From 3M Health Information Systems
Improving Outcomes by Focusing on Results: Readmissions, Complications and CMS
It’s been nearly a century since Dr. Ernest Codman championed an “end results system” to track and measure hospital outcomes to determine the effectiveness of treatment and improve patients’ lives. Within the last decade, outcome measurement has gained momentum as the health care industry seeks to improve quality of care/patient outcomes and reduce health spending through initiatives such as pay-for-performance or value-based purchasing.
A recent issue of JAMA featured four articles and two editorials on the use of readmissions as both a quality indicator and pay-for-performance target. Two main issues were discussed. The first was the correlation between post-discharge complications manifesting after surgery and the likelihood of readmission (it is high).¹,² The second was whether enrollment in the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) was an indicator of better readmission performance (it was not).³,4
Discussion of these issues from the authors and in editorials highlighted challenges in the way the Centers for Medicare & Medicaid Services (CMS) levies financial penalties against hospitals with excessive readmission rates and with the idea of readmissions as an outcome measure. One editorial asserted that CMS does not adjust for complexity of illness [which we would soften to adjust “adequately”] or low socioeconomic status, which unfairly penalizes hospitals that care for these types of patients–typically academic medical centers and safety net hospitals.5 The other pondered the limitations of hospital efficacy in tackling readmissions when those most committed to improvement fare no better than their peers.6
We would like to respond to these articles by highlighting our experience implementing the Potentially Preventable Readmissions (PPR) classification system over the past 10 years in the U.S., working in multiple states on initiatives with Medicaid programs and commercial payers. We have found that there are several essential ingredients for improving outcomes. These include:
- using clinically-valid, transparent classification models which can be replicated and updated with input from providers;
- incorporating appropriate risk adjustment, which takes into account patient complexity, into these models;
- holding providers accountable for multiple measures to drive quality improvement at the system level;
- conducting collaborative initiatives facilitated by one entity; and
- ensuring appropriate financial incentives.
As stated in the Federal Register, “the success of any payment system that is predicated on providing incentives for cost control is almost totally dependent on the effectiveness with which the incentives are communicated.” Effective communication means that the health professional team (including consumers if they so wish) needs to be able to view an overall provider performance score, drill down into the data, and examine the classification logic down to the individual patient. We annually update the PPR software (and all the other Potentially Preventable or PPE Classification Systems). With a categorical, transparent model such as PPRs, health professionals can and do provide constructive suggestions on logic changes, which enables the model to evolve to better achieve the goal of improving outcomes. Unfortunately, they cannot examine regression-based models in the same manner; thus, they have no way of knowing whether regression-based models, such as the NSQIP, adjust for complexity of illness. Unconscionably, although CMS levies penalties on hospitals for higher-than-expected readmission rates, health professionals cannot even replicate the CMS readmissions model, nor can they provide suggestions on its widely-acknowledged deficiencies (pointed out by Leape) in adjusting these rates for complexity of illness.
Like the authors of the JAMA articles we have also observed the strong correlation between complications and readmissions. However, we believe that keeping measures separate, as with Potentially Preventable Complications (PPC) and PPRs, raises awareness among providers and holds them accountable for multiple quality issues at the system level. Our view is that some patients are more likely to end up with a complication or readmission given their complexity (or characteristics) and this, in turn, drives complication and readmission rates. It is for this reason that we create detailed logic and risk-adjustment, applied independently, within both PPC and PPR logic when assessing complication and readmission rates. Accounting for patient complexity through exclusions and recognizing the link the between readmissions and complications “penalizes hospitals twice for post-operative complications” as asserted in the Merkow article. In fact, it could be argued that hospitals are penalized three times since CLABSI (central-line associated blood stream infection) and post-operative sepsis appear in both CMS’ hospital-acquired conditions reduction program (HACRP) and value-based payment (VBP) payment adjustments.
State Medicaid programs and related initiatives, such as the RARE (Reducing Avoidable Readmissions Effectively) project in Minnesota, continue to have success using categorical classification models such as PPRs.7 However, providing data back to the provider (a la NSQIP), typically by the payer, is necessary but not sufficient. The provider needs to be able to replicate and view the model as noted earlier. Another essential element for improving outcomes, also demonstrated in the Minnesota RARE project, is the necessity for an entity to foster collaboration between providers and disseminate best practices. And, lastly, over the long term, there need to be financial incentives to improve outcomes. Without financial incentives for improving outcomes, initiatives are unlikely to succeed. These incentives and collaboration need to apply to more than the actions taken during the hospital stay—they need to be applicable to all patient care and treatment for a minimum of 15 days post discharge. While today most of my patients discharged from the hospital have an appointment to see me (all too often weeks after discharge), there is no attempt at coordination between the hospital and the primary care physician. It is unreasonable to focus readmissions only on what occurs during the hospital stay.
A sense of urgency and buy-in by health care leaders, at the policy, payer and provider levels, are vital to challenging the status quo—key features underpinning the JAMA articles. We believe that Codman was right, but without the essential ingredients highlighted here, his promise to improve outcomes will never be completely realized. Our viewpoint is that classification systems such as PPRs will slowly but surely pave the way and accelerate the movement to pay for better outcomes.
Richard Fuller, MS, is an economist with 3M Clinical and Economic Research.
Norbert Goldfield, MD, is medical director, Clinical and Economic Research, 3M Health Information Systems.
1Dimick JB, Ghaferi AA. Hospital readmission as a quality measure in surgery. JAMA. 2015;313(5):512–3. doi:10.1001/jama.2014.14179.
2Merkow RP, Ju MH, Chung JW, et al. Underlying Reasons Associated With Hospital Readmission Following Surgery in the United States. JAMA. 2015;313(5):483. doi:10.1001/jama.2014.18614.
3Etzioni DA, Wasif N, Dueck AC, et al. Association of Hospital Participation in a Surgical Outcomes Monitoring Program With Inpatient Complications and Mortality. JAMA. 2015;313(5):505. doi:10.1001/jama.2015.90.
4Osborne NH, Nicholas LH, Ryan AM, Thumma JR, Dimick JB. Association of hospital participation in a quality reporting program with surgical outcomes and expenditures for Medicare beneficiaries. JAMA. 2015;313(5):496–504. doi:10.1001/jama.2015.25.
5Leape LL. Hospital Readmissions Following Surgery. JAMA. 2015;313(5):467. doi:10.1001/jama.2014.18666.
6Berwick DM. Measuring surgical outcomes for improvement: was Codman wrong? JAMA. 2015;313(5):469–70. doi:10.1001/jama.2015.4.