The comprehensive model of information seeking (CMIS) is a well-known framework to predict health information seeking by a combination of health beliefs and medium-related factors. Despite being proposed almost three decades ago, few efforts have been made to systematically summarize CMIS scholarship. To fill this gap in the literature, 36 meta-analyses were first conducted to identify the bivariate relationships between variables in the CMIS. These meta-analytic data were then used to test path models evaluating the role of health beliefs and medium-related factors. The results showed that the models containing only factors related to the communication medium, only health factors, and a modified version of the CMIS fit the data relatively well. The original CMIS did not demonstrate an acceptable model fit. Theoretical and practical implications are discussed.