There is at least one aspect of “Obamacare” that my mother-in-law and I can firmly agree on: hospitals should not get paid for frequent readmissions.
The Hospital Readmission Reduction Program (HRRP), enacted by the Centers for Medicare and Medicaid Services (CMS) in 2012 with the goal of penalizing hospitals for excessive readmissions, has great face validity – and noble intentions. Does it also have a potentially disastrous downside?
On one side of the coin, the HRRP has been a remarkable success. It moved the national needle significantly on readmission rates. Yes, there are some caveats about increases in observation status patients and other shifts that could account for some of the difference, but it is fairly uncontroversial that overall there are fewer 30-day readmissions across the country following initiation of HRRP. That is perhaps encouraging evidence of the potential positive impact that policy can make to drive changes for specific targets.
However, there is also a murkier – and more controversial – side. There have been a number of studies that have suggested reductions in readmission rates may have been associated with an increase in mortality in some patient groups. You discharge a patient and hope they won’t return to the hospital, but perhaps you should be more careful what you actually wish for.
Overall, the evidence of an association between readmissions and mortality has been complicated and conflicting. Headlines have alternatingly raised alarm about increased deaths and then reassured that there has been no change or perhaps even some concordant improvements in mortality. Not necessarily surprising, considering that these studies are all unavoidably of observational design and use different criteria, datasets and analytic models, which then drive their seemingly conflicting results.
A new article published this week in the Journal of Hospital Medicine enters into this fray. The researchers examined the potential association between changes in rates of COPD readmissions and 30-day mortality following HRRP introduction. While the initial HRRP program and subsequent analyses included patients with congestive heart failure, acute myocardial infarction, and pneumonia, the program was extended in 2014 to include patients with COPD. So, what happened in this patient group?
Through a number of statistical gymnastics, which as a non-statistician I am having difficulty truly wrapping my head around, the researchers seem to have found a number of important insights:
- The all-cause 30-day risk-standardized readmission rate declined from 2010-2017
- The all-cause 30-day risk-standardized mortality rate increased from 2010-2017, and the rate of increase in mortality appears to be accelerating
- Hospitals with higher readmission rates prior to COPD readmission penalties had a lower rate of increase in mortalities
- Hospitals that had a larger decrease in readmission rates had a larger rate of increase in mortality
These researchers could not evaluate data at the patient level and could not adjust for changes in disease severity. However, taken together these findings suggest that something bad may be truly happening here.
The authors of this study also point out that the associations with increased mortality have larger been seen in patients with congestive heart failure – and now in patients with COPD – which are both chronic diseases characterized by exacerbations, as opposed to acute myocardial infarction and pneumonia which are episodic and treatable. Perhaps in those types of disease, efforts to avoid readmissions may be more universally helpful. Maybe.
Even if it is challenging for me to adjudicate the complicated methods and results of this study, I find it concerning that there is “biological plausibility” for this association. Hospitalists know exactly how this might have happened. Have you heard of the pop-up alerts that fire in the emergency department to let the physicians know that this patient was discharged within the past 30 days? Ya’know that alert is not meant to tell you what to do, but you just might want to consider trying to discharge them or at least place them in observation – use your clinical judgment, if you know what I mean. Within the past decade, observation units quickly cropped up all over the country, often not staffed by hospitalists nor cardiologists, where patients with decompensated heart failure, chest pain, and/or COPD, can be given Lasix and/or nebulizer treatments – at least just enough to let them walk on back out that door without a hospital admission.
At the end of the day, whether or not mortality rates have truly increased in the real-world, this well-intentioned program seems to have serious issues. As Ashish Jha wrote in 2018, “Right now, a high-readmission, low-mortality hospital will be penalized at 6 to 10 times the rate of a low-readmission, high-mortality hospital. The signal from policy makers is clear – readmissions matter a lot more than mortality – and this signal needs to stop.”
I have noticed the same in practice as the goal of avoiding 30 day readmissions became more important than what was best for the patient. Apparently, CMS has changed their criteria to what counts as an admission with the end result of a patient being switched to observation status. This all needs to change and after experience on both sides of a patient receives a treatment they could only obtain in a hospital it should be an admission.
I wonder whether the increased use of palliative care helps to explain the data. As such, the association between lower readmission rate and increased mortality may be correlative, not causative.
Patients with end-stage COPD and CHF readmit to the hospital at an alarming rate. Often, we are not able to change the trajectory of these diseases. Many institutions have increased palliative services help address these issues. With increase palliative care services, the mortality rates may potentially increase, as there would be fewer patients returning to the hospital. In turn, perhaps these patients are having a more humane deaths at home.