Antitrust Analysis of Hospital Networks and Shared Services Arrangements

Share This Page


The American Hospital Association, Sixth Annual OGC/Allied Legal Counsel Seminar

New York, N.Y.

Robert F. Leibenluft, Former Assistant Director

It is a pleasure to be here to discuss antitrust issues of concern to hospitals. I would like to focus my remarks on two issues: (1) the recent revision of the Federal Trade Commission/Department of Justice Statements of Enforcement Policy in Health Care, particularly as the revisions apply to hospitals; and (2) how we analyze agreements among hospitals to combine, coordinate, or allocate clinical services, an issue that AHA raised in the course of our revision effort this past summer. Before I begin, I should explain that I am speaking only for myself, and that my views do not necessarily reflect those of the Commission, any Commissioner, or the Bureau of Competition.

I hope to continue the productive dialogue between the Federal Trade Commission and the American Hospital Association that began some time ago. Responding to the rapid changes in the health care marketplace poses challenges for us as well as for hospitals. We consider it a very important part of the FTC's mission to clarify to hospitals and other health care providers the requirements of the antitrust laws as those providers respond to market demands in their local communities. At the same time, we need your input to help us understand the competitive issues facing you, and the ways you feel it is necessary to respond to those issues. AHA staff and members were very helpful in providing information concerning the revision of our enforcement guidelines that I will discuss today, and in explaining to us issues of concern to AHA members. We look forward to working with the AHA and its member hospitals as they try to sort out how the general principles contained in that document apply in particular settings. Beyond that, we are aware that there are some questions on which many hospitals feel more guidance is needed. The shared services arrangements I will discuss later are a prime example of that. And we look forward to learning more about the types of arrangements that hospitals wish to undertake, and in providing additional guidance on the appropriate antitrust analysis of those arrangements.

FTC/DOJ Enforcement Policy Statements

I will not review today the history of the FTC/DOJ enforcement policy statements in health care. As you know, in 1993 and 1994 the agencies issued statements concerning a number of types of provider conduct, and this past August we revised two of the statements -- one covering physician networks and the other covering physician-hospital organizations (PHOs) and other health care multiprovider networks.

One of the major purposes of the revisions was to clarify when agreements among otherwise competing providers on the prices at which they will sell their services through a network will be considered per se illegal price-fixing agreements, and when they will be considered ancillary to a potentially procompetitive joint venture and analyzed under the rule of reason. The 1993 and 1994 policy statements relied heavily, but not exclusively, on financial risk-sharing among network participants, through mechanisms such as capitation or substantial risk withholds, as an indicator of integration sufficient to bring related price agreements within the rule of reason. The new statements continue to focus on integration among network participants that has the potential for generating efficiencies, but make it clear that financial risk-sharing as that term was used in the prior statements is not the only type of integration that will be recognized. The over-arching principle articulated in the statements is that price agreements that accompany and are reasonably necessary to substantial integration that is likely to result in significant efficiencies are subject to review under the rule of reason.

Last fall, FTC staff began an inquiry designed to gather information from many segments of the industry about various types of provider networks and the efficiencies that could flow from them. We were particularly interested in how rapid changes in health care markets have affected the kinds of arrangements that buyers seek in the market, and whether antitrust enforcement policy -- or uncertainty or misperceptions about that policy -- impeded the development of new ways of meeting purchasers' needs for high quality, cost-effective health care.(1) Among other things, we wished to assess concerns expressed by some observers that the policy statements' emphasis on financial risk-sharing was driving market participants toward risk-sharing arrangements regardless of whether those structures best served consumers' needs, and were stifling the development of other types of potentially procompetitive arrangements.

We spoke with a large number of individuals and groups, including consumer groups, self-insured employers, purchaser coalitions, representatives of providers, economists, attorneys and consultants who advise providers seeking to form networks, and insurers and managed care plans. In addition, drafts of the revised policy statements were shared with a large number of individuals, and their comments were considered in finalizing the document. We met several times with hospital representatives, including the AHA, and their input was very helpful.

