Use of natural language processing in electronic medical records to identify pregnant women with suicidal behavior towards a solution to the complex classification problem

• Published in: European journal of epidemiology Vol. 34; no. 2; pp. 153 - 162
• Main Authors: Zhong, Qiu-Yue, Mittal, Leena P., Nathan, Margo D., Brown, Kara M., González, Deborah Knudson, Cai, Tianrun, Finan, Sean, Gelaye, Bizu, Avillach, Paul, Smoller, Jordan W., Karlson, Elizabeth W., Cai, Tianxi, Williams, Michelle A.
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Science + Business Media 01.02.2019; Springer Netherlands; Springer Nature B.V
• Subjects: Adaptive algorithms; Algorithms; Cardiology; Codification; Data Mining; Diagnostic systems; Electronic Health Records; Electronic medical records; Epidemiology; Female; Humans; Infectious Diseases; International Classification of Diseases - standards; Language; Medical records; Medicine; Medicine & Public Health; Mental health; Natural Language Processing; Oncology; Pregnancy; Pregnancy Complications; PSYCHIATRIC EPIDEMIOLOGY; Public Health; Self destructive behavior; Suicidal behavior; Suicidal Ideation; Suicides & suicide attempts; Electronic medical; Classification algorithm; Natural language processing; Suicidal behavior; Pregnant women; Records
• ISSN: 0393-2990; 1573-7284; 1573-7284
• DOI: 10.1007/s10654-018-0470-0

We developed algorithms to identify pregnant women with suicidal behavior using information extracted from clinical notes by natural language processing (NLP) in electronic medical records. Using both codified data and NLP applied to unstructured clinical notes, we first screened pregnant women in Partners HealthCare for suicidal behavior. Psychiatrists manually reviewed clinical charts to identify relevant features for suicidal behavior and to obtain gold-standard labels. Using the adaptive elastic net, we developed algorithms to classify suicidal behavior. We then validated algorithms in an independent validation dataset. From 275,843 women with codes related to pregnancy or delivery, 9331 women screened positive for suicidal behavior by either codified data (N = 196) or NLP (N = 9,145). Using expert-curated features, our algorithm achieved an area under the curve of 0.83. By setting a positive predictive value comparable to that of diagnostic codes related to suicidal behavior (0.71), we obtained a sensitivity of 0.34, specificity of 0.96, and negative predictive value of 0.83. The algorithm identified 1423 pregnant women with suicidal behavior among 9331 women screened positive. Mining unstructured clinical notes using NLP resulted in a 11-fold increase in the number of pregnant women identified with suicidal behavior, as compared to solely reliance on diagnostic codes.