JMIR Publications

ABSTRACT

Background:

Prior literature suggests that psychosocial factors adversely impact health and healthcare utilization outcomes. However, psychosocial factors are typically not captured by the structured data in electronic medical records (EMR) but recorded as free text in different types of clinical notes.

Objective:

Propose a text-mining approach to analyze electronic medical records to identify older adults with key psychosocial factors that predict adverse health care utilization outcomes – measured by 30-day readmission. The psychological factors are appended to the LACE Index for Readmission to predict readmission risk.

Methods:

We conduct a retrospective analysis using EMR notes of 43,216 hospitalization encounters in a hospital from 01 Jan 2017 to 28 Feb 2019. We employed text-mining techniques to extract psychosocial topics which are representative of these patients and test the utility of these topics in predicting 30-day hospital readmission beyond the predictive value of the LACE Index for Readmission.

Results:

We observed that the added text-mined factors improved the AUROC of the readmission prediction by up to 8.46% and this improved predictive accuracy is higher for geriatric patients compared to other patient cohorts.

Conclusions:

The results show the feasibility of extracting psychosocial factors from EMR clinical notes and the value of these notes in improving our readmission risk prediction. Psychosocial profiles of patients can be curated and quantified from text-mining clinical notes and these profiles can be successfully applied to AI models to improve readmission risk prediction. Clinical Trial: N.A.