Health Affairs, 23, no. 3 (2004): 210-221 doi: 10.1377/hlthaff.23.3.210 © 2004 by Project HOPE	New Online   House Health Reform Bill   Paying for Reform   Vetting AHIP's Report   HIV/AIDS Costs   Brief: Insurance Reform   HA Blog Top 10

This Article Abstract Reprint (PDF) Supplemental Exhibit Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Schlesinger, M. Articles by Gusmano, M. Search for Related Content PubMed PubMed Citation Articles by Schlesinger, M. Articles by Gusmano, M. Related Collections Access To Care Health Promotion/Disease Prevention Managed Care Managed Care - Consumers Managed Care - Regulatory Issues

Health Tracking

MARKETWATCH

A Broader Vision For Managed Care, Part 3: The Scope And Determinants Of Community Benefits

Abstract

 
Managed care plans have been encouraged to address the health of the communities in which they are located. This paper presents the first nationally representative portrait of health maintenance organizations’ (HMOs’) community benefit activities, based on survey data from 1999. We found that HMOs were engaged in a wider variety of community involvements than were identified in past legal decisions and legislation defining "community benefits" for health plans. The scope of community activities was broader for HMOs that enroll Medicaid recipients, are influenced by local business leaders, operate under nonprofit ownership, and are located in states with community benefit reporting laws.

Some HMOs have engaged in projects with broad community benefit.³ However, documentation of such activities is largely either anecdotal or drawn from small demonstration projects.⁴ It has thus been impossible to assess whether a community orientation represents a broad, realistic aspiration for the managed care industry. Little is known about the scope of community benefit activities among HMOs or factors that encourage or discourage HMOs from broadening their purview beyond their enrollees.⁵

This paper addresses this gap in our knowledge. Building on a previously reported conceptual framework developed to capture the full range of community benefits and public goods, we provide the first nationally representative portrait of HMOs’ community benefit activities. Drawing on a 1999 survey, we document the scope and magnitude of these activities and plan executives’ attitudes about them. Our analysis explores plan characteristics and environmental factors that encourage or discourage community benefit activities. We also suggest policy implications for public officials concerned with the role of managed care in U.S. health care.

Pathways To Community Involvement

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
To survey HMOs’ community benefit activities, we first identified four perspectives on the topic in the literature, as discussed in two previous papers.⁶ Each suggests a distinctive form of community benefit.

Legal-historical perspective. The term community benefit derives from the expectations that apply to "charitable" health care organizations under tax law.⁷ Past legal actions suggest policies and practices that HMOs might adopt to satisfy legal requirements, such as providing free care and subsidized premiums.⁸

Market failures perspective. Economic theories of nonprofit organizations highlight circumstances under which organizations’ costs and benefits differ from the impact of their decisions on others.⁹ Plans might bear costs that could be displaced ("externalized") elsewhere (such as to families or community organizations) or might provide services for which they cannot fully recover the economic return (as with public goods such as medical research).

Community health perspective. Derived from public health and health services research, this perspective points to ways that HMOs might contribute to a robust health care system that can serve the full range of communities’ changing health needs.¹⁰ Thus, plans could share resources and expertise in addressing problems and collaborate with safety-net providers. Also included are health promotion activities aimed beyond enrolled populations.

Healthy community perspective. Derived from communitarian philosophy and social epidemiology, this perspective emphasizes fostering more-effective local decision making in allocating health resources, as well as attention to social and environmental factors that affect population health.¹¹ HMOs might work with the local nonmedical infrastructure or work to shift public priorities to address underlying causes of disease. The healthy community perspective puts primacy on the preferences of community residents.¹²

Each perspective suggests pathways through which HMOs could affect the well-being of communities. The full extent of an HMO’s community involvement can be assessed only if all dimensions suggested by the four perspectives are measured. We hypothesize that the extent of such involvement depends on characteristics of both HMOs and their environments.

