Interpreting epidemiologic studies of diet and health is often complex because it involves many methodological issues and sources of uncertainty. Pertinent examples are false-positive and false-negative results, that is, findings that erroneously indicate the presence or absence, respectively, of association. Our objective in this commentary is to help nutrition and other health professionals interpret epidemiological studies when faced with this task. Our aim is not to provide an exhaustive discussion of all the methodological nuances of nutritional epidemiology; rather, we focus on some of the salient factors when interpreting findings from dietary studies, and the role that uncertainty (namely, from false-positives) plays, using relevant examples from the past few decades.
EpidStat in the News
EpidStat presented "Survival Synthesis: Methods for Meta-analysis of Survival Rates and Aggregation of Survival Data" at a poster session for the International Conference on Health Policy Statistics on January 10, 2018 in Charlston, SC. The poster details the methods used in Bylsma et al. Arteriovenous Fistulae for Haemodialysis: A Systematic Review and Metaanalysis of Efficacy and Safety Outcomes. European Journal of Vascular and Endovascular Surgery, 2017. We discussed the process of digitizing survival curves from the published literature and simulating individual patient data to derive survival statistics that were not published with the article. We conclude that while labor-intensive, the methodology makes the best use of available study data.
EpidStat presented "Prevalence of KRAS and BRAF Mutations in Metastatic Colorectal Cancer (mCRC) by Tumor Location: A Review and Meta-analysis" on January 20, 2018 at the ASCO Gastrointestinal Cancers Symposium. Results from 44 studies comprising nearly 16,000 mCRC patients demonstrated that both KRAS and BRAF mutations were significantly more prevalent among right-sided colon tumors than left-sided tumors. These results may aid clinicians and researchers in the search for target novel therapeutics.