Steven J. Kern, Esquire
This letter responds to your request on behalf of Children's Healthcare, P.A., ("CHPA") for a statement pursuant to the Department of Justice Business Review Procedure, 28 C.F.R. §50.6, of the Department's present enforcement intentions regarding CHPA's proposal to form a network of pediatricians practicing in southern New Jersey for the purpose of negotiating contracts and contracting with managed care health plans.
Provider owned and controlled networks can offer additional competitive options to consumers. For this reason, the Department supports the formation of properly structured provider networks. The Department and the Federal Trade Commission jointly issued the Statements of Enforcement Policy and Analytical Principals Relating to Health Care and Antitrust (hereinafter "Health Care Antitrust Statements") in September 1994. Statement 8 of the Health Care Antitrust Statements established antitrust safety zones for exclusive physician networks that comprise 20% or fewer of the physicians in each specialty who practice in the relevant geographic market and share substantial financial risk and for nonexclusive physician networks that comprise 30% or fewer of the physicians in each specialty who practice in the relevant geographic market and share substantial financial risk.
The Statements noted that those percentages were antitrust safe harbors. That means that not only would physician networks that meet these criteria be acceptable (except in extraordinary circumstances), but that the Department also would not object to networks with higher percentages if anticompetitive consequences were unlikely in a particular situation or if efficiencies clearly outweigh the risk of competitive harm. In order to facilitate formation of procompetitive networks, the Department makes available a business review procedure through which individual guidance can be given to parties attempting to form a network. Under the business review procedure, the Department has issued favorable letters for a number of risk-sharing physician networks with percentages higher than 30% of the physicians in the relevant market. See, e.g., California Chiropractic Association (Dec. 8, 1993) (50%); Collaborative Provider Organization (July 6, 1994) (six specialties in which percentage of physicians was above 30%); International Chiropractor's Association of California (Oct. 27, 1994) (50%); Physician Care, Inc. (Oct. 28, 1994) (percentages in certain specialties "significantly higher than 30 percent"); Mid-South Physician Alliance, Inc. (Mar. 30, 1995) (42%). The Department has also issued favorable business reviews for provider networks with high provider market shares where the network uses a "messenger model" (see Statement 9 of the Health Care Antitrust Statements at 94-96) under which individual providers do not agree on or jointly negotiate prices or other significant terms of competition. See, e.g., Georgia Preferred Podiatric Network (Nov. 3, 1995) (88% eligible); Preferred Podiatric Network (Sept. 14, 1994) (54% eligible). In addition, the Department has made clear through its enforcement actions that, in at least some circumstances, price-setting provider networks in which the market share of the owning physicians does not exceed the safety zone limits may subcontract with additional physicians, thereby allowing the network to offer a provider panel with up to 100% of the physicians in the market, as long as those subcontracts are appropriately structured to minimize competitive concerns. See, e.g., United States v. Health Choice of Northwest Missouri, Inc., (E.D. Mo., consent settlement pending).
In particular cases, however, physician networks can be anticompetitive and harm, rather than help, consumers. For the reasons explained in detail below, the Department has reviewed CHPA's proposed physician network under the rule of reason and has concluded that it is likely to have anticompetitive effects that are not outweighed by procompetitive efficiencies. Consequently, the Department cannot issue a favorable business review letter for CHPA's planned physician network if it is implemented as proposed.
From CHPA's business review request, we understand that CHPA would be formed as a professional services corporation with a membership comprising approximately 65-70 pediatricians maintaining individual or group practices in seven counties in southern New Jersey.(1) Although some prospective members may be board-certified in one or more pediatric subspecialty, CHPA intends to contract with managed care plans to provide basic health care to children of plan enrollees.
Each member physician would maintain an independent practice and would be able to contract individually with healthcare plans or to participate in other networks, subject, however, to an obligation to grant CHPA a right of first refusal to negotiate with any payer that seeks to contract, or to renew an existing contract, for the physician's services.(2) Despite this restriction, CHPA describes the venture as non-exclusive.
CHPA also has stated that it intends to incorporate risk sharing into its structure and operations. CHPA plans to solicit and accept full-risk capitation contracts, but it will also offer healthcare plans the option of a contract with a "discounted" fee schedule together with a withhold fund into which a defined percentage of all fees will be deposited and which will be distributed to CHPA's members only if predetermined cost containment goals are met. Under the fee schedule option, CHPA's members will be paid the lesser of the amount from the fee schedule or the member's usual and customary charges.