Hypothesized plan influences on community benefit. Past research suggests that HMOs’ willingness to pursue community activities might be affected by their legal form of ownership (for-profit or not-for-profit); by the extent to which policies are set locally (whether the plan is a subsidiary of a multi-state corporation), which might enable the plan to capture some return from their community involvement and orient the plan to the healthy community perspective; and by the extent of their Medicaid enrollment, which might encourage programs targeted to disadvantaged groups.

HMOs’ ability to pursue community benefit activities may be facilitated by size (perhaps economies of scale exist in producing community benefits) and age (collaborations may take time to develop). Their ability to finance community benefit activities may be negatively affected by the extent of their Medicaid involvement, because low premiums may introduce constraints on resources.¹³

Hypothesized environmental influences on community benefit. HMOs’ community benefit orientations may be affected by three community characteristics. First, the greater the competitive pressures in a market, the fewer resources will be available for community activities. Second is regulation, which comes in two forms. State regulation designed to stimulate community benefits might be expected to stimulate more activity. At the time of our survey, only Massachusetts and Minnesota had community benefit provisions for HMOs. Most state regulation of plans is intended to protect individual enrollees or ensure due process for health professionals.¹⁴ We hypothesize that such regulation diverts resources away from community-oriented activities.

The third environmental influence on HMOs is external actors: the consumers and purchasers; stakeholders from local communities; and investors, state and federal officials, and corporate offices for HMOs affiliated with multiplan organizations from outside the local community. One might expect corporate offices to increase standardization across metropolitan areas and lessen HMOs’ focus on particular communities. It is unclear what influence the other doctors will have on community activities.

Measuring HMOs’ Involvement With Local Communities

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
To measure HMOs’ community benefit activities and explore the factors that encourage or inhibit community involvement, we sought data from a representative national sample of HMOs in a survey fielded during March–November 1999.

Sampling strategy. Because the survey focused on relationships between HMOs and communities, our primary sampling units were metropolitan statistical areas (MSAs). A random sample of forty MSAs from twenty-nine states was drawn, stratified by size. When appropriately weighted, our results are representative of the practices of the urban managed care industry in 1999. Within each MSA, we identified all HMOs (commercial and Medicaid) that had been in operation for at least two years and had at least 10,000 enrollees in the MSA.¹⁵ HMOs were identified from industry directories, local telephone directories, and contacts with state and local informants. The age and size criteria were to exclude HMOs that had a minimal local presence or that were too new to have had coherent community practices during the study period.

We asked about HMOs’ community-related activities in the sampled MSA. In states with more than one MSA in the sample, a single licensed entity may have operated in more than one sampled city. For survey purposes, each of these embodiments was treated as a distinct plan. One hundred eighty HMOs were eligible for inclusion in the study; of these, 112 responded to the survey, an overall response rate of 62 percent. Responding HMOs were similar to national averages in age, size, and ownership status.

Instrument design and data collection. To construct our survey instrument, we developed measures for every type of community benefit, with guidance from a national advisory panel from the managed care industry and academe.¹⁶ We divided the content into three instruments to interview the chief executive officer (CEO), medical director, and the person responsible for community affairs. (A fourth instrument was used with the twenty HMOs that supported some community activities through a separate foundation.) We did so because no one person is in a position to provide reliable information about all pertinent practices and policies.

Our primary goal was to capture the full range of HMOs’ community benefit activities. Although divided among different instruments, the policies and activities are listed in Exhibit 1 in the four perspectives listed earlier, where they were summed to provide measures of "breadth."¹⁷

Preliminary interviews and pretests suggested that sound information on resource commitments could be gathered for only seven dimensions of community benefit. We obtained data on spending for these areas from each plan (and from foundations where such entities existed).¹⁸ We refer to these variables as measures of the "depth" of community benefit commitments.¹⁹

To facilitate the analysis of factors associated with broader or deeper commitments to community benefit, the survey collected three additional types of information. The first involved HMOs’ operational definitions of their community, including whether certain activities were targeted within the MSA and on what basis. The second focused on executives’ views of a plan’s social responsibilities. The final ancillary questions were descriptive, including information about the plan, its enrollees, market conditions, and the extent of various influences on its policies and practices.

