OBJECTIVE: The authors aimed to assess the adequacy of traditional billing codes for capturing behavioral health crisis events and to develop a comprehensive claims-based methodology for identifying such events among Medicaid beneficiaries. METHODS: In this retrospective analysis of Medicaid claims data collected from 2016 to 2021, the authors conducted a five-step, multisource approach that incorporated traditional crisis service codes, behavioral health-related encounters in emergency department and urgent care settings, diagnosis codes indicating suicide attempts or intentional self-harm, naloxone administration, and emergency procedures associated with behavioral health diagnoses. Diagnosis, service, and place-of-service indicators were used to identify crisis service events beyond those captured by traditional billing codes. RESULTS: In 2021, traditional crisis billing codes identified 645,940 beneficiaries compared with >2.3 million identified with the expanded approach, capturing <28% of individuals experiencing a behavioral health crisis. From 2016 to 2021, the rate of individuals experiencing a crisis service event increased by 6.6% (from 31.8 to 33.9 per 1,000 beneficiaries), with the largest increases observed for naloxone administration (50.0%) and suicide attempt or self-harm events (14.8%). Compared with traditional codes, the expanded method identified a higher proportion of rural residents, older individuals, and individuals without a formal behavioral health diagnosis. CONCLUSIONS: Reliance on traditional crisis billing codes was found to significantly underestimate the true burden of behavioral health crises among Medicaid populations. A multisource approach incorporating diagnosis, service, and place-of-service indicators provides a more accurate and inclusive framework for crisis surveillance, with important implications for evaluating and strengthening behavioral health crisis systems.
Karakus et al. (Wed,) studied this question.