Public and private insurers are increasingly paying health care providers based on the quality of care they provide rather than volume. This shift has generated some concerns about how performance on quality is measured and how rewards or penalties should be structured. This month’s Narrative Matters essay, “Mr. G and the Revolving Door: Breaking the Readmission Cycle at a Safety-Net Hospital,” authored by Carolyn Dickens and colleagues, provides a window on some of the controversy surrounding paying safety-net providers for quality performance.

The essay tells the story of health care providers at a Chicago safety-net hospital and a 41-year homeless man who they refer to as Mr. G. The patient has a history of incarceration, and was diabetic, a schizophrenic, and a substance abuser. He was admitted and readmitted to the hospital repeatedly for a heart condition, making him what some refer to as a “super utilizer” of health care services.

Mr. G’s serial admissions were typically occasioned by his failure to consistently take prescribed medication, his poor diet, continued substance abuse, or a combination of those factors. With the help of hospital staff, he was discharged, at various points, to an inpatient substance abuse center, a homeless shelter for persons with medical needs, and a nursing home, though he didn’t last long at any of these facilities. After returning to the streets one last time, he ended up back in the hospital, where he died from pneumonia at the age of 42.

Dickens and colleagues observe that their hospital is penalized for patients like Mr. G under the Medicare Hospital Readmission Reductions Program (HRPP) — notwithstanding their efforts to follow evidence-based practices to reduce readmissions. While Mr. G’s insurance situation at the end of his life was not clear (he was at one point uninsured but apparently later gained coverage), the authors raise an important point that has frustrated some providers and policymakers alike since the HRPP began in 2013: Although the program adjusts for patients’ clinical characteristics in assessing penalties for hospital readmission, it does not consider other factors that can also affect patient outcomes, such as poverty, health literacy, and the availability of social support and housing. Indeed, some have charged that the HRPP unfairly penalizes hospitals that take care of large numbers of poor patients — that is, safety-net hospitals.

Research On The Hospital Readmission Penalty Program

The HRRP was authorized in 2010 as part of the Affordable Care Act (ACA) in an effort to tie Medicare reimbursement to hospital performance on quality measures. In fiscal year 2017, the HRRP includes payment penalties of up to 3 percent of Medicare base operating payments for hospitals with excess readmissions. The formula for computing the penalties is complicated, but in essence for selected medical conditions (including heart failure, one of Mr. G’s medical conditions), it penalizes hospitals whose risk-adjusted 30-day Medicare readmissions rate are greater than the national average. The evidence so far suggests that the program is working: Fee-for-service Medicare readmissions rates have declined since the program began, and a 2016 study provides evidence suggesting this reduction has not translated into higher observation stays.

Despite this success, the HRRP has been criticized on several fronts. Vulnerable patients are more likely to be readmitted to the hospital, and certain hospitals treat a higher share of such patients than others. Research bears this out. The current risk-adjustment methodology for the readmissions measures includes “clinically relevant” factors such as patient age and co-morbidities but, as mentioned, it does not account for any measures of patient socioeconomic status (SES), such as poverty status and social support. On the flip side, others, including CMS, argue that including SES risk adjustment is tantamount to letting hospitals “off the hook” for providing a lower quality of care.

Patients like Mr. G who are discharged to fragile environments with limited social supports may find it more difficult to comply with post-discharge instructions, medication regimens, or go to follow-up appointments — factors that are outside the hospital’s control. Safety-net hospitals treat more economically vulnerable patients, which research has shown contributes to a higher readmissions rate for these hospitals.

One recent study showed that including the fairly limited patient-level SES measures available in Medicare claims data explains about 25 percent of the higher Medicare readmissions rates at safety-net hospitals (defined by the Medicare DSH ratio), while another found that including community-level SES measures reduced variation in readmissions rates across hospitals. Even after accounting for SES measures in risk adjustment, however, the former study found that safety-net hospitals still had higher readmissions rates than their non-safety counterparts, which could reflect meaningful differences in quality of care between safety-net hospitals and their counterparts.

Given that safety-net hospitals overall have higher readmission rates, it would follow that they would also incur greater penalties under the HRPP. The data suggest otherwise, however. To date, safety-net hospitals’ HRRP penalties have not been significantly higher than their counterparts. Further, the full impact of the HRRP penalties on a hospital’s bottom line depends both on the size of the penalty and also how Medicare payments fit within a hospital‘s overall revenue streams.

