Published in the January 2017 issue of Today’s Hospitalist
CRITICS HAVE LONG COMPLAINED that the federal government’s relentless focus on potentially avoidable readmissions needs a rethink. New research now adds another question mark to the approach taken by the Centers for Medicare and Medicaid Services (CMS).
The most common criticism is that the Hospital Readmissions Reduction Program, which penalizes hospitals for high 30-day readmission rates, is unfair to facilities that treat the most challenging patients. Critics have said that penalties don’t account for problems associated with poverty—polysubstance abuse, inadequate primary care access—that can drive readmissions. (Passed last month, the 21st Century Cures Act will for the first time require Medicare to consider patients’ socioeconomic status in assessing readmission penalties.)
Add to that criticism the data from a new analysis of more than 4,450 acute-care facilities, which found that hospitals with the highest hospital-wide readmission rates are more likely to have the lowest mortality rates for patients with three common conditions.
“We need to be very careful about overblowing the importance of readmissions as a quality metric.”
~ Daniel J. Brotman, MD, Johns Hopkins Hospital
“The fact that mortality and readmission were, in some instances, inversely related, should raise some eyebrows about how well readmissions function as a quality metric,” says Daniel J. Brotman, MD, director of the hospitalist program at Baltimore’s Johns Hopkins Hospital.
He is the lead author of that study, which was published in the September issue of the Journal of Hospital Medicine. The research compared hospital-wide 30-day readmission rates with 30-day mortality rates for Medicare patients with discharge diagnoses of stroke, heart failure, COPD, MI, pneumonia and CABG.
For heart failure, COPD and stroke patients, hospitals in the top (worst) tertile for readmissions were more likely to be in the lowest (best) tertile in terms of mortality. Researchers found no significant association between readmission and mortality rates for MI, pneumonia or CABG.
In light of those findings, Dr. Brotman objects to the CMS using hospital-wide readmission rates as one measure to calculate its overall hospital quality star ratings.
Those star ratings weight readmission rates the same as mortality rates in the hospital summary scores. The ratings, which were released for the first time in July 2016 and reflect hospital performance on 64 measures, are based on a one-to-five scale. Each hospital’s star rating is posted on the Hospital Compare Web site.
“I don’t claim that, in general, hospitals with higher readmission rates deliver higher quality care than those with lower readmission rates,” says Dr. Brotman. “But I do think readmissions are so multifactorial that they fail the sniff test as a quality measure.”
Dr. Brotman spoke with Today’s Hospitalist.
What prompted you to conduct this study?
I have been working on a multidisciplinary institutional initiative to try to understand and prevent readmissions, and I have been struck by how challenging this has been. I query the providers in my group about their readmissions on a daily basis, and the overwhelming picture I get is that the vast majority of readmissions are not something we could have prevented. They are generally due to social factors or disease progression, or the care provided outside the hospital.
As we examined our processes for preventing readmissions, I heard anecdotes about how those processes might sometimes actually lead to readmissions. For example, if we do a post-discharge phone call and realize a patient is not doing very well, we bring her back to the hospital. That might not have happened without that call.
Sometimes the hospital is the safest place to care for a recently discharged patient who is struggling. I think we need to be very careful about overblowing the importance of readmissions as a quality metric.
Do you think the focus on readmissions by the CMS and others is misplaced?
Readmissions provide us with a very helpful lens for examining potential care defects. Historically, for instance, we physicians have not done as good a job as we could with care transitions.
A lot of good has come from people focusing on improving handoffs and post-discharge care. We did a study in the June issue of the Journal of Hospital Medicine that looked at the turnaround time for discharge summaries.
We found that the longer you wait to do your discharge summary, the higher the readmission rate. I think it’s unconscionable that some providers do discharge summaries weeks after a patient has left the hospital. Instead, they should be done contemporaneously so other physicians have access to that information.
So I do not take the stance that we should not measure readmissions or address the care processes involved in preventing readmissions. In fact, I think we should. But I think the pendulum is swinging too far in framing readmissions as a horrible evil.
Should financial penalties for high readmission rates be discontinued?
I think readmissions are a lot like length of stay. There are plenty of reasons a patient may spend a long time in a hospital, some of which are good and some not so good. Patients may stay in the hospital for a long time because they are very sick and need a lot of intensive care or because they have complications. Nobody should say that a short length of stay is a measure of good quality of care.
Like length of stay, high readmission rates are multifactorial but are certainly an indication of resource utilization, some of which may be excessive. I don’t necessarily disagree with payers indirectly penalizing institutions that have unusually high readmission rates, and bundled payments are one way to do that.
But to conflate readmission rates with quality—as the CMS has done with it star rankings—is fundamentally problematic. The reason we thought to do this study was to see whether readmissions track with what is the end-all and be-all of quality metrics, which is mortality rates. Dead patients can’t get readmitted, and sometimes readmission prevents death.
How would you like to see the CMS focus on readmissions?
Readmission rate measurement could be truly quality-based if it was viewed the same way that hospital-acquired venous thromboembolism (VTE) is looked at: by coupling an outcome with a care process.
Just looking at a hospital’s VTE rate is inherently problematic because you can’t assume the adverse event is due to a care defect. So you can ask two questions, “Did the patient have a thromboembolic event?” and “Was that associated with the failure to prescribe prophylaxis?” That makes it truly a quality measure.
If a patient was readmitted and you could prove that providers failed to communicate or that nobody checked to see if the patient could afford the prescribed medications, you could in fact ascertain that some readmission component was truly preventable. However, measuring those factors is a lot harder than measuring whether somebody got VTE prophylaxis. But there are certainly precedents for looking at care processes, not just outcomes.
Lola Butcher is a freelance health care writer based in Springfield, Mo.