Statistical methods. To construct descriptive statistics, we weighted our raw findings to be representative of the portion of industry that operates in urban settings, based on the size characteristics of the MSAs we sampled. Because we identified more theoretically significant variables than can be estimated in models with 112 observations, we first used bivariate analyses to identify factors that appear most closely associated with community benefits.²⁰ We then used this more limited set of variables to estimate regression models to predict involvement in community-oriented activities.²¹

Study Findings

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
Attitudes. CEOs’ responses to eight prescriptive statements about HMOs’ community responsibilities are presented in Exhibit 2. Orientations favorable toward community benefits would be reflected in agreement with the first four statements and disagreement with the second four. Clearly, plan officials have disparate views.

Breadth of activity. The frequency with which HMOs reported the various forms of community benefits is presented in Exhibit 1. Each is listed under the perspective with which it is most closely connected.²² Every form was reported by at least a handful of HMOs, but none was addressed by all.

A few activities were reported by more than two-thirds of HMOs. These "consensus" forms of community benefit came from all four perspectives and include (1) health education aimed at the public, (2) support for continuing medical education for affiliated physicians regarding health problems with large social costs (such as domestic violence, mental illness and substance abuse, and unintentional injuries), (3) sharing plan data with researchers, (4) making donations to community-based organizations and activities, and (5) supporting school-based health initiatives. Other common practices include support for research and collaborations with local health care providers "involving benefits that are not limited to your enrollees."

Noteworthy patterns emerge from examination of the four perspectives in Exhibit 1. Although tax-exemption policy for health care organizations has been largely derived from the hospital context, it is still surprising that, apart from broad-based health education, community benefit activities derived from tax policy (the legal-historical perspective) were pursued by comparatively few HMOs in 1999.²³ Because this perspective captures practices that have been articulated in legal contexts as benefiting the community, it is striking both that so few plans were active in these areas and that so many plan activities have been overlooked in the legal context.

Also striking is HMOs’ apparent commitment to the healthy community perspective. This perspective’s emphasis on social determinants of health has been endorsed by the World Health Organization and many academic researchers, but the U.S. health care system has been accused of ignoring it.²⁴ Yet a large number of HMOs reported involvement with school systems, shelters, and social service agencies in activities associated with non-medical determinants of community health.

The paucity of several forms of community service is also noteworthy. Perhaps out of concern for legal liability, HMOs seemed reluctant to take actions that would benefit the community after identifying "poor clinical practices that are resulting in adverse patient outcomes." Fewer than 20 percent regularly reported these observations to state agencies.²⁵ Although 40 percent of respondents acknowledged a responsibility to protect the health care safety net, only a quarter of HMOs provided assistance to community health centers. Finally, HMOs did not appear willing to address the needs of family caregivers by adapting utilization review criteria. The breadth of community benefit activity varied widely among HMOs (data not shown). Six percent engaged in thirty or more forms of community benefit, while 12 percent were involved in ten or fewer. HMOs also varied markedly in their commitment to particular versions of community benefit.

Depth of involvement. Spending per enrollee is shown in Exhibit 3 for the seven community benefit activities for which we could measure resource commitment. Because we could not determine which activities fell within these programs, some HMOs’ spending for the activities in the lower six rows may include expenditures shown in the top row. The extent of this double counting is unknown, so our estimates should be treated as upper bounds on the true levels of spending on these particular activities. On the other hand, spending for some forms of community benefit (such as advocacy) is probably not included in any of these measures.

Perhaps the most striking feature of Exhibit 3 is the modesty of resources committed to community benefit activities. Even were there no overlap between formal community benefit programs and the six areas of spending, the average plan spent just $12 per enrollee each year. To put this in context, the typical annual premium per enrollee for employer-sponsored health insurance in 1999 was just over $2,000.²⁶ Thus, the value of spending for activities that we have measured was approximately 0.6 percent of HMOs’ revenues in 1999. But even this modest spending adds up to sizable aggregate expenditures ($720 million each year, based on roughly sixty million Americans enrolled in HMOs in 1999). And, we could not obtain data on resource commitments for many community benefit activities.

