This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

M E D I C A R E
I M P R O V I N G  C A R E  

26 April 2005  
  

Challenges In Improving Care For 
High-Risk Seniors In Medicare

Lessons and observations from past field demonstrations.

By Marsha Gold, Timothy Lake, William E. Black, and Mark Smith

ABSTRACT:

Despite strong interest in improving care for high-risk elders, demonstration projects typically show negative results. This paper examines one large foundation-sponsored initiative to gain insight on why success often is so elusive. The findings indicate that specific flaws in concept, design, and implementation each make it more challenging for demonstrations to achieve their intended goals, especially those involving cost and utilization reductions. We speculate that part of the reason for this is that organizational and political processes lead to fundamentally conservative demonstrations that assume that small amounts of resources directed at incremental change can be effective in generating substantial change in organizations and can do so rapidly.

There is much interest in improving care delivery for people with chronic illnesses, particularly those whose complex needs contribute extensively to the costs of health care. Indeed, such improvements factor high among the targeted priorities recently identified by the Institute of Medicine.¹ The Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003 also includes a number of new initiatives to improve quality and care management in Medicare.²

Achieving improvements in care for frail elders is very difficult, and multiple previous efforts have failed to affect Medicare costs and health care outcomes. Prominent early examples include the National Long-Term Care Channeling and Social Health Maintenance Organization (S/HMO) demonstrations, which introduced care management in an effort to generate change. The evaluations of these programs showed that they failed to reduce costs and, in the case of the S/HMO, had inconsistent effects on health and functional status and quality of care.³ Medicare demonstrations of case management in the early 1990s also showed little evidence of success when subjected to formal evaluation.⁴ Independent reviews of innovative programs targeting chronic illness for the elderly also have found limited or inconsistent evidence of success.⁵ Evaluation findings from the Program of All-Inclusive Care for the Elderly (PACE) demonstration were more positive, but this demonstration targeted people at very high levels of frailty and involved developing new organizational forms rather than overlaying change onto existing organizations and care delivery systems.⁶

Without an understanding of why these efforts were unsuccessful, policies and programs aimed at improving care for elderly patients with complex needs are likely to continue to disappoint. Such awareness is particularly relevant in the current policy environment, because Medicare is a federal program and its beneficiaries have complex needs that account for a large share of the nation’s health care costs.⁷

The challenges associated with creating change are considerable. As Gerard Anderson and James Knickman articulate, care systems tend to be fragmented. Insurance coverage focuses on services that resolve acute conditions rather than those that slow the progression of disease or maintain mobility, and it targets single conditions instead of comorbidities. Payments typically target acute care episodes, with little or no provision for ongoing care and counseling, and long-term care is considered a social rather than medical service.⁸ These fragmented funding streams have carried over into managed care.

A total overhaul of the system presents obvious and major difficulties, but evidence suggests that even limited change is challenging. In a recent national survey, for example, half of all medical groups or independent associations with twenty or more physicians reported the use of four or fewer of the sixteen common care management processes.⁹

In an effort to learn more about why success is so elusive, this paper analyzes one foundation’s major effort to improve care for high-risk seniors. Our analysis suggests that outcomes varied across projects but that flaws in conceptualizing the intervention, unrealistic expectations of success, and overly optimistic views about the ease of implementation all limited positive outcomes of demonstrations. At the end of the paper we consider why such outcomes are particularly common in evaluations of demonstrations in this area.

Overview Of The Program For Elders In Managed Care

The California HealthCare Foundation’s (CHCF’s) Program for Elders in Managed Care (PEMC) was authorized in 1997 as the foundation’s first program. Beginning in 1998, two rounds of awards were made. In total, the PEMC funded nine implementation grants with up to $1 million for three years, seven of which ultimately were implemented and are discussed here.¹⁰ Of these seven, five focused generally on elders at high risk, and two focused on improved care for dementia. In addition, the PEMC provided support for several operational activities that were disease-specific but also focused on care for high-risk elders.

