New Marketplace

Not All High-Risk Populations Are the Same

Article · May 5, 2016

Managing the care of high-cost patients is a key concern of physicians and health systems that are forming accountable care organizations (ACOs) and entering into alternative payment contracts tying reimbursement to performance on cost trends and quality measures.1 The logic is simple: given that a small percentage of patients (often those with complex or multiple medical conditions) account for the majority of health care spending,2 directing additional resources and services toward patients who are likely to incur high costs and experience poor outcomes — a strategy known as high-risk care management — could substantially reduce costs and improve quality. Faith in this proposition has led to widespread adoption of high-risk care-management programs by ACOs.2

Successfully structuring these programs requires targeting the particular drivers of excess care utilization by high-cost patients; only programs that closely match delivery interventions to specific clinical needs have succeeded.3 Prevailing approaches for managing the care of high-cost patients focus on improving adherence and disease management for multiple co-occurring conditions — a strategy developed and tested among Medicare patients.1 Can tactics honed among the elderly be successfully applied to other high-cost populations?

Health services researchers have documented that across populations covered by different health care payers, small groups of patients are responsible for outsized portions of health care costs.2 Less is known, however, about variation in clinical characteristics and care-utilization patterns among payer-defined groups. To further characterize this variation, we analyzed 2014 claims data for the costliest 1% of patients in each payer category whose care is managed by Partners HealthCare, a large integrated delivery system in Massachusetts (see table). Because of the structure of U.S. health care financing, these payer-defined populations are helpful surrogates for clinically distinct subgroups of patients and reflect the locus at which alternative payment contracts are negotiated.

Selected Utilization and Health Status Information for High-Cost Patients at Partners HealthCare

Selected Utilization and Health Status Information for High-Cost Patients. Click To Enlarge.

The costliest 1% of Medicare patients had an average of eight co-occurring chronic conditions. Most had cardiovascular risk factors, and more than half had end-stage sequelae of ischemic heart disease, congestive heart failure, or chronic kidney disease. These patterns argue for the use of disease management and care coordination to improve care and reduce cost. For example, the use of nurse care managers who work with high-risk patients to coordinate care among providers, monitor and track outcomes, and engage patients in disease management has been successful in some cases.1 We also found that 20% of spending in this group is attributable to post-acute care, which suggests the need for strategies aimed at high-value post-acute and skilled-nursing care.

In the Medicaid population, high-cost patients also had several co-occurring chronic conditions (an average of five) but there was a striking prevalence of mental health disorders. A quarter of the patients had been diagnosed with depression, another quarter with anxiety, and almost one fifth with bipolar disorder. Given the burden of mental health disorders and current inadequacies in outpatient detection and management,4 it would make sense for accountable care strategies for this population to improve access to mental health care and to integrate mental health services into broader care-coordination and disease-management models. The demographics of this population also necessitate care-management strategies focused on the socioeconomic context of care delivery and health maintenance5 and targeting the substantial cognitive and physical disabilities affecting certain subpopulations (e.g., dually eligible Medicare and Medicaid beneficiaries).

For high-cost patients in commercial plans, the picture was quite different. These patients had fewer chronic conditions than their Medicare or Medicaid counterparts and were more likely to have disease risk factors than end-stage sequelae. Drivers of high costs in this population included catastrophic injuries, neurologic events, and need for specialty pharmaceuticals — particularly antineoplastics, but also biologics for multiple sclerosis or rheumatoid arthritis. Strategies for managing the care of high-cost commercial patients would therefore need to focus on ensuring appropriate use of specialty pharmaceuticals in treating chronic disease. And although catastrophic events themselves are often unavoidable, care-management models could be developed to coordinate and improve follow-up care in an effort to improve outcomes and prevent unnecessary care utilization.

These patterns underscore the challenges and opportunities inherent in caring for high-cost patients. Though there are some evidence-based tactics for addressing the drivers of high costs and poor outcomes among various subpopulations,1,3,4 tremendous heterogeneity in clinical characteristics and care needs complicates efforts to develop integrated strategies. Interventions that are appropriate and effective for one group will often do little to improve care and reduce costs for others.

That reality has important implications for health systems and physicians. As reform activities shift payment away from fee-for-service models, the incentives to improve care for high-cost patients will continue to grow. Specific patient subgroups and clinical trends will vary with the demographics, payer mix, and catchment area of a given physician group or health system. Sustaining performance across populations will require a diversified approach tailored to the particular clinical realities of target populations.

Such tailoring will complicate strategic planning and investments in population health management and accountable care capabilities. Even for health systems that have experience with alternative payment models, assuming responsibility for new populations will require an in-depth understanding of the clinical characteristics and care-utilization patterns of high-risk subgroups and identification of evidence-based programs and tactics for managing their care. The required up-front investments will be substantial. Thus, it will be important for payers to understand that demonstrated competence in caring for specific high-risk populations does not obviate the need for contracts to provide adequate financial incentives and security to support investments in new care-management capabilities.

The implications for the roles of individual physicians are quite different. Depending on their specialty and panel composition, physicians may see patients ranging from the full spectrum of their health system’s high-cost patients to a single subgroup. Physicians in ACOs, therefore, have a critical role in engaging patients and matching them with specific programs according to clinical need. Furthermore, front-line clinicians can help system leaders identify and test new strategies for high-cost patients and provide insights into care needs at a level of nuance and granularity that cannot be gleaned from claims or electronic health data.

