Reducing Excess Readmissions: Promising Effect of Hospital Readmissions Reduction Program in US Hospitals

Ning Lu; Kuo-Cherh Huang; James A. Johnson


Int J Qual Health Care. 2016;28(1):53-58. 

In This Article


Characteristics of Hospitals

Among eligible hospitals included in our study, 25.4% were located in the rural areas, 59.1% were small hospitals with the number of beds fewer than 200 and 25.1% were safety-net hospitals (Table 1). Majority of the hospitals (91.6%) were a member of Council of Teaching Hospital of the Association of American Medical Colleges. For FY 2013, the percentage of hospitals that were identified as having excess readmissions by CMS was 47.5% for PN, 50.2% for AMI and 48.4% for HF. These percentages did not change significantly for the next two fiscal years of 2014 and 2015 (Table 2). Due to the underlying structure of HRRP, the percentage of hospitals that were identified as having excess readmissions will remain about 50% and face a financial penalty even as hospitals overall improve over years.

As shown in Table 2, hospitals with excess readmissions for the applicable conditions differed significantly from hospitals without excess readmissions by characteristics. Hospitals with excess readmissions tended to be large, located in metropolitan areas, not-for-profit, safety-net and hospitals with larger percentages of Medicaid patients.

Effect of HRRP on Excess Readmisisons

To examine the effect of HRRP on excess readmissions, we focused our further analyses on hospitals with excess readmissions by restricting our sample to hospitals that had readmission ratios >1 in FY 2013. We found that the excess readmissions ratios reduced significantly after FY 2013 for PN, AMI and HF (Table 3). Specifically, the results showed a significant downward trend in readmission ratios for PN from FY 2013 to 2014 (ratio reduction 0.016, 95% CI: 0.013–0.018, P < 0.001), from FY 2014 to 2015 (ratio reduction 0.019, 95% CI: 0.015–0.023, P < 0.001) and with the greatest reduction from FY 2013 to 2015 (ratio reduction 0.035, 95% CI: 0.030–0.039, P < 0.0001). For AMI, the readmission ratios were reduced 0.053 (95% CI: 0.043–0.064) from FY 2013 to 2014 (P < 0.001), 0.028 (95%CI: 0.018–0.039) from FY 2014 to 2015 (P < 0.001) and 0.082 (95% CI: 0.069–0.094) from FY 2013 to 2015 (P < 0.001). For the same period, the readmission ratios reduction for HF was 0.013 (95% CI: 0.009–0.018, P < 0.001), 0.021 (95% CI: 0.016–0.026, P < 0.0001) and 0.034 (95% CI: 0.028–0.04, P < 0.001), respectively. These reductions in readmission ratios signified significant decreases in excess readmissions of 3.3% for pneumonia, 7.6% for acute myocardial infarction and 3.2% for heart failure after 3 years of implementing the HRRP.

Effect of HRRP on Excessive Readmissions According to Hospital Characteristics

To examine whether the effect of HRRP on reducing excess readmissions differed according to hospital characteristics, we included the two-way interaction terms in repeated-measures analysis of variance models. We categorized the proportions of Medicare and Medicaid patients into quartiles with the 1st quartile being the lowest and 4th being the highest proportion of percentage of Medicare and Medicaid patients admitted to a hospital. We had hypothesized that the HRRP might have a greater effect on safety-net hospitals or on hospitals with a higher proportion of Medicare or Medicaid patients. The results of our analyses demonstrated a stronger response of safety-net hospitals to the effect of HRRP on excess readmissions for AMI (P < 0.05, Fig. 1). However, the effect of HRRP on excess readmissions for PN and HF did not differ according to whether a hospital was safety-net hospital or not. In addition, we found no evidence that the effect of HRRP was associated with the proportion of Medicare or Medicaid patients admitted to a hospital for any of the three conditions under study (P > 0.05).

Figure 1.

Differential effect of HRRP by safety-net hospitals for AMI.

As for other related hospital characteristics, the results of our analyses indicated that the hospital size, location and the type of ownership were associated with a differential response to the effect of HRRP on readmissions for AMI (Table 4). Specifically, small hospitals (<200 beds), public hospitals and hospitals located in rural areas had a stronger response to the effect of HRRP than their counterpart hospitals. For heart failure, we found the effect of HRRP only differed by the hospital ownership with a stronger HRRP effect on public hospitals (Table 4). For pneumonia, no differential effect of the HRRP on excess readmissions was found associated with any hospital characteristics under study.