Methodological biases in estimating the relationship between alcohol consumption and breast cancer: the role of drinker misclassification errors in meta-analytic results.

BACKGROUND: While alcohol consumption has been linked to breast cancer in women, few studies have controlled for possible biases created by including former or occasional drinkers in the abstainer reference group. We explored the potential for such misclassification errors as sources of bias in estimates of the alcohol-breast cancer relationship. METHODS: Meta-analyses of population case-control, hospital case-control, and cohort studies to examine relationships between level of alcohol use and breast cancer morbidity and/or mortality in groups of studies with and without different misclassification errors. RESULTS: Of 60 studies identified, only 6 were free of all misclassification errors. The abstainer reference group was biased by the inclusion of former drinkers in 49 studies, occasional drinkers (<10 g ethanol [EtOH] per week) in 22 and by both these groups in 18. Occasional drinkers were also mixed with light or hazardous-level drinkers in 22 studies. Unbiased estimates of the odds ratio (OR) for breast cancer were 1.011 (95% confidence interval [CI]: 0.891 to 1.148) among former drinkers (n = 11) and 1.034 (95% CI: 1.003 to 1.064) among occasional drinkers (n = 17). Hazardous-level drinking (>20 g < 41 g EtOH/d) was not significantly associated with breast cancer in studies with occasional drinker bias. However, in studies free from occasional drinker bias, the OR for breast cancer was 1.085 (95% CI: 1.015 to 1.160) for low-level (<21 g/d) drinkers (n = 17), 1.374 (95% CI: 1.319 to 1.431) for hazardous-level drinkers (n = 26), and 1.336 (95% CI: 1.228 to 1.454) for harmful-level (>40 g/d) drinkers (n = 9). CONCLUSIONS: While the great majority of studies of the alcohol-breast cancer link include misclassification errors, only misclassification of occasional drinkers was found to bias risk estimates significantly. Estimates based on error-free studies confirmed that low, hazardous and harmful levels of alcohol use each significantly increase the risk of breast cancer.

A systematic review and meta-analysis of alcohol consumption and injury risk as a function of study design and recall period.

BACKGROUND: It is well established that alcohol consumption is associated with an increased risk of injury. This systematic review and meta-analysis addresses important methodological issues commonly encountered in the alcohol and injury field by delineating the effect of study design and alcohol consumption recall period on effect size magnitude and by conducting gender-specific analyses. METHODS: We performed meta-analyses using random-effect models. Data sources were peer-reviewed studies on alcohol and injury from 1970 to 2009 from MEDLINE, PsychInfo, and on-line journals. Case-control or case-crossover emergency department (ED) studies reporting injury risk from alcohol consumption 6 hours before injury were included. RESULTS: The overall odds of injury were 2.799 (2.214 to 3.538, p < 0.001). For case-crossover studies, the odds were 3.815 (2.646 to 5.499, p < 0.001); for ED case-control studies, the odds were 1.977 (1.385 to 2.821, p < 0.001); and for population case-control designs, the odds were 3.145 (1.583 to 6.247, p < 0.005). The "usual frequency" recall period yielded an odds ratio of 4.235 (2.541 to 7.057, p < 0.001), compared to 2.320 (1.789 to 3.008, p < 0.001) for all other methods. There were significant differences in odds ratio magnitude when comparing studies by design and recall period. Females had higher odds of injury than males, 2.285 (1.361 to 3.836, p < 0.005) versus 1.071 (0.715 to 1.605, p = 0.737). CONCLUSIONS: Study design and alcohol consumption recall period have significant effects on effect size magnitude in estimating the risk of injury from alcohol consumption 6 hours prior to injury. For the "usual frequency" case-crossover design, significant moderator effects were found, resulting in overestimates of injury risk from alcohol. ED case-crossover designs tend to overestimate risk, and ED case-control designs tend to underestimate. We provide recommendations for future ED research.

Adolescent smoking: effect of school and community characteristics.

BACKGROUND: A substantial challenge in addressing adolescent tobacco use is that smoking behaviors occur in complex environments that involve the school setting and larger community context. PURPOSE: This study provides an integrated description of factors from the school and community environment that affect youth smoking and explains variation in individual smoking behaviors both within and across schools/communities. METHODS: Data were collected from 82 randomly sampled secondary schools in five Canadian provinces (British Columbia, Manitoba, Ontario, Quebec, Newfoundland, and Labrador) during the 2003-2004 school year. Cross-sectional data were obtained from students; school administrators (school-based tobacco control policies and programs); and from observations in the community. In 2009, hierarchic logistic regression was used to model the role of individual, school, and community variables in predicting student smoking outcomes. RESULTS: Students who attended a school with a focus on tobacco prevention (OR=0.87, 95% CI=0.81, 0.94) and stronger policies prohibiting tobacco use (OR=0.92, 95% CI=0.88, 0.97) were less likely to smoke than students who attended a school without these characteristics. A student was more likely to smoke if a greater number of students smoked on the school periphery (OR=1.25, 95% CI=1.07, 1.47). Within the community, price per cigarette (OR=0.91, 95% CI=0.84, 0.99) and immigrants (OR=0.99, 95% CI=0.98, 0.99) were inversely related to students' smoking status. CONCLUSIONS: The results suggest that school and community characteristics account for variation in smoking levels across schools. Based on the current findings, the ideal school setting that supports low student smoking levels is located in a neighborhood where the cost of cigarettes is high, provides tobacco prevention education, and has a policy prohibiting smoking.

The influence of school policies on smoking prevalence among students in grades 5-9, Canada, 2004-2005.

INTRODUCTION: School characteristics may account for some of the variation in smoking prevalence among schools. The purpose of this study was to investigate the relationships between characteristics of school tobacco policies and school smoking prevalence. We also examined the relationship between these characteristics and individual smoking status. METHODS: Tobacco policy data were collected from schools in 10 Canadian provinces during the 2004-2005 school year. Written tobacco policies were collected from each school to examine policy intent, and school administrators were surveyed to assess policy enforcement. Students in grades 5 through 9 completed the Youth Smoking Survey to assess smoking behaviors and attitudes. We used negative binomial regression and multilevel logistic regression to predict the influence of school policies on smoking behavior at the school and student levels. RESULTS: School policies that explicitly stated purpose and goals predicted lower prevalence of smoking at the school and individual levels. Policies that prohibited smoking on school grounds at all times predicted lower smoking prevalence at the school level but not at the individual level. CONCLUSION: For maximum effectiveness, school smoking policies should clearly state a purpose and goals and should emphasize smoking prohibition. These policies can help reduce smoking prevalence among youths and are part of a comprehensive school approach to tobacco control.

Women's responses to information on mammographic breast density.

The objective of this study was to determine the negative and positive outcomes of providing mammographic breast density (MBD) information to participants of a screening program. A randomized experiment was conducted with a sample of 618 women 50 years or older with MBD greater than 50% of breast volume. The intervention consisted of reporting the presence of MBD in the screening mammography results letter that was sent along with an information pamphlet. Compared to the controls, more women in the intervention group described the term breast density correctly and recognized it as a risk factor for breast cancer. Although at the 4-week follow-up the intervention group indicated that they were "very likely" to have an annual clinical breast examination more frequently than controls, no differences were detected at 6 months. There were no significant differences on other behavioural or psychological measures, although at the 4-week follow-up the control group perceived their risk for breast cancer, relative to other women their age, as "a lot lower" than did women in the intervention group. The results demonstrate a feasible and non-threatening way to provide women with important personalized information about breast cancer risk.