In addition, CHPA anticipates cost savings from its proposed operations. CHPA expects to achieve these savings through the development of managed care standards, joint purchasing, and the sharing of some administrative expenses.
CHPA believes that the competitive effect of the proposed organization should be evaluated in a relevant geographic market reaching beyond the counties in which its members practice and defined as the Greater Delaware Valley, consisting of southern New Jersey, southeastern Pennsylvania and northern Delaware, all areas that send patients to major specialty care pediatric hospitals in Philadelphia. It asserts that within this tri-state area, CHPA's membership would represent less than 20 percent of the pediatricians with active hospital staff privileges.
Furthermore, CHPA maintains that family practitioners and other primary care physicians who treat children are good substitutes for pediatricians and should therefore be included in the relevant service market. If so, even if the geographic market were defined more narrowly to include only the seven counties where CHPA's physicians actually practice, CHPA asserts it would still comprise only 26 percent of pediatricians and family practice doctors with active hospital staff privileges.
Statement 8 of the Health Care Antitrust Statements refers to a physician contracting organization or "network" of independent medical practices as a "physician network joint venture" ("PNJV"), and states that a PNJV will be reviewed under a rule of reason analysis either if the physicians in the joint venture share substantial financial risk or if the combining of the physicians into a joint venture enables them to offer a new product producing substantial efficiencies.(3) If either of these qualifiers is met, the next step under the analysis is to determine whether the planned PNJV is likely to have a significant anticompetitive effect and, if so, whether that harm is outweighed by procompetitive efficiencies created by the venture.
After careful consideration of the information you have provided, as supplemented by our own inquiries,(4) the Department believes that the PNJV, as proposed, would likely violate the antitrust laws. Although CHPA has not provided complete details about its proposed risk-sharing measures (for example, it has not specified what percentage of its members' fees would be deposited in the withhold fund and has not explained what it intends to discount to arrive at its discounted fee schedule), for purposes of this letter we have assumed that CHPA would implement the proposed PNJV with adequate, substantial risk sharing. Thus, we have analyzed the PNJV under the rule of reason.
Our investigation has revealed that in the southern New Jersey area to be served by CHPA, family practitioners are not acceptable substitutes for pediatricians in the development of managed care physician networks and that markets for basic pediatric services are significantly more localized than CHPA asserts. As a result, in several south New Jersey communities, CHPA would achieve high levels of concentration in the relevant service market and would be able to exercise market power to the detriment of consumers. This potential anticompetitive harm is not outweighed by proposed possible efficiencies.
RULE OF REASON ANALYSIS
The first step in our rule of reason analysis of a PNJV is to identify the relevant service market in which the PNJV's member-physicians compete. CHPA contends that the relevant service market is broader than just pediatricians--that it also includes general practitioners, board-certified family practitioners and other doctors who may include children among their patients, or who would be willing to do so if CHPA demanded supracompetitive contract terms for pediatricians. After investigating this issue, we cannot conclude that family practitioners or primary care providers are acceptable substitutes for pediatricians in this market. While some general and family practitioners do accept children as patients, our investigation has revealed that many health plan enrollees (and their employers) expect their health plan to provide pediatricians for child care, except in very rural communities where few or no pediatricians are available. We also have found that family doctors who accept children as patients frequently decline to treat infants or children below a certain age, and often limit the number of children they will accept as patients.
Moreover, our investigation, including interviews with the medical directors and provider contracting officials of health benefits plans, leads us to the conclusion that it is impossible successfully to market a health plan that requires (or provides incentives to) its enrollees to use doctors other than pediatricians to care for their children, especially in the more affluent and populous areas of southern New Jersey. In fact, pediatricians are one of three physician practice areas (the other two being OB-GYNs and family doctors) that form the core of any marketable plan, and that provide the benchmark against which all plans are measured by employers and enrollees. Our investigation reveals that health benefits plans would not be able to sufficiently substitute family doctors or others for pediatricians even if all pediatricians in the relevant geographic market (however defined) raised those fees significantly.
In short, because health plan enrollees expect their plan to offer pediatricians to care for their children, primary-care pediatricians are an essential component of health benefits plan physician networks, and other types of primary-care physicians are not reasonable substitutes. Consequently, primary care provided by pediatricians appears to be the relevant service market for analyzing the competitive effects of CHPA's proposal.
The next step in a rule of reason analysis is to determine the geographic market or markets in which the PNJV's share of available providers should be measured. As outlined in the Health Care Antitrust Statements at p. 72, for each relevant service market, the relevant geographic market will include all physicians whom health benefits plans and their subscribers consider to be good substitutes for physicians participating in the joint venture. CHPA asserts that the relevant geographic market in which to assess its PNJV is the Greater Delaware Valley--consisting of southern New Jersey, southeastern Pennsylvania and northern Delaware--because major Philadelphia pediatric hospitals draw patients from throughout that area.
