Predictors of readmission in patients with congestive heart failure

Repeated hospitalizations among patients (pts) with congestive predictors o f readmission. Methods: Inclusion criteria: admitted to University Hospital with a primary diagnosis of CHF between 10/1/94-9/30/95: lived in Jefferso n county. Exclusions: cardiac transplant during study period; major comorbi dity (e.g. malignancy, advanced renal failure). Predictors of readmission w ere determined by stepwise logistic regression analysis and predictor of ti me to readmission with Cox Proportionate Hazards modeling p<0.05 was consid ered statistically significant. Results: Mean age of the 237 pts was 66.5 y rs; 56% women. Mean left ventricular ejection fraction (LVEF) was 29%; 96% were in NYHA Class III/IV. Mean length of stay was 5 days; 52 pts (22%) had >1 admission. CHF etiologies: Ischemic (42%) hypertensive (37%), idiopathi c (12). Demographic characteristics and insurance status did not predict re admission risk. Predictors of readmission in the logistic and Cox models we re similar. Increased risk of readmission was associated with myocardial is chemia (logistic OR 42.7), past NYHA Class III and IV (OR 32.8), plasmatic creatinine at discharge (OR 1.9) and continued smoking (OR 3.26). History o f CABG was associated with a decreased risk of rehospitalization (OR 0.12). Beta-blocker use was associated with decreased risk, but did not achieve s tatistical significance. ACE-1 use (prescribed in 78% of pts), did not cont ribute to the model. Diabetes Mellitus and a lowe LVEF were more frequent i n the readmitted group, but they did not predict readmission. Conclusion: C HF pts who have evidence of ischemia, advanced symptoms, renal dysfunction, and who continue to smoke are at increased risk for hospital readmission. Pts with these characteristics should be identified prior to hospital disch arge and considered for intensive outpatient intervention.