In one way or another, the reform of American health insurance has been a leading political issue for much of this decade. Five years ago, the critical question was whether President Clinton's proposal for universal coverage through "managed competition" would be enacted (Hacker 1997). Today, the debate focuses on the quality of health insurance for those who have it. The rhetorical centerpiece is no longer "managed competition." It is "managed care"-a blanket expression denoting a mix of changes in private insurance that many Americans view with anxiety. And the question now preoccupying health policy analysts, as this special issue indicates, is how to make sense of the seeming political backlash against these developments.

The premise of our commentary is that this question cannot be answered as currently formulated. The term managed care, much like that ubiquitous reform phrase of the early 1990s, managed competition, is a confused assemblage of market sloganeering, aspirational rhetoric, and managerial jargot that sadly reflects the more general state of discourse about American medical institutions. Because managed care is an incoherent subject, most claims about it suffer from incoherence as well. Moreover, to incorporate "managed care" and other similar terms into health policy research is, in effect, to presuppose answers to many of the most pressing questions raised by the recent evolution of medical care in the United States.

Hence our reflections are of two sorts. The first part of our commentary briefly discusses the context in which marketing slogans about medical care have emerged. The second turns to analysis of the term managed care in particular and tries to separate out the diverse trends that it is meant to capture. Our main argument is that scholars should shun industry-promoted slogans and instead develop more precise and neutral conceptual tools with which to evaluate specific changes in reimbursement methods, managerial techniques, and organizational forms.

Medical Care and the Rise of "Corporatespeak"

The management discussion of many of the major topics in modrn medical care is marked by fads, sloppiness, and confusion. Marketing hyperbole and managerial jargon dominate contemporary reflections on topics like the management of care and its costs, quality, and organization.

Health policy audiences will be familiar with some of the shifting fashions in managerial commentary. Once, management by objective (MB0) and zero based budgeting (ZBO) were all the rage. In recent years, "corporatespeak" shifted to such expressions as total quality management (TQM), integrated delivery systems (IDS), and, in the case of this issue's focus, managed care. For a time, big was better. Politicians as well as managers embraced larger scale operations. Then, within a few years, small was beautiful. Divestiture, devolution, decentralization, and specialization suddenly became the watchwords of managerial correctness. The favored relations among managers and employees have ranged from simple hierarchies with strict divisions of labor to cooperative teams, from models emphasizing adversarial combat to those featuring bonding mechanisms. Within these broader notions of organizational design, a dizzying array of techniques ranging from "just in time" inventory management to statistical quality assurance, have been offered as catch-all solutions to managerial malaise. In contemporary discussions of quality in medicine, the much-heralded technical panaceas include "outcomes measurement," "integration," "coordination," and "evidence-based medicine."

Expressions like these are all slogans-persuasively defined terms that imply success by their very formulation. Note, for example, that we do not hear of "unmanaged care," "disintegrated delivery systems," or "non-evidence-based medicine." The absence of such categories illustrates the extent to which terms of this sort are idealizations, rather than accurate descriptions. Yet these persuasive definitions carry with them real truth-claims and normative connotations. And because they do, they have the potential to shape our perceptions not merely of the desirability, but of the very character of the organizational realities to which we apply them.

Of course, the claims and connotations are not always positive. With the emergence of public concern about recent changes in American medical care, "managed care" has mutated from a term of approval into one of opprobrium. The danger to coherent thought, however, is the same in either case. The categories that we use to understand organizational change should not prejudge its desirability, nor should they reflect uncritically the allegations of its critics and defenders. They should tell us about the structure and behavior of an organization, not whether it is good or bad, successful or unsuccessful, benevolent or sinister. Precisely because much of the language used to describe American medicine today is meant to convince rather than explain, even thoughtful observers often end up endorsing claims whose validity they should be assessing.

The Managed Care Example

Our argument is straightforward: By adopting the marketing jargon of corporate medical care, analysts risk adding credence to the claims and associations that come with it. Yet we also wish to emphasize an additional risk posed by unreflective reliance on persuasive definitions like "managed care"-namely, that scholars will fail to understand the developments that they seek to explain. For not only do these slogans embody often questionable claims; they also represent poor conceptual tools for identifying and explaining what is distinctive about recent organizational changes. Nothing illustrates this better than the term "managed care."