We obtained a broad range of reactions and views. Not surprisingly, different segments of the health care industry -- consumers, physicians, hospitals, non-physician providers, employers, and HMOs and payers -- have different perspectives and primary concerns. Moreover, markets vary tremendously across the country. While we found no direct evidence that the existing policy statements in any significant way prevented payers from obtaining what they wanted in the market, it became apparent that there were areas of the guidelines that could be clarified. These included questions concerning the appropriate treatment of networks that do not assume financial risk, the types of mechanisms through which financial risk could be shared, and the purpose and meaning of the antitrust safety zones for physician networks. The revisions address all these areas.

The revised policy statements emphasize that the same antitrust principles that govern all other industries also apply to health care providers -- and our goal is to be neither more strict nor more lenient in the application of these principles to health care than to other industries. The revisions also make clear that integrated provider arrangements that offer the potential for creating significant efficiencies for consumers, regardless of their precise form, will be evaluated under the rule of reason. One example of such integration -- termed "clinical integration" -- is discussed in the statements, and involves a network implementing active and ongoing systems to evaluate and modify practice patterns by network participants and that create a high degree of interdependence and cooperation among the providers to control costs and assure quality. Such systems could include, for example, mechanisms to monitor and control utilization of services in order to control costs and assure quality, selectively choosing providers in order to further the network's efficiency goals, and significant investment of monetary and human capital in building the capability to realize the efficiencies sought by the venture.

This type of clinical integration was of interest to many of the people to whom we spoke, and it is one example of an arrangement that does not involve financial risk-sharing but receives rule of reason treatment. However, the statements make it clear that we will evaluate other types of efficiencies. The statement on multiprovider networks --which covers networks involving hospitals, non-physician health professionals, or ancillary service providers, as well as networks including a variety of types of providers, such as PHOs -- recognizes that some of the mechanisms that may produce efficiencies for physician networks may not be relevant to other types of networks; conversely, these other networks may employ mechanisms that produce efficiencies that would not be relevant to physician networks. Thus, we will consider the particular nature of the services provided by the network in assessing its potential to produce efficiencies. Certainly, the types of efficiencies demanded by buyers of those services would be relevant to that inquiry.

At the same time, the statements emphasize that we will look at the substance of the efficiencies, not just form. Our enforcement experience has taught us that providers sometimes attempt to dress up a cartel that seeks to defeat consumer choice in the garb of a provider network. Arrangements designed primarily to impede or prevent competitive forces from operating in the market, rather than to achieve efficiencies, will continue to be condemned summarily, regardless of the form in which the arrangement is clothed, or its declared purposes.

The statements contain several new hypothetical examples designed to illustrate these points. There are two similar examples, one involving a physician network and one involving a PHO, that involve substantial clinical integration. In these examples, the networks commit to a serious effort to establish goals relating to the quality and utilization of services, to monitor participants' performance concerning these goals, and to take concrete steps, where necessary, to modify participants' actual practices. In addition, the networks develop specific mechanisms, hire staff, and obtain information processing capabilities necessary to implement these programs. The statements conclude that these arrangements offer significant potential for creating efficiencies, and that the price agreements among the providers are reasonably necessary to the realization of those efficiencies, so that the networks are subject to evaluation under the rule of reason.

There also is an example applying the rule of reason to a network that has risk-sharing contracts, but also does some business on a non-risk sharing basis, using the same provider panel, the same fee schedule, and the same utilization management mechanisms that are involved in its capitation contracts. A similar arrangement, involving hospital laboratories, received a favorable FTC staff opinion this past summer.(2) Mayo Medical Laboratories proposed to establish regional or state-wide networks of hospital laboratories in order to compete with large commercial laboratories for broad-area contracts to provide outpatient laboratory services to managed care plans. It was anticipated that most network contracts would be on a capitation basis, and the networks planned to develop a number of mechanisms, including testing algorithms to eliminate unnecessary testing and physician office visits, designed to help them manage risk under such contracts. The networks also proposed to bid on requests for proposals on a fee-for-service basis, and to provide the same integrated laboratory services, involving the same mechanisms that would be used under the capitation contacts. Based on the description of the proposed networks in the advisory opinion request, the staff concluded that the proposed networks appeared to involved substantial integration creating the likelihood of significant efficiencies, and the opinion letter analyzed both the capitated and fee-for-service contracts under the rule of reason.

