NLP to Identify Opioid Abuse - MUSC

​​Melax Tech developed portable NLP tools to recognize uncoded opioid overdose related concepts from Electronic Health Records (EHR) notes in collaboration with the Medical University of South Carolina (MUSC).

Opioid misuse and dependence continue to be a significant and growing cause of preventable morbidity and mortality in the United States. Over the last 10 years, the number of individuals presenting to emergency departments (EDs) as a result of intentional or unintentional opioid-related overdose(OD) has dramatically increased. Many of these individuals pass away or have a repeat OD presentation within a year and public health costs associated with opioid-related OD are high.

Melax Tech, in collaboration with the Medical University of South Carolina (MUSC), developed a clinical NLP pipeline to recognize opioid overdose disorder (OOD) cases from ED notes. This pipeline contains machine learning, deep learning and context rules to recognize uncoded OOD cases at sentence level of EHR notes. The pipeline is highly portable and can be adapted to other organizations with small amounts of customizations.

The NLP pipeline has already been adopted by the Accrual to Clinical Trials (ACT) Network, which aims to support the design of trials for survivors of opioid overdoses presenting to emergency departments (ED). Four institutions (Medical University of South Carolina (MUSC), Dartmouth Medical School (DMS), University of Kentucky (UK), and University of California San Diego (UCSD)) worked to adapt the pipeline to the ACT network.

Details of the OOD NLP solution and ACT network progress can be found in this paper: Lenert LA, Zhu V, Jennings L, McCauley J, Obeid J, Ward R, Hassanpour S, Marsch LA, Hogarth M, Shipman P, Harris DR. Enhancing Research Data Infrastructure to Address the Opioid Epidemic: The Opioid Overdose Network (02-Net). medRxiv. 2021 Jan 1.