A new study in The New England Journal of Medicine is shedding light on some of the complexity around Medicare’s readmissions reduction program, and reinforcing the importance of federal officials’ exercising very good discernment in how they measure hospitals’ success in the program, as it evolves forward.
The article, entitled “Effect of a Hospital-wide Measure on the Readmissions Reduction Program,” was written by Rachael B. Zuckerman, Ph.D., Karen E. Joynt Maddox, M.D., M.P.H., Steven H. Sheingold, Ph.D., Lena M. Chen, M.D., and Arnold M. Epstein, M.D. As the authors write, “The Hospital Readmissions Reduction Program (HRRP) was established as part of the Affordable Care Act (ACA) in 2010 to give hospitals incentives to reduce readmissions within 30 days after hospital discharge. However, a number of concerns have been raised about the implementation of the HRRP. In particular, the current measures of readmissions include only a limited number of clinical conditions; our previous work,” they write, “showed that decreases in readmissions were greater for targeted than for nontargeted conditions, which suggests that there is a potential benefit to be gained from including a wider range of conditions.”
The authors go through considerable detail in sharing their analysis of data that began to be gathered in October 2012, and which has led to Medicare payment cuts for “higher-than-expected 30-day readmission rates after hospitalizations for acute myocardial infarction, heart failure, or pneumonia. Penalties were expanded to include total hip and knee replacement and chronic obstructive pulmonary disease in fiscal year 2015. All penalties are based on performance over a 3-year period; HRRP penalties in 2015 were based on readmissions for hospitalizations from July 2010 through June 2013.”
The authors note that, “Although the ACA states that penalties are to be based on excess readmissions for ‘specified conditions,’ there has been support to move to a hospital-wide readmission measure that provides incentives for improvement across a wider range of conditions. The president’s budgets in 2016 and 2017 and the Medicare Payment Advisory Commission (MedPAC) have all proposed transitioning to a hospital-wide measure,” they write. “A hospital-wide readmission measure has also been endorsed by the National Quality Forum and is reported on the Hospital Compare website of the Centers for Medicare and Medicaid Services (CMS). The hospital-wide measure includes many conditions and could have sufficient sample sizes with 1 year of data instead of 3 years, allowing the penalties to reflect changes in performance with a much smaller time lag, and it may include additional hospitals that do not meet the volume threshold for the condition-specific measures.”
Essentially, though, the authors argue, shifting to a hospital-wide measure would be a big mistake. “In this study,” they state, “we focused on how a transition from the first five specific conditions included in the HRRP to a single, hospital-wide measure might change the number of hospitals meeting the volume threshold for readmissions measures and the magnitude of penalties overall, at both safety-net hospitals and other (i.e., non–safety-net) hospitals. We also examined the effect of this transition under three alternative policy options: first, if overall penalties were kept budget-neutral; second, if penalties were calculated with adjustment for Medicaid enrollment; and third, if penalties were applied to groups of hospitals stratified according to the proportion of Medicaid and low-income Medicare patients treated.”
Here's what the authors did in order to simulate the results of a shift to a hospital-wide measure: using the 2014 measure update specifications, they calculated risk-adjusted 30-day readmission rates, excess readmission ratios, and standard errors from bootstrapping for the five conditions in the fiscal year 2015 HRRP on the basis of data from fiscal years 2011 through 2013; they also calculated 30-day readmission rates and excess readmission ratios for the hospital-wide measure using 1 year of data (fiscal year 2013) and 2014 measure specifications (as reported on Hospital Compare).
Could safety-net hospitals really be put in peril?
What the researchers found was this: “The hospital-wide readmission rate and most of the condition-specific rates were in the range of 15 to 20 percent, with the exceptions of total hip or knee replacement, for which the estimate (±SE) was 4.9±0.09 percent, and heart failure, for which the estimate was 22.2±0.02 percent. Safety-net hospitals consistently had slightly higher readmission rates than other hospitals, although these differences were small.” But, they write, “As compared with the use of condition-specific measures, moving to a hospital-wide measure would be associated with substantially higher mean annual penalties across all hospitals (the mean was computed with hospitals that had no penalty included); penalties would increase to 0.89±0.01 percent of base DRG payments, or $393,000±14,000, but only 43 percent of hospitals would be penalized. Moving to the hospital-wide readmission measure would also substantially increase the disparity between safety-net and other hospitals: the mean penalty as a percentage of base DRG payments would be 0.41 percentage points ($198,000) higher among safety-net hospitals.”
What’s more, the study’s authors found, if conditions of budget neutrality were added into the equation, “[T]he mean penalty among safety-net hospitals [would be] $68,000 higher than that among other hospitals. Adjustment for Medicaid enrollment without budget neutrality decreased the differential in the mean penalty as a percentage of base DRG payments between safety-net hospitals and other hospitals to 0.29 percentage points ($154,000). Assigning penalties within strata according to DSH index decile,” they found in their analysis, “resulted in a mean penalty as a percentage of base DRG payments that was 0.14 percentage points lower among safety-net hospitals than among other hospitals, although the penalties in absolute dollars were still slightly higher among safety-net hospitals ($21,000), because the penalized hospitals have higher total payments from Medicare.”
Yes, that’s a lot of numbers and percentages. But the bottom line is clear: at least in absolute-dollar terms, shifting from very specific measures to a broad, hospital-wide measure, could prove harmful to the nation’s safety-net hospitals—the very institutions that need help and support in the current climate of reduced reimbursement and the shift over to broad value-based payment in U.S. healthcare.
The reality is that there is so much that remains unknown about how federal healthcare officials will use the levers of reimbursement, regulation, and other incentives, in order to try to push providers forward into providing value for payment. The intentions are largely good ones, as is the broad vision. And yet, as with so many things, the devil is always in the details when it comes to reimbursement changes.
And if safety-set hospitals are meaningfully disadvantaged by any newly proposed changes to hospital payment, then federal healthcare officials need to do a quick “regroup” and rethink their strategies or their tactics—because safety-net hospitals are going to be needed more than ever in healthcare, as more Medicaid and uncompensated patients come into those institutions for care.
Measurement, as all those who have worked with it can attest, is inevitably a tricky thing. Let’s just hope that federal healthcare officials think through any HRRP program-related changes very, very carefully, as any changes detrimental to safety-net hospitals could end up having a domino effect across U.S. healthcare.