{"stable_id":"81c7b1bc482d6d8f","slug":"optum-care-risk-bias","company":"Large US health system / Optum algorithm","workflow_name":"Commercial cost-proxy risk ranking","function_code":"enrollment","pattern_codes":["queue_elimination","continuous_decisioning"],"changed_assumption":"A predictable proxy can still allocate scarce care unjustly.","evidence_strength":"verified","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 independent, 1 peer reviewed; publication outcomes are verified.","caveat_summary":"Counterfactual repair in one system; vendor disputed use characterization.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Care-management program","action":"Has a limited number of intensive nursing and coordination slots to allocate.","actor_type":"control"},{"actor":"Clinicians and program staff","action":"Use clinical knowledge to identify patients with complex needs.","actor_type":"human"}],"hinge":"A predictable proxy can still allocate scarce care unjustly.","after":[{"actor":"Commercial risk algorithm","action":"Ranks patients by predicted future cost as a proxy for health need.","actor_type":"ai"},{"actor":"Program staff and clinicians","action":"Use the score among selection inputs and allocate scarce extra care.","actor_type":"human"}],"decision_mode":"retained","decision_marker":"Human authority remains at the consequential boundary"},"before":[{"order":1,"actor":"Care-management program","action":"Has a limited number of intensive nursing and coordination slots to allocate.","handoff_to":"Patient-selection process","control":"The Science study does not provide a dated pre-algorithm workflow; this is a bounded resource-allocation baseline."},{"order":2,"actor":"Clinicians and program staff","action":"Use clinical knowledge to identify patients with complex needs.","handoff_to":"High-risk care program","control":"Human judgment, constrained by program capacity."}],"after":[{"order":1,"actor":"Commercial risk algorithm","action":"Ranks patients by predicted future cost as a proxy for health need.","handoff_to":"Top-percentile enrollment queue","control":"Top 97% risk threshold at the studied site; proxy choice embeds unequal access and spending."},{"order":2,"actor":"Program staff and clinicians","action":"Use the score among selection inputs and allocate scarce extra care.","handoff_to":"Nursing and care-coordination team","control":"Clinicians can supplement the score, but scaled allocation is score-shaped; health-need counterfactual exposed bias."}],"decision_rights":"Programs own enrollment; clinicians retain judgment.","exception_path":"Clinical judgment can add patients, but the queue is score-shaped.","removed_work":["Some manual population prioritization"],"outcomes":[{"metric":"Black share identified for extra care","baseline":"17.7% under cost proxy","result":"46.5% if ranked by actual health need","period":"Published 2019","scale":"43,539 White and 6,079 Black patients","attribution_caveat":"Counterfactual repair in one system; vendor disputed use characterization.","evidence_label":"verified"}],"executive_lesson":"Target validity matters more than headline predictive accuracy.","anti_pattern":"Do not call cost prediction inaccurate; the target was wrong.","questions_for_leaders":["Where is the operating threshold set and who can override it?","What measured result would trigger rollback or retraining?","Which residual decisions must remain human-owned?"],"collections":["queue-eliminated","negative-results"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/optum-care-risk-bias"}