and lag times and the spread in implementation of the intervention dictate the pattern. In later years mortality is affected by the changing demographic structure, age-specific mortality rates and future trends in risk factor prevalence that are in the model. Historical testing of the risk factor smoking, using a 1966 Ontario database, shows good agreement between PREVENT estimates and observed total mortality. The agreement between predicted and observed mortality is not as close for lung cancer, IHD and COLD.

Two models of risk for sexual behaviour were developed using stepwise discriminant analysis. The first model assessed the predictive strength of demographic, knowledge, attitudinal, and behavioural variables in determining which adolescents "ever had sexual intercourse" compared to adolescents who "never had sexual intercourse". The second model assessed the same variables to determine which adolescents were more likely to have had an increased number of partners. For males, the model of risk for "ever had sexual intercourse" yielded nine variables which accounted for 27.6% of the discriminatory power of the model. The nine determinants in order of entry to the model were: alcohol use, cigarette use, cannabis use, positive attitude towards casual sex, high self-esteem/self-image, average mark last term less than 60%, weekly church attendance, poor relationship with parents. and do not know mother's birthplace. For females, the model of risk for "ever had sexual intercourse" yielded fourteen variables which accounted for 27.5% of the discriminatory power of the model. The fourteen determinants in order of entry to the model were: cigarette use, positive attitude towards casual sex, cannabis use, alcohol use, subjects in high school were preparing the student for work, urban/rural, live with "other/mom only/dad only", finish education at high school, do not know mother's birthplace, father's occupation was unemployed, poor relationship with parents, average mark last term less than 60%, low knowledge score, and low sexual worry. Predictive models of health, notably the Health Belief Model, The Theory of Reasoned Action and the Social Learning Theory, have been developed within specific theoretical frameworks and have shown promise in delineating the determinants of behaviour. (Abstract shortened by UMI.)

Objective. To determine the prevalence of cardiac risk factors in offspring of patients with premature ischemic heart disease, their awareness of these risk factors, and to find out what proportion of offspring have had risk factor assessment including cholesterol screening and what proportion have adopted risk factor reducing strategies. Main results. The median age of the offspring was 20 years (range 2-39). Among the late adolescent and young adult progeny 37% were smokers, 31% were overweight and 30% exercised less than three times per week. Although 78% had been examined by a physician in the preceding three years, only 97 (44%) reported having had a blood cholesterol measurement performed during that time and only five of the 97 actually knew their cholesterol levels. Fifty-seven percent of the males had had a blood pressure measurement in the previous year. Although all respondents were aware that eating fatty food could contribute to heart disease and cholesterol elevation, few (13%) recognized the role of heredity as a causal factor for heart disease and only 22% reported that they would know how to lower their blood cholesterol. Conclusions. The low rates of cardiovascular risk factor assessment and management identified in this survey represent missed opportunities for primary prevention. (Abstract shortened by UMI.)

Background. Guillain-Barre syndrome (GBS) is of public health interest in Canada, as well as the rest of North America, for two main reasons. It is occasionally a vaccine-associated adverse event and is also a differential diagnosis of poliomyelitis. Objectives. (1) To ascertain the incidence of GBS in the Canadian provinces of Ontario and Quebec for the years 1983-1989, inclusive. (2) To demonstrate the feasibility of measuring the incidence of GBS through internal record linkage of Canadian hospital-service data. Results. 1,302 and 1,031 records representing GBS incident admissions in Ontario and Quebec, respectively, were identified through the record-linkage procedure. The mean annual GBS incidence after age-and-sex-standardization to the 1986 Canadian census population was 2.02 per 100,000 person-years in Ontario and 2.30 in Quebec. The incidence was higher in older age-strata in both provinces (70-80 years), and was higher in males (M:F = 1.1). Reviews of charts of incident admissions of GBS cases reveal that 26.2%-32.6% of Ontario cases and 21.0%-24.0% of Quebec cases may be false positive diagnoses. No possible false negative cases were identified through chart review. Cross linkage of records belonging to the other province with records from the other dataset revealed 0.5% false negative misclassification of Ontario incident admissions and 1.8% for Quebec. Mortality figures obtained from CMDB were in both provinces less than those obtained in the hospital service data, indicating that it is unlikely a significant number of GBS cases die before reaching hospital. Conclusions. (1) It is possible to internally link records in the HMRI and Med-Echo databases into personal histories (cases) of a condition. (2) The high percentage of false positive misclassifications discovered on examination of incident admissions raises concern about the validity of HMRI and Med-Echo data for epidemiological purposes. (Abstract shortened by UMI.)

