Abstract P44: Assessing the Risk of Hospital Readmission in Patients with Congestive Heart Failure: A Comparison of 4 Risk Scores

• Published in: Circulation Cardiovascular quality and outcomes Vol. 4; no. suppl_2
• Main Authors: Njeim, Mario, Alaali, Yathreb, Assaad, Mahmoud, Ambulgekar, Nikhil, Bou Malham, Sarah, Abas, Mustafa, Hudson, Michael
• Format: Journal Article
• Language: English
• Published: 01-11-2011
• ISSN: 1941-7713; 1941-7705
• DOI: 10.1161/circoutcomes.4.suppl_2.AP44

Abstract only
Background:
Readmission within 30 days of a prior hospitalization for congestive heart failure (CHF) has emerged as a major marker of quality improvement and payment reform. Whereas a number of interventions are available to improve outcomes and reduce readmission rates, there remains an urgent need for accurate risk stratification tools that could be used to build cost-effective and personalized CHF management programs.
Methods:
We conducted a retrospective cohort study on patients discharged from a major urban hospital between June 2009 and October 2009 with a primary diagnosis of CHF. Patients were followed until their first all-cause hospital readmission. We made minor modifications to 4 different readmission risk scores in order to assess and compare their ability to estimate the risk of hospital readmission at 30 days from discharge. Three of the scores (Chin and Goldman 11 point scoring, Philbin and DiSalvo 15 point scoring and Krumholz et al 4 point scoring) were retrieved from reviewing the literature. The fourth tool (Henry Ford score) is an internal non-validated computed calculator based on 8 predictive factors, previously derived within our institution. The scores were compared by logistic regression and area under the ROC curve.
Results:
285 patients were included in the study. The readmission rate at 30 days was 23.5% (67 of 285). A logistic regression revealed that both Chin and Goldman and Philbin scores were poor and statistically non-significant predictors of 30-day all-cause readmission (p-values = 0.784 and 0.165 respectively). When compared by logistic regression and area under the ROC curve, Henry Ford tool (p-value = 0.003; AUC = 0.64) performed the best followed by the Krumholz score (p-value = 0.017; AUC = 0.60).
Conclusions:
In an era where there is an urgent need to improve and personalize heart failure management programs, simple scores can be used to stratify patients according to their risk of re-hospitalization. Krumholz and Henry Ford scores are both accurate tools for the prediction of 30-day all-cause readmission. Krumholz score relies on 4 easily retrievable risk factors while the Henry Ford tool is a computed model that integrates 8 different factors.