Authors: Hansen, R; Zeng, P; Ryan, P; Gao, J; Sonawane, K; Teeter, B; Westrich, K; Dubois, RW.
Distributed data network drug safety analyses are difficult to interpret in the face of heterogeneity, or individual treatment effects.
Understanding and reporting of the potential contributions of various factors to heterogeneity in studies that utilize distributed data networks is important, and pooling of results may not be appropriate in light of high heterogeneity. The researchers suggest that a formal assessment of heterogeneity should be conducted, and approaches illustrated in this study can be useful to explore heterogeneity. These approaches may have parallel value in non-health care examples that involve similar types of heterogeneity.