Effects of Physician Volume on Readmission and Mortality in Elderly Patients with Heart Failure: Nationwide Cohort Study

Abstract

Purpose: Readmission and mortality rates of patients with heart failure are good indicators of care quality. To determine whether hospital resources are associated with care quality for cardiac patients, we analyzed the effect of number of physicians and the combined effects of number of physicians and beds on 30-day readmission and 1-year mortality.

Materials and methods: We used national cohort sample data of the National Health Insurance Service (NHIS) claims in 2002-2013. Subjects comprised 2345 inpatients (age: >65 years) admitted to acute-care hospitals for heart failure. A multivariate Cox regression was used.

Results: Of the 2345 patients hospitalized with heart failure, 812 inpatients (34.6%) were readmitted within 30 days and 190 (8.1%) had died within a year. Heart-failure patients treated at hospitals with low physician volumes had higher readmission and mortality rates than high physician volumes [30-day readmission: hazard ratio (HR)=1.291, 95% confidence interval (CI)=1.020-1.633; 1-year mortality: HR=2.168, 95% CI=1.415-3.321]. Patients admitted to hospitals with low or middle bed and physician volume had higher 30-day readmission and 1-year mortality rates than those admitted to hospitals with high volume (30-day readmission: HR=2.812, 95% CI=1.561-5.066 for middle-volume beds & low-volume physicians, 1-year mortality: HR=8.638, 95% CI=2.072-36.02 for middle-volume beds & low-volume physicians).

Conclusion: Physician volume is related to lower readmission and mortality for heart failure. Of interest, 30-day readmission and 1-year mortality were significantly associated with the combined effects of physician and institution bed volume.

Keywords: Readmission; heart failure; mortality; physician volume.

Fig. 1. Kaplan-Meier curves of 30-day heart failure all-cause readmission and 1-year mortality according to number of physicians and number of beds.

Fig. 2. HR and 95% confidence intervals for 30-day readmission and 1-year mortality associated with combined effects of number of physicians and number of beds: adjusted for patient characteristics and hospital characteristics. ^*p<0.05. HR, hazard ratio.