This population-based multi-centered case-control study of incident postmenopausal breast cancer was carried out in Montreal. Case subjects included all incident histologically confirmed breast cancer among postmenopausal women, age 51--75 years, diagnosed in 1996 and 1997 in all major hospitals. Control subjects had other histologically-confirmed sites of cancer from the same hospital, and were approximately frequency-matched by age to the case subjects. Detailed information was obtained on active smoking and exposure to environmental tobacco smoke (ETS) in the home before the age of 18 and in the workplace as an adult. Unconditional logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI) between the risk of postmenopausal breast cancer and various indices of smoking and exposure to ETS. The OR for breast cancer among women with exposure to active smoking only was 0.71 (9596 CI: 0.34--1.48), for early exposure to ETS only, the OR was 1.09 (95% CI: 0.65--1.85), and for occupational exposure to ETS only, the OR was 1.00 (95% CI: 0.53--1.90). When occupational and early exposures to ETS were combined, the OR was 1.40 (95% CI: 0.81--2.40). In summary, I found no associations between the incidence of postmenopausal breast cancer and active smoking or exposure to ETS.

Background. Practice guidelines for screening mammography diverge on the recommended age at which systematic screening should begin for women aged 40 to 69 years. 'Routine' guidelines recommend screening starting at 40 years; 'age-selective', at 50 years and only for younger women who have a first-degree family history of breast cancer. Objectives. (1) To estimate and compare recently-licensed family physicians' adherence to 'routine' and 'age-selective' guidelines in women aged 40 to 69 years, (2) To identify intrinsic physician and practice environment factors which predict referral rates for women aged 50 to 69 years as well as age-selective referral. Study Design. A census sample of 743 recently-licensed family physicians was followed for 18-months of incipient medical practice. Mammograms were identified from physician claims to the Regie de l'assurance-maladie du Quebec for every clinically eligible woman aged 40 to 69 years who had a complete physical examination by a study physician in a community setting (primary care practice). Adherence to 'routine' guidelines was measured as the percent of women with a screening mammogram referred by the study physician in the study period. Adherence to 'age-selective' guidelines was referral rates in older (50-to-69 years) minus younger (40-to-49 years) women. Multiple linear regression was used to identify predictors of mammography referral. Results: Study physicians referred 6% (SD: 10) of younger women and 12% (SD: 15) of older women in their primary care practice. The correlation or ranks of adherence to 'routine' and 'selective' guidelines corresponded poorly (Spearman r = 0.51). Factors positively associated with higher referral rates in women aged 50 to 69 years were female gender, prevention competence, later medical school graduation, the combination of comprehensive inquiry and continuity care, and the extent of community-based care. Negatively associated factors were higher patient volume and shared primar

Background. The early diagnosis and treatment of breast cancer has become an important health care concern. A recent study reported the median waiting time for breast cancer surgery in Quebec was 34 days with 14% of women waiting in excess of 90 days. Understanding the determinants of long waiting is essential to develop optimum interventions to reduce delay. Objective. The purpose of this study was to identify the determinants of waiting time to surgery among women with primary breast cancer in Quebec between 1992 and 1997. Methods. The target population was all women 20 years and older diagnosed with primary breast cancer in Quebec between 1992 and 1997. The data was compiled from physician fee-for-service claims maintained by the Regie de I'assurance maladie du Quebec (RAMQ); the Quebec hospital discharge database (MedEcho), and the 1991 Canadian census. Waiting time was defined as the number of days from the initial breast diagnostic procedure to the first definitive surgical treatment. Three-level hierarchical linear models were used for statistical analysis. Findings . Overall, 13,383 women with primary breast cancer treated by 614 surgeons in 107 hospitals were identified. No statistically significant variation of waiting time was found among hospitals. Longer waiting times for breast cancer surgery were observed for women 50 to 64 years of age, without comorbidity, with history of benign breast disease, living in the lower education areas, having surgery at day-surgery setting, having surgery in more recent years, or having surgery performed by younger a surgeon (20 to 49 years old). Women who had surgery performed in a teaching hospital had longer waiting times and this effect was larger when mastectomy was performed. These results could be used to identify women and care delivery practices at higher risk for delays which could be the focus of interventions.

