Systematic Error and Confounding: Meta-Analyses of Alcohol and Disease

系统误差和混杂因素：酒精与疾病的荟萃分析

• 批准号: 8331473
• 负责人: KAYE M FILLMORE
• 金额: $ 50.16万
• 依托单位: SCIENTIFIC ANALYSIS CORPORATION
• 依托单位国家: 美国
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-09-15 至 2014-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8331473
• 关键词: Accounting; Address; Age; Alcohol consumption; Alcoholic Beverages; Alcoholic beverage heavy drinker; Alcohols; Applications Grants; Attention; Award; Brain hemorrhage; Case-Control Studies; Categories; Cessation of life; Characteristics; Cholelithiasis; Chronic Disease; Clinical; Cohort Studies; Colorectal Cancer; Confounding Factors (Epidemiology); Consumption; Coronary heart disease; Country; Data; Dementia; Disease; Disease Association; Disease Outcome; Elements; Epidemiologic Studies; Epidemiology; Female Breast Carcinoma; Future; Gender; Grant; Health; Health Benefit; Health Policy; International; Ischemic Stroke; Lead; Light; Malignant neoplasm of lung; Masks; Measurement; Medical; Meta-Analysis; Methods; Morbidity - disease rate; Non-Insulin-Dependent Diabetes Mellitus; Outcome; Pattern; Population; Public Health; Quality Control; Renal carcinoma; Research; Research Design; Resolution; Risk Factors; Science; Sentinel; Shapes; Source; Staging; Testing; United States National Institutes of Health; Update; Work; alcohol epidemiology; attributable mortality; burden of illness; clinical practice; cohort; design; drinking; drinking behavior; evidence base; improved; malignant breast neoplasm; malignant stomach neoplasm; mortality; premature; prevent; prospective; protective effect

DESCRIPTION (provided by applicant): 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.

描述（由申请人提供）：当偏倚和混杂的系统性错误在多项流行病学研究中重复出现时，可能导致错误的结论。该应用程序解决了队列和病例对照研究中的偏倚和混淆，这些研究探讨了酒精使用与一系列哨兵疾病之间的关联，并有可能对其进行纠正。如果不加以纠正，这些错误会模糊对酒精与疾病关系的科学认识，损害公共卫生政策的质量和准确性，并导致误导性信息的传播。 2009年，我们的研究团队获得了NIH挑战资助（CG）。成功的CG应用程序代表了一个全面的研究计划的一小部分。本申请填写了总体计划，并为其添加了新的和重要的元素。首先，我们测试了关于酒精作为与7种哨点疾病相关的发病率和死亡率以及全因死亡率的风险因素的流行病学研究中可能存在的系统性错误的假设：（i）2型糖尿病，（ii）胆石症，（iii）女性乳腺癌，（iv）胃癌，（v）肾癌，（vi）肺癌，（vii）冠心病（用新的研究更新我们以前的发现），和（viii）全因死亡率（也更新我们以前的研究结果）。这些疾病中的每一种都在过去的综述中被确定为与酒精消费有以下关系之一：（i）J形曲线（对“轻度/中度”饮酒者的“保护”作用）;（ii）统计学显著的正线性曲线（对中度和重度饮酒者的有害影响）;（iii）混合交叉研究效应（无影响）。除了我们在CG中测试的4种疾病结局外，还检查了这些情况。其次，我们采用了一种新的方法，可以解释，至少部分地，不同的饮酒类别内的混杂因素的不平等分布，纠正结果，因为我们目前知道他们。第三，我们测试新开发的假设“不平等”的混杂因素（这已经出现了CG）使用的数据从一个大型队列研究。第四，我们利用我们的研究结果，重新估计在人口水平上各类饮酒的潜在危害和益处。荟萃分析将用于确定研究设计中存在的系统性错误、潜在混杂因素测量的质量和全面性或研究中这些因素的组合是否会使结果过度偏向：（i）对某些疾病产生明显的酒精使用保护作用;（ii）掩盖潜在的统计学显著性关联;（iii）导致无法明确解决的“混合”结果;（iv）影响已建立的酒精疾病协会的规模和意义。 这项拟议的工作旨在促进对酒精导致一系列慢性病的认识，并改进对酒精所致死亡率和发病率的估计。它应该澄清临床实践建议的证据基础，并有助于理解国家和国际战略，以减少酒精饮料的危害。