Impact Of Plan And Community Characteristics

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
Plan characteristics. Plan characteristics hypothesized to mediate provision of community benefits included ownership type, local control, enrollment of Medicaid recipients, and plan size and age. Comparing the level and form of community benefits in terms of these variables revealed several noteworthy patterns.²⁷ First, plan characteristics were rather poor at predicting the depth of involvement, as measured by spending per enrollee on the four most extensive activities: the formal community benefit program, research, donations to community agencies, and subsidized premiums. However, as expected, locally controlled HMOs had higher community-oriented spending: They spent five times more on formal community benefit programs than did HMOs that were subsidiaries of larger corporate entities. More surprisingly, smaller HMOs had three times the level of community benefit spending per enrollee, which suggests that there may be diseconomies of scale in community engagement.

Plan characteristics were strongly associated with breadth of community involvement, and these relationships were more consistently as hypothesized. Local control and nonprofit ownership were the two most consistently important factors, being associated with broader involvement under all four community benefit perspectives.²⁸ Also, as anticipated, HMOs that enrolled Medicaid recipients tended to be more oriented to charitable and redistributive activities and thus were more engaged in the legal-historical and healthy community forms of community benefit.

Community characteristics. Regarding hypothesized community characteristics that could influence HMOs’ community benefit activities, three measures stood out in our bivariate analyses: the influence of investors, the influence of local business leaders, and the existence of a community benefit reporting requirement in the state. Indeed, these external factors appear more closely related to the depth of community benefits than any plan characteristics were.²⁹

Investors’ influence. Although HMOs reporting that investors were influential were significantly more likely than other HMOs to support research, in all other dimensions investor-influenced HMOs were less likely to provide community benefits, particularly in the market failure and healthy community paradigms. Strikingly, among for-profit HMOs, plans whose executives reported strong investor influence were significantly less likely than other for-profit HMOs to provide several forms of community benefits.

Local business leaders’ influence. However, profit-oriented influence does not appear antagonistic to HMOs’ community-focused initiatives under all circumstances. The reported level of influence of local business leaders on HMOs’ boards was the single most consistent factor in predicting higher levels of spending on community benefits. (An exception was plan-funded research.) Their influence was also associated with modestly broader (but not statistically significant) activity in all four community benefit paradigms. Other sources of community influence on plan policies—the local medical community and community-based nonprofit agencies—were not associated with significant variations in community benefit activity (data not shown).

Reporting requirement. Although state requirements to report community benefit activity were not consistently associated with higher spending on community activities (only for donations was there a statistically significant relationship), HMOs in states with those requirements reported broader activity in all four forms of community benefit.³⁰ This appears to be a product of the reporting requirements themselves and not of greater state oversight per se. In states where HMOs reported that regulators had strong influence, no more activity was reported than in other states.

Relative importance. To explore the relative importance of these different influences, we incorporated these variables as explanatory factors in a set of regression models using plan executives’ attitudes toward community involvement and our breadth and depth measures as the dependent variables. Space constraints prevent our reporting the results in detail here.³¹ The two factors that most consistently predicted favorable attitudes toward community benefit were a state reporting requirement and the influence of local business leaders on the board. Regarding community benefit spending, higher spending was predicted by local leaders’ influence, small plan size, and operation in a state with reporting requirements. But no single factor was consistently associated with significantly higher levels of spending. Regression results using our four measures of breadth of involvement were more consistent. Nonprofit ownership and location in a state with reporting requirements were associated with greater activity under all four perspectives. Medicaid participation was associated with engagement in the legal-historical and healthy community perspectives. Influence by local business leaders was again associated with greater activity levels, again emphasizing those two perspectives.

A Complex Picture

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
Our data reveal a complex picture of relationships between HMOs and communities in 1999. On the one hand, a sizable majority of plan executives acknowledged responsibility for community service. HMOs of all sizes, ages, and forms of ownership engaged in community benefit activities. Substantial percentages of HMOs engaged in one or more activities from all four community benefit paradigms. Even the most esoteric forms of community engagement were found in at least a handful of HMOs. On the other hand, the total resources committed to community benefits appear to have been modest. Nor was the commitment to community engagement universal—a minority of HMOs acknowledged no particular responsibility to the communities they served. Among plans that did acknowledge a community obligation, the expression of that commitment varied widely and might have reflected multiple motives—partially to benefit disadvantaged residents, partially to promote future plan growth.