The PEMC was conceived at a time when Medicare managed care was growing rapidly nationally and was expected to accelerate sharply in response to the implementation of Medicare+Choice (M+C) as authorized in the Balanced Budget Act (BBA) of 1997. Although we now know that such expectations proved wildly wrong, they were widely shared at the time.¹¹ In contrast to the rest of the country, California already had what was regarded as a strong and robust managed care market that was actively engaged in Medicare.¹² The CHCF hoped to leverage California’s mature managed care market to test changes that were feasible under the current system of insurance and care delivery.

At its core, the PEMC involved grants to consortia of managed care organizations, providers, and community organizations to improve care for high-risk seniors in Medicare managed care through a variety of diverse interventions. It had what its sponsors believed to be relatively limited but realistic and important objectives. The program was built around a managed care base because the incentives of capitation make it easier for providers to be somewhat flexible in benefits as long as the changes are cost-neutral. By encouraging coordination between medical and social services, the PEMC aimed to provide an additional layer of support to help high-risk elders with multiple chronic conditions address physical and cognitive impairments that complicate needs and impede effective self-care. Having heard criticism that past programs were overly prescriptive, the CHCF designed the PEMC to be flexible so that teams of managed care organizations, providers, and community groups could propose feasible interventions that were both responsive to their needs and resources and consistent with the program’s goals. Potential grantees were required to test and fund an independent evaluation of their interventions, with a strong preference for randomized designs.

Despite the good-faith efforts of grantees to implement the interventions as planned and reasonable success at this goal, almost none of the interventions resulted in improvements on measures of health status, quality of life, service use, or cost of care.¹³ Because improving care for high-risk seniors remains a concern, this paper seeks to identify why outcomes were not stronger and what lessons can be learned that might lead to different results in the future.¹⁴

Summary Of PEMC Interventions And Outcomes

The PEMC funded two types of operational projects. The first, and most typical, were various forms of case management interventions that focused on high-risk beneficiaries. The second were projects targeted at a specific condition (for example, dementia). Exhibit 1 provides additional project-specific information.

General interventions focused on high-risk people. The PEMC’s core projects involved identifying seniors at risk and developing care plans that included the use of community-based providers. Five of the seven original projects reached operational status: (1) Jewish Family and Children’s Services (JFCS); (2) Kaiser Tri-Central (KTC); (3) Sharp HealthCare; (4) PacifiCare/Secure Horizons; and (5) Long-Term Care Group (LTCG)/CalPERS.¹⁵ Of these five, all but one (LTCG) involved fairly traditional case management interventions that differed in their structure and intensity; one (KTC) also included expanded benefits for home and community-based services, along with an intensive case management option. The LTCG promoted the use of exercise and lifestyle change to lower-risk seniors who were enrolled in a managed care plan and had purchased long-term care insurance. Although specific outcome measures differed by site, each involved assessing relatively short-term (for example, twelve-month) changes in health status/quality of life, use of services, and spending. These outcomes, it was felt, provided necessary evidence about the potential positive return on investment to generate long-term support from health plans or other organizations (such as CalPERS, the California Public Employees’ Retirement System). With a few limited exceptions, however, the evaluations found little or no change on these measures within this time frame. After the PEMC ended, only the Sharp program was maintained, although tools and insights from the demonstration have influenced care in at least two additional sites (KTC and Secure Horizons).

Projects targeting a specific condition. Four PEMC-funded grants targeted a specific condition. The Institute on Aging and University of California, Los Angeles (UCLA) Neuropsychiatric Institute (UCLA NPI) focused on dementia.¹⁶ Project IMPACT, which involved the PEMC partnering with the Hartford Foundation’s national program, focused on the treatment of elders with depression, and the UCLA Division of General Internal Medicine (GIM) has developed and is testing diabetes guidelines to better support care for older adults.¹⁷ Although evaluations of each of these demonstrations are not necessarily complete, their findings tend to be more encouraging than those from the general case management demonstrations. Findings from one dementia demonstration (UCLA NPI) provide some evidence that the demonstration resulted in care that was more consistent with dementia care guidelines than previously.¹⁸ National findings from Project IMPACT also are positive; they show, for example, significant reductions in depressive symptoms and improved quality of life, although California-specific results are not yet available.¹⁹

Why Were PEMC Projects Not More Successful?

