Predicting early psychiatric readmission with natural language processing of narrative discharge summaries

Abstract

The ability to predict psychiatric readmission would facilitate the development of interventions to reduce this risk, a major driver of psychiatric health-care costs. The symptoms or characteristics of illness course necessary to develop reliable predictors are not available in coded billing data, but may be present in narrative electronic health record (EHR) discharge summaries. We identified a cohort of individuals admitted to a psychiatric inpatient unit between 1994 and 2012 with a principal diagnosis of major depressive disorder, and extracted inpatient psychiatric discharge narrative notes. Using these data, we trained a 75-topic Latent Dirichlet Allocation (LDA) model, a form of natural language processing, which identifies groups of words associated with topics discussed in a document collection. The cohort was randomly split to derive a training (70%) and testing (30%) data set, and we trained separate support vector machine models for baseline clinical features alone, baseline features plus common individual words and the above plus topics identified from the 75-topic LDA model. Of 4687 patients with inpatient discharge summaries, 470 were readmitted within 30 days. The 75-topic LDA model included topics linked to psychiatric symptoms (suicide, severe depression, anxiety, trauma, eating/weight and panic) and major depressive disorder comorbidities (infection, postpartum, brain tumor, diarrhea and pulmonary disease). By including LDA topics, prediction of readmission, as measured by area under receiver-operating characteristic curves in the testing data set, was improved from baseline (area under the curve 0.618) to baseline+1000 words (0.682) to baseline+75 topics (0.784). Inclusion of topics derived from narrative notes allows more accurate discrimination of individuals at high risk for psychiatric readmission in this cohort. Topic modeling and related approaches offer the potential to improve prediction using EHRs, if generalizability can be established in other clinical cohorts.

Predictive modeling for classification of positive valence system symptom severity from initial psychiatric evaluation records

“…Our NLP methods could be adapted to further explore these problems and potentially expand upon the current proposed solutions. For example, it has been shown that NLP can aid in identifying psychiatric patients at risk of early readmission from narrative discharge summaries [22]. By modifying the Sectionizer component to handle narrative discharge summaries and removing the PVS filtering in the Question-Answer-Extraction component, we could apply our methods to the task of predicting psychiatric readmission from narrative discharge summaries.…”

Depression Severity Estimation from Multiple Modalities

“…Wang et al [53] applied data mining techniques to build models which achieved a precision of 80% for detecting de-pression based on sentiment analysis of users on a Chinese micro-blogging platform. Rumshisky et al [42] demonstrated through a study on 4687 patients that NLP techniques such as topic modeling can be used to improve prediction of psychiatric readmission.…”

Predicting mental conditions based on “history of present illness” in psychiatric notes with deep neural networks

“…Rumshisky et al [38] predict early readmits (within 30 days) to inpatient psychiatric units through topic models built with discharge summaries. Jackson et al [17] identify over 40 key symptoms (e.g., aggression, apathy, irritability, and stupor) of severe mental illness based on discharge summaries from nearly 8000 patients visiting a UK based mental healthcare provider using SVM models.…”