Caveats. Our findings and conclusions need to be considered in light of some methodological caveats. Because this information proved to be time-consuming to collect, we could obtain data from only about a hundred HMOs. Although we believe that our sample is representative of the industry in urban areas, its modest size might have led us to miss certain characteristics of HMOs or their environments that could have influenced community engagement.³²

A second concern is that the data were self-reported. Respondents might have felt pressure to overstate their activities. Conversely, some activities might have gone unreported because respondents were not aware of them. The validity of the survey measures merits exploration in future research.

Third, we could obtain quantifiable measures of commitment for only a subset of the forty-five measures in the survey. Thus, our estimates of spending on community benefit (Exhibit 3) almost certainly underestimate the resources that HMOs committed to their communities in 1999. Conversely, some reported activities may have been perfunctory or symbolic.

Finally, the data were collected five years ago in an industry that has been rapidly changing. Since most of these changes involve relationships between health plans and physicians (for example, utilization review methods, financial incentives) we do not expect them to have affected plans’ connections to their communities.

Salient implications. Despite these limitations, this study provides the first representative portrait of HMOs’ community benefit activities and a base upon which advocates of greater community engagement can build. In this regard, two implications merit further discussion.

Need to encourage HMO–community connections. First, the greatly expanded regulation of managed care has focused almost exclusively on HMOs’ relationships with individuals (consumers and providers) and does little to encourage links between HMOs and communities. In our judgment, this focus is shortsighted. Many of the most important determinants of population health are unrelated to clinical medicine. Although HMOs cannot be responsible for remolding society in a healthier image, nonetheless they can (and some do) pursue initiatives that use social determinants of health as starting points.

Our findings suggest that HMOs can do more to address the health needs of their communities. If HMOs’ spending on community benefits rose from its current average to the level of the plans in the upper 5 percent of the spending distribution ($28 per year, on average), one would see an increase of 125 percent in spending on community benefit activities (in aggregate, more than $1 billion per year). Plan performance from the two states with community benefit reporting laws suggests that these sorts of regulations are effective at elevating community benefit activities among plans’ priorities and in broadening the scope of HMOs’ community benefit involvement.

If policymakers were to take more seriously plans’ potential to affect the communities they serve, they would also need to rethink the legal concept of community benefit in the managed care context. The conventional ways of measuring community benefit—captured in our study by the legal-historical perspective—are simply too narrow. Of the eleven forms of community engagement that were reported by more than half of HMOs, only one came from the legal-historical perspective. Clearly a more comprehensive view of community benefit under managed care is needed.

Need to be aware of impact of HMO industry changes. A second concern is that in ignoring community benefit, policymakers have been insensitive to changes in the managed care industry that could undermine the current impetus to engage with communities. Our findings suggest that an industry of locally controlled, stable, nonprofit HMOs will perform differently in community benefit terms than would an industry of HMOs operated by for-profit corporations, with limited local autonomy, strong pressures from investors, and turbulent ownership. This latter scenario combines several factors that our analyses have shown to have independent negative consequences for community involvement. Yet it characterizes today’s managed care industry.

The growth of investor ownership was not happenstance, nor some inevitable consequence of an evolving industry. Rather, it was a conscious public policy goal in the early 1980s, and such policies are still being pursued, with little awareness of their consequences.³³ In our assessment, these policies ought to be reconsidered, in light of their consequences for community health and well-being.