In support of its geographic market assertions, CHPA provided to us the residence zip codes of patients that had been seen by three of the large pediatric practices that are participants in CHPA. All three practices are located within fifteen miles of each other in relatively urban, affluent portions of Burlington and Camden Counties. The zip codes cover a wide geographic area, suggesting the possibility of broad geographic markets. This type of data, however, has significant limitations that must be considered in context with other evidence. First, such data does not reveal either the number of visits by each patient or the nature of the patients' illnesses.(5) Second, such data provides no information regarding the number of families living inside that area that may have selected as their primary-care pediatrician a doctor whose practice is located elsewhere.(6) Finally, and most significantly, such data tells us little about the willingness of families that choose to use a pediatrician in a particular location in southern New Jersey to switch to a pediatrician in a different location.
CHPA is organizing to provide basic health care for children, most of which is provided by pediatricians at their offices, and the balance of which is provided by those pediatricians at local hospitals where they regularly practice--for example, the in-hospital examinations most hospitals require before discharging newborns, attendance at Caesarean section deliveries, and in-hospital treatment for children who are not severely ill and do not require tertiary-level services. Our investigation, including extensive interviews with knowledgeable market participants currently operating in southern New Jersey, makes it clear that the relevant geographic markets are far smaller than either the Greater Delaware Valley or the seven-county area in which CHPA's members practice. Rather, because most plan enrollees appear reluctant to travel long distances to obtain basic health care for their children, the relevant markets appear to consist of at most two or three small adjacent cities, or other localized areas probably not more than fifteen miles in diameter. For example, payers generally considered the Cherry Hill area to be a distinct local market in which they must offer pediatricians because plan enrollees within that area expect to be able to obtain basic health care for their children within the Cherry Hill vicinity. One plan explained that it strives to offer at least one or two pediatricians' offices within a seven or eight mile distance of any enrollee's home address.
Our investigation shows that, due to the preferences of plan enrollees, managed-care health plans could not successfully market a plan that required (or offered incentives to) enrollees located in a narrow geographic area, such as Cherry Hill, to use a primary care pediatrician located in another area within southern New Jersey, let alone in Philadelphia or in Delaware. For that reason, plans would absorb a significant rate increase by the pediatricians in one such geographic area rather than attempt to do without (or to reduce their use of) pediatricians in that area. In short, our investigation shows that existing and forming health plans need to contract with pediatricians in each of many relatively narrow geographic markets throughout southern New Jersey, and that they could not practically alter this practice to avoid a significant rate increase by pediatricians in any one of the narrow geographic areas.
In evaluating the possible competitive effects of the proposed PNJV, we note that even in hypothetical local geographic markets drawn more broadly than the just-described appropriate markets, CHPA would represent approximately 50 to 75 percent of the primary-care pediatricians in at least several important local markets for pediatric services in southern New Jersey. For example, in the Cherry Hill/Voorhees/Haddonfield/Marlton/Medford area its share would be 77%.(7) In the nearby Woodbury/Woodbury Heights area its share is 67%, and in Atlantic City, 50%. With market shares this high, CHPA likely would be able successfully to demand supracompetitive rates because succumbing to supracompetitive contract demands by CHPA would be less detrimental to health benefits plans than the alternative of offering what would be a much less marketable (or unmarketable) physician network in key areas of southern New Jersey.
CHPA contends that because it is "nonexclusive," it will be unable to impose supracompetitive contract terms on managed care customers: because its member-physicians would be free to contract with customers on competitive terms--either individually, or by participating in competing physician network organizations--CHPA will not pose any competitive risks. This assertion, however, is based on the assumption that CHPA would actually be nonexclusive in practice.(8) The right of first refusal provision, and other language in CHPA's proposed membership agreement and organizational documents, raise a significant question as to whether CHPA as proposed would in fact operate as a nonexclusive network. Moreover, in order to determine whether a PNJV is truly nonexclusive it is necessary to look beyond whether the bylaws and membership agreements are facially nonexclusive, and attempt carefully to determine whether a proposed PNJV's members are in fact likely to contract directly with managed care health plans or to participate in competing physician network organizations on competitive terms.
