A comparison of the results of prospective and retrospective cohort studies in the field of digestive surgery.

Abstract

We compared the results of prospective and retrospective cohort studies in the field of digestive surgery to clarify whether the results of prospective cohort studies were more similar to those of randomized controlled trials (RCTs). We conducted a secondary analysis of the results to compare the results of RCTs with those of cohort studies in meta-analyses of 18 digestive surgical topics. The data from the prospective and retrospective cohort studies were combined. The summary estimates of each design were compared with those of RCTs. We used the Z score to investigate discrepancies. Twenty-nine outcomes of 11 topics were investigated in 289 cohort studies (prospective, n = 69; retrospective, n = 220). These were compared with the outcomes of 123 RCTs. In comparison to retrospective studies, the summary estimates of the prospective cohort studies were more similar to those of the RCTs [19/29 (prospective) vs. 10/29 (retrospective), P = 0.035). Five of the 29 outcomes of prospective studies and 6 of 29 outcomes of retrospective studies (P = 0.99) showed significant discrepancies in comparison to RCTs. In the digestive surgical field, the results of prospective cohort studies tended to be more similar to those of RCTs than retrospective studies; however, there were no significant discrepancies between the two types of cohort study.