Participants in a randomized controlled trial had longer overall survival than non-participants: a prospective cohort study.

PURPOSE: While some studies show improved outcomes in clinical trial participants as compared to non-participants, existence of such a trial effect has not been proved precisely. METHODS: This was a prospective cohort study to compare the prognoses for participants in the randomized controlled trial (SELECT BC) and non-participants. SELECT BC compared S-1 and taxane as first-line treatment for metastatic breast cancer. Non-participants were all patients who met the eligibility criteria of SELECT BC and who had been requested to participate in that trial by attending doctors and declined. The study aimed to compare the prognoses between participants and non-participants. The primary endpoint was median overall survival. RESULTS: The median OS in participants was significantly superior to that in non-participants with a statistically significant difference (36.8 months vs. 25.2 months. HR 1.48, p = 0.022). A similar result was obtained when only patients who received the same chemotherapy (S-1 or taxane) used in SELECT BC after declining participation were assumed as non-participants (36.8 months vs. 22.0 months. HR 2.03, p = 0.006). CONCLUSIONS: This study may suggest the existence of a trial effect, in which, for a given treatment, participation in a clinical trial is associated with a better outcome.