Mahalo

Mahalo Health (Mahalo Recruit) — Protocol-to-chart matching asset built inside authorised EHRs: every protocol's inclusion/exclusion criteria translated into transparent, editable rule-based screening logic (compound boolean logic, time windows, negations) and evaluated against live FHIR chart data on a daily refresh. The byproduct assets are (a) a library of normalised, machine-executable eligibility rule-sets per protocol/indication, (b) ranked candidate shortlists with the specific chart evidence attached to each criterion hit, and (c) the human-reviewer verdicts (confirmed / rejected) rec…

Mahalo offers (Alternative, Reference) — No published count. Structural estimate: one rule-set per active protocol/indication plus one eligibility evaluation per patient per protocol per day — so a site with ~50k-500k screened charts across a handful of concurrent studies yields 10^6-10^7 daily match evaluations, of which the durable asset is the rule library plus the reviewer-confirmed shortlists (order 10^3-10^5 reviewed candidates per site-year)..

Training and evaluating clinical NLU/LLM systems that convert narrative protocol criteria into executable logic (the rule library is the target output, the protocol text the input); benchmarking AI-vs-clinician chart-review accuracy on eligibility decisions with evidence-grounded gold labels; trial feasibility and enrollment forecasting; protocol criteria quality analysis (which phrasings produce ambiguous or un-executable logic); synthetic-patient and cohort-generation tooling validated against real per-criterion counts; automated chart-review QA for study teams.

The data is with 3 years of history.

Coverage spans US; Health Care Technology; alternative, reference; equities.