Although the exact provenance of "managed care" is uncertain, the term came into widespread usage only in the past decade. The expression does not appear once, for example, in Paul Starr's exhaustive 1982 history, The Social Transformation of American Medicine. Nor can it be found in other prominent books on U.S. health policy written before the early 1980s, including Lawrence Brown's classic 1983 work on the health maintenance organization (HMO) legislation of 1973, Politics and Health Care Organizations. The phrase first appeared in the New York Times in 1985 but surfaced in only a handful of articles during the decade. In the l990s, however, Times articles mentioning the phrase exploded, increasing from 27 in 1990 to 287 in 1994 to 587 in 1998. Because "managed care" has become a household term, it is difficult to recognize how recently it entered American discourse.

What exactly managed care is has never been entirely clear, however, even among its strongest proponents. To some, the crucial distinguishing feature is a shift in financing from indemnity-style fee-for-service reimbursement, in which the insurer is little more than a bill payer, to capitated payment. Yet there is nothing intrinsic to fee-for-service payment that requires that reimbursement be open-ended or insurers passive, and many, if not most, health insurance plans labeled "managed care" do nor rely primarily on capitation. To others, the distinctive characteristic is the creation of administrative protocols for reviewing and sometimes denying care demanded by patients or medical professionals. But such microlevel managerial controls are not universal among so-called managed care health plans either, and in fact may be obviated by payment methods, such as capitation or regulated fee-for-service reimbursement, that create more diffuse constraints on medical practice. Finally, to some, what distinguishes managed care is the establishment of integrated networks of health professionals from which patients are required to obtain care. Yet some so-called managed care plans have no such networks, and what is called a network by many plans is little more than a list of providers willing to accept discounted fee-for-service payments. That hardly represents the dense "integration" celebrated by managed care enthusiasts.

Perhaps the most defensible interpretation of "managed care" is that it represents a fusion of two functions once seen as separate: the financing of medical care and the delivery of medical services. This, at least, provides a reasonably accurate description of the most familiar organizational entity that marched under the managed care banner in the early 1980s-the HMO. When the majority of health insurers used fee-for-service payments and placed few restrictions on patient or provider discretion, it was at least possible to identify a small subset of health plans that existed outside this insurance mainstream, however poorly the expression "managed care" described such plans. Today, however, that is decidedly no longer the case. Only 2 percent of private health plans in 1997 conformed to the traditional model of fee-for-service indemnity insurance. Another 16 percent used fee-for-service payment but employed some form of utilization review, such as precertification (HIAA 1997). Thus between 80 and 98 percent of today's private health insurers appear to fall into the general category of managed care. The category does not, in other words, offer any guidance as to how to distinguish among the vast majority of contemporary health plans.

The standard response to this problem has been to subdivide the managed care universe into a collage of competing acronyms: HMOs, preferred provider organizations (PPOs), exclusive provider organizations (EPOs), and the like. This is the approach taken by Jonathan Weiner and Gregory de Lissovoy in their oft-cited 1993 article "Razing a Tower of Babel: A Taxonomy for Managed Care and Health Insurance Plans," which represents perhaps the best recent explication of the conventional method of categorization. Weiner and de Lissovoy argue that what usually distinguishesmanaged care plans from (plans) that are more traditional is that there is a party that takes responsibility for integrating and coordinating the financing and delivery of services across what previously were fragmented provider and payer entities" (Weiner and de Lissovoy 1993: 78). They then proceed to divide this broad category into five mutually exclusive types of managed care plans: fee-for-service plans with utilization review (what they call "managed indemnity plans" (MIPs)), PPOs, EPOs, open-ended HMOs (O/HMOs), and regular HMOs. Although Weiner and de Lissovoy propose a fairly complicated scheme for distinguishing among these five plan types (reproduced in Table 1), the crucial distinguishing features are twofold: (1) whether plans require their patients see certain specified medical providers (EPOs and regular HMOs do, MIPs do not, and PPOs and O/HMOs do but penalize patients who receive care from providers outside the network), and (2) whether physicians bear financial risk (only in HMOs do they do so, Weiner and de Lissovoy argue, because HMOs rely on capitation). With the exception of MIPs, Weiner and de Lissovoy dub all these plans "integrated delivery systems."