We also added a hypothetical that directly addresses an issue of concern to some hospitals: the small rural PHO that includes all or most of the physicians on the medical staff of the hospital. In that example, the only hospital in a rural county and all its physicians form a PHO to try to contract with managed care plans that are steering some of their employees who live in the rural area to providers located in the nearest city, with which the managed care plans have contracts. The managed care plans have been unwilling to contract with the hospital and physicians individually, but express an interest in contracting with the PHO under a risk sharing contract that involves not capitation or risk withholds, but a bonus for meeting certain utilization targets. Physician participation in the PHO is nonexclusive. Under the circumstances described in the example, the analysis concludes that the arrangement would not limit competition in any market, and that the agencies would not challenge the formation of the PHO.

The revised policy statements provide additional examples of financial risk-sharing, and like the earlier versions, emphasize that other forms of acceptable risk sharing may well exist. Percentage of premium or revenue arrangements, bonuses based on meeting utilization targets, and some global or all-inclusive case rates are specifically discussed, and there is a new hypothetical involving a PHO that provides bone marrow transplants and related cancer care on a per case basis.

There also is a clarification concerning the various forms of the "messenger model" that can be used by networks that do not involve substantial integration among members to avoid agreements on price-related terms while reducing the cost of contracting with payers. We have attempted to make clear that the critical question is whether there is a horizontal agreement on price or price-related terms, not the form of the mechanism used.

In revising the guidelines, we did not change the safety zones for physician networks; nor did we add any additional safety zones. Safety zones of "per se" legality are rare in antitrust law, because they preclude the detailed analysis of the specific conduct and market characteristics that often is necessary to assess the impact of an arrangement on competition. Because the safety zones need to be simple to apply, they are based on very rough approximations of market share, which typically are only a starting point for assessing market power and may not fully reflect competitive conditions in the market. For these reasons, the safety zones apply only to a limited class of networks that can be presumed lawful without further examination. We learned through our discussions with the industry, however, that the safety zones were misunderstood by some to be limits on legal, or at least, non-risky behavior, and that they could result in channeling networks into a particular model that may not best meet consumers' needs, just to assure easy approval of the enforcement agencies. As we have tried to make clear, the demands of the market, expressed through consumer preferences, should direct the development of delivery systems. Antitrust law is intended to prevent distortions of the market, not to prescribe how ventures should be organized. The revisions emphasize principles over numbers, and explain that networks outside the safety zones often may be legal. Overall, we hope that providers will focus less on the safety zones, and more on developing innovative ways to serve consumers better.

The enforcement agencies are committed to efficient and effective antitrust enforcement, while still allowing innovation in response to market demands. The guideline revisions clarify that rule of reason analysis likely will apply to many forms of provider collaboration. Those arrangements will need to be evaluated individually to determine their actual effect on competition. We realize that this may not always provide the clear black and white line that many providers would like. Bright-line rules that are easily understood offer certainty, but they do not easily accommodate future market developments, and they risk discouraging innovation. Instead, through the use of hypotheticals and an explanation of the underlying antitrust principles, we hope that we have provided sufficient guidance, while still allowing the flexibility that will allow health care providers to respond creatively to market demands.

Shared Services Arrangements

During the guidelines revision process the AHA expressed interest in more guidance about hospital services joint ventures, including agreements among hospitals to share existing services and agreements to create "centers of excellence" by combining services. Today I would like to describe the guidance that we already have given, and discuss the issues that may be raised by other ventures and how we would analyze them.(3)

As a starting point I should emphasize that, as in other health care arrangements that we examine, we apply to hospital service joint ventures the same antitrust principles that govern arrangements in other sectors of the economy. Neither we nor the Department of Justice has brought a case challenging a shared services joint venture; nor have the agencies issued many advisory opinion or business reviews on the subject. Since antitrust often requires a careful analysis of a specific factual context, experience with a real example in an investigation or advisory opinion generally is the best way to clarify our analysis of a particular type of arrangement. However, joint ventures among competitors are common in other industries, and antitrust law has developed ways of analyzing their competitive implications. The potential of such ventures to reduce costs is well recognized. At the same time, it is important to preserve vigorous market competition in order to spur efforts to reduce costs and improve products or services.

