Readmission Penalties: Quality Improvement or Surtax?
CMS is now in the third year of penalizing hospitals if they re-admit too many patients within 30 days of hospital discharge. According to Kaiser Health News, this year 2,610 hospitals are being penalized, which represents 75% of those receiving money through the Inpatient Prospective Payment System (IPPS) who are not otherwise exempt. 39 hospitals, including one in my area, are receiving the maximum 3% penalty. Is this a performance improvement measure or a surtax?
Krumholz and Bernheim, who are affiliated with the Center for Outcomes Research and Evaluation at Yale-New Haven, have described how their work led them to identify that readmission rates within 30 days of hospital discharge as a suitable target for quality improvement efforts.
… readmission rates were exceedingly high for common diagnoses, such as acute myocardial infarction, heart failure, and pneumonia, and differences in patient populations did not seem to explain the variation in performance by the hospital…The development of performance measures always involves choices that merit discussion. While developing these readmission measures, we made difficult decisions about which variables to include in the models, which outcomes to count and over what length of time, and which type of mathematical models to use.
Cassel and associates recently wrote on “getting more performance from performance measurement” and discuss the readmission penalty program. They noted:
When such programs began, some observers expressed concern about the metrics’ focus on a small subgroup of patients and raised questions about how many readmissions were really preventable, whether hospitals could address the root causes of readmission, whether focusing on readmissions might distract organizations from other important efforts, and whether the programs might disadvantage hospitals serving poorer populations…Yet recognizing that measurement and payment-reform programs have to start somewhere and can be improved over time, readmissions-reduction programs have moved ahead and continue to evolve.
They note recent data released by CMS shows that the readmission rate, which had been averaging 19%, had dropped to 17.5% in 2013. This decline is presented as evidence the program is working to improve care and prevent “unnecessary” readmissions.
I looked recently at the last two years of admission and re-admission data for my chronic dialysis population and it was evident that a small number accounted for a disproportionate share of the admissions (and re-admissions.) I have not seen similar hospital studies, but I would predict similar findings there as well.
So what are hospitals doing? The answer is almost anything they deem likely to impact readmission rates. One common approach is to make sure a recently discharged patient is seen by his or her primary care physician promptly after discharge, and that the primary care doctor have access to the results of tests, medication lists, and a discharge summary when the patient is seen. CMS has even started a new program in 2015 to allow primary care physicians to charge an extra fee for that visit, provided the patient is not readmitted within 30 days. But the question remains: can hospitals address the root causes of readmission?
A recent study by Kind and associates makes some interesting observations that may bear on this question. They looked at readmission rates as a function of neighborhood, as opposed to individual, socioeconomic status, using census tracts as the unit of analysis. They note determining an individual patient’s socioeconomic status is not done routinely in clinical care, so studies looking at this issue often use insurance status or ZIP code of primary residence as a proxy. Of course, in a Medicare population, insurance status provides no information. They used a previously published “area deprivation index,” or ADI, which consists of 17 factors determined in census data, and widely used in health care planning studies. This study considered the issues of nesting patient level data within hospitals and also looked at the most deprived 15% of tracts compared to the other 85%, and also looked at the deprived areas in thirds. They found readmission rates went up the more the “ADI” went up. The map they published is instructive. The highest ADI census tracts are heavily concentrated in Alabama, Mississippi, Arkansas, West Tennessee, “Appalachia,” areas with large American Indian reservations and inner city ghettos—all of the places that typically come in last on measures of wealth and health. And, as noted in the Kaiser Health News Report, almost all the hospitals in some states, including Tennessee, were subjected to some degree of readmission penalty.
