One of the biggest, yet perhaps not fully appreciated, ramifications of health care reform is a major focus on population health, assigning responsibility for the care of groups to Accountable Care Organizations (ACOs) and expanding coverage to a broader segment of the uninsured.
ACOs will experience the challenges of other “managed health” organizations, including the co-morbid risks of our rapidly aging population. ACOs will be accountable for both the costs and quality of care that they manage for their service populations. Especially in terms of quality measurement, there is a need for a much for sophisticated view of how the presence of other conditions can impact measures to ensure that ACO performance is equitably assessed and compensated.
With this new focus on populations and care, accountability is a growing appreciation of the influence of co-morbidity on the effectiveness of care. The aging of the baby boomer population has introduced a growing strain on U.S. healthcare that ripples through many areas, exerting serious effects on our ability to financially support and effectively offer the care that is being demanded. Baby boomers are predictably showing an increasing prevalence of chronic disease – and chronic disease rarely occurs as single isolated diseases.
[Most popular: How to harness Big Data for improving public health.]
Multi-morbidity is the 800-pound gorilla of effective health care, omnipresent but somehow unseen. Disease management programs have classically focused on one disease at time; there are compelling reasons for this. First, the evidence base is built around single diseases. By design, the gold standard of evidence randomized clinical trials requires a very homogeneous population. Complicating factors, like co-morbidity, are viewed as a problem, and these patients are often excluded from trials. Unfortunately, real patients offer unique presentations of disease that are influenced by an array of associated conditions, thus the evidence does not always speak to patients under treatment.
Further, physicians are often neither trained nor funded to focus beyond a single disease that best captures the reason for a particular visit. Multi-morbidity is the norm and not the exception among the chronically ill and aging population, however, and treating one isolated disease at a time for these persons is simply bad medicine with the potential for adverse drug interactions and contraindicated care.
The potential problems of this approach were the topic of a classic article by Boyd, et al., in JAMA. Boyd and colleagues applied evidence-based care to a prototypical, but fictional, 79 year-old elderly woman with five co-occurring common chronic conditions. Slavish adherence to clinical practice guidelines would have involved the prescribing of a complex therapeutic regimen that included 12 medications costing around $406 per month. There was a strong potential that these medications could have adverse interactions. The concordant and discordant associated conditions that may profoundly affect treatment effectiveness do not get the attention that they may warrant. In our current fragmented health care system, it is sometimes assumed that another physician will address these other conditions. Health care reform emphasizes shared accountability, insisting that every physician have some “skin in the game” related to the overall health and functioning of their patients.
The jury is still out on the effectiveness of disease- and care-management programs to improve health and to better control costs. This is undoubtedly due in part to the fact that diseases do not present themselves in some standard and predictable way. Therefore, the ultimate goal of therapy should move from, for example, treating diabetics, to treating persons with diabetes.
The revolution in e-Health, another important theme in health reform, will permit a much more finely delimited appreciation of diseases, identifying particular subgroups of patients who share a principal diagnosis but who would benefit from more tailored disease- and care-management. As a consequence, disease-management and care-management programs will deliver more of their original promise.
[Q&A: NwHIN as a bridge between private, public health in southeastern Minnesota.]
Two long-term projects inside the Johns Hopkins Bloomberg School of Public Health demonstrate what’s been done to date – and the great promise for health IT to advance coordinated-care, in this instance particular to multi-morbidity patients.
The Guided Care paradigm developed by physicians at Johns Hopkins, makes multi-morbid patients explicit targets of care management. Guided Care deploys specially trained Guided Care Nurses (GCNs) who oversee the care of 50 to 60 individuals with multi-morbidity. The GCN is resident in a primary care practice, helps coordinate the care of two to five primary care physicians, and deploys a multi-faceted approach to care that includes home visits, patient empowerment/self-management, regular health coaching, care transition facilitation, and caregiver education and support. An early assessment of the feasibility of Guided Care by Boyd et al in The Gerontologist found physicians very receptive to this process and some suggestion that patients and families experienced benefits.
Dr. Barbara Starfield created the Johns Hopkins University’s Adjusted Clinical Groups (ACG) System over 20 years ago, with the implications of multi-morbidity as a central ACG System design imperative. ACGs, from the onset, categorized care into a relatively small set of homogeneous resource groups. Organized around groups of sometimes seemingly different clinical conditions, ACGs powerfully explained how medical resources were expended. ACGs offered a simple and elegant way to adjust for the level of disease burden in populations, making possible a range of applications.
Much has changed over the ensuing 20 years, and particularly gratifying is that medicine is now catching up to Dr. Starfield’s perspective on the importance of addressing the implications of multi-morbidity. With this recognition comes even greater challenges, namely that improved and expanding data have made it possible to assess care in terms of both efficiency and quality (i.e., value) and to permit the ACG System to be used in a far more proactive fashion through predictive modeling. Predictive modeling permits the identification of potential future health risks, facilitating initiatives targeting the minimization of this risk. The ACG System’s use of new data streams, such as pharmacy claims, has reduced the latency of the period from risk discovery to action.
Physician thought leaders recognize that effective care- and disease-management cannot be supported by simple predictive risk scores alone. Over the past years, resources have been dedicated to develop an array of clinical markers that provide context to these scores and enhance the ability of care managers to more quickly hone in on those at the greatest need. Among this growing list include a summary coordination of care score, an array of metrics to capture issues associated with gaps in medication possession, and flags for unexpectedly high pharmacy utilization. Further, as noted earlier, clinical performance is now being assessed both in terms of cost and quality. Predictive modeling needs to expand beyond costs to predict health outcomes. While current data streams provide limited information related to outcomes, it is a safe assumption that the advent of new data streams drawn from electronic health records and other sources will dramatically alter the horizon of what is predictable and what can be managed.
In the coming years there will be a rapidly growing need to organize vast amounts of new information from EHRs, PHRs, and genomic profiles into a form that supports quick and effective decision-making. The implications of multi-morbidity will only become more salient as these new data sources permit greater individualization of care.
David Bodycombe, an assistant scientist in the Health Services Research and Development Center, earned a bachelor's degree from Johns Hopkins University in 1971, a master of science degree from Georgetown University in 1977, and a doctor of science from the Johns Hopkins School of Public Health in 1993.