Analysis of particular joint ventures requires careful consideration of the precise ways in which the ventures restrict competition among the participants and in the market as a whole, the efficiencies that the ventures may produce, and the relationship of the competitive restrictions that flow from the ventures to the achievement of the efficiencies. To pick just one example, a number of years ago the Commission considered a joint venture between GM and Toyota to produce small cars in a plant in California. The venture was designed to provide GM ready access to small cars that it would market independently of Toyota, and to permit GM to learn about Toyota's production methods. After careful analysis of the surrounding circumstances, the Commission permitted the venture to proceed. However, the order imposed two types of restrictions. First, in order to preserve GM's incentives to produce its own small cars, the duration and maximum output of the venture were limited. Second, the order restricted the exchange between the parties of commercial information that was not essential to the success of the joint venture.(4)

The health care policy statements contain a specific discussion of hospital joint ventures to provide specialized clinical or other expensive health care services.(5) As the statement notes, such ventures may create procompetitive efficiencies that benefit consumers. These efficiencies may include reducing the cost of providing a service; improving quality; or enabling hospitals to share the costs of developing a service, thereby permitting it to be offered at all. On the other hand, by their very nature these ventures often eliminate present or possible future competition between the participating hospitals, and must be scrutinized carefully under the antitrust laws.

The threshold inquiry with any shared services arrangement is whether the venture involves integration among the participants that is likely to produce significant efficiencies in the provision of the affected services. Thus, the first step in the analysis is to determine whether the venture involves real integration, or is simply an agreement to restrict competition and decrease output. Thus, as the statement notes, an agreement among two hospitals, each of which can profitably provide both open heart-surgery and a burn unit, that in the future open-heart surgery will be offered only at one hospital while the burn unit is offered only at the other, would, standing alone, be considered an illegal market allocation agreement.(6) On the other hand, an agreement among hospitals to jointly finance, build, and operate a new open-heart surgery or burn unit, where neither hospital had provided the service before, clearly involves significant integration among the parties.

As a general rule, the antitrust rule of reason is applied to integrated joint ventures. This analysis focuses on whether the joint venture may reduce competition substantially, and if it might, whether it is likely to produce procompetitive efficiencies that outweigh the anticompetitive potential. Let me make clear that the potential of a venture to provide services at a lower unit cost than the venturers could independently does not end our analysis. Cooperation often produces economies of scale, and that is important. Nonetheless, intense competition is one of the strongest motivators for hospitals to find ways to provide services that are both more cost effective and more attractive to consumers. Thus, cooperation that reduces competition may result in a loss of consumer benefit that outweighs the gains from the cooperation. Moreover, once such competition is lost, it may be very difficult for it to be reintroduced into the market.

The policy statement outlines four stages in a rule of reason analysis. In a particular case it might not be necessary to conduct all four stages of the inquiry. In some cases the central question of anticompetitive effect may be fairly clear on its face. For example, if a legitimate joint venture faces significant competition from other providers in the same market, and it limits competition among the joint venturers only in ways that are reasonably necessary for the venture's attainment of its legitimate objectives, then the venture is not likely to pose competitive problems. On the other hand, if the venture eliminates existing or likely future competition among the joint venturers and there are few or insignificant other competitors in the market, the anticompetitive potential is obvious; and if potential efficiencies are slight or could readily be attained by significantly less anticompetitive means, the venture can be condemned without extended analysis.

The first step in the rule of reason analysis involves defining the product and geographic markets in which the joint venture operates. This involves identifying providers that offer services that patients or other customers would consider a good substitute for those provided by the venture.(7) It should be noted that patients may be willing to travel farther for highly specialized services than for routine care, so the geographic market for some such services may be broader than the market for primary medical care.