This thesis comprises three independent but interrelated manuscripts. (1) The results from seven epidemiological studies of adult cancers in relation to residential exposure to power frequency magnetic fields (MF) indicated that the association between leukemia and MF has been inconsistent. It also indicated that the risks of brain tumors and breast cancer, the other cancers frequently suspected of being associated with occupational exposure to MF, were rarely investigated. Based on these epidemiological results, the analysis of the roles of chance and bias, and the criteria for causal inferences, it appears that the evidence is not strong enough to support the putative causal relationship between residential MF and adult cancers. (2) A case-control study, using matching on date of birth, sex, and date of diagnosis, was therefore carried out in northern Taiwan to further evaluate the risks of adult ($ sbsp{=}{>}$ 15 years of age) leukemia, brain tumors, and female breast cancers in relation to residential exposure to 60 Hz MF. Cases were newly diagnosed cancers reported to the cancer registry between 1987 and 1992 and controls were incident cancers from sites other than those previously suspected of being associated with MF during the same period. Assessment of MF in the residences occupied by the study subjects at the time of diagnosis was performed by modeling power information of high-voltage transmission lines. The results were based on the separate analysis of 708 leukemia, 455 brain tumors, and 1,562 female breast cancers. The risk of leukemia for exposure to MF $>$ 0.2 $ mu$T relative to the reference level ($<$0.1 $ mu$T) was significantly elevated (odds ratio = 1.51, 95% confidence interval 1.05-2.19). A dose-response relationship showed a gradient increase in relative risk estimates for leukemia with MF. The relative risk estimates for brain tumors and female breast cancers were slightly elevated, but were statistically compatible to null. (3) To valida

Exposure to drugs is a complex phenomenon. A method to characterize long term patterns of drug use amenable to analysis in the framework of risk assessment was developed. The impact of patterns was assessed with the effect of non-steroidal anti-inflammatory drugs (NSAIDs) on upper gastro-intestinal bleeding (UGIB). A hospital based case-control study was conducted (244 cases age 68 and over, 615 matched controls). Objective data on all medications dispensed during the three years preceding admission were obtained from the Quebec prescription database. Patterns were expressed in terms of recency, duration and switches. Past use of NSAIDs was associated with a lower risk of UGIB. The risk decreased as the duration of use increased. Patients who switched products were at greater risk. Concomitant use of protective agents, history of gastropathy and current smoking increased the risk significantly. Patterns of drug utilization should therefore be considered to identify groups at high risk for adverse event.

Research has indicated that the adjustment of psychiatric patients is adversely affected by family criticism. The present study elaborated on this theme by focusing on fifty schizophrenic patients who were formally interviewed, using a standard visual analog scale format, about their perception of family criticism. Two important conceptual distinctions were proposed. The first was the extent to which criticism was perceived as "constructive" or "destructive", and the second was the extent to which the criticism was perceived to be "person directed" or "behavior directed". Regression analyses were performed in order to determine the predictive utility of the various dimensions of criticism on patient outcomes. The results indicate first, that criticism is not always perceived as negative and secondly, patients who perceived high levels of negative criticism (destructive, person directed) had relatively poor outcomes. Those who perceived high levels of positive criticism (constructive, behavior directed) had more positive outcomes. The results therefore indicate that criticism is multidimensional. This awareness of different meanings of criticism may lead to early identification of adjustment difficulties and subsequent interventions to prevent them.

