The opioid epidemic in the US, coupled with emerging data on the association between opioid use and suicidal behavior, highlights the urgent need for research that identifies patients at greatest risk and provides strategies to mitigate this risk. However, the development of risk algorithms that provide clinical guidance for providers with patients using opioids who may have underlying mental disorders or comorbid diagnoses has been nonexistent. Recent work has demonstrated the compelling need for a better understanding of, and clinical decision support tools to address, suicide risk among prescription opioid drug users. The proposed study provides an opportunity to address gaps in both the identification and clinical management of suicide risk in patients using prescription opioids. We propose to use transfer learning approaches to identify the clinical and demographic characteristics associated with elevated risk of suicidal behavior among prescription opioid users; to develop clinical phenotypes of patients with higher risk of suicidal behavior associated with prescription opioids, and to incorporate these phenotypes in a clinical decision support platform that can be used for identification and intervention at the point of care; and to conduct a pilot study implementing and evaluating the impact of the clinical decision support platform in the existing clinical workflow in 3 diverse clinical settings. Our approach is specifically designed to address the incomplete picture of patient risk in existing models due to the fragmentation of clinical care across settings. We do this by developing algorithms that can statistically ?borrow? information from more comprehensive patient datasets and apply it to more limited datasets in a particular healthcare setting. The proposed work will draw on comprehensive clinical data from a mature health information exchange containing more than 2.3 million patients across the spectrum of clinical care to develop a statistically robust method to measure suicide risk associated with prescription opioid use. The clinical decision support tool developed under this proposal will provide a generalizable platform that could be extended to other opioid related risks, e.g., OUD and overdose. The potential public health significance of this study is substantial. The fragmentation of the healthcare system, particularly in relation to patients' behavioral health needs, highlights the critical need to cultivate comprehensive, system-wide approaches to identifying and managing at patients using prescription opioids who are at risk of suicide.
Suicide and prescription opioid use are two of the most serious public health problems facing the United States. The proposed project will demonstrate an innovative method for identifying patients using prescription opioids who are at risk of suicidal behavior.
