RESOLVING DISCREPANCIES AMONG STUDIES - THE INFLUENCE OF DOSE ON EFFECT SIZE

In conducting reviews or meta-analyses, epidemiologists frequently mus t reconcile conflicting results. This paper addresses heterogeneity in nonexperimental studies. The emphasis is on simple exploratory method s rather than formal approaches. Five examples illustrate how quantita tive cent cordance among studies is possible, even when measured effec ts appear discrepant. The examples concern ethylene oxide and leukemia s, methylene chloride and liver cancer, saccharin and bladder cancer, prenatal lead exposure and birthweight, and aspirin and bleeding tende ncies in labor and delivery. Data examined here indicate that differen ces in dose levels frequently explain heterogeneous effect measures, o ften outweighing other sources of variability among studies. We presen t simple methods for combining dose information from the study of inte rest with dose response data from other epidemiologic studies or anima l studies to derive plausible hypothesized effect levels. These plausi ble effect: sizes are the measures of association that would be predic ted, for the actual exposures, by extrapolating from other studies wit h possibly differing exposure levels. Post hoc power calculations and comparisons of confidence intervals for overlap to reconcile ''positiv e'' and ''null'' studies may be misleading, since these approaches ass ume a uniform true association obscured by random fluctuations only. W henever it can be estimated, a plausible effect size should be the sta rting point to assess findings of either positive or null studies. Wit hout such calculations, comparisons among conflicting studies may not be meaningful.