March 23, 2016

Readmission prediction face-off

Using patients’ health records to assess preparedness for hospital discharge is more effective at predicting readmission or death than commonly used questionnaires.

For any hospitals that may want to survey newly discharged patients about their preparedness for discharge, there are multiple questionnaires available for use. In the Journal of Hospital Medicine, Amanda Mixon, M.D., MS, MSPH, and colleagues report that, when used strictly to assess risk of unplanned readmission or death, two commonly used discharge preparedness questionnaires paled in comparison to an administrative algorithm that mainly uses data from health records.

Some 1,239 recently discharged cardiovascular patients were surveyed by phone using the 11-item B-PREPARED questionnaire and the 3-item CTM-3 questionnaire. Using data from these patients’ health records, investigators concurrently ran an algorithm called LACE: Length-of-stay, Acuity, Comorbidities, ED visits. Then they counted eventual readmissions and deaths, with cutoffs at 30 days and 90 days post-discharge.

Among these three measures, only the LACE algorithm was independently associated with readmission or death at both 30 and 90 days. B-PREPARED bore independent association at 90 days, but only weakly by comparison with LACE.

Joining Mixon in the study were Kathryn Goggins, MPH, Susan Bell, M.B.B.S., MSCI, Eduard Vasilevskis, M.D., MPH, Samuel Nwosu, MS, Jonathan Schildcrout, Ph.D. and Sunil Kripalani, M.D., MSc. The study was supported by the National Institutes of Health (HL109388) and the U.S. Dept. of Veterans Affairs (HSR&D CDA 12-168).

Send suggestions for articles to highlight in Aliquots and any other feedback about the column to aliquots@vanderbilt.edu