CMS's 30-Day Post-MI Readmit Metric Undercut by NCDR Analysis

Fran Lowry

April 28, 2017

The risk-adjusted 30-day readmission rate for patients treated for acute MI is not a good reflection of hospital quality of care, nor is it closely related to clinical outcomes within the same period or over the next year, conclude researchers[1].

Their analysis, based on data collected at centers participating in the National Cardiovascular Data Registry (NCDR)/Acute Coronary Treatment and Intervention Outcomes Network (ACTION) Registry–Get With the Guidelines (GWTG) effort, was published April 26, 2017 in JAMA Cardiology.

The 30-day readmission metric is used controversially by the US Centers for Medicare and Medicaid Services (CMS) to judge hospitals' quality of care. It needs to be tweaked, first author Dr Ambarish Pandey (University of Texas Southwestern Medical Center, Dallas) told heartwire from Medscape.

"The current CMS readmission metric that it uses to penalize hospitals for excess 30-day readmission does not correlate with long-term clinical outcomes at 1 year and also does not correlate with the quality of care that these patients receive during their index hospitalization," Pandey said.

"Also, more important, there is an inequitable distribution of penalties, such that hospitals that treat a greater proportion of socially or medically disadvantaged patients are penalized unfairly despite comparable quality of care," he said.

"I think the burden of the readmission penalty is substantial, so while a lot of hospitals were getting penalized it was unclear whether or not it was a good metric to begin with. Validating the metric should have been an important step before implementation."

The analysis from Pandey and his colleagues included centers that participated in the CMS Hospital Readmissions Reduction Program between July 1, 2008, and June 30, 2011. Among the 380 hospitals that treated 176,644 patients with MI during this period, 43% had higher-than-expected 30-day readmission rates.

The proportion of black patients was higher among hospitals that had higher 30-day readmission rates than in those with lower readmission rates (7.6% vs 4.5%; P=0.01).

Similarly, the proportion of patients with signs of heart failure at admission and those with bleeding complications were greater in hospitals that had higher 30-day readmission rates.

There was no significant association between adherence to MI performance measures and higher-than-expected 30-day readmission rates.

Clinical outcomes data at 1 year were available for 51,453 patients. The risk for mortality or all-cause readmission within 1 year was higher for patients treated at hospitals with higher 30-day readmission rates.

This association was largely driven by readmissions that occurred early after discharge but was not significant 30 days after discharge, Pandey said.

"When we compared the quality of care as well as the long-term clinical outcomes across all hospitals, we found no meaningful differences in the quality of care provided, as measured by using the performance metrics that ACTION uses for myocardial infarction, nor was the 1-year mortality different across all of the hospitals," he said.

Hospitals Can Be "Penalized Unfairly"

"We have various concerns about the fair and equitable allocation of CMS penalties for readmission," Pandey said.

"We see that half of those that take care of large numbers of socioeconomically challenged patients, those with more African Americans, and hospitals that take on the sickest patients are penalized even though the quality of care they are providing and their long-term outcomes are not different across all of the hospitals. True, they had more readmissions, but those did not really matter for long-term mortality.

"The basic premise of the CMS program was that using 30-day readmission as a metric should identify hospitals that do a good job," he said. "But what we see here is that the 30-day readmission does not really track with the quality of care that patients get and also does not track with long-term outcomes."

He said the metric should be changed to account for differences in disease severity and socioeconomic status, so that hospitals with a higher burden of such patients are not penalized unfairly.

"Also, factors that determine 30-day readmission may not necessarily be linked with the quality of the care provided during the index hospitalization. There are a lot of things that happen after a patient is discharged that may be driving the readmission rate and not be modifiable by the quality of care provided," according to Pandey.

"Our hope is that given the evidence that we and others are generating, the CMS will take heed and try to modify or improve the current existing metric by accounting for social-economic status and disease-severity differences so that it captures long-term clinical outcomes a little better," he said.

"An Unfortunate Misunderstanding"

"This study has an unfortunate misunderstanding of the 30-day readmission measure," Dr Harlan Krumholz (Yale University, New Haven, CT), one of the developers of the metric, told heartwire .

"The measure was not designed to reflect in-hospital processes, which have largely topped out due to the marked increase in quality over the past 15 years, or mortality, which is predicted by different features from [those of] readmission and has different causal factors," Krumholz said.

"Best we can tell, readmission has to do with largely unmeasured hospital events, including patient preparation for discharge, transitional care, coordination and collaboration among providers, communication between patients and their clinicians and others, and even perhaps the degree to which errors or poor nutrition or inactivity or poor sleep occurred during the hospitalization," he said.

"Anyone who has ever been in the hospital knows this. You go home and your doctors haven't talked to each other, they don't get the information from the hospital, you don't really quite know what's up, you came in on certain medications and you leave on other medications, and through the fog of illness you can't quite remember what anybody told you. This is why we created this readmission measure. Measuring whether people die or not, or whether they get aspirin or not, is not giving you insight into whether the doctors are collaborating or communicating or whether people are ready to go home.

"Those of us who developed the model do not expect it to correlate with the old process measures that we introduced 20 years ago or mortality, in fact. If it correlated highly with mortality you would not need to measure readmission," he said.

As for long-term readmission risk, "there is no surprise that the further you get out, the less the readmissions have to do with the hospitalization," Krumholz said., "In fact that supports the idea that the 30-day readmission is reflecting something distinctive. The effect of hospitalization gets diluted because all sorts of other things intervene that have nothing to do with the hospitalization, and the number of subsequent admissions starts to dwarf the ones that occurred in 30 days."

On other findings in the analysis, Krumholz said, "Do the authors think that it is not plausible that the quality of care for black patients in this country might be worse than for white patients? It has always puzzled me that people are so quick to blame the patients for their risk."

He continued, "Of course there are vulnerable populations, but many studies show that there may be racial disparities in quality. This may be particularly true in issues regarding coordination and collaboration. Structural racism in our health systems exist. To suggest they don't would be an unfortunate position for our future progress toward equity."

The study was supported by the American College of Cardiology Foundation's National Cardiovascular Data Registry. Pandey reported no relevant financial relationships; disclosures for the coauthors are listed in the paper. Krumholz reported no relevant financial relationships.

For more from, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.