Text-mining of psychosocial information from patients’ clinical notes improved prediction of readmission risk

​Key Findings

1. Psychosocial information extracted from patients’ clinical notes using text-mining improved prediction of readmission risk

2. Combining psychosocial information captured by non-physicians and physicians improved prediction the most.

3. Improvements in predicting readmission risk were observed for all types of patients, with the highest improvements observed for geriatric patients.

SOURCE:
GERI

DOI:
GERI Research Brief 2023-09-001 [PDF, 699 KB]

AUTHOR(S):
Dr Jonathan Gao