- Jennifer Shaw, Senior Researcher, Southcentral Foundation, Anchorage, AK
- Jaedon Avey, Senior Researcher, Southcentral Foundation, Anchorage, AK
- Troy Wolcoff, Clinical Supervisor, Southcentral Foundation, Anchorage, AK
Background: The suicide rate among Alaska Native and American Indian people in Alaska is twice the overall state rate and three times the national rate for Alaska Native and American Indian people. Recent research has shown that healthcare systems are critical sites for preventing suicide. Self-report screening methods (such as PHQ-9 or C-SSRS) are necessary but not sufficient to prevent suicide in health systems. Recently, machine learning techniques have been applied to routinely-collected EHR data to detect suicide risk. Setting and Methods: In 2019, Southcentral Foundation (SCF), an Alaska Native-owned healthcare system, validated a suicide risk prediction algorithm developed by the Mental Health Research Network (MHRN). In 2020, interviews were conducted with patients, providers, and leaders to identify stakeholder concerns and priorities for implementation of the algorithm. Transcribed data were thematically analyzed within and across participant sub-groups. Results: The MHRN algorithm accurately identified risk of suicide attempt within 90 days of a primary care visit among SCF patients with behavioral health diagnoses (AUC=0.826; 95% CI 0.809-0.843). As of March 1, 2021, 14 SCF patients and 15 providers and leaders had been interviewed. All participant groups favored the EHR-based approach to suicide risk detection. Patients expressed concern about who would have access to EHR data for this purpose and how it would be used. For some, acceptability was predicated on prior patient consent for using health records for suicide risk detection and prevention. All participants regarded the patient-provider relationship and trust as critical factors in successful implementation of the algorithm, as well as robust EHR data system and accessible, behavioral health care. Conclusions: Preliminary data analyses suggest strong but conditional stakeholder support for implementation of an EHR-based algorithm suicide risk detection in an Alaska Native owned health system.
- Identify the characteristics of SCF's EHR-based approach to suicide risk detection
- Analyze the study performed by SCF on attitudes of patients, providers, and leaders to the EHR-based approach to suicide risk detection
- Determine the key components of the successful deployment of an EHR-based approach to suicide risk detection in a tribal health system