Statistical Reliability of a Diet-Disease Association Meta-analysis

Risk ratios or p-values from multiple, independent studies – observational randomized can be pooled to address a common research question in meta-analysis. However, reliability of should not assumed as claimed risk factor−disease relationships may fail reproduce. An evaluation was undertaken published meta-analysis cohort examining diet−disease associations; specifically between red and processed meat six disease outcomes (all-cause mortality, cardiovascular all cancer breast incidence, colorectal type 2 diabetes incidence). The number hypotheses examined were counted 15 random base papers (14%) 105 used the Test statistics (relative values with 95% confidence limits) for 125 results converted p-values; p-value plots examine effect heterogeneity p-values. possible large, median = 20,736 (interquartile range 1,728–331,776). Each plot selected health effects showed either pattern (p-values > 0.05), two-component mixture (small < 0.001 while other appeared random). Given potentially large numbers studies, questionable practices cannot ruled out explanations some test small Like original findings meta-analysis, our concludes that do support evidence an association investigated.