A 2013 MedPAC analysis, for example, estimated that major teaching hospitals had higher average HRRP penalties as a share of Medicare base operating payments, but the penalties’ effect on hospital revenue was relatively small because these hospitals receive significant funds from other sources, such as indirect medical education, DSH, and outlier payments. Some evidence also suggests that the influence of patient SES on the penalty gap between safety-net hospitals and other hospitals has narrowed somewhat since surgical conditions (hip and knee replacements) were added to the HRRP in fiscal year 2015.

Reducing Adverse Effects On Safety-Net Hospitals

While the HRPP penalties may not be appreciably higher at safety-net hospitals, by eating into their already thin profit margins, they could potentially limit these hospitals’ ability to implement the very changes needed to reduce readmissions. This begs the question: Is there a way to lessen the adverse impact of the HRPP penalty on safety-net hospitals without tacitly endorsing lower quality care for vulnerable populations?

MedPAC, for one, has proposed a revision to the method for calculating readmissions penalties. Rather than including patient SES in the risk-adjustment step, which MedPAC argues would take years to develop empirically and could mask true quality disparities, MedPAC suggests grouping hospitals into peer groups based on their share of low-income Medicare patients and then set readmissions targets for each peer group. Put another way, hospitals with similar shares of low-income patients would be compared with each other instead of all hospitals. This is one proposal that could strike a balance between the need for transparency about quality of care disparities and the additional challenges safety-net hospitals face in caring for vulnerable populations.

Other efforts to minimize the adverse effects on safety-net hospital could focus on the structure of the HRPP. The 30-day window for measuring readmissions, for example, is arbitrary. Some have suggested that a shorter timeframe (such as three or seven days) would more closely reflect the quality of care provided by the hospital, rather than the composition of the hospital’s population or resources in the local community.

The selection of quality measures could also be tweaked. Certain quality measures may be more sensitive to patient SES and thus could have a greater impact on safety-net providers. Efforts to build the evidence base on associations between quality measures and patient SES is already underway through support provided by the Improving Medicare Post-Acute Care Transformation (IMPACT) Act of 2014. Such information will help policymakers thoughtfully evaluate measures and potential impacts on providers.

Yet another approach might be to leave the HRRP as is but provide additional support (financial or otherwise) to safety-net hospitals and other institutions that are grappling with complex, challenging patients like Mr. G. Such support could be used to address social determinants of health — building linkages within the community with housing, criminal justice, outpatient medical and mental health providers, and social support agencies. Indeed, new initiatives in several states and from the Centers for Medicare and Medicaid Innovation are aiming to break down entrenched silos.

Other Payers Move The Needle On Hospital Readmissions

In addition to Medicare, private insurers and Medicaid are increasingly paying health care providers based on their performance as measured by the quality of care they provide. Medicaid’s move to pay for health care based on value rather than volume is particularly important for safety-net hospitals, given that Medicaid accounts for such a large share of these hospitals’ revenue.

Reducing hospital readmission rates is one of several value-based initiatives that state Medicaid programs have recently targeted. For example, one goal of New York’s 2014 “Delivery System Reform Incentive Payment” (DSRIP) waiver, a component of the state’s broader Section 1115 waiver, is to reduce hospital readmissions. As part of New York’s DSRIP initiative, participating organizations, many of which are safety-net hospitals, are incentivized to make (and held accountable for making) fundamental changes in how they do business.

This includes increasing coordination of care between health care providers and social services organizations, and promoting the integration of medical and behavioral services. Having a “medical village,” as envisioned in New York’s DSRIP initiative, may have helped a complex patient like Mr. G. In a similar vein, CMS will also soon begin testing the Accountable Health Communities Model, which promotes partnerships among medical care, social services, public health, and community-based organizations in the hopes of reducing excess health care use and generating Medicare and Medicaid savings.

Final Thoughts

Without a doubt, payers have a growing desire to pay for value in health care, not volume. While certainly an important and laudable goal, it is critical as we move forward that we find that sweet spot where we don’t reward safety-net hospitals for providing poor quality of care to poor people, but we also don’t unfairly punish them because they take care of poor, medically complex patients.