Should these data impact the readmission penalty program? Krumholz and Bernheim defended CMS’ decision not to adjust readmission penalties for socioeconomic status. Their original studies suggested “safety-net” hospitals were failing to deliver “standard” care as often as other hospitals, so the concern was that an adjustment would perpetuate substandard treatment. But, they note:
Patients situations are not likely to improve by imposing financial penalties on these [safety-net] institutions, and that is a reason for caution in the payment policy. It may also be an acknowledgment that some patient populations have challenges that have to do with social context…It remains our view that concerns about the effects of penalty programs on safety-net hospitals should be addressed directly through payment policies rather than through changes in the measures.
The Kaiser Health News article, though, reported “…roughly 2 million patients return a year, costing Medicare $26 billion. Officials estimate $17 billion of that comes from potentially avoidable readmissions…” The officials did not say where their estimate came from, but it suggests they believe readmission rates should be substantially less than 10%. Since nothing in the medical literature supports such an ambitious goal, it seems to me this is a kind of administrative surtax.
The IPPS provides a lump sum payment for each Medicare patient’s admission based upon the principal diagnosis, so one way to “game” the system is to break up one long admission into a series of shorter admissions, thereby generating more revenue. Of course, that sort of gaming, while theoretically possible, requires collusion not only between the hospital and the doctors, but also the patients. While this serves as one of the justifications CMS provides for its decisions, I doubt this is a major issue. On the other hand, it is certainly true that before the readmission penalties were implemented, hospitals had no real motivation to invest time, energy, and resources into discharge planning or case management.
As I consider the issue from my local perspective, I have concerns. Of the 17 counties served by my medical community, almost all have at least one census tract listed among the highest ADI. By definition, these residents are not going to have commercial insurance, so the local hospitals are proportionately more dependent upon government revenue to stay in business. As a taxpayer, I want government officials to do their fiduciary duty, and, since CMS is the largest payer for healthcare services, it is certainly their prerogative to change the rules, even if it forces institutions such as hospitals to undergo major, transformational changes. But is this the a case of the “kudzu principle,” the law of unintended consequences, at work? As currently being applied, the program is regressive. Those who can least afford to pay the tax are those most likely to incur the penalty.
So what can be done? If you are a “conservative,” or one who favors market solutions, these counties are poor, so the business opportunity is not there. Our region has insufficient numbers of primary care resources, and it is unlikely the private sector will make sufficient investments to change this. If, on the other hand, you are a “liberal” and favor governmental (societal) solutions, you still have the problem of a low tax base. One reason Tennessee has not participated in the Affordable Care Act is fear that it will lead to a sustained increase in the state’s Medicaid (Tenncare) budget. Already, this part of the budget has crowded out resources for other valued programs such as education, infrastructure repair, or prisons. We certainly need coordinated efforts by government at all levels combined with the private sector to make any progress. But our track record obtaining sustained cooperation is not encouraging. Stay tuned.
3 January 2015
 Rau, Jordan. Medicare Fines 2,610 Hospitals in Third Round of Readmission Penalties. 2 October 2014.
http://kaiserhealthnews.org/news/medicare-readmissions-penalties-2015/. Accessed 3 December 2014.
 Krumholz HM, Bernheim SM. Considering the Role of Socioeconomic Status in Hospital Outcomes Measures. Ann Intern Med 2014;161:833-834. doi:10.7326/M14-2308.
 Cassel CK, Conway PH, Delbanco SF, Jha AK, Saunders RS, and Lee TH. Getting More Performance from Performance Measurement. N Engl J Med 2014;371(23):2145-2147. doi:10.1056/NEJMp1408345.
 Kind AJH, Jencks S, Brock J, Yu M, Bartels C, Ehlenbach W, Greenberg C, Smith M. Neighborhood Socioeconomic Disadvantage and 30-Day Rehospitalization: A Retrospective Cohort Study. Ann Intern Med 2014;161;765-774. doi:10.7326/M13-2946.
On Dying in America
The IOM report "On Dying in America" makes recommendations for change that physicians should embrace.
Rationing is a dirty word, but a necessary part of healthcare reform. A consideration of the implication of this for practicing physicians.
Shared Decision Making
A consideration about decision making at the person, group and organizational levels.