Risk model predicts 30-day readmission for heart failure
Where a patient lives, his or her race, discharge blood values and previous hospitalization record were among variables that most predicted 30-day heart failure readmission risk, a Boston research group found. The model they developed, simplified from a list of 25 variables, had high accuracy compared with commonly used standards.
Lisa M. Fleming, MD, of the Division of Cardiology at Beth Israel Deaconess Medical Center in Boston, and colleagues wrote that finding best methods to predict risk and intervene early is important. Readmission reduced patients’ quality of life and placed longevity at risk. At the institutional level, they noted, readmission rates reduce healthcare quality performance ratings and place healthcare institutions at risk of not being reimbursed.
Researchers reviewed the hospital patient registry from October 2007 through August 2011. The team acquired a body of 3,413 index heart failure admissions and related data. They were able to break the 25 variables down into three groups: demographic variables, dispositional variables and clinical variables. These were combined into a series of models to determine optimal prediction of 30-day readmission risk. The research team sought a prediction model for readmission that included admission, discharge and sociodemographic characteristics.
They found the best model incorporated discharge creatinine, hematocrit, troponin and hyponatremia; zip code; race; discharge hour; and number of hospitalizations in the year prior. While Fleming et al noted n-terminal probrain natriuretic peptide had a high predictive value, too few patient records included this variable, and it was not included in the final model. Likewise, marital status was not included due to small numbers of unmarried men in the sample.
Fleming at al also noted variables like age, gender, atrial fibrillation, diabetes, discharging department and diuretic use on the day of discharge were not as predictive of risk.
The model had immediate clinical utility, they wrote, allowing providers to target focused interventions on high-risk patients at discharge.
The findings were published in the November issue of the American Journal of Cardiology.