A 250-patient synthetic community rheumatology panel, screened against three lupus protocol designs by the deterministic Amyn matching engine. Every verdict is criterion-level and cites its evidence. Time-to-eligibility is arithmetic on documented medication dates, not a predictive model.
Eligible patients found in the same 250-patient panel, by protocol design. The broad pragmatic design finds more than twice the patients of the narrow design.
Patients excluded by exactly one criterion (everything else passing), for the selected protocol. The underrepresented share of each bar shows where a criterion relaxation buys inclusion.
Share of the eligible pool by group, against each group's share of US lupus burden from published epidemiology. Gaps are the integrity metric Amyneion reports on every trial.
Patients not eligible today whose blocking criterion resolves by a known date: biologic washouts ending, steroid tapers completing. This is recruitment as a forecast, not a search.
Top matches ranked by disease activity, with the evidence a navigator sees before the first call. Barrier flags route each patient to the right CARRY package: rides, home draws, Spanish-language outreach.
The COMPILE simulator re-runs matching with one criterion removed. This is the protocol design feedback sponsors get before site selection.