case_id stringclasses 10
values | indication stringclasses 6
values | phase stringclasses 2
values | site_plan stringclasses 10
values | population_assumptions stringclasses 10
values | eligibility_design stringclasses 10
values | feasibility_signals stringclasses 10
values | recruitment_risk_draft stringclasses 1
value | gold_risk_level stringclasses 2
values | gold_failure_mode stringclasses 10
values | gold_correct_action stringclasses 10
values |
|---|---|---|---|---|---|---|---|---|---|---|
CRCM-001 | Oncology | Phase II | 30 sites 6 months | High prevalence; rapid referrals | Requires biomarker + prior line + ECOG 0-1 | Screen fail 55%; biomarker positivity 18% | Low risk | high | assumption_overreach | Revise prevalence and screen fail model. Expand sites and loosen criteria or extend timeline |
CRCM-002 | Immunology | Phase III | 40 sites 10 months | Stable patient pool | Excludes common comorbidity; long washout | Competing trial nearby; drop out risk | Low risk | high | competitive_displacement | Add differentiated site strategy and increase sites. Reduce washout where safe |
CRCM-003 | Neuro | Phase II | 20 sites 8 months | Caregivers available | Requires weekly visits; imaging every 4 weeks | Travel burden high; missed visit trend | Low risk | medium | visit_burden | Reduce visit frequency or add home support. Add travel support |
CRCM-004 | Metabolic | Phase III | 25 sites 12 months | Primary care will refer | Complex labs; frequent fasting visits | Referral rate historically low | Low risk | high | referral_gap | Build referral pipeline. Simplify visit schedule. Add outreach |
CRCM-005 | Infectious | Phase II | 15 sites 6 months | Hospital CAP volume adequate | Enrollment within 24h; consent barriers | Night enroll limited; delayed consent | Low risk | medium | operational_window | Add 24/7 screening and consent workflow. Extend window if clinically safe |
CRCM-006 | Oncology | Phase III | 50 sites 14 months | Biomarker prevalence 40% | Strict tissue requirements | Prevalence actually 15%; tissue insufficient | Low risk | high | biomarker_miscalibration | Rebaseline prevalence. Add pre screen registry. Broaden tissue options |
CRCM-007 | Cardio | Phase III | 35 sites 9 months | High adherence expected | Requires daily diary; device upload | Compliance low in similar studies | Low risk | medium | adherence_burden | Reduce diary burden. Use passive capture. Add adherence support |
CRCM-008 | Immunology | Phase II | 18 sites 7 months | Low screen fail expected | Excludes prior biologics | Most patients already exposed | Low risk | high | eligibility_mismatch | Relax exclusion or shift to earlier line population. Reforecast |
CRCM-009 | Neuro | Phase III | 60 sites 18 months | Broad population | Requires MRI and specialist visits | Limited MRI slots; long waits | Low risk | high | capacity_constraint | Secure imaging capacity. Add sites with MRI access. Extend timeline |
CRCM-010 | Metabolic | Phase II | 12 sites 5 months | High volunteer interest | Intensive run in; diet stabilization | Run in drop 35% | Low risk | medium | run_in_attrition | Shorten run in or add support. Adjust enrollment targets |
Clinical Recruitment Coherence Mapping v0.1
Purpose
Detect when population and feasibility assumptions will break recruitment.
Model task
Return one JSON object
- risk_level
low, medium, high - failure_mode
one allowed label - correct_action
one short paragraph
Scoring
0 to 100
- risk accuracy 30
- failure mode accuracy 35
- action similarity 25
- format pass 10
Run
python scorer.py --predictions predictions.jsonl --test_csv data/test.csv
- Downloads last month
- 11