The second step involves assessment of the competitive effects of the venture. In this stage, we look at the extent to which the venture would eliminate existing or potential competition among the joint venturers, and the extent of competition from other providers in the market. A venture may restrict competition in a number of ways: it can eliminate one of the venturers as an existing competitor in the market; or it can eliminate future entry by one of the venturers that likely would have occurred but for the venture; or the cooperation of the venturers in the joint undertaking can spill over into markets where the venturers continue to compete separately, and result in diminished competition there. On the other hand, if the venturers produce a service that neither could provide independently, or if they jointly produce the service but market it independently, competition may not be significantly restrained. Other factors bearing on the likely competitive effects of a venture include the extent and nature of competition from others in the market and the ease of entry into the market by outsiders.

If there is reason to believe that the venture restricts competition significantly, the analysis proceeds to an evaluation of the venture's potential to create efficiencies, and a balancing of the efficiencies against the likely anticompetitive effects. In the context of health care joint ventures, efficiencies might include making the venture economically possible by spreading its cost over a greater volume of business, and the improvement of quality flowing from the providers' performing a larger number of procedures. In some cases, for example, studies suggest that within a certain range, increased patient volumes may be associated with improved outcomes for patients. We also consider whether the efficiencies sought by the venture can reasonably be achieved through means significantly less restrictive of competition.

Even if the venture as a whole passes muster under the rule of reason analysis, we also look at whether the venture involves collateral agreements that unreasonably restrict competition without contributing significantly to the legitimate purposes of the venture. A very clear case of a collateral agreement would be if the venturers agreed on the price of services they sold individually outside the joint venture. For example, a joint venture among two hospitals to provide open-heart surgery would not justify an agreement on the price to be charged for other surgical services, and that agreement would be subject to challenge even if the joint venture to provide open-heart surgery itself were lawful. A harder question is posed by agreements by the joint venturers not to compete with the venture. In some cases, such agreements may be reasonably necessary for the success of the joint venture. In other cases, the breadth or duration of the noncompete agreement may be excessive in light of the legitimate needs of the venture, and subject to challenge. For example, a joint venture between two hospitals to open a primary care clinic in an underserved rural area might well be procompetitive; but the arrangement would not likely justify an agreement among the participants that they would not establish competing clinics in other areas where separate clinics would be economically viable.

The policy statement includes one example of a services joint venture that clearly does not pose competitive concerns. It involves a joint venture by the only two general acute care hospitals in a relatively small community to provide open-heart surgery services that neither hospital provides, or plans to provide, on its own. Moreover, present demand in the community will support only one open-heart surgery unit. The two hospitals share the cost of establishing the program, and share expenses and revenues from it. The hospitals do not exchange competitively sensitive information, and they do not enter into any collateral agreements that reduce competition and are not reasonably necessary to achieve the goal of the venture. In this instance, the venture does not reduce competition for open-heart surgery, since neither hospital now provides the services, and having two competing units in the area is not economically viable. Accordingly, the venture would not be challenged.

In assessing the competitive effects of the venture and in weighing those effects against the venture's promised efficiencies, we would consider whether the venture sets the "retail" price of the service, as opposed to selling it to the parent hospitals at a set price, and leaving it to them to determine the price that is charged to patients. Joint production of a product or service may provide efficiencies that do not depend on joint marketing of that product. I will describe in a moment some examples of hospital joint ventures that did not involve joint pricing. There also are many examples outside the health care field. The National Cooperative Research and Production Act of 1993 provides certain protections from the antitrust laws to joint ventures formed for certain purposes. Significantly, however, the Act does not cover either joint pricing of the joint venture output (except to the joint venture parties) or market allocation agreements among competitors.(8) Separate retail pricing and marketing, of course, also were features of the General Motors/Toyota consent agreement I mentioned earlier. In any case, we would need to make an individualized analysis of whether any joint marketing, if it existed, was reasonably related and necessary to the production of the efficiencies promised by the joint venture; and, on the other hand, whether separate pricing is likely, given the overall incentives and market position of the parties, to preserve effective competition among them in selling the joint venture service.