A provisional bottom line. A skeptic might wonder, "How much does this all matter to the health of the average American?" In the absence of evaluations of the effectiveness of HMOs’ community activities, there can be no concrete answer. But the issues that are addressed through HMOs’ initiatives are as consequential as some issues that have been aggressively addressed by states in regulating HMOs.³⁴

Also, although HMOs’ investments in community benefit activities in 1999 were quite modest relative to their spending on medical care, these investments came with very little governmental encouragement. Were community benefits to be embraced as an appropriate focus of health policy, the scope of community engagements and the resources HMOs devote to them would likely grow. If more states adopted policies that support community benefit, HMOs’ community involvement would likely be of greater consequence. Should this happen, HMOs could more fully live up to their name, not only for their enrollees but for everyone in their communities.

Editor's Notes

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 
Mark Schlesinger (mark.schlesinger{at}yale.edu) is an associate professor at Yale and Rutgers Universities. Brad Gray is director of the Division of Health and Science Policy at the New York Academy of Medicine. Michael Gusmano is associate director of the World Cities Project, International Longevity Center–USA, in New York City. Parts 1 and 2 of this series of papers were published in 1998.

This project was supported by a grant from the Robert Wood Johnson Foundation. The authors thank Gerard Carrino, Mary Duncan, Vincent Antonelli, and Jennifer Stuber for research assistance. The instruments used in the study were developed with the assistance of an advisory panel that included Howard Bailit, Charles Bell, Marsha Gold, Paul Halverson, Michael Herbert, Anthony Kovner, Roz Lasker, Sheila Leatherman, Susan Pauker, Mary Pittman, J. David Seay, Cary Sennett, Soshanna Sofaer, T.J. Sullivan, and Dan Wolfson.

NOTES

Top Editor's Notes Pathways To Community... Measuring HMOs’... Study Findings Impact Of Plan And... A Complex Picture NOTES

 

