{"stable_id":"bfb4dcb5dde322ef","slug":"copenhagen-ohca-dispatch-rct","company":"Copenhagen Emergency Medical Services","workflow_name":"Real-time ML listening and dispatcher cardiac-arrest alerts","function_code":"engineering","pattern_codes":["creator_to_judge"],"changed_assumption":"A live emergency-call alert may not improve dispatcher recognition despite higher standalone model sensitivity.","evidence_strength":"verified","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 peer reviewed; publication outcomes are verified.","caveat_summary":"High-quality randomized negative result. The model alone was more sensitive but materially less specific and had lower positive predictive value than dispatchers.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"Medical dispatcher","action":"Listen to the caller, recognize suspected cardiac arrest, and initiate the protocol.","actor_type":"control"},{"actor":"Dispatcher","action":"Coach CPR and dispatch resources when cardiac arrest is recognized.","actor_type":"control"}],"hinge":"A live emergency-call alert may not improve dispatcher recognition despite higher standalone model sensitivity.","after":[{"actor":"Speech-recognition ML model","action":"Listens to ongoing calls and flags suspected cardiac arrest.","actor_type":"ai"},{"actor":"Medical dispatcher","action":"Considers the alert and decides whether to recognize OHCA and initiate the standard response.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"Medical dispatcher","action":"Listen to the caller, recognize suspected cardiac arrest, and initiate the protocol.","handoff_to":"Caller and emergency response","control":"Standard dispatch protocol"},{"order":2,"actor":"Dispatcher","action":"Coach CPR and dispatch resources when cardiac arrest is recognized.","handoff_to":"Caller and responders","control":"Human clinical judgment"}],"after":[{"order":1,"actor":"Speech-recognition ML model","action":"Listens to ongoing calls and flags suspected cardiac arrest.","handoff_to":"Medical dispatcher","control":"Model alert"},{"order":2,"actor":"Medical dispatcher","action":"Considers the alert and decides whether to recognize OHCA and initiate the standard response.","handoff_to":"Caller and responders","control":"Dispatcher retains all response rights"}],"decision_rights":"The model alerts; the dispatcher decides whether to declare suspected OHCA and start CPR coaching and dispatch.","exception_path":"Dispatchers ignore false alerts and follow standard protocols; no response is automatically triggered by the model.","removed_work":["None demonstrated; the alert added information but did not significantly improve the human outcome"],"outcomes":[{"metric":"Dispatcher recognition of confirmed out-of-hospital cardiac arrest","baseline":"90.5% with standard protocol and no model alert","result":"93.1% with the alert, P=.15; no statistically significant improvement","period":"2018-09-01 through 2019-12-31","scale":"169,049 calls screened; 5,242 suspected calls randomized; 654 confirmed OHCA cases","attribution_caveat":"High-quality randomized negative result. The model alone was more sensitive but materially less specific and had lower positive predictive value than dispatchers.","evidence_label":"verified"}],"executive_lesson":"A better classifier is not a better operating model. The alert's low precision and insertion into a time-critical conversation prevented technical sensitivity from becoming human performance.","anti_pattern":"Selecting on model sensitivity while ignoring alert precision and operator response.","questions_for_leaders":["Does the alert improve the human decision?","What false-positive rate is tolerable in a time-critical queue?"],"collections":["negative-results"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/copenhagen-ohca-dispatch-rct"}