Early hospital readmissions (readmission within 30 days) are common, costly and potentially preventable. Between 18 and 20 % of Medicare beneficiaries discharged from a hospital in 2003 and 2004 were readmitted within 30 days, costing the Medicare program an estimated $17 billion annually. A variety of interventions to reduce readmission have been tested, with mixed results. In FY2013, CMS began assessing financial penalties on hospitals with unplanned readmission rates for congestive heart failure (CHF), acute myocardial infarction (AMI) and pneumonia (PN) that exceeded rates expected for their patient population under the Hospital Readmissions Reduction Program (HRRP). Almost half of all U.S. hospitals face penalties under the program. The novelty of financial penalties in hospital reimbursement, along with their size and scope, make investigation of HRRP's impact a critical priority. Thus, we propose the following questions: Q1: Which hospitals are getting penalties and how are their readmission rates changing over time? Using descriptive, stratified analyses, we will examine penalties and readmission rates over time for various types of hospitals, including those with high, low and no penalty, those serving large low-income and/or minority populations, safety net and financially troubled hospitals. Descriptive analyses will provide context for subsequent analyses and timely, clear information about whether and how the policy should be changed. Q2: What is the impact of HRRP on Medicare readmission rates targeted by the program? Using hierarchical generalized linear models where patient, hospital and market characteristics influence condition- specific readmission rates we will examine readmissions before and after the policy change. Our primary focus will be CHF, AMI and PN, with a secondary focus on new conditions targeted in FY 2015: chronic obstructive pulmonary disease, coronary artery bypass graft surgery, and percutaneous coronary interventions. Q3: Do we observe any spillover effects of HRRP on other readmission rates? If hospitals find it unethical, impractical, or unprofitable to treat Medicare CHF/AMI/PN patients differently from other patients, we may see spillover effects. We will look for these among readmissions for similar conditions to those covered by HRRP and among non-Medicare readmissions for the same conditions. For comparison, we will also look at readmissions for dementia and back pain, two clinically unrelated conditions not covered by HRRP. Using 2010-2014 Health Care Utilization Project (HCUP) State Inpatient Databases from 9 states, publicly reported penalties, American Hospital Association Annual Survey of Hospitals data, Medicare hospital data, Area Resource File and Census data, we will address study questions. We use HCUP data since they are released earlier than Medicare claims and include non-Medicare patients. Our combined descriptive and multivariate approach facilitates timely policy-oriented publications (Q1) and rigorous assessment of HRRP's effect on targeted and non-targeted readmissions for Medicare and non-Medicare populations (Q2 and Q3).
We propose to examine the impact of Medicare's Hospital Readmission Reduction Program (HRRP) by focusing on which hospitals are getting penalties and how their readmission rates are changing over time. Using hierarchical generalized linear models, we will also examine the impact of HRRP on Medicare readmission rates targeted by the program and look for evidence of spillover effects on other readmission rates (e.g. similar conditions, non-Medicare readmissions). The novelty of financial penalties in hospital reimbursement, along with their size and scope, make investigation of HRRP's impact a critical policy priority.
|Bazzoli, Gloria J; Thompson, Michael P; Waters, Teresa M (2018) Medicare Payment Penalties and Safety Net Hospital Profitability: Minimal Impact on These Vulnerable Hospitals. Health Serv Res 53:3495-3506|
|Thompson, Michael P; Waters, Teresa M; Kaplan, Cameron M et al. (2017) Most Hospitals Received Annual Penalties For Excess Readmissions, But Some Fared Better Than Others. Health Aff (Millwood) 36:893-901|
|Thompson, Michael P; Kaplan, Cameron M; Cao, Yu et al. (2016) Reliability of 30-Day Readmission Measures Used in the Hospital Readmission Reduction Program. Health Serv Res 51:2095-2114|