In support of its claim that CHPA would likely be nonexclusive in practice, CHPA points out that, before it was formed, many of its members contracted directly with managed care health plans or participated in a preexisting multispecialty IPA. This, however, does not establish that CHPA's members will not now decide to pursue supracompetitive contract demands by bargaining collectively through CHPA. On the contrary, other information gathered in our investigation suggests a significant danger that they may make such supracompetitive demands.(9)
First, CHPA documents strongly suggest that one of the primary objectives of CHPA is to obtain and exercise enhanced bargaining power in negotiations with health plans by presenting a united front in the pediatricians' dealings with those customers. Second, in anticipation of CHPA's formation and operation, a number of pediatricians who were committed (or prospective) members of CHPA declined to sign individual contracts with managed care plans and indicated that they were waiting until CHPA could negotiate the contracts for them. Some pediatricians told managed care plans that CHPA precluded them from participating in the plans independently. Other pediatricians signed individual contracts but told the plans that they would replace their individual contracts with CHPA contracts as soon as CHPA was operating. Third, it appears that CHPA's organizers already have negotiated with managed care health plans on behalf of some or all of its members and have demanded substantial increases in rates. Finally, at least one managed care health plan official was told by a member of CHPA that southern New Jersey pediatricians were organizing and would soon be in a position to dictate terms to managed care health plans. All of this evidence supports the conclusion that the high shares that CHPA would possess in particular geographic markets could and likely would be used to raise prices to the ultimate detriment of consumers.
CHPA has asserted that any power it might gain over price through the formation of its PNJV would be disciplined by the entry and potential entry of pediatricians who would provide alternative sources of supply. In support of this assertion, CHPA cited a number of hospital-sponsored and other health plans that CHPA believed were bringing new physicians into southern New Jersey. The evidence available to us, however, indicates that it is unlikely that a supracompetitive rate increase for pediatric services in southern New Jersey markets would be defeated by the entry of either newly certified pediatricians or pediatricians now practicing elsewhere.
With one exception, the health plans cited by CHPA either are staff-model plans that are hiring and bringing in new physicians as employees (who will not be available to contract with the network-model health plans served by CHPA) or are not bringing in new doctors from outside the area.(10) In fact, each of the network-model health plans we interviewed stressed that the prospect of entry by outside or newly-certified pediatricians would not deter CHPA from imposing a supracompetitive rate increase. Because many health plan enrollees seek treatment for their children from established pediatricians with whom they are familiar, plans must offer provider networks with such physicians in order to obtain enrollees; "importing" or relying on new entrants in lieu of contracting with the established local pediatricians is not feasible. Furthermore, few, if any, new or outside pediatricians would be likely to enter a market under circumstances that could alienate a large portion of the established local pediatricians, since new entrants depend on the support (back-up coverage, consultations and referrals) of established local pediatricians. Additionally, an effort to significantly expand the number of local pediatricians without the cooperation of a large proportion of local private pediatric groups would at best be an extremely slow process, since pediatricians beginning practice or relocating to a new area strongly prefer to do so as part of an existing medical group.
Finally, against the prospect of a significant anticompetitive outcome, we weigh, under the rule of reason, any substantial efficiencies that are likely to result from the formation of CHPA and could not be achieved without it. CHPA claims that its proposal will achieve efficiencies from the risk-sharing measures it will implement and through the development of practice procedures, sharing of administrative expenses and joint purchasing.
We have considered possible efficiencies arising from implementation of CHPA's risk-sharing proposal, but upon careful review, we have concluded that any such efficiencies are minimal and would probably not be passed on to consumers. It is unlikely that CHPA, or any similar specialty physician venture, could realistically offer a "full-risk" capitation agreement that would cover a broad range of services beyond the primary care pediatric services offered by CHPA's members. In any event, the potential efficiency benefits would not outweigh the high risk of competitive harm posed by CHPA's large share of the pediatricians in local markets. Given that very substantial risk, it is unlikely that CHPA would pass on to health care consumers any efficiency benefits it might be able to achieve by such an arrangement.
To the extent that CHPA intends to offer capitation contracts limiting its risk to the services provided by its members, CHPA could provide that type of capitation arrangement with a much smaller share of the pediatricians in any local market, significantly reducing the risk of anticompetitive harm while achieving the same benefits. In addition, the benefits from capitation arrangements that CHPA may offer are limited since health plans are already contracting on a capitated basis with individual pediatricians and group pediatric practices. As a result, whatever efficiencies that might be achieved from CHPA's risk-sharing measures cannot justify the substantial likelihood that the proposed PNJV will injure competition to the detriment of consumers.
