Trends in incidence and risk factors for drowning among infants aged less than 1 and toddlers aged 1 to 4 in Canada from 1991--98 were compared to other injury deaths. Incidence, risk factors and in-hospital mortality of infant and toddler hospitalizations due to near drowning from 1994--98 were compared to other injuries. Drowning rates decreased by 79% among infants, from 1.4 per 100,000 person-years during 1991--94 to 0.3 during 1995--98 (0.001 < p < 0.0025 by chi2) and by 38% among toddlers, from 3.2 to 2.0 (p < 0.0005 by chi2). The rate of near drowning hospitalization among children aged 0--4 decreased by 30% from 1991--92 to 1997--98 (0.01 < p < 0.02 by chi2 for trends). Near drowning was the source of 5% of infant and 28% of toddler in-hospital injury deaths in 1994--98. The case-fatality ratio of near drowning hospitalizations was the highest of all injuries with 7% mortality among infants and 12% among toddlers. The decrease in incidence of drowning among infants and toddlers was not paralleled by a similar dramatic decrease in the incidence of other injury deaths in the same period in Canada, nor of near drowning hospitalization, and could be linked to prevention interventions from the Canadian Red Cross Society.

The objective of this thesis was to assess the acceptance of an online-questionnaire in a survey of young adults in southern Germany. The online-questionnaire was implemented in a study of respiratory diseases. Each of 280 subjects, aged 18 to 20 years, was assigned randomly to one of two groups who responded using (1) a paper-based questionnaire and (2) an otherwise identical questionnaire that was to be completed on the Internet. This group also received a copy of the printed questionnaire. Only eight subjects made use of the online-questionnaire, and the remainder of the respondents used the paper-based instrument. The major reason given for not using the online-questionnaire was that subjects considered it too time-consuming (39.7%). Additionally, 27.0% of the subjects did not have access to the Internet at home. In conclusion, the acceptance of an Internet-based questionnaire by adolescents with an Internet coverage of about 60% may still be low.

The objective of this study was to determine the two-year predictors of current smoking (smoked one or more times in the past year) and of ever smoking (lifetime use of cigarettes) among grade 4 students in multiethnic, low-income, inner-city neighborhoods in Montreal. Baseline and two-year follow-up data were collected from 501 children (54.3% girls) who, at baseline, were never-smokers. At the two-year follow-up, 22.0% reported ever smoking, including 7.6% who were current smokers. Gender-specific multivariate logistic regression analyses identified friends smoking (odds ratio (OR) = 3.0 95% confidence interval (1.2--7.1)) and sibling smoking (OR = 6.6 (1.7--22.2)) as independent predictors of ever smoking among boys. Friends' smoking (OR = 4.7 (2.0--10.7)) and school were predictors of ever smoking among girls. Parental smoking also predicted ever smoking (OR = 4.3 (1.6--11.9)) among girls born in Canada only. The only predictor of current smoking among boys was friends' smoking. Predictors of current smoking among girls included Canadian family origin.

Objective. Since the 1980's the incidence of tuberculosis (TB) in Montreal has remained at 11 cases per 100,000. In order to improve TB prevention and control programs, we sought to identify predictors of tuberculosis drug-resistance and to describe the epidemiology of TB drug resistance on the island of Montreal. / Study design. Retrospective descriptive analysis Study population. All culture proven TB patients reported to the Montreal Regional Health Board aged 0--49 for 1992--1994 and 0--18 years for 1995. / Results. Drug resistant TB was found in 18.3% of culture-proven TB cases. The rate of INH resistance in our study cohort was 10.6%. Two percent of TB cases were found to have MDR-TB. Only 3 TB cases (0.9%) in our study cohort developed acquired drug resistance over the study period. Previous history of TB was associated with a 3.9 times greater risk of drug resistant TB. / Conclusions. Drug resistance is a significant problem in Montreal that continues to hinder TB treatment and control. Previous history of tuberculosis is a strong predictor of drug resistance. In addition, immigration from individual countries was not associated with an increase in the rate of drug resistance. Nonetheless, country-specific drug resistance rates may serve to predict the likelihood of drug resistant TB among the foreign-born in Canada.

We performed a large (n = 878), multicenter, hospital-based, cross-sectional study to measure the health-related quality of life (HRQOL) of healthy participants treated for dyslipidemia and patients with coronary heart disease (CHD). The HRQOL measures included a nonpreference-based measure (SF-36 Health Survey) and three preference-based measures (Rating Scale (RS), Time Trade-off (TTO) and Standard Gamble (SG)). / The adjusted mean HRQOL, of healthy participants undergoing treatment for dyslipidemia was 2.8 (p = 0.02) and 3.3 (p = 0.02) points lower, when compared to a similar group of participants not being treated for dyslipidemia, on the RS and the SF-36 General Health Perception (GHP) subscale, respectively. These differences were unlikely to be due to confounding by comorbidity, age, gender and body mass index. No significant differences were detected on the TTO and SG scales. / For each preference-based scaling technique, the adjusted HRQOL mean scores obtained from patients diagnosed with angina, myocardial infarction, or angina and myocardial infarction were similar. Patients with congestive heart failure reported the worst HRQOL on all scales. / The test-retest reliability, over a 3 to 6 week period, was acceptable for all scaling techniques and the majority of participants reported consistent scores at the test and the retest assessments. Correlation between the preference-based measures and each of the SF-36 subscales varied from poor to moderate. Compared to the TTO and the SG, the RS was the most highly correlated with the different aspects of the HRQOL measured by the SF-36 Health Survey and had the highest ability to discriminate CHD patients with various physical disabilities and participants reporting specific number of health problems. / This study provides a complete set of preference-based measures for use in cost-effectiveness analysis of CHD primary prevention. It suggests that the impact of detecting and treating dyslipidemia on the participant's HRQOL may be small but significant from a public policy point of view. Further research should be done confirming these results and elucidating the causes and the consequences of this negative effect on HRQOL of healthy individuals treated for dyslipidemia.

