Published protocols
Twenty-four protocols outlining COS development studies were identified (table S2)(21-44) including: 14 (58%) in pregnancy and childbirth, 3 (13%) benign gynaecology and subfertility, 6 (25%) newborn and neonatal health and 1 (4%) in oncology (table S2). The scope of the proposed COS including the health condition, population, intervention and setting was clearly specified in 20 (83%). Ten (42%) identified the steering group for the protocol. Twenty-one (87%) described their intention to carry out a systematic review of published research to identify potential core outcomes; three (13%) referred to already completed systematic reviews. Eight (33%) described the use of qualitative methods to identify potential core outcomes: 6 described interviews or focus groups alongside qualitative reviews and 2 proposed only qualitative reviews. Eleven (46%) protocols discussed the impact of missing data and/or attrition bias. All intended to identify core outcomes using a modified Delphi method and 23 proposed a consensus development meeting (96%). Overall, only four protocols (17%) completely fulfilled COS‐STAP criteria (table S3).