CHPA also expects to achieve efficiencies by developing practice procedures, but these, too, could be achieved without the large pediatrician concentrations contemplated by CHPA. In fact, one health benefits plan in southern New Jersey already has developed practice procedures in cooperation with its contracting physicians. In addition, practice procedures could be developed jointly without incorporation of the price-setting activities proposed by CHPA.
Another efficiency benefit expected by CHPA is savings through the sharing of administrative expenses. CHPA has not specified the administrative services for which its members will share costs, but whatever savings might realistically be expected, it appears they could also be achieved if the proposed PNJV were limited to a smaller share of the pediatricians in any relevant market, or if the sharing of administrative expenses were not combined with price-setting activities. In either event, the savings could be achieved without the substantial risk of competitive harm presented by the current proposal.
Finally, anticipated savings from joint purchasing cannot justify the formation of a price-setting network of pediatricians that is likely to possess significant market power. Like many other groups, New Jersey pediatricians (or pediatricians together with other physicians) could achieve the benefits of volume purchase discounts by organizing a joint purchasing cooperative without the risk of anticompetitive harm presented by the current proposal.
For the reasons explained in this letter, we conclude that CHPA's PNJV, as proposed, would likely injure competition, harm consumers, and therefore violate the antitrust laws. Unless CHPA modifies its proposal in a manner that sufficiently addresses the foregoing concerns, the Department is likely to challenge implementation of CHPA's proposed PNJV.
This statement is made in accordance with the Department's Business Review Procedure, 28 C.F.R. §50.6, a copy of which is enclosed. Pursuant to its terms, your business review request and this letter will be made publicly available immediately. Your supporting documents will be publicly available within 30 days of the date of this letter unless you request that any part of the material be withheld in accordance with Paragraph 10(c) of the Business Review Procedure.
1. The seven counties are Atlantic, Burlington, Camden, Cape May, Gloucester, Ocean and Salem.
2. The members would, however, be permitted to honor any existing contracts until their expiration date.
3. Health Care Antitrust Statements at 71.
4. The Department interviewed officials responsible for network development at the network-model health plans that are either currently operating or attempting to develop provider networks in southern New Jersey. Together, these plans account for a very large proportion of the individuals enrolled in network-model health plans in southern New Jersey. We also spoke with officials at several large teaching hospitals and medical centers in Philadelphia and New Jersey, and reviewed documents and information submitted by CHPA, its individual members and its counsel. Finally, we spoke at length with Dr. John Tedeschi, founder of CHPA, and counsel.
5. Thus, it is not possible to tell how much of that data reflects a regular patient base, as opposed to patients who visited those groups only once or occasionally either because they happened to be visiting that area when they needed to see a pediatrician, or because they were referred to one of those three large, sophisticated pediatric groups for a particular condition or procedure, for a second opinion, or for a particularly serious condition, possibly one requiring hospitalization at a hospital where those pediatricians practice.
6. Other evidence we have obtained indicates that it is not unusual for patients living outside urban areas to travel to see a pediatrician in a more urbanized area, but that few, if any, of the enrollees who live in the more urbanized areas are amenable to traveling to a more rural area for pediatric care.
7. These figures actually may substantially understate CHPA's share of available pediatricians. A spot check of pediatricians on staff at hospitals in the Cherry Hill area who are not proposed participants in CHPA indicates that a significant number work for staff-model health plans, or for hospitals, and are not available for recruitment by network-model health plans.
8. Even a nonexclusive network can be anticompetitive when the network includes a large portion of the available providers in a relevant market. When a network has a large percentage of available providers, these providers face significant incentives to change their contracting patterns so that the network becomes de facto exclusive or to contract outside the network only on noncompetitive terms. This is particularly so in single-specialty networks and in multi-specialty networks that include a high percentage of available providers in many or all of the component specialties.
9. While providers' participation in other networks and in individual contracts with health benefits plans are factors to be considered in determining whether a network is exclusive or nonexclusive (see Health Care Antitrust Statements at 69), these factors are significantly more useful in evaluating an existing network than in assessing the likelihood that a prospective network will be nonexclusive.
10. The one exception is a network-model health plan that has established medical office sites at a number of locations throughout New Jersey, each of which is occupied by a different local group practice that has agreed to serve that plan exclusively, in addition to treating its own private indemnity patients. In some instances where the group practice did not include a pediatrician, the plan assisted the group in recruiting a pediatrician to join or be employed by the group practice. However, the health plan emphasized that based on its experience, it would not respond to a supracompetitive rate increase for established local pediatricians by attempting to recruit pediatricians from outside local markets, because it does not believe that it could successfully employ such a strategy.