The PEMC appears from the evaluation to be well run and financed by a foundation with a strong, clear vision. The program was supported by grantees who worked hard to translate ideas into reality, a program office to help them do so, and an outside panel of experts positioned to provide access to existing knowledge in this area. Yet few of the projects resulted in meaningful change on the outcomes measured, and while some important tools, guidelines, and education remain, most efforts have not been sustained. With the benefit of hindsight, the PEMC’s success was limited by flaws in concept, problems in translating concept to effective design, and problems in execution.

Concept. Like most demonstration projects, the PEMC and its grantees expected a lot from a relatively limited program, a point to which we return later as we draw broader national lessons. Under the PEMC, grantees received up to $1 million for work that would span three years and also include a rigorous evaluation of results. In the context of most of these large organizations, this amount of funding is relatively invisible and dwarfed by revenues from patient services, whose incentives serve to guide the development of the organizations in which the PEMC projects were embedded. Accurately perceiving the market, PEMC grantees defined success for most projects as the ability to yield “big” results that would improve health status and functioning, reduce use, or cut plans’ costs and do so rapidly. Today such a standard remains typical of what purchasers demand and health plans say they need to strive for, but evidence to support such expectations is limited.²⁰

Potential conceptual flaws in some PEMC interventions also lessened the chances for success. In some projects—for example, Sharp HealthCare—the intervention differed little from the care already received by those in the control group, and utilization was already low, which made further reductions unrealistic. In others, getting participants to use offered services was an issue. For example, encouraging KTC enrollees to purchase a service benefit or those in the LTCG to participate in the education classes offered was problematic. The PEMC experience highlights potential conflicts between what beneficiaries and professionals perceive they need. (There may be reluctance by beneficiaries to accept home-based services except when in a crisis.) Interventions also may be insufficiently targeted to people who were most ready to receive them.

Design. The PEMC was developed by a new foundation and was implemented relatively rapidly. Technical assistance in the form of an experienced program office was not established until the program was well under way. Although foundation staff used experts to review proposals and make funding recommendations, hindsight reveals that not all of the tested interventions were as operationally well-defined or strong as they might have been. For example, the sophistication of eligibility criteria to screen for high-risk participants varied greatly across sites. At least two sites—Sharp and LTCG—concluded that better targeting might have improved the outcomes. Although evidence indicates that provider involvement—generally viewed as key—was part of the PEMC, it was limited in some projects, and securing it was a challenge for projects not based in provider organizations. The short time frame for demonstrating outcomes, the low prevalence of some outcomes, and the limited size of treatment groups likely also affected the results. Unfortunately, small numbers might be unavoidable in some interventions, particularly if they target those at highest risk. Expert panels, eager to be responsive to potential funders’ interests, might be too reluctant to provide needed criticisms of well-intentioned efforts in ways that could have improved some interventions’ effectiveness.

Implementation. Both internal and external factors hindered implementation.²¹ Approval of the demonstration by institutional review boards (IRBs) took more time than many expected.²² Because multiple entities were involved, some projects required approval from three or four IRBs. Recruiting patients quickly was a challenge in almost every project. Even when recruitment targets were met, project schedules were often delayed, which left less time to show effective outcomes. Externally, the environment for managed care was changing rapidly in California, as it was across the country. Although California was spared the massive withdrawals from the Medicare+Choice market that occurred in some parts of the country, medical groups’ financial status and stability were major issues in California that disrupted the system. At least four PEMC projects were adversely affected because they lost the support of major providers involved in their projects.

Lessons For The Future

The PEMC experience is highly relevant to today’s efforts to fundamentally improve care for high-risk seniors. The PEMC is not the first demonstration in this area unable to reduce the costs of care through the provision of care management. We conclude by highlighting five issue areas worthy of consideration in developing effective future interventions.

The challenge of establishing realistic outcome goals. Demonstration projects often emerge when support for more massive change does not exist, which leads to an interest in testing and introducing change on a smaller-scale, more incremental basis. Some demonstrations can smooth the way for larger change—for example, Medicare’s nationwide adoption of diagnosis-related groups (DRGs) was viewed more positively when policymakers could point to New Jersey’s experience with this system even though the methods used in New Jersey differed from what was implemented nationally for Medicare. But many demonstrations appear to generate little broader change in the health care system. It is beyond the bounds of this study to detail precisely why this is so; however, we speculate that part of the reason may be that political processes result in pressure to enlarge demonstration goals beyond what is reasonable to expect as politicians and organizations aim to “sell” the concept.