Our approach to some of these issues is illustrated more specifically by an FTC staff opinion letter issued in 1995 involving a proposed joint venture between two hospitals in Chattanooga, Tennessee, to build and operate a new hospital specializing in obstetrical and related gynecological services.(9) One of the joint venturers, Erlanger Medical Center, was the largest general acute care hospital in the area and the leading supplier of obstetrical services. The other joint venturer, Memorial, was the third-largest hospital firm in the market, and historically had been the most direct competitor of Erlanger. Memorial had not offered obstetrical services since 1982. Both hospitals were located in central Chattanooga. The proposed joint venture, Women's East, would establish a 28-bed hospital located in a suburban area, specializing in routine, low-risk childbirths and related gynecological services, and which would focus on delivering services in a combined labor, delivery, recovery and post-partum setting. Women's East would sell services directly to patients and health plans, as well as making services available to each of the hospitals for resale to health plans as part of the overall package of general hospital services covered by each hospital's contracts with those plans.

The central issue examined in the staff opinion letter is the likely competitive effect of the joint venture. Since Memorial did not provide OB services, the venture did not eliminate existing competition. Memorial was, however, a potential competitor in the market -- it had strong incentives to provide OB services in order to be able to offer a full range of services to managed care plans, it was Erlanger's most direct competitor, and it had considered again providing OB services by itself. In that particular context, the loss of a potential competitor could be of serious concern, because the market appeared to be highly concentrated, and certificate of need regulation made entry by other firms extremely unlikely. The staff concluded, however, based on a full review of the facts (some of which are nonpublic) and its understanding of the market based on prior investigations, that any elimination of competition was outweighed by the venture's procompetitive potential. The venture was likely to enhance competition in the larger market for hospital services by allowing Memorial to offer OB services of lower cost or higher quality than it could offer on its own, and both hospitals had significant competitive incentives to pass these efficiencies on to consumers.

The letter also considered whether the agreement between the two hospitals in establishing and managing the joint venture was likely to spill over into markets in which they were competitors. Based on consideration of a number of factors, including the continuing incentives of the parties to compete vigorously across the full range of hospital services, the relatively narrow scope of the joint venture compared to the scope of services where the hospitals competed with one another, and protections against the exchange of competitive information through the joint venture, the staff concluded that the joint venture was not likely to endanger competition. The letter also noted that the contract defined the maximum prices that Memorial would have to pay for the venture's services, and that there were no constraints on the prices at which Memorial could resell those services to payers.

Another example of a different type of joint venture entered into by a hospital is provided by a staff opinion letter we issued involving cooperation between a hospital and a rural primary care clinic in Colorado.(10) In order to provide services to patients in the area, the two parties cooperated in staffing a primary care clinic. They shared a common facility and many of the expenses associated with the clinic, but did not share revenues, and did not combine their operations into a single business. The two parties agreed on the hours during which each party's clinic would be open, on which party would see which types of patients during hours when both clinics were open, and on referral of patients for follow-up care. There was no agreement on prices to be charged by either party, and no restriction on the ability of either party to offer services at other locations, and the arrangement would not override a particular patient's preference for obtaining services from one or the other provider. Moreover, similar services were being offered by a number of other providers in the same area.

The staff concluded that the parties were not jointly operating a single clinic, but were coordinating, and sharing some expenses of, separate businesses operated from shared premises. Nonetheless, the staff did not treat the joint operating agreement as an illegal market allocation scheme. Instead, the letter found that the arrangement did not significantly restrict competition between the parties, and that any incidental restriction of competition was legitimately ancillary to the purpose of assuring that the clinic was open during all normal business hours without either party having to provide full-time coverage with its personnel.

It also might be instructive to discuss briefly a hypothetical example of a service joint venture that the AHA posed to us earlier this year. The hypothetical does not include enough facts for me to fully analyze the venture, but I can indicate the questions we would ask and the issues we would consider.

