When systematic errors of bias and confounding are repeated across multiple epidemiological studies, they may lead to false conclusions. This application addresses bias and confounding in cohort and case-control studies of associations between alcohol use and a range of sentinel diseases with the potential of correcting for them. Uncorrected, these errors obscure scientific understanding of the alcohol-disease relation, undermine the quality and accuracy of public health policy and lead to the dissemination of misleading information. In 2009, our research team was awarded an NIH Challenge Grant (CG). The successful CG application represented a small portion of a comprehensive research plan. The present application fills out the total plan, and adds new and important elements to it. First, we test hypotheses regarding possible systematic errors in epidemiological studies of alcohol as a risk factor for morbidity and mortality associated with 7 sentinel diseases as well as all-cause mortality: (i) Type 2 diabetes, (ii) cholelithiasis, (iii) female breast cancer, (iv) stomach cancer, (v) renal cancer, (vi) lung cancer, (vii) coronary heart disease (updating our previous findings with new studies), and (viii) all-cause mortality (also updating findings from our previous studies). Each of these diseases has been identified in past reviews as having one of the following relations to alcohol consumption: (i) a J-shape curve (a 'protective'effect for 'light/moderate'drinkers);(ii) a statistically significant positive linear curve (a detrimental effect for moderate and heavier drinkers);and (iii) a mixed cross-study effect (no effect). These conditions are examined in addition to the 4 disease outcomes we are testing in the CG. Second, we apply a new method that may account, at least in part, for the unequal distribution of confounders within distinct drinking categories, correcting results as we currently know them. Third, we test newly developed hypotheses regarding 'unequal'confounders (which has since emerged from the CG) using data from a large cohort study. Fourth, we utilize our findings to re-estimate potential harms and benefits of various categories of drinking at the population level. Meta-analyses will be used to determine whether the presence of systematic error in study design, the quality and comprehensiveness of potential confounder measurement, or combinations of these in studies can unduly bias results towards: (i) creating apparent protection from alcohol use against some diseases;(ii) masking potentially statistically significant associations;(iii) contributing to 'mixed'outcomes that defy a clear resolution;(iv) influencing the size and significance for established alcohol-disease associations. This proposed work is intended to advance the understanding of alcohol's contribution to a range of chronic diseases and improve estimates of alcohol-attributable mortality and morbidity. It should clarify the evidence basis for advice from clinical practice and contribute understanding to national and international strategies to reduce the harm from alcoholic beverages.

Public Health Relevance

The proposed study uses meta-analysis (comparing many relevant studies) to examine the two most prevalent biases in alcohol-related medical epidemiology - systematic errors and confounding. These biases often prevent science from confidently reaching conclusions about the associations between alcohol consumption and disease outcome. Correction of these biases may clarify these relationships and improve the quality and accuracy of clinical advice and public health policy.

Agency
National Institute of Health (NIH)
Institute
National Institute on Alcohol Abuse and Alcoholism (NIAAA)
Type
Research Project (R01)
Project #
1R01AA019939-01A1
Application #
8183364
Study Section
Behavioral Genetics and Epidemiology Study Section (BGES)
Program Officer
Zha, Wenxing
Project Start
2011-09-15
Project End
2014-06-30
Budget Start
2011-09-15
Budget End
2012-06-30
Support Year
1
Fiscal Year
2011
Total Cost
$503,203
Indirect Cost
Name
Scientific Analysis Corporation
Department
Type
DUNS #
044875854
City
San Francisco
State
CA
Country
United States
Zip Code
94107