Predicting hospital readmission in patients with heart failure: Usefulness of psychosocial factors not included in established risk scores

Background: Frequent readmissions are a major healthcare problem. Heart failure (HF) is the leading cause of readmission. We incorporated psychosocial factors into the In-Patient Evaluation Center (IPEC) and Yale risk scores to evaluate if they provide better prediction of readmission within thirty days than the risk scores alone.

Methods: The additional factors selected were smoking, alcohol use, living alone, depression/anxiety, and enrollment in Telehealth and the Visiting Nurse Services. Data were collected using patient surveys and medical record review. Receiver Operating Characteristic (ROC) curve analyses were performed by comparing the Area Under the Curve (AUC) for the two risk scores as well as augmented risk scores that included the aforementioned six factors to determine their accuracy in predicting hospital readmission. All comparisons were to chance (AUC=0.5).

Results: We evaluated 83 discharged HF patients. The AUC (with standard error) for the IPEC and Yale risk scores were 0.5608 (0.0791) and 0.5547 (0.0723). After the addition of the covariates, the augmented IPEC and augmented Yale AUC's were 0.6563 (0.0748) and 0.6428 (0.0702). When compared to chance, the IPEC (p= 0.44) and the Yale (p=0.45) scores were not statistically different; the augmented IPEC (0.03) and the augmented Yale (0.04) were statistically different.

Conclusion: Psychosocial factors significantly improve the prediction of readmission. This could lead to developing more accurate risk score calculators and programs to target psychosocial issues relevant to readmission. This improvement in care may lead to a decrease in readmissions, which would benefit both HF patients and the healthcare system.