Elsevier

American Heart Journal

Volume 164, Issue 3, September 2012, Pages 365-372
American Heart Journal

Clinical Investigation
Congestive Heart Failure
Predicting the risk of unplanned readmission or death within 30 days of discharge after a heart failure hospitalization

https://doi.org/10.1016/j.ahj.2012.06.010Get rights and content

Background

The accuracy of current models to predict the risk of unplanned readmission or death after a heart failure (HF) hospitalization is uncertain.

Methods

We linked four administrative databases in Alberta to identify all adults discharged alive after a HF hospitalization between April 1999 and 2009. We randomly selected one episode of care per patient and evaluated the accuracy of five administrative data-based models (4 already published, 1 new) for predicting risk of death or unplanned readmission within 30 days of discharge.

Results

Over 10 years, 59 652 adults (mean age 76, 50% women) were discharged after a HF hospitalization. Within 30 days of discharge, 11 199 (19%) died or had an unplanned readmission. All 5 administrative data models exhibited moderate discrimination for this outcome (c-statistic between 0.57 and 0.61). Neither Centers for Medicare and Medicaid Services (CMS)–endorsed model exhibited substantial improvements over the Charlson score for prediction of 30-day post-discharge death or unplanned readmission. However, a new model incorporating length of index hospital stay, age, Charlson score, and number of emergency room visits in the prior 6 months (the LaCE index) exhibited a 20.5% net reclassification improvement (95% CI, 18.4%-22.5%) over the Charlson score and a 19.1% improvement (95% CI, 17.1%-21.2%) over the CMS readmission model.

Conclusions

None of the administrative database models are sufficiently accurate to be used to identify which HF patients require extra resources at discharge. Models which incorporate length of stay such as the LaCE appear superior to current CMS-endorsed models for risk adjusting the outcome of “death or readmission within 30 days of discharge”.

Section snippets

Study setting

The province of Alberta has a single-payer, government-funded health care system that provides universal access to over 3.7 million people for hospital, emergency department (ED), and physician services. This study received ethics approval from the Health Ethics Research Board at the University of Alberta.

Data sources

This study used de-identified data from 4 administrative databases maintained by Alberta Health to create our study cohort including (1) the Discharge Abstract Database, which records the

Results

During our study, 59 652 adults in Alberta were discharged after being hospitalized with a diagnosis of HF listed as either the most responsible or a secondary diagnosis. Death or unplanned readmission occurred in about one fifth of patients in the first 30 days after discharge (Table I, Table II), with most of these events being readmissions (and 18.6% of the readmissions were for HF as most responsible diagnosis). Patients who were subsequently readmitted or had died within 30 days of

Discussion

This is the first study to examine the discriminative ability of the CMS-endorsed prediction models and the LACE/LaCE indices in a broad heart failure population and for the composite outcome of “death or unplanned readmission within 30 days of discharge.” Although LACE and LaCE demonstrated only moderate discrimination for predicting 30 day unplanned readmission or death, thereby limiting their use for patient-level prediction, both indices were significantly better able to predict the

Disclosures

Drs McAlister, Kaul, and Ezekowitz receive salary support from Alberta Innovates-Health Solutions and this project was funded by the Canadian Institutes of Health Research and Pfizer Canada through a peer-reviewed operating grant. The funding agencies did not have input into study design, data collection, interpretation of results, or write up/approval for submission.

Acknowledgements

This study is based in part on data provided by Alberta Health. The interpretation and conclusions contained herein are those of the researchers and do not necessarily represent the views of the Government of Alberta. Neither the Government of Alberta nor Alberta Health express any opinion in relation to this study.

References (28)

  • D.S. Lee et al.

    Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model

    JAMA

    (2003)
  • R. Amarasingham et al.

    An automated model to identify heart failure patients at risk for 30-day readmission or death using electronic medical record data

    Med Care

    (2010)
  • B.G. Hammill et al.

    Incremental value of clinical data beyond claims data in predicting 30-day outcomes after heart failure hospitalization

    Circ Cardiovasc Qual Outcomes

    (2011)
  • H.M. Krumholz et al.

    An administrative claims model suitable for profiling hospital performance based on 30-day mortality rates among patients with heart failure

    Circulation

    (2006)
  • Cited by (0)

    View full text