1. See, for example, R. Bowser and L. Gostin, "Managed Care and the Health of a Nation," Southern California Law Review 72, no. 5 (1999): 1209–1295;[Web of Science][Medline] and D. Kindig, Purchasing Population Health: Paying for Results (Ann Arbor: University of Michigan Press, 1997).
2. See, for example, D.M. Lawrence, P.H. Mattingly, and J.M. Ludden, "Trusting in the Future: The Distinct Advantage of Nonprofit HMOs," Milbank Quarterly 75, no. 1 (1997): 5–10;[CrossRef][Web of Science][Medline] and Coalition for Nonprofit Health Care, State Law Approaches to Ensuring the Social Accountability of Nonprofit Health Care Organizations (Washington: CNHC, 1999).
3. See, for example, G.P. Mays et al., "How Managed Care Plans Contribute to Public Health Practice," Inquiry 37, no. 4 (2000/2001): 389–410; and P. Halverson et al., "Not-So-Strange Bedfellows: Models of Interaction between Managed Care Plans and Public Health Agencies," Milbank Quarterly 75, no. 1 (1997): 113–138.[CrossRef][Web of Science][Medline]
4. National Center for Chronic Disease Prevention and Health Promotion, "Special Focus: Public Health and Managed Care," Chronic Disease Notes and Reports 10, no. 1 (1997): 1–37.
5. GlennMays and colleagues assess factors that influence community involvement but focus only on collaborations between plans and local public health departments. See Mays et al., "How Managed Care Plans Contribute to Public Health Practice."
6. M. Schlesinger and B. Gray, "A Broader Vision for Managed Care, Part 1: Measuring the Benefit to Communities," Health Affairs 17, no. 3 (1998): 152–168;[Abstract] and M. Schlesinger et al., "A Broader Vision for Managed Care, Part 2: A Typology of Community Benefits," Health Affairs 17, no. 5 (1998): 26–49.[Abstract]
7. D. Fox and D. Schaffer, "Tax Administration as Health Policy: Hospitals, the Internal Revenue Service, and the Courts," Journal of Health Politics, Policy and Law 16, no. 2 (1991): 251–280.[Web of Science][Medline]
8. M. Schlesinger, B. Gray, and E. Bradley, "Charity and Community: The Role of Nonprofit Ownership in a Managed Health Care System," Journal of Health Politics, Policy and Law 21, no. 4 (1996): 697–752.[Web of Science][Medline]
9. Ibid.; M.V. Pauly, "Health Systems Ownership: Can Regulation Preserve Community Benefits?" Frontiers of Health Services Management 12, no. 3 (1996): 3–34;[Medline] and C. Madden and A. Katz, Community Benefits and Not-for-Profit Health Care: Policy Issues and Perspectives, Report prepared for the Catholic Health Association (Seattle: University of Washington School of Public Health and Community Medicine, 1995).
10. Kindig, Purchasing Population Health; and J. Showstack et al., "Health of the Public: The Private-Sector Challenge," Journal of the American Medical Association 276, no. 13 (1996): 1071–1074.[Abstract/Free Full Text]
11. D.G.Whiteis, "Unhealthy Cities: Corporate Medicine, Community Economic Underdevelopment, and Public Health," International Journal of Health Services 27, no. 2 (1997): 227–242; [Medline] J. McKnight, The Careless Society: Community and Its Counterfeits (New York: Basic Books, 1995); and M. Marmot and R.G. Wilkinson, Social Determinants of Health (New York: Oxford University Press, 1999).
12. J. Lomas, "Reluctant Rationers: Public Input to Health Care Priorities," Journal of Health Services Research and Policy 2, no. 2 (1997): 103–111.
13. M.S. Sparer and L.D. Brown, "Uneasy Alliances: Managed Care Plans Formed by Safety-Net Providers," Health Affairs 19, no. 4 (2000): 23–35;[Abstract] and B.H. Gray and C. Rowe, "Safety-Net Health Plans: A Status Report," Health Affairs 19, no. 1 (2000): 185–193.[Abstract]
14. T.E. Miller, "Managed Care Regulation: In the Laboratory of the States," Journal of the American Medical Association 278, no. 13 (1997): 1102–1109.[Abstract/Free Full Text]
15. Medicaid-only plans were identified separately because they are not always licensed as HMOs by states. See Sparer and Brown, "Uneasy Alliances"; and Gray and Rowe, "Safety-Net Health Plans." For both commercial HMOs and Medic-aid-only plans, we contacted plans identified as operating in sample MSAs and asking membership offices for the number of enrollees living in the MSA’s counties. Since they sometimes supplied rough estimates, we erred on the side of including plans whose estimated local enrollment was near the 10,000-member threshold.
16. See acknowledgments for a list of members.
17. These indices, which were simple sums, reflect the four perspectives for community benefit derived from the legal, medical, and health services literatures. They are conceptual in origin. Thus, we did not necessarily expect the indices to have the internal consistency of typical psychometric indices. As discussed in our earlier papers, the legal-historical perspective is especially likely to be inconsistently embodied in plans’ practices because it evolved piecemeal through a disjointed legal decisions, rather than as the result of coherent assessment of plans’ impact on communities. That expectation was borne out in the properties of the index. The internal coherence of the legal-historical index, as measured by Cronbach’s alpha, is only 0.25 (a low score on a measure that ranges from zero to one, with higher values reflecting greater intercorrelation among the component measures). By contrast, the alpha statistics for the market failure, community health, and healthy community indices are 0.60, 0.65, and 0.73, respectively.
