P R O F I T S T A T U S :
H O W N O N P R O F I T S M A T T E R
20 June 2006
How Nonprofits Matter In
American Medicine,
And What To Do About It
Reports of the demise of nonprofits in U.S. health care are premature.
by Mark Schlesinger and Bradford H. Gray
ABSTRACT:
Skeptics question nonprofit health care on the grounds that nonprofits fail to distinguish themselves from their for-profit counterparts and do not reliably provide community benefits commensurate with their tax subsidies. Drawing on the most recent and comprehensive evidence, we assess these charges, judging them to be either wrong or incomplete. Although conventional critiques are therefore unconvincing, there are nonetheless important challenges facing the nonprofit sector in American medicine. To address these, we propose reformulating ownership-related policies to define both the appropriate forms of community benefit and the appropriate mix of ownership in terms of local markets and communities. [Health Affairs 25 (2006): w287–w303 (published online 20 June 2006; 10.1377/hlthaff.25.w287)]
The legitimacy and favored tax treatment of nonprofit medical care organizations have come under fire from politicians and academics, as for-profit care has expanded for some services (Exhibit 1). Some policymakers doubt that nonprofits reliably contribute community benefits commensurate with the value of their tax exemptions as charitable organizations. Sen. Chuck Grassley (R-IA), for example, justified his recent “investigation” of ten nonprofit hospitals by stating that “tax-exempt status is a privilege. Unfortunately, some charities abuse that privilege.”1 Some academics question whether nonprofit ownership matters. As Jill Horwitz notes, “Many scholars claim that the diversity of corporate form is essentially a fiction. …While the particular arguments vary, the message is simple. The not-for-profit form does not matter for the public good or, in many cases, matter at all.”2
View larger version
[in this window]
[in a new window]
We focus in this paper on two skeptical readings of the evidence about nonprofit medical care. First, because findings from empirical comparisons of nonprofit and for-profit performance are seen as “mixed and inconsistent,” many scholars argue that for-profits and nonprofits are “similar health services organizations.”3 Second, because organizational practices vary so much among non-profits, there are concerns that some organizations should be held more accountable or stripped of their tax benefits. (Elsewhere we address a third charge, that nonprofits have lost public legitimacy.)4
These criticisms are repeated so frequently in academic and policy discourse that one might presume their accuracy. But each distorts the evidence about ownership-related differences, perhaps fostering counterproductive policy responses. After reviewing the evidence, we conclude that although the critiques of nonprofit health care contain elements of truth, each is misleading.
We seek here to clarify the real issues and offer new policy approaches that (1) separate questions regarding appropriate tax exemption policy and optimal ownership mix, (2) develop a process to more sensibly account for “community benefit” activities, and (3) promote a beneficial mix of nonprofit and for-profit ownership at the community level.
Does Ownership Matter? Variation In Nonprofit And For-Profit Health Care
As distinct legal forms, nonprofit and for-profit ownership lead to different mixes of monetary and nonpecuniary incentives for administrators and staffs, different sources of capital, and different influences on governance. Whether these differences translate into distinctive behavior has been examined in about 275 empirical studies covering hospital care, psychiatric services, nursing home care, home health care, treatment of end-stage renal disease (ESRD), hospice care, rehabilitative services, managed care plans, preventive examinations, and various ambulatory services. For some services, for-profit ownership is dominant; in others, nonprofits predominate (Exhibit 1). Studies have examined many attributes of services: cost, quality, accessibility for indigent clients, trustworthiness of organizational practices, pricing policies, and stability of service provision over time.
Supporters and critics of nonprofit health care agree that ownership-related differences regarding cost, quality, and accessibility vary greatly across studies. For critics, variation suggests that ownership can’t count for much if it does not predict consistent differences between nonprofit and for-profit practices. As John Colombo testified in a recent congressional hearing, “Empirical studies on quality of care, costs of care, and free care for the poor show decidedly mixed results…. All we can conclude is that nonprofits in some markets in some measures outperform for-profits, and that in other markets on other measures, for-profits outperform nonprofits.”5
This interpretation overlooks the possibility that legal form might interact (in systematic, not random ways) with other factors to affect organizational performance on activities that promote the public good. Differences might be expected between services that are well insured (such as hospice care) and services for which millions of patients lack adequate coverage (such as hospital care), between services whose benefits go beyond patients and services that help only patients or their families, and between poor and wealthy communities.
