Use of core outcome sets was low in clinical trials published in major medical journals.

OBJECTIVES:To examine current practices in late-phase trials published in major medical journals and examine trialists' views about core outcome set (COS) use. STUDY DESIGN AND SETTING:A sequential multi-methods study was conducted. We examined late-phase trials published between October 2019 and March 2020 in JAMA, NEJM, The Lancet, BMJ, and Annals of Internal Medicine. The COMET database was searched for COS potentially relevant to trials not reporting using a COS; overlap of trial and COS outcomes was examined. An online survey examined awareness of, and decisions to search for and use a COS. RESULTS:Ninety-five trials were examined; 93 (98%) did not report using a COS. Relevant COS were identified for 31 trials (33%). Core outcomes were measured in 9 (23%) studies; all trials measured at least one core outcome. Thirty-one trialists (33%) completed our survey. The most common barrier to COS use was trialist's own outcome preferences and choice (68%). The most common perceived facilitator was awareness and knowledge about COS (90%). CONCLUSION:COS use in this cohort of trials was low, even when relevant COS were available. Increased use of COS in clinical trials can improve evaluation of intervention effects and evidence synthesis and reduce research waste.