It also might be difficult to reconcile potential discrepancies between what the market or policy environment demands and what interventions can realistically achieve. For example, research to date indicates that achieving financial savings from disease management programs (DMPs) is difficult and could be far harder with elderly people who have complex chronic conditions and situations.²³ In the case of the PEMC, well-designed screening tools, the support of credible sponsors (as with CalPERS), and active involvement of the medical staff (as in the Kaiser Permanente group and Sharp HealthCare) led to stronger interventions, although they were still not sufficient to meet the large goals set for them.

The challenge of effectively targeting interventions to problems. If political pressures lead funders to expect more from demonstrations than they can deliver, organizational processes may at the same time lead to a diffusion of demonstration concepts, as potential grantees aim to meld demonstration concepts with the reality of individual organizational needs and constraints. In the PEMC interventions we studied, it was often not clear which problem was being addressed and by what logic the intervention was likely to achieve the outcomes targeted. Thus, the PEMC was characterized as a program for “frail elders,” yet its content included elements relevant to the treatment of both chronic disease and functional impairments among a range of elders with varying degrees of risk.²⁴ The specific needs to target and the treatments with which to do so were not always explicitly defined by grantees, nor was the category of “high-risk seniors” or “frail elders.” Further, the common practice nationally of distinguishing interventions as “disease management” or “case management” modes provides little guidance on how to prioritize and develop interventions for high-risk people who might have elements of both sets of problems (see Exhibit 2). For example, guidelines could be useful to provide structure for all types of interventions, not just disease-specific ones, and information on other concurrent conditions is often essential in any kind of treatment. The different funding streams that apply to medical and social interventions also complicate targeting care for high-risk elders who require both medical and social services.

If an intervention’s objective is to save money in the medical care system, policymakers and those sponsoring interventions arguably might need to be more ready to accept that (1) there may be important problems whose solutions will not save the medical care system money, and (2) there are limits to how far the medical care system can likely go on its own toward solving problems whose origins are external to that system and located in the broader social fabric. Effective cost-saving interventions probably need to be well targeted to those who spend a lot on medical care; do so for reasons that are understood and in areas where evidence exists on effective practice; and can, with some reasonable amount of help, be cared for equally effectively for less money.

It is disconcerting that these criteria leave out a number of people with important and urgent needs. But identifying such situations could be the first step in considering how the problems they reflect might more realistically or effectively be addressed.

Paying attention to fundamentals. Experienced evaluators likely will agree that almost any program they have examined took longer to implement than sponsors anticipated and that operational challenges, anticipated and unanticipated, constrained ultimate program success. Yet politicians and managers, facing pressure for rapid and notable results, often underestimate implementation requirements.

The PEMC demonstrates that chronic care improvement efforts should be carefully targeted and that substantial investment and time are required to recruit and retain patients while encouraging the use of appropriate care. Day-to-day efforts at communication among providers, patients, and those who support them must be well structured and supportive of the intervention. Ample time at the front end is important, to establish new channels and reinforce existing ones, and to negotiate the potentially lengthy series of administrative approvals required to support patient care interventions in today’s environment.

Developing an appropriate role for evaluation. The PEMC, like many current Medicare disease management demonstrations, used the “clinical gold standard” randomized controlled trial (RCT) to assess the success of interventions. PEMC participants were supportive of this approach, even though the outcomes were disappointing. Many noted that a rigorous evaluation could identify when an intervention did not work, enabling the diversion of efforts to other approaches.

Yet such evaluations take time, may be expensive, and may lack critical information to explain why success was not more forthcoming. Some also argue that they may be poorly suited to situations in which the research premise of holding “all else constant” is violated, even in randomized designs, because (1) controlled comparisons are insufficient to account for the kind of continuous change that occurs in the real world as interventions are adapted in response to learning, or (2) trends for both groups are disrupted by major environmental events such as the collapse of some of the medical groups that participated in the demonstration.

In recent years, Donald Berwick and others have proposed an alternative strategy to encourage and evaluate improvements in care, called the continuous quality improvement (CQI) model.²⁵ This type of model aims to encourage organizational change in complex systems through innovation with rapid-cycle feedback that can help systems learn from experience and adapt.

