Education, Teaching, and Quality Improvement |

The Cleveland Clinic Prediction Model for Real-Time Estimation of 30 Day All-Cause Readmission Risk for Patients Admitted With Pneumonia FREE TO VIEW

Umur Hatipoglu, MD; Kevin Chagin, MS; Dhruv Joshi, MD; Brian Wells, MD
Author and Funding Information

Cleveland Clinic-Respiratory Institute, Cleveland, OH

Chest. 2015;148(4_MeetingAbstracts):469A. doi:10.1378/chest.2272873
Text Size: A A A
Published online


SESSION TITLE: Education, Research, and Quality Improvement

SESSION TYPE: Original Investigation Slide

PRESENTED ON: Sunday, October 25, 2015 at 07:30 AM - 08:30 AM

PURPOSE: The Affordable Care Act funded the Centers for Medicare and Medicaid Services (CMS) to establish the Hospital Readmission Reduction Program (HRRP). Pneumonia was selected for the HRRP due to its prevalence, high cost of care, and the existence of a readmission quality metric endorsed by the National Quality Forum. An accurate prediction model for the risk of 30-day readmission after a hospitalization for pneumonia could improve the allocation of post-discharge interventions to the highest risk patients. The purpose of this study was to create such a model.

METHODS: The model was created using a dataset of 1295 patients admitted to the Cleveland Clinic Main Campus with pneumonia between January 1, 2010 and December 31, 2012. Candidate variables were determined after literature review and limited to structured variables consistently available in the electronic health record (EHR). Harrell’s model approximation method was used to rank the variables in the full multiple regression model. The final model was created via backward elimination at a cutoff that maximized the cross-validated concordance statistic. The final model was compared with the CMS risk prediction model among patients 65 years of age and older (n=628). Calibration was assessed graphically by plotting predicted probability versus actual probability.

RESULTS: 330 patients (25%) were readmitted within 30 days of discharge. The final model contained 48 of the 136 candidate variables and had a bias-corrected concordance statistic of 0.76 (95% CI: 0.73, 0.79). The model was well calibrated. Number of admissions in the prior six months, age, lowest pulse within the first 24 hours of admission, and lowest hemoglobin within the last 24 hours were the predictor variables with the greatest weight in the model. The discriminative performance of the study model was superior to the CMS model (c-statistic 0.74 versus 0.59, p <0.0001).

CONCLUSIONS: The proposed risk prediction model for pneumonia readmissions is accurate and includes variables that are available real time via EHR at the point of discharge. The results suggest that institutions should not rely on the CMS model for risk stratification in leveraging local quality improvement efforts.

CLINICAL IMPLICATIONS: Accurate risk prediction for 30 day readmission at the point of discharge is feasible and can potentially be used to focus post-acute care interventions in a high-risk group of patients.

DISCLOSURE: The following authors have nothing to disclose: Umur Hatipoglu, Kevin Chagin, Dhruv Joshi, Brian Wells

No Product/Research Disclosure Information




Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543