The hypothetical is as follows: In a six-hospital market, Hospitals A and B both provide acute inpatient pediatric services. The hospitals would like to joint venture to combine their pediatric units. Another hospital in the market also has a pediatric unit. The two hospitals believe that if they combine pediatric services, the increased patient volume will allow them to recruit and support a pediatric physiatrist. Neither hospital could independently support such a specialist. In addition, a consultant's report indicates that increased patient volume should increase efficiency from both an economic and quality point of view. Hospitals A and B intend to market and price the pediatric services jointly.

In this example, the joint venture appears to involve significant integration. But the arrangement would entirely eliminate the existing competition among the two hospitals in pediatric services, leaving only one other competitor. The arrangement would be equivalent to a merger of the two hospitals' pediatric units, and its effects would be analyzed in accordance with the principles outlined in the agencies' merger guidelines, which are used in evaluating full hospital mergers. Among other things, we would need to know is how significant for the market is this reduction of competition. How substantial is the competition between the two hospitals, and how great a factor is the third hospital which also offers pediatric services? Will the combination of the two pediatric units permit the hospitals to exercise market power in pediatric inpatient services? Is there a prospect of timely and effective entry in pediatrics by others if the venture reduces competition significantly? If the two hospitals in the venture plan to operate the unit together, they likely would be located fairly close to one another. As a result, they may be more direct competitors than if they were located in different areas of a large market, and the elimination of that competition may be very significant. On the other hand, if the two hospitals both had small, underutilized pediatric units and faced competition from a large, well-established children's hospital, the combination of the units might benefit competition by making the combined unit a more effective competitor to the dominant hospital in the market. We also would consider whether the joint pricing was necessary to achievement of the venture's efficiencies. On the other hand, if pricing is done separately, do the parties have any real incentives to compete with respect to price? And to what extent would separate pricing be likely to reduce the anticompetitive effects of the venture?

Given the competitive concerns that would be posed by this hypothetical joint venture, we also would take a close look at the efficiencies that the venture was likely to produce. Besides economies of scale, what types of efficiencies are possible, and how significant are they? Certainly, it is not obvious that general pediatric care is likely to experience the significant decreases in mortality and morbidity that are associated in some fields with the performance of a certain minimum volume of specialized procedures. The parties would need to explain in much more detail how the greater patient volumes would increase quality and lower costs. If it turns out that the potential efficiencies are real and significant, we would need to balance those efficiencies against the loss of competition. This can be a very difficult undertaking in some cases, but it is the kind of analysis that frequently must be performed in the analysis of joint ventures in other industries. As part of this assessment, we also would be very interested in the views of payers in the market about the magnitude of the efficiencies and the overall competitive impact of the venture.

The possibility that the venture would permit the hospitals to recruit a pediatric physiatrist does not by itself appear to justify the elimination of competition inherent in the joint venture. It would seem that the hospitals could jointly recruit the physiatrist without combining the whole of their pediatric services, by sharing the expenses of the effort and splitting the individual's services between the two hospitals. Indeed, one of our staff opinion letters approved a proposal by two rural hospitals to jointly purchase services, including clinical services such as laboratory, pathology, radiology and physical and respiratory therapy services.(11) In that instance the hospitals planned to share the services of personnel who could not profitably be employed on a full-time basis at either hospital. Notably, in contrast to the hypothetical, there was no plan to coordinate the offering of clinical services at the two hospitals.

Finally, we would look at the potential impact of this joint venture on competition among participating hospitals with respect to the services they provided outside the joint venture. For example, we would look at whether the structure of the venture permitted one hospital to discipline or discourage competitive behavior by the other.

AHA also offered two other wrinkles on the hypothetical. It asked how our analysis would differ if (1) Hospitals A and B are the only hospitals in the market providing pediatric services; or (2) instead of joint venturing to recruit a specialist, the hospitals wish to joint venture because they are currently losing money providing the services, or project that they will incur such losses in the future?

If the only hospitals in the market combine their services, the effect is totally to eliminate competition and consumer choice with respect to the service. If both hospitals can profitably provide services in the market, it is hard to imagine circumstances that would lead us to conclude that the consolidation would benefit consumers. On the other hand, if both hospitals could not profitably provide the service, then consolidation may be preferable to the other alternatives. We would, however, have to look very closely at the evidence that the services could not be profitably provided by both hospitals in the long run, and at what other alternatives might be available that would have a less detrimental impact on competition. For example, could the parties undertake other types of joint ventures that would permit them to reduce costs, while still competing with each other with respect to clinical services?