Categorization and Confusion

Weiner and de Lissovoy's taxonomy, if nothing else, conforms to popular usage. It introduces a new and more comprehensible plan moniker "open-ended HMOs," to substitute for the commonly used yet confusing label "point-of-service" (POS) plan. But, otherwise, it simply offers a fuller definition of the most common names already used by industry actors. Although Weiner and de Lissovoy are right to simplify the jumble of health plan slogans, the complicated scheme they come up with does not so much "raze a tower of Babel" as rehabilitate it.

Note first that Weiner and de Lissovoy's scheme actually tells us relatively little about each type of health plan. If a plan places financial risk on sponsors, for example, it may be a MIP, PPO, EPO, or even a traditional fee-for-service plan. If it puts intermediaries at financial risk, it may be any of the plan types. We are told that if a plan has a network of providers it is an "integrated medical system." But what integration means in this context is unclear, especially since it is a characteristic apparently shared by all but one of the plan types. (Why MIPs are not considered integrated medical systems is also unclear, since they are counted as managed care plans and, according to Weiner and de Lissovoy's definition, the essence of managed care is the "integration" of medical care.) Virtually, the only clear criterion offered by the scheme is that if medical providers bear risk, then a plan is an HMO of some sort.

And even this distinction is problematic. As Weiner and de Lissovoy note, many different types of health plans are experimenting with ways to shift risk onto providers through payment methods, profit sharing, and bonus schemes. Virtually all health financing methods, even a system of national health insurance, place some risk on providers. Rather than say risk bearing is present or absent, it is far more instructive to identify the locus of risk, whether it be all providers within a geographic area (as in a national health insurance scheme with a global budget), a specific group of provider (such as an HMO's medical group), or an individual professional (as in many of the most recently developed incentive arrangements).

The central problem with Weiner and de Lissovoy's taxonomy and, indeed of most commentary about health insurance, is the tendency to confuse reimbursement methods, managerial techniques, and organizational forms. For example, fee-for-service, a payment method, is often contrasted with "managed care," which is presumably an organizational form. In Weiner and de Lissovoy's scheme, MIPs are distinguished from traditional fee-for-service plans by their reliance on a particular managerial technique, namely utilization review. In contrast, PPOs and EPOs are distinguished from MIPs by their particular organizational form, namely, their reliance on a network of participating providers. And HMOs are distinguised from all these plans by their particular payment method namely capitation.

The practice of conflating organization, technique, and incentives leads to unnecessary confusion. It means that when we contrast health plans we are often comparing them across incommensurable dimensions (arguing, for example, that an HMO is somehow more "managed" than a fee-for-service plan with utilization review even when the latter may use much stricter controls on individual treatment decisions). It means, too, that we may be tempted to presume necessary relationships between particular features of health plans (such as their payment method) and specific outcomes that are alleged to follow from these features (such as the degree of integration of medical finance and delivery), even though such outcomes usually result from a complex of financial, organizational, and administrative factors. Finally, it encourages a wild-goose chase of efforts to come up with black-and-white standards for identifying plan types. As health plans employ increasingly diverse payment methods and organizational forms, the search for the "essence" of a particular plan will become all the more futile.

For this reason, we believe that health policy scholars will increasingly find that to say something meaningful about the structure and operation of health plans, they will have to look beyond broad plan labels and focus more intensively on the constituent features of the plans themselves. Three such features seem to us to be most critical: (1) the degree of risk sharing between providers and the primary risk-bearing agent (such as a health plan or a self-insured employer), (2) the degree to which administrative oversight constrains clinical decisions, and (3) the degree to which enrollees in a plan are required to receive their care from a specified roster of providers.