Objectives. Few studies have identified longitudinal predictors of smoking cessation in disadvantaged communities. This study identified predictors of cessation in a 5-year longitudinal cohort of adults aged 18-65 years and living in low-income, inner-city neighborhoods of Montreal, Canada. / Methods. Secondary analysis of data from the non-randomized evaluation of Coeur en Sante St. Henri, a community-based intervention program designed to decrease cardiovascular disease risk (CVD) factors. Data on lifestyle behaviors were collected in telephone interviews of a representative sample of residents at baseline and five years later. Independent predictors of cessation were identified among 303 subjects who smoked at baseline, using multiple logistic regression. / Results. After 5 years, 20% of baseline smokers reported quitting including 22% of female smokers, and 17% of male smokers. From among 7 potential predictors only two were retained in multivariable analysis, including having a post-secondary or higher education relative to secondary school or less (OR=1.88, 95%CI: 1.01-3.51), and number of cigarettes smoked per day (OR=0.95, 95%CI: 0.91-0.98). / Conclusions. Few predictors of cessation emerged in this disadvantaged community. It is notable that even in a disadvantaged community, increased education predicts cessation. Improved understanding of the mechanisms by which education leads to higher quit rates may help the development of cessation programs targeting disadvantaged communities.

The epidemic of Type 2 diabetes in Aboriginal peoples can be curbed by promoting healthy physical activity habits early in life. Purpose . To identify childhood physical activity correlates. Methods . 383 4th--6th grade Mohawk children (193 boys, 190 girls) completed questionnaires and had anthropometric measures taken during 2 serial cross-sectional surveys as part of the Kahnawake Schools Diabetes Prevention Project (1994--1997). Physical activity was measured using the Weekly Activity Checklist. Demographics, community activities, parental variables, self-efficacy, sedentary activity, and overweight were correlates assessed using ordinal logistic regression. Results. 29.7, 33.4, and 36.8 percent were inactive, moderately active, and very active, respectively. Odds ratios and 95% confidence intervals [OR(95%CI)] of independent correlates for boys included year of measurement (1996 vs. 1994) [1.7(1.3,2.2)], older age [0.7(0.5,0.9)], and enrollment in school #1 [2.0(1.1,3.6)], lessons [2.8(1.3,5.9)] and summer sports [1.1(1.0,1.1]. Independent correlates for girls were year of measurement [1.5(1.1,2.1)] and parental physical activity [2.0(1.2,3.3)]. Conclusions. Parental models are stronger influences for girls' activities while community factors influence boys' activities.

Cancers of the upper aero-digestive tract (UADT) rank as the fifth most common neoplastic disease worldwide. Two identified risk contributors are consumption of tobacco and alcohol. Among all other potential etiological factors, diet has long been recognized to play an important role in the development of cancers of the UADT. Data from a multi-centre, hospital-based case-control study conducted in Brazil were used to assess the association of dietary intake with the risk of cancers of the UADT. Dietary assessment was made in terms of estimated intake of nutrients, specific foods and food groups. After adjusting for the effects of alcohol and tobacco consumption as well as empirical confounders, protective effects against cancer of the mouth (Odds Ratio (OR) = 0.61, 95% confidence interval (95% CI): 0.4--1.0) and the pharynx (OR = 0.51, 95%CI: 0.3--0.9) were found for consumption of citric fruits. (Abstract shortened by UMI.)

Atherosclerotic narrowing of the carotid arteries in asymptomatic individuals is associated with an increased risk of cerebrovascular ischemic events. The prevention of stroke in this context requires an understanding of the natural history of cerebrovascular atherosclerosis and the availability of an effective means for the detection and follow-up of carotid stenosis. / Methodological standards for the conduct and reporting of natural history studies of asymptomatic carotid disease are proposed and used to carry out a structured methodological appraisal of the literature. The findings suggest that greater attention should be paid to the design and statistical analysis of natural history studies. / A secondary analysis of the data from a natural history study of asymptomatic carotid disease is undertaken for the purpose of evaluating the prognostic usefulness of repeated duplex ultrasound examinations. Multivariable Cox proportional hazards models with fixed and time-dependent covariates are used to evaluate the predictive ability of carotid stenosis measurements obtained by duplex ultrasound at enrollment and during follow-up. The results of the analysis do not support the use of routine carotid ultrasound in clinical practice to determine prognosis for unselected patients with asymptomatic carotid disease.