Type 2 diabetes and cancer are two chronic diseases that negatively influence the survival of millions of people worldwide. Diabetes has been shown in meta-analyses to be associated with an increased risk of several cancers, but these associations are not yet documented as causal. The evidence for specific diabetes drugs affecting cancer risk is limited, because randomized clinical trials are not designed to evaluate specifically cancer outcomes, and observational studies are often confounded by indications for specific drugs. We propose to better characterize the degree of involvement of type 2 diabetes diagnosis and pharmacologic treatment in cancer initiation or progression. Therefore, we will a) evaluate bias and credibility in the existing meta-analyses of diabetes diagnosis and cancer risk, b) determine whether diabetes diagnosis is associated with risk of prostate cancer using the European Prospective Investigation into Cancer and Nutrition cohort, c) perform a systematic review and meta-analysis of randomized clinical trials and observational studies that have evaluated the association between diabetes medications and cancer risk, and d) evaluate the association between diabetes drugs and cancer risk (separately by site) using the UK-based General Practice Research Database. Cutting-edge evidence synthesis methods (Harbord’s test, excess significance bias test, credibility ceilings) will be used in our first aim, and summary meta-analytic estimates will be calculated in our third aim using random effects models. An incident diabetes drug cohort design with adherence-adjusted intention-to-treat style of analysis will be used in aim 4, while Cox proportional hazards models will be used in both aims 2 and 4. The proposed project will have major impact in the field, because the literature on diabetes diagnosis, treatment on cancer risk will be properly evaluated, and these associations will be further assessed in two large, unique prospective studies.