The implications of such structures for evaluating innovation have not yet been fully considered within clinical settings and the health services research community to anticipate the conflicts likely because CQI methods will not generate the kind of hard evidence that clinicians have come to expect from randomized trials. The PEMC experience suggests that there may be merit in blending both strategies. A rapid-cycle model can be useful in the initial stages of implementation to help define and target the intervention so that the most promising ones can be identified and targeted. Highly promising interventions can then be subject to a more formal test that generates the kind of evidence clinicians and others expect.

Taking advantage of opportunities under MMA. Generating major improvements in care for high-risk elders is difficult without the major system changes called for as part of the chronic care model. Yet massive change in financial incentives and how institutions relate to one another is not likely to occur in the immediate future. MMA and related recent initiatives create opportunities for progress, but the points we have raised previously also suggest that expectations about these opportunities’ outcomes should be realistic.²⁶

For example, under MMA, Medicare is mounting a new pilot that will be large, comprising approximately 270,000 beneficiaries across nine sites. Although modeled on traditional DMPs, pilot programs for beneficiaries with congestive heart failure and diabetes will target those at higher risk and include care for comorbidities and case management. Success will be measured on the entire eligible population, not just those participating. Yet it remains to be seen whether these limited incremental interventions will be sufficiently strong or targeted to generate the financial savings (5 percent) that the Centers for Medicare and Medicaid Services (CMS) expects from them.

MMA also provides the authority to contract with specialized private plans that focus on care for Medicare beneficiaries with special needs, including institutionalized patients and those dually eligible for Medicare and Medicaid. Merging Medicare and Medicaid benefits to support an integrated intervention offers promising potential to blend medical and social services for high-risk seniors. This authority removes an important roadblock to providing more integrated care for high-risk seniors because it merges the Medicare and Medicaid funding streams. Again, however, expectations need to be realistic. In the PEMC, several planning grants were provided to organizations seeking to better managed care for elders who were eligible for both Medicare and Medicaid, but sponsors had a difficult time developing operational plans that could sufficiently merge the cultural and legal distinctions that exist in the two programs.

In sum, there is much to be learned from research on efforts to encourage organizational care in ways that seek to improve care for frail elders. But such findings also challenge policymakers, health care providers, and other stakeholders because they highlight issues that make the pursuit of change challenging.

Research for this paper was supported by the California HealthCare Foundation, a nonprofit philanthropic organization whose mission is to expand access to affordable, quality health care for underserved individuals and communities and to promote fundamental improvements in the health status of the people of California. The authors are grateful to Jan Eldred at the foundation for her guidance and flexibility with this evaluation; to David Reuben, Janet Frank, and Pamela Jackson-McCall in the PEMC’s program office at UCLA for their support and access to information; and to the PEMC grantees and independent evaluators for their openness and cooperation. At Mathematica, Debra Draper was a core researcher on the project for much of the time. Beth Stevens and Craig Thornton provided valuable comments on earlier drafts of this manuscript.