Introduction. Trauma triage protocols are important because they identify, at the injury scene, patients with major injuries requiring transfer to a Level I trauma center, from those with non-major injuries who could be treated at Levels II and III trauma centers. The Pre-hospital Index (PHI) is a physiological injury severity measure which may be used as a trauma triage tool. / Purpose. The purpose of the present study was to: (1) prospectively evaluate the predictive ability of the PHI in identifying trauma patients with major versus non-major injuries, and (2) develop a trauma triage scale which incorporates, along with the PHI, a subset of the variables age, body region injured, mechanism of injury, comorbidity, and time between 911 call and departure of the ambulance from the injury site, so as to improve the predictive ability of the PHI-based triage instrument. / Methods. This study was based on 1,291 trauma patients treated in Montreal between April 1993 and December 1996. A patient was considered to have major injuries if the patient died within seven days since hospital admission, had an intensive care unit admission within seven days, or major surgery performed within four days. Three hypothetical trauma triage protocols were developed using logistic regression analysis; where the model that describes the data best was selected according to Bayes factor approximation. In detecting major versus non-major injuries, sensitivities, specificities, positive and negative predictive values were calculated for all the cutoff points of the PHI and the triage protocols. Also, areas under the Receiver Operating Characteristic (ROC) curves were calculated and compared for these instruments. / Results. The trauma triage protocol which included the variables age, body region injured, mechanism of injury, comorbidity, and PHI produced the best combination of sensitivity and specificity; of 0.95, and 0.24, respectively. This algorithm underwent a significant improvement over the PHI (area under the ROC curve: 0.76 versus 0.66, p < 0.05). / Conclusion. An improvement in the predictive ability of the PHI-based triage instrument was introduced after the addition of the variables age, body region injured, mechanism of injury, and co-morbidity.

Background. Benzodiazepines are sedative-hypnotic medications frequently prescribed in elderly patients for several clinical indications. An association with increased risk for falls has been reported but there is continued debate regarding which specific benzodiazepines are associated with this risk. / Objectives. To estimate the risk of injuries from falls associated with benzodiazepine use in an elderly cohort taking into account patient characteristics and changes in patterns of use over time. / Methods. Using information from provincial administrative health databases, 462,543 community-dwelling, 66 year old Quebec residents were screened for benzodiazepine use in 1989. Subjects who did not use benzodiazepines in 1989 were observed for the next five years to estimate incidence rates and evaluate patient characteristics associated with new use for thirteen benzodiazepines. Patterns of use for incident users were characterized in terms of duration, dose and frequency of switching or adding benzodiazepines. New methods were developed to model the past cumulative dose and duration of benzodiazepine exposure. The impact of benzodiazepine exposure on risk of injury was estimated using Cox proportional hazards analyses with time-dependent covariates to take into account changes in dose and patterns of use. / Results. The overall incidence rate for benzodiazepines was 88.7 per 1,000 person-years, with higher rates in women (95.0) than men (81.8). Predictors of incident use were different in individual products and there were systematic differences between users and non-users. Use of anti-depressants in 1989 was the strongest predictor for incident benzodiazepine use (HR 1.45 to 3.07, p < 0.0001). The median duration for uninterrupted periods of use was 31 days (mean = 75.5 days, sd = 137.2). The mean dose was almost half the recommended maximum adult daily dose and only 8.6% of subjects exceeded the maximum. Older age at date of first prescription significantly increased the likelihood of increasing duration and dose overtime (OR = 1.02, p < 0.0001). All benzodiazepines except clonazeparn were significantly associated with an increased risk of injuries from falls (p < 0.05). The best predictive model for most benzodiazepines included a cumulative measure of duration and current dose. / Conclusion. Benzodiazepines are associated with an increased risk of injuries from falls in elderly patients, however duration of exposure may be more critical than dose. Physiological dependence and withdrawal symptoms appear to play an important role in increasing the risk for many benzodiazepines.

No description available.

This thesis comprises a review of studies on the diffusion of recommended guidelines on preventive practices and three papers evaluating the preventive practice patterns of general practitioners in Torino, Italy. / The first paper presents estimates of preventive practices and perceived effectiveness of preventive interventions. Patterns of practice were found to be consistently similar to those in other studies. / The second and third papers explore an array of determinants for ten primary and secondary preventive interventions. / An analysis of determinants indicates that complex patterns of behaviours are rather condition-specific, while the application of techniques is influenced to a varying extent by organizational and attitudinal factors.

An epidemic of hepatitis B occurring in a rural area of Nova Scotia in 1988 and 1989 was investigated. Illicit injectable drug use (IIDU) was the major determinant of transmission. The epidemic was the first highly visible indication of IIDU in Nova Scotia. A contact-tracing approach was used to identify the cohort of IIDUs. Of 186 IIDUs, 78 had serological evidence of hepatitis B infection. Using epidemiological criteria, it was determined that 57 of these formed a definite cluster of hepatitis B infections. Age, the total number of IIDU-contacts named and the number of hepatitis B seropositive IIDU-contacts named were identified as risk factors for hepatitis B infection. Six cases of delta hepatitis coinfection were found among the cluster cases, with a secondary attack rate estimated at seven percent. Risk-reduction and immunization strategies for the prevention of viral illnesses associated with IIDU were recommended.