Data on the Medicare Readmissions Reduction Program released by CMS on August 1 shows the reduction in their base DRG payments that hospitals will see in the coming year (beginning October 1, 2012) as a result of having "excess readmissions" for their heart failure, acute MI, and pneumonia patients. The program was authorized by the Affordable Care Act (the ACA, Section 3025), which amended Section 1886 of the Social Security Act by adding a new Section (q). Reductions of base DRG payments of up to 1 percent are possible for the coming year, with the ceiling increasing to 2 percent the following year and then 3 percent thereafter. The entire list of hospitals and their respective penalties can be downloaded from http://www.kaiserhealthnews.org/~/media/Files/2012/Medicare%20Readmissions%20Penalties%202013.pdf.
When the readmissions reduction program was first announced, hospitals across the country naturally wanted to avoid the penalties and improve their discharge and follow-up procedures, so they scrambled to find ways to reduce preventable readmissions, focusing on patients treated for heart failure, acute myocardial infarction, and pneumonia, the three conditions targeted by the Medicare program. The results have been disappointing for many hospitals, some of which will see Medicare inpatient revenue loss up to the ceiling level of 1 percent of their base DRG in the coming year.
For a hospital with $50 million in base DRG revenue, for instance, the loss in the coming year can be as much as $500,000 (1 percent) and as much as $1.5 million (3 percent) in 2015. With so much money at risk, it is important for hospitals to clearly understand how the program works. In some very important ways, the program appears to be less than ideal in its approach to reducing readmissions.
On top of the inherent difficulty in reducing readmissions, starting with the definition of readmission, there are features of the program that make it even harder for hospitals to see and get the benefit from the improvements they have implemented. When the Affordable Care Act created the program, it stated, “the term readmission means, in the case of an individual who is discharged from an applicable hospital, the admission of the individual to the same or another applicable hospital within a time period specified by the Secretary from the date of such discharge…”
According to CMS regulations (Federal Register, Volume 76, Number 160, August 18, 2011), “The readmissions measures are designed to measure whether a patient experienced at least one readmission within 30 days of an initial (or “index”) discharge as a single binary (yes/no) event, rather than counting the number of readmissions experienced within 30 days of discharge as separate readmissions. For any given patient, only the first readmission they have will be counted for the Hospital Readmissions Reduction Program. In addition, only one readmission during the 30 days following the discharge from the initial hospitalization will count as a readmission… For any given patient, none of the subsequent readmissions they experience within 30 days after discharge would be counted as a new ‘index’ admission (that is, an admission evaluated in the measure for a subsequent readmission). Any eligible admission after the 30-day time period will be considered a new index admission.”
According to this working definition, then, a patient admitted to any acute care hospital, including one other than the one that discharges the patient, is considered a readmission for the hospital that performed the “index admission.” Since CMS does not inform hospitals when their patients are readmitted to other hospitals, or where and when they have been previously admitted, admitting hospitals don’t know for sure when they admit a patient if that admission represents a readmission for another facility, exempting them from responsibility for additional readmissions within 30 days of the index hospital discharge (not theirs). Furthermore, the hospital doesn’t know whether it will be charged with a readmission if the patient is readmitted, even if to the same hospital in a short period of time. This makes it impossible for hospitals to track their own performance with respect to the Medicare program. CMS has said (CMS-1518-F/1430-F) that it “will consider whether it is operationally possible to provide these data to hospitals and whether sharing these data would be consistent with patient privacy considerations.” Since the readmission penalties are based on comparative data on all hospitals, however, they will be unable to reproduce the calculations. So, for now, no data sharing, and hospitals are in the dark about their performance in the program until they get their annual report.
The Medicare program uses a “risk-adjusted readmission rate” to determine a hospital’s performance. Readmission rates are adjusted for “key factors that are clinically relevant and have strong relationships with the outcome (for example, patient demographic factors, patient co-existing medical conditions, and indicators of patient frailty). This risk-adjustment approach adjusts for differences in the clinical status of the patient at the time of the index admission as well as for demographic variables.” While these adjustments do help level the playing field somewhat for hospitals in different settings, they don’t include adjustments for socioeconomic variables, rural vs. urban settings, and other factors, such as availability of timely follow-up in the medical community, or rate of uninsured or Medicaid in the population served by the hospital—factors that can have a large impact on the rate of readmissions, and put big city hospitals at a disadvantage as well as having other unanticipated effects.
The program counts “all cause readmissions,” not just those clinically related to the index admission. A patient treated for pneumonia is counted as a readmission if he/she is admitted within 30 days to any hospital for any cause whatsoever, including such things as motor vehicle accidents, surgery, and heart attacks. These all-cause readmissions clearly dilute the effectiveness of hospital programs targeting clinical issues related to the index hospital stay. For example, a well-planned program to monitor heart failure patients after discharge, including daily weights, pharmacy consulting, and home nursing visits, would have no effect on readmissions for strokes. If these unrelated readmissions occur with equal frequency to patients discharged from all hospitals, then they will even themselves out over time. But there is no guarantee that this is the case, or that over the short run (such as one year) there won’t be considerable variation that will hurt a hospital financially.
Finally, there appears to be confusing language in the way the ACA (and in turn the Medicare program) account for exclusion of unrelated readmissions. The key section of the ACA (Section 3025.5.A) states, “such endorsed measures have exclusions for readmissions that are unrelated to the prior discharge (such as a planned readmission or transfer to another applicable hospital).” The ACA delegates to the National Quality Forum (the NQF—technically “the entity with a contract under section 1890(a)”) the responsibility for selecting the measures and the criteria to be used for determining readmissions.