NOTES

1. The Institute of Medicine (IOM) names care coordination as a cross-cutting priority for national attention. Chronic disease care heavily dominates the conditions they identify as priorities (from frailty associated with old age to such diagnostic specific conditions as diabetes, ischemic heart disease, and major depression). IOM, Priority Areas for National Action: Transforming Health Care Quality (Washington: National Academies Press, 2003).
2. Alliance for Community Health Plans, The Medicare Prescription Drug, Improvement, and Modernization Act of 2003: A Review of Quality-related Provisions (Washington: Alliance for Community Health Plans, 2004).
3. P. Kemper, “The Evaluation of the National Long Term Care Demonstration: 10. Overview of the Findings,” Health Services Research 23, no. 1 (1988): 161–174; and T. Thompson, “Evaluation Results for the Social/Health Maintenance Organization II Demonstration” (Washington: U.S. Department of Health and Human Services, 2002).
4. See, for example, J.L. Schore, R.S. Brown, and V.A. Cheh, “Case Management for High-Cost Medicare Beneficiaries,” Health Care Financing Review 20, no. 4 (1999): 87–101.
5. C. Boult et al., “Innovative Healthcare for Chronically Ill Older Persons: Results of a National Survey,” American Journal of Managed Care 5, no. 9 (1999): 1162–1172; and D.B. Reuben, “Organizational Interventions to Improve Health Outcomes for Older Persons,” Medical Care 40, no. 5 (2002): 416–428.
6. P. Chatterji et al., Impact of PACE on Participant Outcomes (Cambridge, Mass.: Abt Associates, July 1998); and A. White, The Effect of PACE on Costs to Medicare: A Comparison of Medicare Capitation Rates to Projected Costs in the Absence of PACE (Cambridge, Mass.: Abt Associates, July 1998).
7. For an excellent discussion of the characteristics of high-risk elders and the challenges in caring for them, see C. Thornton et al., Constrained Innovation in Managing Care for High Risk Seniors in Medicare+Choice Risk Plans (Washington: Mathematica Policy Research, January 2002).
8. G. Anderson and J.R. Knickman, “Changing the Chronic Care System to Meet People’s Needs,” Health Affairs 20, no. 6 (2001): 146–160.
9. The survey asked about use of case management, physician feedback, a disease registry, clinical practice guidelines, and self-management skills for four conditions: asthma, congestive heart failure, depression, and diabetes. L.R. Casalino et al., “External Incentives, Information Technology, and Organized Processes to Improve Health Care Quality for Patients with Chronic Diseases,” Journal of the American Medical Association 289, no. 4 (2003): 434–441.
10. The other two grants were terminated. One program faced delays in obtaining the federal waiver needed to implement the demonstration, resulting in termination (the program ultimately received a waiver and now operates independently of the PEMC). The other involved a grant to a relatively new managed care initiative in a relatively rural area. The grant was terminated after the participating managed care plan severed links with the participating independent practice association (IPA) after the IPA encountered financial problems. The PEMC also included a limited companion program of planning grants available to programs applying for implementation grants in the second round. Of ten organizations awarded planning grants in the first cycle, three later received an implementation grant in round two. Those receiving planning grants viewed them as important in building effective operational projects. Many of the programs that proceeded directly to implementation said that, in retrospect, they had underestimated what this required and would have benefited by additional time to plan.
11. M. Gold, “Medicare+Choice: An Interim Report Card,” Health Affairs 20, no. 4 (2001): 120–138.
12. M. Gold and T. Lake, Medicare+Choice in California: Lessons and Insights (Washington: Henry J. Kaiser Family Foundation, September 2002).
13. M. Gold, T. Lake, and W. Black, Evaluation of the Program for Elders in Managed Care: Final Report (Submitted to the California HealthCare Foundation, 20 August 2004). This paper builds upon the insights of this cross-cutting evaluation of the PEMC program that the CHCF commissioned from MPR in spring 2002 (toward the end of the program). The evaluation was designed to complement grantees’ independent outcome evaluations by providing a programwide perspective to generate new insights and a vehicle for integrating results across the project. At the CHCF’s request, MPR relied heavily on a review of secondary sources developed over the course of the program (including the independent evaluation of each grantees’ outcomes) but conducted semistructured interviews with the grantees and other program participants to validate or elaborate on secondary sources.
14. S.M. Foote, “Population-based Disease Management under Fee-for-Service Medicare,” Health Affairs, 30 July 2003, content.healthaffairs.org/cgi/content/abstract/hlthaff.w3.342 (22 March 2005).
