Fig. 1: Consolidated Standards of Reporting Trials (CONSORT) diagram.

The BA arm and control arm were common across the four 2 × 2 trials and are shown by the dotted squares in trials 2 to 4. The prespecified control condition was the delayed treatment. We conducted sensitivity analyses by replacing the delayed treatment with the health information arm and the self-check arm (Methods). aParticipants working on a three-shift schedule were excluded in the primary and sensitivity analyses for trial 4 (199 (5.1%) of the total sample were shift workers). AT, assertion training; BA, behavioral activation; BI, behavior therapy for insomnia; CR, cognitive restructuring; PHQ-9, Patient Health Questionnaire-9; PS, problem solving.