The NQF found only two reasons for exclusion of readmissions on the basis of being “unrelated to the prior discharge": Patients who are readmitted following acute MI for coronary artery bypass graft, or PTCA. In fact, though, these are planned readmissions, not unrelated readmissions. A PTCA or CABG is obviously related to an admission for acute MI. Furthermore, there are no exclusions for any unrelated or planned readmissions following discharge for heart failure or pneumonia. As stated in the Federal Register entry cited above, “The NQF… has endorsed ‘measures of readmissions’” for each of these three conditions, and those NQF-endorsed measures “have exclusions for readmissions that are unrelated to the prior discharge (such as a planned readmission or transfer to another applicable hospital).”
This statement is not really accurate.
The language on exclusion of unrelated readmissions found in the Affordable Care Act is ambiguous (at best) since it uses examples that do not illustrate the premise. Consider: If the intent of the language is to truly exclude “readmissions that are unrelated to the prior discharge” (an appropriate approach, considering the purpose of the program), the examples given, “planned readmission or transfer to another applicable hospital,” are not examples of unrelated readmissions at all. A planned readmission for a procedure is completing the patient’s treatment plan and clearly is related to the prior admission. A “transfer to another applicable hospital” is not an unrelated readmission, either. It is difficult to understand how a transfer can be construed as an unrelated readmission, but is it appropriate, nonetheless, not to count the transfer as a readmission against the sending hospital?
So what was intent of the legislation? Did Congress intend to exclude unrelated readmissions, a policy that would encourage hospitals to focus on providing optimum follow-up care for the designated conditions, or did it intend to count all readmissions, regardless of whether they are related to the index admission or potentially avoidable? One clue is the use of the term “such as.” The Act calls for exclusion of unrelated readmission “such as a planned readmission or transfer to another applicable hospital,” but doesn’t limit the exclusion to those two examples. Since, as explained above, those are not examples of unrelated readmissions at all, there appears to be a need for a technical clarification of the Act. The language would make more sense and would be more clinically appropriate if the words “such as” were deleted, and it required instead that the measures have “exclusions for readmissions that are unrelated to the prior discharge, planned readmissions, and transfers to another applicable hospital.”
As a result of the NQF’s extremely limited interpretation of the exclusion for unrelated admissions, essentially all admissions in the 30-day period following an index discharge are counted against the discharging hospital. This has to be frustrating for hospitals that have created readmission reduction programs just to find out that an unknown number of accountable readmissions are actually far beyond their ability to control. What they thought was an index admission to their facility may actually be a readmission to another, and a readmission to their own hospital may count as a readmission for their own data collection, but not count as a readmission under the Medicare program, because the patient had been discharged from another hospital within 30 days for “an applicable condition.”
With the planned expansion of NQF-approved conditions in 2015 (adding COPD, PTCA and coronary artery bypass graft surgery), an even larger proportion of hospital admissions will be counted as readmissions for hospitals elsewhere, whether clinically related to the first admission or not, and hospitals will find their data correlates less and less with the Medicare readmissions program.
CMS’ position on criticisms on the “unrelated” policy is that they are required by the ACA to follow the recommendations of the NQF exactly as published. According to CMS (as published in the Federal Register concerning the adoption of the NQF-endorsed measures): “If we were to modify the endorsed measures, we are concerned that they would no longer be considered ‘endorsed.’” Excluding readmissions for planned cardiac procedures makes sense, but the language of the Act does not limit exclusions to the two examples of planned readmissions and hospital transfers.
CMS adds, “…we asked clinical experts to identify planned readmissions for these conditions, that is, those which would not count as a readmission, after an admission for HF or PN… No such related, planned procedures were identified as occurring commonly after the index admissions for HF or PN at the time of the development of the Hospital IQR Program measures.” Was this the appropriate question for CMS to ask its experts? The question implies that the distinction between a related and an unrelated readmission is whether the readmission is for a “common planned procedure.” Is that what “unrelated” really means?
In the Federal Register entry, CMS acknowledges this concern was raised by many who commented on the proposed rule. “Many commenters,” they wrote, “stated that the current set of existing exclusions for unrelated readmissions did not meet Congress’ intent, which they believed requires additional exclusions for certain readmissions… The statute does not state that the measures must account for all possible unrelated readmissions. (Italics added.) Moreover, adding exclusions would be inconsistent with the statute, which requires us to adopt the measures as endorsed by the NQF, and the endorsements currently include specific exclusions for unrelated readmissions, which include transfers.”
It is intuitively obvious that improving the discharge process represents an improvement in our fragmented health care system and that these efforts are to be commended, supported, and expanded. If the Medicare Readmissions Reduction Program promotes this end, then it will have served a valuable purpose. Some changes in the program that will encourage even more effective efforts on the part of hospitals and physicians could help reduce readmissions by keeping patients healthier at home.
About the Author
Steven J. Meyerson, MD, is a Vice President of Accretive Physician Advisory Services®. He is Board Certified in Internal Medicine and Geriatrics. He has recently been the medical director of care management and a compliance leader of a large multi hospital system in Florida. He has distinguished himself by creating innovative service lines and managing education for Accretive PAS®.
Contact the Author
To comment on this article please go to firstname.lastname@example.org