Clinical decision support applications that use natural language processing to identify heart failure (HF) patients and calculate readmission risk helped reduce 30-day mortality while improving discharge rates, according to a study published online in the Journal of the American Medical Informatics Association (JAMIA).
The researchers, from Intermountain Healthcare and the University of Utah, were able to identify HF patients earlier by using natural processing, and with that information--as well as data from the electronic health record--created a predicative score of the risk for each patient to be readmitted within 30 days.
Once the patients with HF risk and readmission risk were identified, the researchers were able to track the patients' care through a multidisciplinary care process pathway (CPP).
Reducing heart failure readmissions has been a struggle for hospitals. Nearly 1 in 4 patients are readmitted to hospitals within 30 days of discharge due to the condition, according to research published in the Journal of Cardiac Failure.
Use of the CPP "reduced 30-day mortality and significantly increased patient discharges to home care rather than to a specialized nursing facility," the authors found. "The results of this study also suggest that using a real-time, structured, and multidisciplinary CPP for HF patients who are at risk for poor outcomes can significantly reduce mortality by improving the use of ideal postacute care," they conclude.
Running natural language processing in the EHR also worked for Atrius Health, a Boston-based not-for-profit alliance of medical groups, which used the process to run risk analyses on patients in order to flag those at risk for issues such as diabetes and obesity, FierceHealthIT previously reported.
To learn more:
- here's the study abstract