18. Some plans use closely linked foundations as a conduit for spending on community activities. Our interviews indicated that these operated largely under the direction of plan executives and were funded primarily through resources transferred from the plan. Only one foundation had been established in the ownership conversion of the related HMO. Although its spending could arguably be excluded, it represents a tiny part of the overall activities reported here. Excluding this foundation’s data would not materially alter the results. When a foundation’s geographic purview covered multiple MSAs, we asked respondents for expenditures targeted to the MSA being studied.
19. The linkage between resource commitments and socially valued outcomes is more direct for some activities than for others. For example, the benefits produced by an immunization program can be measured reasonably well by the number of doses of vaccine purchased and distributed. By contrast, the impact of a plan’s environmental advocacy or technical support for safety-net agencies cannot be readily assessed by the inputs to these activities.
20. We included only those that were significant at the .10 level. Excluded as a result was our measure of intensity of competition. The explanatory variables in the regressions included five HMO characteristics (size, age, Medicaid enrollment, nonprofit ownership, and local control) and three environmental characteristics (influence by investors, influence by local business interests, and presence of state reporting law). At a reviewer’s request, we also included the HMO’s model type—a dichotomous variable of whether it was a group/staff model (twenty-seven of our plans had this structure). Model type was not consistently related to community benefit activities in either bivariate relationships or multivariate models. All variables were converted into dichotomous variables for analysis: (1) 37 percent of the HMOs had been in operation for five years or fewer in the MSAs under study, (2) 44 percent had some Medicaid enrollees, (3) 59 percent reported that local business leaders had strong influence over local plan policies, (4) 32 percent operated under nonprofit ownership, (5) 73 percent were subsidiaries of a larger corporation, (6) 7 percent operated in states with community benefit reporting laws, and (7) 20 percent were "small plans" (10,000–25,000 enrollees in the MSA under study). If sample size had been larger, it would have been desirable to use a hierarchical model to control for local effects.
21. Because our dependent variables varied in their form, we used two techniques to estimate the models. Attitudes toward community benefit activities were categorical. When these are the dependent variables, the regressions were estimated using ordered logistic techniques. Our measures of depth and breadth are continuous; when these are the dependent variables, the regressions are estimated using ordinary least square techniques.
22. Where an activity falls under more than one perspective, we assigned it to the perspective with which it is most closely connected in the literature.
23. Schlesinger et al., "A Broader Vision, Part 2," 29–31.
24. See, for example, Marmot and Wilkinson, Social Determinants of Health; and R.G. Evans, M.L. Barer, and T.R. Marmor, eds., Why Are Some People Healthy and Others Not? The Determinants of Health of Populations (New York: Aldine De Gruyter, 1994).
25. This reluctance to respond more aggressively to observed patterns of substandard treatment has been documented in other segments of the managed care industry. See M. Schlesinger, B. Gray, and K. Perreira, "Medical Professionalism under Managed Care: The Pros and Cons of Utilization Review," Health Affairs 16, no. 1 (1997): 106–24.[Abstract]
26. Henry J. Kaiser Family Foundation/Health Research and Education Trust, Employer Health Benefits: 1999 Annual Survey (Menlo Park, Calif.: KFF/ HRET, 1999), 22.
27. Space limitations preclude presenting the complete findings here. They are available online as a supplemental exhibit, at content.healthaffairs.org/cgi/content/full/23/3/210/DC1.
28. For additional analysis, see M. Schlesinger, S. Mitchell, and B. Gray, "Measuring Community Benefits Provided by Nonprofit and For-Profit HMOs," Inquiry 40, no. 2 (2003): 114–132.[Web of Science][Medline]
29. See supplemental exhibit online, as described in Note 27.
30. This could be a reporting artifact. Plans required by a state to report on their community benefit activities might have documented them more carefully and thus reported them in our survey more completely than did plans from states with no legal reporting requirement.
31. Results are available from Mark Schlesinger, mark.schlesinger{at}yale.edu.
32. The small sample size also precludes us from exploring one methodological issue arising from the structure of the industry. The inclusion of MSAs in the same states meant that multiple plans from the same organization were included in some states in the sample, perhaps double counting the responses of those plans if they followed the same policies everywhere. We would like to have performed sensitivity analysis to check on whether this affected our results, but the numbers were too small. We did, however, find that for HMOs that had multiple plans within states, there were significant differences in depth of community benefit activities but not in breadth. Ultimately, in research that focuses on policies and practices at the local level, there may be no sampling strategy that satisfactorily reconciles geography with plans’ organizational structure.
33. Schlesinger et al., "Charity and Community."
34. A.A. Noble and T. Brennan, "The Stages of Managed Care Regulation: Developing Better Rules," Journal of Health Politics, Policy and Law 24, no. 6 (1999): 1275–1305.[Web of Science][Medline]

                      What's this?

This article has been cited by other articles:

				M. Schlesinger and B. H. Gray How Nonprofits Matter In American Medicine, And What To Do About It Health Aff., July 1, 2006; 25(4): W287 - W303. [Abstract] [Full Text] [PDF]