Dataset Viewer
Auto-converted to Parquet Duplicate
id
stringclasses
6 values
prompt
stringclasses
6 values
gold_json
stringclasses
5 values
reference_notes
stringclasses
6 values
__index_level_0__
stringclasses
6 values
Quad enrollment analysis. Oncology trial. Screen failure rate drops from 42% to 18% at two high-enrolling sites. Inclusion labs missing for 12% of enrolled patients. Recruitment targets increased by sponsor. Governance: interim analysis in 3 weeks. Determine enrollment drift risk and action. Return strict JSON only.
{"enrollment_drift_risk":true,"risk_type":"criteria_relaxation_under_pressure","driver_nodes":["screen failure drop cluster","missing inclusion labs","recruitment target increase","imminent interim analysis"],"recommended_action":"screening_audit_and_retraining","action_detail":["audit recent enrollments","restore lab ...
confidence":0.89}
Criteria drift.
ENR-001
Quad enrollment analysis. Cardiology trial. One region enrolls faster with lower baseline risk scores. Site selection favored high-volume centers. Screening logs incomplete. Governance: DSMB review next month. Determine risk and action. JSON only.
{"enrollment_drift_risk":true,"risk_type":"site_selection_bias","driver_nodes":["fast enrollment region","lower baseline risk","site selection bias","incomplete screening logs"],"recommended_action":"baseline_and_screening_review","action_detail":["review baseline data","complete screening logs","rebalance enrollment",...
confidence":0.87}
Site bias.
ENR-002
Quad enrollment analysis. Neurology trial. Eligibility deviations cluster at sites with staffing shortages. Screening visit time reduced. Training refresh overdue. Governance: regulator inquiry pending. Determine risk and action. JSON only.
{"enrollment_drift_risk":true,"risk_type":"staffing_screening_gap","driver_nodes":["eligibility deviation cluster","staffing shortage","reduced screening time","overdue training"],"recommended_action":"staffing_and_training_intervention","action_detail":["increase screening staff","deliver refresher training","audit el...
confidence":0.86}
Staffing node.
ENR-003
Quad enrollment analysis. Pulmonary trial. Enrollment rises after advertising campaign. Baseline severity scores lower. Screening documentation incomplete. Governance: interim efficacy tied to enrollment pace. Determine risk and action. JSON only.
{"enrollment_drift_risk":true,"risk_type":"advertising_induced_population_shift","driver_nodes":["enrollment surge","lower severity baseline","incomplete screening docs","efficacy tied to pace"],"recommended_action":"population_and_doc_audit","action_detail":["audit baseline severity","complete documentation","pause ta...
confidence":0.85}
Advertising pressure.
ENR-004
Quad enrollment analysis. Metabolic trial. Inclusion criteria modified at one site without approval. Screening logs delayed. Recruitment bonus introduced. Governance: compliance audit scheduled. Determine risk and action. JSON only.
{"enrollment_drift_risk":true,"risk_type":"unauthorized_criteria_change","driver_nodes":["criteria modification","screening log delay","recruitment bonus","audit scheduled"],"recommended_action":"criteria_and_log_reconciliation","action_detail":["restore approved criteria","update logs","audit enrolled patients","prepa...
confidence":0.88}
Criteria change.
ENR-005
Quad enrollment analysis. Vaccine trial. Enrollment cluster shows younger patients only. Age eligibility documentation incomplete. Outreach focused on one demographic. Governance: regulator review upcoming. Determine risk and action. JSON only.
{"enrollment_drift_risk":true,"risk_type":"demographic_selection_bias","driver_nodes":["age cluster","incomplete age docs","targeted outreach","regulator review"],"recommended_action":"demographic_and_doc_audit","action_detail":["verify age eligibility","broaden outreach","update documentation","prepare regulator brief...
confidence":0.87}
Demographic node.
ENR-006

Clarus Clinical Quad Coupling Enrollment Criteria Drift Site Selection Bias Screening Pressure v0.1

Purpose
Detect enrollment population drift driven by four interacting nodes.

Quad nodes

  • Criteria relaxation or documentation gap
  • Site selection or recruitment bias
  • Screening workflow pressure
  • Governance or interim timing pressure

Input
One vignette.

Output
Strict JSON only.

Required keys

  • enrollment_drift_risk
  • risk_type
  • driver_nodes
  • recommended_action
  • action_detail
  • rationale
  • confidence

Files
data/train.csv
data/test.csv
scorer.py

Run scoring
Create JSONL predictions and run:

python scorer.py --gold_csv data/test.csv --preds_jsonl preds.jsonl

Downloads last month
5