TECH Signal 487
Dataset: Dead mental health startups, 2000-2026, coded on 18 fields
↗
A new CC-BY-4.0 dataset of 542 digital mental-health companies that exited the market between 2000 and 2026 has been released.
Engineers can use the raw CSV/JSON to run their own survival analyses, benchmark business-model assumptions, or train models on real-world outcomes. The data include detailed fields such as payer type, funding, clinical evidence, and a narrative of each company’s fatal mistake, which can surface hidden risk factors. However, the set only covers firms that have already left the market and is skewed toward the US/UK, so any predictive use must treat the numbers as comparative, not absolute probabilities.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
The download bundle contains CSV and JSON tables, a rationale file for every label, a data dictionary, and a CC-BY-4.0 license, requiring no signup.
Four classification axes (product type, entity type, care mode, clinician-in-the-loop) were generated by LLM agents, with the full reasoning provided for auditability.
Analysis of the data shows payer model is the strongest survival predictor, while having a medical co-founder does not affect exit rates.
THE CLUSTER