Clearly, analysis of hospital shared services arrangements can be a very fact-intensive undertaking. In order to provide more specific guidance to hospitals about service joint ventures, we need a better understanding of what types of ventures hospitals are interested in undertaking, and of the forces driving the ventures and the efficiencies that are sought. I urge you to discuss these issues further with us, either on an informal basis or through the vehicle of an advisory opinion request involving a specific proposal that the hospitals propose to undertake.(12)

* * * * * * *

In closing, let me reiterate my commitment to working with you as we seek to apply established antitrust principles to new health care institutions and relationships. In this time of tremendous change, none of us can foresee clearly where the market will take us the future. I hope to build on the good working relationship we already have developed with the AHA as all us of respond to the challenges before us.

1. See Mark D. Whitener, FTC Bureau of Competition Deputy Director, Remarks before the American Enterprise Institute for Public Policy Research (December 5, 1995) (discussing "Antitrust, Medicare Reform and Health Care Competition.")

2. Letter from Robert F. Leibenluft, FTC Bureau of Competition Assistant Director, to George A. Cumming, Jr. (July 17, 1996) (Mayo Medical Laboratories).

3. This discussion considers joint ventures to provide clinical services. Joint ventures to produce or purchase administrative or maintenance services or inputs into the services that the venturers sell in competition with one another usually will not raise competitive concerns. See "Joint Purchasing Arrangements Among Health Care Providers" in U.S. Department of Justice/Federal Trade Commission Statements of Antitrust Enforcement Policy in Health Care at 53 (August 1996) (hereinafter cited as Health Care Statements).

4. General Motors Corporation, 103 F.T.C. 374 (1984) (consent order); set aside order, October 29, 1993, 58 Fed. Reg. 64,950 (December 10, 1993).

5. A separate statement covers the related area of hospital joint ventures involving high-technology or other expensive health care equipment. "Hospital Joint Ventures Involving Specialized Clinical or Other Expensive Health Care Services" in Health Care Statementsat 31. The statement provides a safety zone for joint ventures to purchase, operate and market new or existing equipment if the venture includes only the number of hospitals whose participation is necessary to support the equipment; other hospitals that could not support the equipment individually or through formation of a competing joint venture also can be included. Ventures that do not fall within the safety zones will be analyzed under the rule of reason, which considers the nature and competitive effects of the venture, including efficiencies that the ventures could produce.

6. Health Care Statements at 32, n.9.

7. This step of the analysis is similar to that outlined in the 1992 Department of Justice and Federal Trade Commission Horizontal Merger Guidelines 1.1 and 1.2 (product and geographic market definition), 4 Trade Reg. Rep. (CCH) 13,104 (April 2, 1992).

8. 15 U.S.C. 4301 (b) (1993). The legislative history of the Act also indicates that it is not intended to cover joint ventures "to provide what are simply services, such as health care or legal services, that are unconnected to any concrete technological innovations, or to joint ventures solely to purchase medical equipment." S. Rep. No. 51, 103d Cong., 1st Sess. 9 (1993).

9. Letter from Mark J. Horoschak, FTC Bureau of Competition Assistant Director, to Carlos C. Smith, Esq. and Edward N. Boehm, Esq., (May 31, 1995)

(Erlanger Medical Center/ Women's East Inc.)

10. Letter from David Pender , FTC Bureau of Competition Deputy Assistant Director, to Richard J. Sahli, (November 8, 1995) (Columbine Family Health Center).

11. Letter from Mark J. Horoschak, FTC Bureau of Competition Assistant Director, to Derrell O. Fancher (June 20, 1995) (Elmore Community Hospital, Inc.).

12. Some hospital joint ventures may seek immunity under state hospital cooperation laws, if those laws meet the strict requirements of the state action doctrine. See, e.g., S. Vance, Immunity for State-Sanctioned Provider Collaboration after Ticor, 62 Antitrust L.J. 409 (1994).