Patient Prediction Model Trims Avoidable Hospital Readmissions


A new prediction model that uses a familiar phrase can help identify potentially avoidable hospital patient readmissions, according to a report in JAMA Internal Medicine.

The retrospective cohort study, "Potentially Avoidable 30-Day Hospital Readmissions in Medical Patients," used a model dubbed HOSPITAL to create a score that targets patients most likely to benefit from pre-discharge interventions. The model is based on seven factors: hemoglobin at discharge, discharge from an oncology service, sodium levels at discharge, procedure during the index admission, index type of admission, number of admissions in the prior 12 months, and length of stay. The HOSPITAL score had fair discriminatory power (C statistic 0.71) and good calibration, the authors noted.

"By definition, these [interventions] are expensive and you really want to reserve them for the patients that are most likely to benefit," says study co-author Jeffrey Schnipper, MD, MPH, FHM, director of clinical research and an associate physician in the general medicine division at Brigham and Women's Hospital in Boston.

The study identified 879 potentially avoidable discharges out of 10,731 eligible discharges, or 8.5%. The estimated potentially avoidable readmission risk was 18%. Dr. Schnipper says that in absolute reduction, the model could cut 2% to 3% of readmissions.

"This is an evolution of sophistication in how we think about this work," Dr. Schnipper adds. "Not all patients have a preventable readmission. Maybe some of those patients are more likely to benefit. The next step is to prove it. That's the gold standard and that’s our next study." TH

Visit our website for more information on 30-day readmissions.

Next Article:

   Comments ()