Medical & Healthcare Data
This is the general shelf for clinical and healthcare datasets that don't fit neatly into imaging or pure biology — patient-level records, risk-monitoring frameworks, and treatment-safety data, built the way a hospital's own data warehouse would structure it.
A recurring example: a synthetic cohort built around clozapine, an antipsychotic that requires mandatory blood-count monitoring under an FDA Risk Evaluation and Mitigation Strategy (REMS) because of its neutropenia risk. The dataset models the mathematics of that monitoring window directly, and ships in OMOP Common Data Model format — the same schema real hospital data warehouses use — so it drops into an existing pipeline instead of demanding you build one around it.
What this category is for
- Risk-monitoring and pharmacovigilance data — datasets built around a real regulatory or clinical monitoring requirement, not an invented one.
- OMOP-formatted cohorts — structured to match standard healthcare-data tooling, with honest labelling of which fields are genuinely mapped to standard vocabularies and which are left unmapped rather than falsely mapped.
- Composite risk cohorts — datasets that combine more than one clinical process at once (for example, treatment benefit and treatment-induced harm), where the correct statistical structure between the two is part of what you're buying.
Nothing here is presented as more validated than it is. Where a dataset extrapolates beyond the range of the source literature, that's stated in the passport, not left for you to discover in production.