One rationale for the nonprofit sector involves its capacity to respond more flexibly than can government to variation in the types of benefits that communities value.6 Precisely because the public goods that are valued vary across services, among communities, and over time, a responsive nonprofit sector will produce corresponding variations in performance. It is therefore crucial to understand the nature of these variations. We examine the evidence in the sections that follow.
Variation across services. Much apparent inconsistency in the effects of ownership emerges when scholars carelessly combine findings based on different health services or performance measures. By contrast, meta-analyses that aggregate studies involving a single service and a single well-defined outcome find consistent ownership-related differences. Thus, there are higher mortality rates in for-profit hospitals and renal dialysis facilities.7 Higher prices are found in for-profit hospitals.8 There are higher rates of adverse events in for-profit nursing homes.9 And larger barriers to access for indigent patients are found in for-profit psychiatric facilities.10
Nonetheless, many ownership-related differences vary greatly across services. We illustrate with research comparing nonprofit and for-profit hospitals and nursing homes in terms of economic performance, quality of care, and accessibility for indigent patients. Exhibit 2 summarizes results from 162 studies that use sophisticated methods (multivariate models or matched samples) to account for confounding factors. Because some studies report multiple outcomes, we have 215 distinct comparisons, grouped by type of outcome, type of service (hospitals or nursing homes), and whether the analyses indicate a statistically significant advantage to nonprofit or for-profit providers.11
View larger version
[in this window]
[in a new window]
The impact of ownership differs strikingly between hospitals and nursing homes. Consider first costs and efficiency (that is, the provision of services in a manner that minimizes expenditure of resources; see Exhibit 2 Notes). There is overwhelming evidence that for-profit nursing homes have lower costs and greater efficiency: Twenty studies support this conclusion; one other found no statistically significant difference. In the eight most sophisticated comparisons of technical efficiency, seven found for-profits to be significantly more efficient. Among hospitals, however, costs and efficiency results are more mixed but predominantly favor nonprofits. Among the most sophisticated models of technical efficiency, five found greater efficiency among nonprofits, three found no statistically significant differences, and three found for-profits were more efficient. Consequently, although it is difficult to answer conclusively whether or not ownership matters for hospital costs, its effects are clearly different for hospitals than nursing homes.
Differences are equally striking for some dimensions of accessibility and quality of care. For example, although ownership seems unrelated to the propensity to treat Medicaid patients in hospitals, for-profit nursing homes are the most likely to treat Medicaid recipients. Ownership-related differences in adverse health events within facilities are far more pronounced in nursing homes than hospitals, particularly when inpatient mortality is included as an adverse event.
The pattern illustrated by our comparison of hospitals and nursing homes is a general one. Our literature review found some differences between nonprofits and for-profits regarding cost, quality, or accessibility for every service studied. However, the effects of ownership manifest in different ways for different services.
Four other attributes are related to ownership in a more consistent manner. First, for-profit organizations more aggressively mark up prices over costs and otherwise maximize revenue. This pattern has been documented among community general hospitals, nursing homes, psychiatric hospitals, drug treatment centers, rehabilitation facilities, and health plans.12 Second, nonprofit organizations appear more trustworthy in delivering services, being less likely to make misleading claims, to have complaints lodged against them by patients, and to treat vulnerable patients differently from other clientele.13 Third, nonprofits are typically the incubators of innovation (for example, health maintenance organizations, or HMOs, during the 1930s or hospice three decades ago), using philanthropy and cross-subsidies to finance the development of services for which there is not yet a market.14 Fourth, nonprofit health care providers are slower to react to change, expanding capacity less quickly when demand rises and dropping services or withdrawing from markets less frequently when profitability declines.15
The first three consistent differences stand as advantages for nonprofits. In a health care system beset by increasing costs and ill-informed consumers, marking up prices and exploiting vulnerable patients are hardly desirable attributes. Pioneering new services is a vital societal function, notwithstanding concerns that innovation might further inflate costs.
The normative implications of the dynamic differences between nonprofit and for-profit behavior are less clear. Rapid adaptation to changing conditions can be an important societal benefit, when health needs change or health policies are revised. For-profit providers have a clear edge in these circumstances.16 But in other circumstances, rapid response seems a liability. Providers that constantly alter their service mix or market areas can disrupt vital relationships between patients and providers, and changing insurer practices can undermine patients’ financial security. Recent experience with private health plans in the Medicare+Choice program illustrates such concerns.17 Frequent plan withdrawals and unstable benefits—both more pronounced among for-profit plans—left millions of seniors confused, without medical care, and with uncovered expenses.