Focusing on high-cost patients has become an attractively simple approach to improving care and reducing costs. But this policy panacea is challenged by the reality that patient demographics, health needs, and utilization patterns vary substantially among populations. Optimizing investments in this area will require improving analysis of which patients are amenable to care-delivery interventions and prioritizing interventions according to the specific needs of subpopulations.

The views expressed in this article are those of the authors and do not necessarily represent the views or policies of their institutions.


From Harvard Medical School (B.W.P., S.K.C.), Harvard Business School (B.W.P.), Partners HealthCare (S.K.C.), and Massachusetts General Hospital (S.K.C.) — all in Boston.

1. Hong CS, Abrams MK, Ferris TG. Toward increased adoption of complex care management. N Engl J Med 2014; 371: 491-3.
2. Cohen S, Uberoi N. Differentials in the concentration in the level of health expenditures across population subgroups in the U.S., 2010. Statistical brief 421. Rockville, MD: Agency for Healthcare Research and Quality, 2013.
3. Brown RS, Peikes D, Peterson G, Schore J, Razafindrakoto CM. Six features of Medicare coordinated care demonstration programs that cut hospital admissions of high-risk patients. Health Aff (Millwood) 2012; 31: 1156-66.
4. Druss BG, Walker ER. Mental disorders and medical comorbidity. Research Synthesis Report 21. Princeton, NJ: Robert Wood Johnson Foundation, 2011.
5. Coughlin TA, Long SK. Health care spending and service use among high-cost Medicaid beneficiaries, 2002-2004. Inquiry 2009-2010; 46: 405-17.

This Perspective article originally appeared in The New England Journal of Medicine as “ACOs and High-Cost Patients.”

New call for submissions ­to NEJM Catalyst

Now inviting longform articles


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »

More From New Marketplace
Health Care Costs - Cost Transparency Is the Main Change Needed to Support Patients in Lowering the Cost of Care

Survey Snapshot: Is Transparency the Answer?

NEJM Catalyst Insights Council members say that while transparency might be necessary, we have to change the way care is delivered.

Many Stakeholders Share Responsibility for Lowering the Cost of Care

New Marketplace Survey: Patients Lack Information to Reduce the Cost of Care

Efforts to mitigate rising health care costs have left most industry participants — patients, providers, and payers alike — frustrated at the lack of progress. While there is plenty of blame to go around, a significant open question remains: how responsible are patients for reducing costs?

Hip or Knee OA Care Management - Musculoskeletal Institute and Bundled Payment Program at Dell Medical

Bundled Payments Are Moving Upstream

Dell Medical School at The University of Texas at Austin has improved outcomes and reduced costs for musculoskeletal care with a multidisciplinary Musculoskeletal Institute and bundled payment model.

Elements of a Sustainable Complex Care Management Contract

Sustainable Financing for Complex Care Management Is Critical to a Value-Driven Health Care System

Care management should be payer-agnostic at its core.

Comparison of Certain Model Features in Blue Cross NC Blue Premier vs Next-Generation ACO Model vs BCBSMA Alternative Quality Contract

Engineering a Rapid Shift to Value-Based Payment in North Carolina: Goals and Challenges for a Commercial ACO Program

We believe North Carolina can be a model for the nation.

Small Molecule Drugs Facing Generic Competition - Orphan and Non-Orphan Drugs - Orphan Drug Act

It’s Time to Reform the Orphan Drug Act

Three proposals for improving the law to reflect 21st-century drug development practices.

Three-Part Pricing of PCSK9 Inhibitors

A New Model for Pricing Drugs of Uncertain Efficacy

Are we paying too much for new drugs before we know how well they work? This innovative pricing model proposes postponing major rewards until efficacy is established — which could help both patients and payers while still paying back investments on the most effective drugs.

what does quality measurement in health care mean

Buzz Survey Report: Addressing the Problems of Quality Measurement

An independent NEJM Catalyst report sponsored by University of Utah Health on patient involvement in quality measurement.

Average HOOS and Average KOOS for patients undergoing hip and knee replacement at CJRI

Building a “Hospital-within-Hospital” Model for Joint Replacements

The Connecticut Joint Replacement Institute has demonstrated that formerly competing independent providers can unite on a common vision to yield drastic improvements in quality, safety, and costs.

Discharge Rates and Follow-Up Internval Dashboard for One Provider at MGH Dermatology

A Successful Pilot to Improve Access by Adjusting Discharge and Follow-Up Rates

Actionable data and modest financial incentives can help motivate clinicians to adjust their behavior around scheduling follow-up appointments.


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »


Transforming Specialty Practice in Pursuit of…

Despite significant primary care reform around patient-centered medical home models, specialty care remains fragmented, with…

Bundled Payments

61 Articles

Bundled Payments Are Moving Upstream

Dell Medical School at The University of Texas at Austin has improved outcomes and reduced…

New Marketplace Survey: Patients Lack Information…

Efforts to mitigate rising health care costs have left most industry participants — patients, providers,…

Insights Council

Have a voice. Join other health care leaders effecting change, shaping tomorrow.

Apply Now