A new dataset details 542 digital mental health organizations that exited the market between 2000 and 2026, including shutdowns, bankruptcies, acquisitions, pivots, and consolidations. The data is coded across up to 18 fields, such as business model, funding raised, reason for leaving, and clinical evidence. It also includes four independent classification axes: product type, entity type, care mode (replacing or wrapping around clinicians), and whether a human clinician was involved.

Analysis reveals that who pays for mental health services significantly predicts survival, with institutional payers showing higher survival rates than consumers. Having a clinician co-founder did not impact exit rates. The classification was assisted by LLM agents, with the rationale for each label provided for transparency and review.

The dataset is not a random sample, as it exclusively includes companies that have left the market. Approximately 67% of the sample comprises companies from the US and UK. The data is available for use under a CC BY 4.0 license, with a 406-page report offering further findings and analysis.