15. Jewish Family and Children’s Services (JFCS) was led by an experienced community-based organization collaborating with twenty-four affiliated medical groups. See JFCS, SeniorCare Final Report: Cost Effective Management of Frail Elders (Report submitted to the CHCF, 28 December 2001). Kaiser Tri-Central (KTC) involved managed care organizations (MCOs) collaborating with community-based organizations. See Partners in Care Foundation, Kaiser Permanente Community Partners Project: Final Report (Report submitted to CHCF, October 2003). Sharp HealthCare’s intervention is based in a large established group practice. See R. Newcomer et al., “Outcomes of an Early Intervention Case Management Program in Three Medical Groups,” Evaluation and the Health Professions 27, no. 4 (2004): 323–348; and Sharp HealthCare, “Frail and Elderly Case Management Project 2000–2003” (Final report submitted to CHCF, 29 January 2004). The PacifiCare/Secure Horizons intervention was led by an MCO in collaboration with multiple community groups and affiliated providers. See K.H. Wilber et al., “Partnering Managed Care and Community-based Services for Frail Elders: The Care Advocate Program,” Journal of the American Geriatrics Society 51, no. 6 (2003): 807–812; and K.H. Wilber and G. Shannon, Partnering Managed Care and Home and Community-based Services: The Care Advocate Program (Final report submitted to CHCF, 14 November 2003). The Long-Term Care Group (LTCG)/CalPERS project was a collaboration between the California Public Employees’ Retirement System (CalPERS) and the LTCG, the agency that administers the long-term care benefit for CalPERS members. See S.K. Holland et al., “Preventing Disability through Community-based Health Coaching,” Journal of the American Geriatrics Society 51, no. 2 (2003): 265–269; and S.K. Holland et al., Preventing Disability through Community-based Health Coaching: The CalPERS Health Matters Program (Final report submitted to CHCF, 28 October 2003).
16. The Institute on Aging intervention, part of a seven-site national initiative, was implemented by a community-based organization in collaboration with the Brown and Toland Medical Group. See D. Coon, G. Dowling, and L.Z. Feigenbaum, Northern California Chronic Care Network for Dementia Project (Revised final report submitted to CHCF, 26 January 2004). The UCLA Neuropsychiatric Institute (NPI) intervention was conducted by a task force of provider organizations and community agencies. See K.I. Connor and B. Vickrey, San Diego Alzheimer’s Disease Management Collaborative Initiative (Final report submitted to CHCF, 28 February 2004).
17. On Project IMPACT, see J. Unutzer et al., “Depression Treatment in a Sample of 1,801 Depressed Older Adults in Primary Care,” Journal of the American Geriatrics Society 51, no. 4 (2003): 505–514. More recent analysis indicates that these findings remain after twenty-four months. The UCLA GIM was pilot-tested in two sites: Sharp HealthCare and the West L.A. Veterans Administration.
18. See Connor and Vickrey, “San Diego Alzheimer’s Disease Management Collaborative Initiative.”
19. Unutzer et al.. “Depression Treatment.”
20. See, for example, Congressional Budget Office, An Analysis of the Literature on Disease Management Programs (Washington: CBO, 13 October 2004).
21. Some potential barriers to success proved less problematic than they might have been. Because it was a feature of the program from the start, randomization went smoothly in most sites, and data challenges appear fairly limited. Although there are inherent challenges in melding the disparate cultures of medical and social service providers, grantees generally were able to coordinate their efforts across the two. The most concrete problem was getting data from community-based organizations that are not set up to communicate with large insurance companies.
22. The exceptions included one project in a provider organization that was classified as a “quality improvement project” rather than a clinical trial; another had an experienced insider advising on the sequence of steps that would most effectively gain approval promptly; and another used its planning grant to “jump start” the process.
23. CBO, An Analysis of the Literature.
24. We decided to use “high-risk seniors” because colleagues assumed that a focus on “frailty” meant that these were long-term care demonstrations of social interventions, when, in fact, they were located in medical care organizations and tended to have a medical component (although it was not always well developed).
25. D.M. Berwick, “Continuous Improvement as an Ideal in Health Care,” New England Journal of Medicine 320, no. 1 (1989): 53–56; and D.M. Berwick, “Developing and Testing Changes in Delivery of Care,” Annals of Internal Medicine 128, no. 8 (1998): 651–656.
26. See Alliance for Community Health Plans, The Medicare Prescription Drug, Improvement, and Modernization Act of 2003; and Centers for Medicare and Medicaid Services, “Medicare Program; Voluntary Chronic Care Improvement under Traditional Fee-for-Service Medicare,” CMS-5004-N 2004, Federal Register 69, no. 79 (2004): 22065–22079.

Marsha Gold (mgold{at}mathematica-mpr.com) is a senior fellow; Timothy Lake, a senior researcher; and William Black, a research analyst at Mathematica Policy Research in Washington, D.C. Mark Smith is president and chief executive officer of the California HealthCare Foundation in Oakland.

DOI: 10.1377/hlthaff.w5.199
©2005 Project HOPE–The People-to-People Health Foundation, Inc.