The Relationship Between the Hospital Readmission Reduction Program and Healthcare Acquired Infections

Previous studies showed that the Hospital Readmission Reduction Program (HRRP) had significantly lowered readmissions and proposed several explanations. One possible mechanism was reducing healthcare-acquired infections (HAIs). Because of the significant associations between HAIs and readmissions, hospitals could have adopted stronger infection control protocols and other direct measures to reduce HAIs. Various quality improvements resulting from the HRRP (e.g., better discharge instructions and follow-up) could have also indirectly reduced HAIs. Therefore, we hypothesize that implementation of the HRRP is associated with a statistically significant reduction in HAIs among Medicare patients with targeted conditions, relative to non-targeted controls. We use the Nationwide Readmissions Database (NRD) to compare six types of HAIs in the pre-policy (2010–2011) and post-policy (2012–2014) periods: Clostridium difficile (C. diff) infection, surgical site infection (SSI), sepsis, catheter-associated urinary tract infection (CAUTI), central line bloodstream infection (CLABSI), and any of the aforementioned infections (i.e., “overall infections”). We use the difference-in-differences (DID) method to compare HAIs for the HRRP-targeted conditions (i.e., the treatment group of Medicare patients with acute myocardial infarction, heart failure, or pneumonia) with three control groups: Medicare patients with gastrointestinal (GI) conditions, Medicare patients with non-targeted conditions, and privately insured patients aged 45 to 65 with the same targeted conditions. We conduct various sensitivity analyses to test the robustness of the findings, including alternative definitions of readmissions, infections, readmission days, and comorbidity measures; models controlling for hospital-level fixed effects; the propensity score matching method and its related weights; and triple difference (DDD) models. All models controlled for patient and hospital characteristics and time-fixed effects. We consistently find that HRRP is associated with significant HAI reductions. Overall infection rates decreased significantly more in the HRRP-targeted conditions than in the control groups, especially for C. diff, sepsis, and CAUTI. However, these findings could have been expanded and, perhaps, improved with better and more granular data. Studies using alternative databases with explicit hospital identifiers are needed to examine the relationship between the HRRP and HAIs by penalty status. Future research should also investigate the direct and indirect mechanisms underlying the HAI reductions and other aspects of patient safety and quality-of-care indicators that may have been affected by the HRRP.