Background: Many clinical trials have assessed treatments for depressive disorders and bipolar depression. However, whether, and which, assessed outcome domains really matter to patients, informal caregivers, and health-care professionals remains unclear.; Methods: We did an international online survey in French, German, and English. Participants were adult patients with a history of depression, informal caregivers, and health-care professionals, recruited by purposeful sampling. To identify outcome domains, participants answered four open-ended questions about their expectations for depression treatment. We disseminated the survey without restriction via social media, patient and professional associations, and a media campaign. Four researchers independently did qualitative content analyses. We assessed data saturation using mathematical models to ensure the comprehensive identification of outcome domains.; Findings: Between April 5, 2018, and Dec 10, 2018, 1912 patients, 464 informal caregivers, and 627 health-care professionals from 52 countries provided 8183 open-ended answers. We identified 80 outcome domains related to symptoms (64 domains), such as mental pain (or psychological or psychic pain, 523 [17%] of 3003 participants) and motivation (384 [13%]), and functioning (16 domains), such as social isolation (541 [18%]). We identified 57 other outcome domains regarding safety of treatment, health care organisation, and social representation, such as stigmatisation (408 [14%]).; Interpretation: This study provides a list of outcome domains important to patients, informal caregivers, and health-care professionals. Unfortunately, many of these domains are rarely measured in clinical trials. Results from this study should set the foundation for a core outcome set for depression.; Funding: Fondation pour la Recherche Medicale and NIHR Oxford Health Biomedical Research Centre.