The purpose of the present study was to predict children's posthospitalization behavioural difficulties consequent to day surgery. Temperament characteristics and characteristics of the mother-child relationship were used to predict children's behavioural difficulties at three days, three weeks and two months posthospitalization. The mothers of the 57 subjects rated their child's pre- and posthospitalization behaviour on the Vernon Hospital Questionnaire. The Behavioural Style Questionnaire was used as the temperament measure and the Parenting Stress Index was used to assess the degree of stress within the mother-child relationship. Repeated measures analysis of variance on the Vernon Hospitalization Questionnaire indicated that there were significant differences among children's scores at the pre- and posthospitalization periods. Subsequent analyses indicated significant differences on four of the six subscales of this questionnaire. HMRs were conducted on children's behavioural difficulties with three temperament variables and the total stress variable entered as predictors. The results of these HMRs indicated that mood and total stress predicted behavioural difficulties at pre- and two month posthospitalization, whereas adapt and total stress predicted behavioural difficulties at three days and three weeks posthospitalization. Assessment of the predictive accuracy of the sample regression equations by the jackknife procedure indicated that the regression equations provide accurate predictors. Recommendations are proposed concerning how to identify children at risk for posthospitalization behavioural difficulties.

The purpose of this study was to develop and test a measure of quality of life meant for use in an outpatient cardiac population. This study was conducted in three phases. In the first, the aspects (elements) of daily living which should be included in a measure of life quality were determined. A total of 114 subjects were asked to indicate every item in a list of elements to be important or not important to the quality of daily living of cardiac outpatients. The subjects were comprised of: (a) clinic patients in whom permanent cardiac pacemakers had been implanted, (b) clients attending exercise classes at the cardiac rehabilitation and prevention centre, (c) staff of the University of Ottawa Heart Institute working in the pacemaker clinic and the rehabilitation and prevention centre, and (d) family members of the patients/clients. In the second phase a review, of existing quality of life questionnaires, was done to find questions which related to the twenty-two elements considered by the patients/clients to make important contributions to the quality of their daily living. The questionnaire, called the Quality of Life Assessment Package (QLAP), consists of the Duke Activity Status Index, in its entirety, and sections from a number of other questionnaires. The reliability and validity of the QLAP was tested in the third and final phase of the study. Two hundred and twenty-two subjects completed a demographic profile, the QLAP, and a modified version of Spitzer's Quality of Life Uniscale. An analysis of variance, and post-hoc testing showed that total scores on the QLAP differed significantly between chronically ill, acutely ill and healthy subjects. In conclusion: (a) the QLAP is internally consistent (reliable), (b) the QLAP has content validity, (c) the criterion related validity has been supported, and (d) evidence has been provided to support the construct validity of the QLAP. (Abstract shortened by UMI.)

In 1983, the Ontario government introduced legislation which mandated the restraint of all child passengers under 50 pounds in restraint systems appropriate to their weight. The efficacy of such restraint systems has been proven, and standards for their manufacture have been set by the government of Canada. The objectives of this study were to determine whether Ontario's child restraint legislation has resulted in significant and sustained reductions in the incidence and severity of injury, as well as changes in the anatomical patterns of fatal and non-fatal injury. Data sources included the Hospital Medical Records Institute (HMRI) and Transport Canada's Traffic Accident Information Data Bank (TRAID). The study used the interrupted time-series design and autoregressive integrated moving average (ARIMA) modelling to analyze a variety of rates and indicators calculated for children under 5, as well as for a control group of adults 20-44 years old. The study period encompassed five pre-law years and five post-law years, extending from 1979 to 1988. Results suggested that there were no significant changes in either the incidence, severity or anatomical patterns of injury in children covered by the legislation. Two possible explanations for the apparent lack of an intervention effect are that either the pre- to post-legislation increase in (proper) usage rates was not large enough, or that any positive effects of the law were too gradual to be detected. These conclusions are supported by evidence that there was a relatively high usage of child restraints prior to the legislation, and that rates of proper wearing of restraints did not exceed 50% in the years following implementation. Child restraint laws cannot achieve their objectives without a higher level of sustained compliance. Several means of achieving this objective are presently being implemented by the so-called "95 by 95" program, i.e., the National Occupant Restraint Program that is operating in all provinces, and whose stated goal is a 95% wearing rate of restraints in all age groups by the year 1995. (Abstract shortened by UMI.)

Thee Ontario Community Health Profile was developed in the early 1990s in response to a need for better measurement of community health in the province. This thesis conceptually and empirically evaluates the profile's relevance, representativeness and practicality for measuring community health in Ontario public health units. Based on a review of the literature and various models of community health is defined as more than simply the absence of disease or disability among a collection of individuals. Health must encompass both disabling and enabling characteristics---defined here as measures of negative or positive health---that are related to coping ability. Moreover, community health indicators should encompass levels of measurement that are global and environmental (representing health 'of' the community) as well as aggregate measures (representing health 'in' the community). This paper develops a community health framework---with axes representing the definition of health and the level of community---in order to evaluate indicators according to these two important dimensions. Empirical analyses used data from the HEaLth Planning System (HELPS) data set for Eastern Ontario. (Abstract shortened by UMI.)