Designing Programs for Populations with Chronic Care Needs: A Blanket or a Quilt?

Health policy must promote the ability of smaller systems to use sophisticated, rigorous, and less-than-perfect study designs to evaluate the impact of their local programs.

Former Speaker of the House Thomas “Tip” O’Neill once famously observed that “all politics is local.” Although everyone agrees that deferring politically to the precinct level is crucial for the proper functioning of our democracy, the role of local patterns of care in healthcare is far more nettlesome. Should all, some, or no healthcare policy be local?

How and why patients use local healthcare resources is complex,^1,2 and how physicians respond is considerably influenced by prevailing and highly variable standards of care.³ Yet, in spite of all the challenges related to use of healthcare resources, most persons end up being satisfied with the quality of healthcare they receive.⁴ At the same time, however, considerable attention is being paid by policymakers to the downsides of accommodating local care patterns, including unwarranted variation, spotty availability of key services, restrictive insurance networks, and patient passivity. Much of the current debate over health reform has been about how to find the right balance of economic, educational, and regulatory levers that reduce variation, promote the “right” kind of care, preserve choice, increase patient awareness, assist community hospitals, and broaden local care services with an appropriate mix of national standards and local resources.

The report by McEwen et al in this issue of The American Journal of Managed Care on the success of the University of Michigan’s M-CARE diabetes initiative is important reading for stakeholders in chronic care management.⁵ The authors examined changes in key measures of diabetes care associated with this initiative over a 6-year period. Like other researchers who conducted similar studies of the topic, they found several meaningful and clinically statistically significant improvements. What is important in this study is the special mix of tailored services that changed over time, including a registry, guideline dissemination, physician reminders, nursebased care management, systems-level supports for primary care, the patient base, and the design of the insurance benefit for persons with diabetes. Although each of these elements has been readily recognized as among the key ingredients in population-based care management, the details of their design, execution, and interaction with each of the program’s other elements were undoubtedly affected by a changing patient base, physician feedback, advances in health information technology, a research culture, the particulars of the provider network, and market-based shifts in the overall insurance benefit.

Although we understand what individually works in chronic illness care and readers can feel confident about the generalizability of this knowledge in multiple care settings, the details regarding how are a decidedly local and continuously shifting phenomenon. Healthcare policymakers, physicians, regulators, and other stakeholders understand the advantages of blanketing persons who have chronic care needs with evidence-based care; however, assembling those supporting elements in Michigan versus Los Angeles County or the panhandle of Florida may be better thought of as a regional quilt. This is the paradox: the road to less clinical variation at the patient level may be lined with variable and evolving combinations of locally contrived approaches to care. There are undoubtedly other population-based interventions like M-CARE’s under way nationwide, many of which are not necessarily going to make it to the pages of the Journal. Like the patient and provider stakeholders in southeastern Michigan, the stakeholders in these programs undoubtedly believe their special approach has merit and would like to see it continue.

This variability in the approaches to local care has important implications for the architects of healthcare reform.

When you’ve seen one chronic illness care initiative, you’ve seen one chronic illness care initiative. Even the words “chronic care initiative” in the article by McEwen et al speak to the blurring of classic disease management (DM) and the patient-centered medical home (PCMH). Although advocates of both approaches go to some length to precisely define the elements that make each approach unique, M-CARE is teaching us that using a mix of elements from both DM and the PCMH also can work. What is more, this particular intervention went one step further and changed the insurance design, showing that DM and PCMH can be supplemented by other population-based interventions that blur the line between the providers and the insurers.
The success of the mix of care management services is a function of local circumstances. M-CARE didn’t rely on every dimension of the PCMH and DM, but rather chose a limited number of core features (registry, guidelines, informaticsbased decision support, and nurse coaches) that best fit its network. In addition, the countless details on the contents of the registry, the format of the reminders, and the techniques used by the nurses likely were locally tailored to address the practice patterns of southeastern Michigan.
Assessment of outcomes is a function of effectiveness, not efficacy. The concept of effectiveness is underused in the evaluation of clinical outcomes, including the coming tide of comparative effectiveness research.⁶ In typical research settings, investigators are interested in assessing the impact of a single and well-defined intervention by using designs that hold other variables neutral. Unfortunately, these “efficacy” results may not be generalizable to other settings in which patients, physicians, and local systems differ considerably from those found in academia. In contrast, effectiveness research seeks answers that provide insight by using heterogeneous interventions in heterogeneous settings. Although the results may be prone to a variety of errors and biases, there is a role for inquiry that seeks a reasonable degree of assurance rather than ironclad scientific certainty. Furthermore, there is a role for inquiry that does not necessarily result in publication, but rather produces data that can be used by local leadership. When these results appear in the published literature, population health architects gain by being able to better understand how each of the described elements may be fit together to suit their local healthcare challenges. Disease management organizations’ continuing survival may be testimony to the merits of this approach, because they are supremely willing to adapt their programs to their customers’ needs.⁷

What does this locality mean for health reform? Based on the study by McEwen et al, policymakers may be wise not to assume that classically defined DM or PCMH (or all the other terms currently buzzing about our nation’s capital, like P4P [pay for performance] or EHR [electronic health record]) are static tools that are ready to be installed in toto into any national model of care. Rather, the challenge is to enable the M-CAREs of our country to flexibly make their elements available to local health systems. What is more, the mix of population-based care management elements, insurance design, and information technology may be thought of as a decision to be made by local healthcare leaders. There is no single secret sauce. M-CARE not only has chosen what it values, but also has demonstrated that its mix of services delivered its way results in better management of diabetes mellitus.

Finally, although a large comparative effectiveness research function would be of value, health policy needs to promote the ability of smaller systems to use sophisticated, rigorous, and less-than-perfect study designs to evaluate the impact of their local programs.

Author Affiliation: Sidorov Health Solutions, Harrisburg, PA.

Address correspondence to: Jaan Sidorov, MD, MHSA, Sidorov Health Solutions, 413 Village Wy, Harrisburg, PA 17112. E-mail: jaans@aol.com.

1. Forrest CB, Weiner JP, Fowles J, et al. Self-referral in point-of-service health plans. JAMA. 2001;285(17):2223-2231.

2. Liu J, Bellamy G, Barnet B, Weng S. Bypass of local primary care in rural counties: effect of patient and community characteristics. Ann Fam Med. 2008;6(2):124-130.

3. Sirovich B, Gallagher PM, Wennberg DE, Fisher ES. Discretionary decision making by primary care physicians and the cost of U.S. healthcare. Health Aff (Millwood). 2008;27(3):813-823.

4. Blendon RJ, Benson JM. Understanding how Americans view healthcare reform. Health Care Reform 2009. Posted by The New England Journal of Medicine. August 12, 2009. http://healthcarereform. nejm.org/?p=1424. Accessed August 24, 2009.

5. McEwen LN, Hsiao VC, Notal-Kirbey EM, Kulpa GJ, Schmidt KG, Herman WH. Effect of a managed care disease management program on diabetes care. Am J Manag Care. 2009;15(9):575-580.

6. Forrest CB, Shipman SA, Dougherty D, Miller MR. Outcomes research in pediatric settings: recent trends and future directions. Pediatrics. 2003;111(1):171-178.

7. Felt-Lisk S, Mays GP. Back to the drawing board: new directions in health plans’ care management strategies. Health Aff (Millwood). 2002;21(5):210-217.