We should make clear that these three dimensions of variation are not meant to furnish strict criteria for determining whether a plan is an HMO, PPO, or any of the myriad other labels that are commonly used by industry insiders. The difficulties with existing categorization schemes make us skeptical that these broad labels carry much meaning, or that any simple means for distinguishing among them can be found, especially given the rapid pace of change in America medical care. Rather, we wish to challenge the common way of thinking about health insurance. Our argument is that health plans differ across at least three principal dimensions: managerial control of clinical decision making, risk sharing between plan and provider, and limits on patient choice of medical professional. Each of these dimensions crucially affects the trilateral relationship among provider, patient, and plan. We want to emphasize as well that there is no simple relationship between plan label and the placement of a plan along these axes. Staff-model HMOs may seem like the quintessence of "managed care," yet because they place financial constraints at the group level, they do not necessarily concentrate as much risk on physicians as do other network-based health plans, nor do they necessarily entail as much clinical regulation at the microlevel. Microregulation may go hand-in-hand with restrictions on patient choice of provider, but it also may not. In fact, management of individual clinical decisions and the creation of broad incentives for conservative practice patterns may very well be alternative mechanisms for lowering the cost of medical care. Finally, as recent developments in health insurance suggest, greater risk sharing can coexist with almost any set of arrangements. It does not require a closed network, much less strict utilization review. Risk sharing is a product of the payment methods and incentive structures that connect risk-bearing agents and medical providers; it does not exclusively occur in HMOs, nor does it require capitation.

Notice, too that we have made no mention of those popular buzzwords "integration" and "coordination." Movement toward a closed network, toward greater utilization control, or toward increased risk sharing can create the conditions under which integration or coordination may occur. But they do not imply that such integrative activities actually take place. Nor dios the conventional fee-for-service/capitation dichotomy remain a particularly useful means of classification. What is crucial is the incentives that medical providers face. The particular mix of payment methods that creates those incentives is less important and will undoubtedly change as health plans experiment with new reimbursement modalities in the future.

Disaggregating health insurance into its constituent features not only clarifies what health plans do and how they are structured, but also makes it easier to identify the specific trends in medical finance and delivery that are carelessly jumbled together when we speak of such grand events as the "managed care revolution." Although we cannot provide a comprehensive empirical survey in this context, our reading of the evidence leads us to believe that the developments of the past decade have not pushed American health insurance in a consistent direction, much less toward any single organized entity that might be labeled "managed care."

Indeed, movement along the three dimensions that we identify has been halting and inconstant. Through roughly the late 1980s, and increasing number of health plans moved toward closed networks, but in the last decade, there has been a trend toward intermediate levels of compulsion, with formerly closed plans offering opportunities for patients to opt out (with a penalty) and new plans shying away from closed-network structures. Utilization review was also fashionable during the 1980s, but it has fallen into disfavor as plans have moved toward greater reliance on plan-provider risk sharing, which appears to have become more focused at the individual provider level over time. If there has been a general movement in the past two decades-and surely there has been-it has been from plans with little utilization review, no provider networking, and limited risk-sharing toward plans that incorporate some measure of all three. Yet movement along these three dimensions has been neither consistent nor evenly paced, and while it seems likely that the drift will continue toward greater risk sharing, that does not necessarily mean greater reliance on utilization review or closed provider networks.

Conclusion

We have argued that the most striking feature of the debate over managed care is its confusion. Both political actors and commentators appear largely to be trading in slogans and stylized facts, the truth or falsehood of which remains unproven. If this is true, the starting point for a sensible discussion of recent developments is the acknowledgment that many of the categories we are accustomed to employing in our analyses are essentially slogans that are used for self-promotion by actors in contemporary American medicine. In that respect, they are appropriate objects of study in their own right, but they are not analytical terms that can frame our investigations, or at least not without considerable further specification.

Once we address specific features of health insurance, moreover, the category "managed care" becomes ambiguous. The "managed care revolution" is really a set of related trends, few of which are accurately captured by the blanket term. When these trends are distinguished from one another, the evidence suggest that American health insurance has moved simultaneously in several different, perhaps even contradictory, directions in recent years and that many of the changes are longer standing than the rhetoric of managed care celebrants implies.

The rapid changes taking place in American medical care place a special burden on analysts to be precise about the criteria and considerations that underlie their empirical evaluations and, ultimately, their judgments and assessments. Labels and categories are indispensable, but they should be designed to elucidate the techniques, organizational forms, and incentives that characterize alternative health plans, rather than to confirm or deny the claims of industry friends or foes. "Managed care" fails that test, and although we hardly expect our words to be heeded (especially since both of us have reluctantly used the term in our own writings), we think that it, and other terms like it, should be banished from the health care lexicon for good.