Variation across studies. Consider again Exhibit 2. Even for outcomes that show consistent ownership-related differences (for example, quality problems in nursing homes), there is variation across studies. Some of this is a consequence of using statistical significance to identify “meaningful” differences in performance. Even if ownership is related to quality, a certain number of studies will find statistically insignificant results because of imprecisely measured outcomes. At conventional levels of statistical significance, one might expect to find almost exactly the number of studies as in Exhibit 2 that do not show a quality advantage for nonprofit homes.
Some of the remaining variation reflects methodological considerations. Studies of hospital mortality with smaller sample sizes (fewer than a million cases) typically lack the statistical power to detect ownership-related differences.18 Also, comparisons of adverse events in nursing homes fail to detect ownership-related differences when they do not account for case-mix variation among homes.19 Studies also vary in the organizational characteristics that are included in the statistical models: for example, teaching, multifacility system, or religious affiliation. Although ownership-related differences remain statistically significant whether or not one controls for these other attributes, they influence the magnitude of the differences.20
But some variation in findings cannot be attributed to problems of methods or measurement. For example, studies of adverse treatment events in hospitals, or costs for hospital admissions, include multiple high-quality studies that identify nonprofit hospitals as more effective, but also include multiple studies of equal quality that find either no difference in performance or a significant advantage for for-profit facilities. Are these conflicting results the result of subtle methodological differences, or might comparative performance actually vary this much?
We believe that ownership-related differences do vary greatly, responding to the context in which health care is delivered. Some studies compare organizations operating under benign conditions, others in less lucrative markets. If in poor financial health, even public-spirited organizations have limited capacity to generate community benefits.21 This helps explain why studies comparing organizations before and after their conversion from nonprofit to for-profit ownership often find only small differences in accessibility or quality; nonprofits that convert are typically financially stressed, prior to their sale.
The real challenge. The real challenge here is understanding how context affects ownership-related differences. Evidence of these contextual effects has led some skeptics to dismiss nonprofit health care as an anachronism, no longer compatible with a market-driven health care system dominated by large corporations. This seems quite intuitive—if market pressures and corporate hierarchies constrain facility behavior, how much can ownership actually affect cost, quality, or accessibility of care?
The answer, surprisingly, is “quite a bit.” Growing competition and affiliation with multi-unit systems have not reduced ownership-related differences in performance.22 In fact, the gap between nonprofit and for-profit hospitals in the provision of uncompensated care appears to grow as markets become more competitive.23 Ownership-related differences in accessibility, quality, and trustworthiness are larger among independent than system-affiliated providers.24 The major institutional transformations of American medicine over the past few decades seem not to have vitiated the distinctiveness of nonprofit ownership.
Does Ownership Matter Enough? Accountability And Reliability In Nonprofit Health Care
The performance of nonprofits has also been evaluated in relationship to the tax advantages afforded nonprofit enterprise. Policymakers in several states have been skeptical about this, implementing laws or regulations to hold nonprofits more accountable.
Variation in the forms of community benefit. Assessing the full impact that health care organizations have on communities is difficult, because not all community-benefit activities are readily measurable. Some forms draw policymakers’ attention more than others. Caring for indigent patients falls into both categories. One can readily count numbers of uninsured patients (although not all lack the means to pay) or dollars of uncompensated care (although whether to include “bad debt” remains controversial). Before 1969, when the Internal Revenue Service (IRS) adopted the community-benefit standard, indigent care was the primary criterion for federal tax exemption.
Under this criterion, the performance of nonprofit health care appears far from adequate. For nursing homes and health plans, nonprofit ownership is not consistently associated with any propensity to treat low-income patients.25 Even in many hospitals, performance could not in itself justify tax exemptions. If one does not count bad debt, the amount of uncompensated care provided by as many as three-quarters of nonprofit hospitals is less than their tax benefits.26 (Although performance varies across states, even in the most charitable jurisdictions some 20–40 percent of nonprofit hospitals fail to cover the value of their tax benefits.)27
However, indigent care is not the only form of community benefit or charitable activity in health care. For example, a recent study found that nonprofit health plans were significantly more likely than for-profits to support safety-net providers and contribute to community health initiatives that benefit the poor.28
More generally, this study found that although nonprofit plans did not differ from for-profit plans in provision of free or subsidized services, nonprofits were significantly more involved in three other domains of community benefit. A small body of research suggests that nonprofits provide many more of these diverse forms of community benefit.29 One study found that the nonprofit hospitals that were the least involved in free or subsidized treatment were the most engaged in other forms of community benefit.30 Comparable research, however, is not available for other health services, so we cannot now determine whether variation in the form of community benefits is generally consistent with variation in community needs.