IN THE UNITED STATES COURT OF APPEALS No. 98-3123 v. TENET HEALTHCARE CORPORATION, INC., and BRIEF FOR PLAINTIFF-APPELLEE FEDERAL TRADE COMMISSION DEBRA A. VALENTINE JOHN D. GRAUBERT JOHN F. DALY LESLIE RICE MELMAN Federal Trade Commission SUMMARY OF THE CASE AND REQUEST FOR ORAL ARGUMENT This is an appeal from a decision of the district court granting the Federal Trade Commission's and the State of Missouri's motions for a preliminary injunction to prevent defendant Tenet Healthcare Corporation ("Tenet") from acquiring Poplar Bluff Physicians Group, Inc. d/b/a Doctors Regional Medical Center ("DRMC"). Tenet is a for-profit hospital chain of approximately 120 general acute care hospitals, including Lucy Lee Hospital, in Poplar Bluff, Butler County, Missouri. DRMC is the only other general acute care hospital in Butler County. As a result of intense competition between Lucy Lee and DRMC, residents of the Poplar Bluff area currently enjoy quality health care at very low cost. Tenet's proposed acquisition would eliminate that competition. The Federal Trade Commission ("FTC") sought preliminary relief under Section 13(b) of the FTC Act, 15 U.S.C. § 53(b), pending completion of an administrative proceeding to determine whether the proposed transaction violates Section 7 of the Clayton Act, 15 U.S.C. § 18. The State of Missouri sought relief under Sections 4, 4C, and 16 of the Clayton Act, 15 U.S.C. §§ 15, 15c, and 26. On July 30, 1998, after a five-day evidentiary hearing, the district court entered a preliminary injunction. The principal issue for this Court to resolve on appeal is whether the district court made clearly erroneous factual findings in defining the relevant geographic market. Oral argument of no more than 20 minutes for each side is appropriate. PRELIMINARY STATEMENT (i) This is an appeal from a decision of the United States District Court for the Eastern District of Missouri (Honorable Catherine D. Perry). The decision is reported at 1998-2 Trade Cas. (CCH) ¶ 72,227 (July 30, 1998). (ii) This is a suit for a statutory preliminary injunction. The district court had jurisdiction to decide the FTC's motion for preliminary injunction pursuant to 15 U.S.C. §§ 18 and 53(b), and 28 U.S.C. §§ 1337 and 1345. (iii) This Court has jurisdiction over this appeal pursuant to 28 U.S.C. § 1292(a)(1) (review of grant of injunctive relief). Defendants' notice of appeal from the district court's ruling of July 30, 1998, granting a preliminary injunction, was filed on August 10, 1998. STATEMENT OF ISSUES PRESENTED 1. Whether, in defining the relevant geographic market within which to assess the competitive effects of a merger between the only two hospitals in Poplar Bluff, Missouri, the district court committed legal error by considering all relevant evidence of the commercial realities affecting where patients could practicably go in the event of a post-merger price increase.
2. Whether the district court clearly erred in its factual finding that the Poplar Bluff Region was the relevant geographic market, where that finding was based upon both statistical and direct, nonstatistical evidence that consumers who obtain primary and secondary inpatient care in the Poplar Bluff hospitals could not practicably turn to hospitals outside the Region to avoid a post-merger price increase.
STATEMENT OF THE CASE I. Nature of the Case, Course of Proceedings, and Disposition in the Court Below In this action, plaintiffs the Federal Trade Commission ("FTC" or "Commission") and the State of Missouri challenged a plan by the nation's second largest for-profit hospital chain -- Tenet Healthcare Corporation of Santa Barbara, California ("Tenet") -- to secure a virtual monopoly in the provision of acute care hospital services in the area surrounding Poplar Bluff, Missouri, the economic center of a multi-county region in southeastern Missouri. Add. 4a.(1) Tenet already owns one of the two general acute care hospitals in Poplar Bluff, Lucy Lee Hospital, and now proposes to acquire the other, the now-independent Doctors Regional Medical Center ("DRMC"). After hearing the testimony of numerous witnesses, including local employers and health plans, the district court found that these hospitals currently compete vigorously with one another, and that this competition "has benefitted the consumers of the region and has kept prices significantly lower than those in other areas." Add. 9a. This vigorous competition -- and the benefits it brings to health care consumers -- would be lost if the challenged transaction is permitted to go forward. Add. 23a-24a. The Commission and the Missouri Attorney General filed separate actions seeking to prevent consummation of the acquisition, pending completion of a full adjudication in an FTC administrative proceeding to determine whether "the effect of such acquisition may be substantially to lessen competition, or to tend to create a monopoly," in violation of Section 7 of the Clayton Act, 15 U.S.C. § 18. The FTC seeks this relief pursuant to Section 13(b) of the FTC Act, 15 U.S.C. § 53(b). That section authorizes a district court to issue a preliminary injunction when it determines, after "weighing the equities and considering the Commission's likelihood of ultimate success," that such relief "would be in the public interest." The State brought its action under Sections 4, 4C, and 16 of the Clayton Act, 15 U.S.C. §§ 15, 15c, and 26. The district court consolidated the actions, and conducted a five-day evidentiary hearing. More than 40 witnesses testified either live or by sworn declaration or deposition, and approximately 100 exhibits were received into evidence. On July 30, 1998, the court entered a preliminary injunction, having concluded that plaintiffs had shown that they are "likely to succeed on the ultimate issue of whether the merger would have the effect of substantially lessening competition in the relevant market," and that the balance of equities favored grant of the injunction. Add. 4a. In the present appeal from the district court's preliminary injunction order, defendants advance arguments pertaining solely to the court's findings regarding the relevant geographic market. II. Statement of Facts A. Hospital Competition in the Poplar Bluff Region The Hospitals: Tenet's Lucy Lee Hospital and DRMC are both for-profit, acute care hospitals located in Poplar Bluff, Missouri. At 201 and 230 licensed beds, respectively, the hospitals are similar in size and services offered. PX 118 at 7, PA 535; Tr. 130 (Anderson), PA 34. Both hospitals provide a full range of primary and secondary hospital services (e.g., obstetrics, pediatrics, general medicine, and surgery), but not tertiary care (e.g., high risk obstetrics, neonatology, neurosurgery, and heart surgery). Residents of the Poplar Bluff area must use hospitals in Cape Girardeau, St. Louis, and elsewhere for tertiary care. Add. 4a-7a; see Tr. 47 (Miller), PA 7; Tr. 96 (Reynolds), PA 19; Tr. 743 (Till), PA 178. The only other hospitals within approximately 50 driving miles of Poplar Bluff are a 118-bed Tenet hospital and four rural hospitals. The four non-Tenet hospitals are small (25-50 beds), financially distressed hospitals that offer a much narrower range of services than Lucy Lee and DRMC. Add. 5a-6a, 16a; see PX 65 ¶ 6, PA 338; PX 75 ¶¶ 2, 4, PA 347-48; PX 97 at 35-37, PA 385-87; PX 167 at 21102, 21105, PA 580, 581; PX 376, PA 721; PX 391, Attach. 4, PA 724-26. Aside from small rural facilities, the nearest non-Tenet hospitals in the area are over an hour's drive from Poplar Bluff, and major tertiary referral centers in St. Louis and Memphis are roughly a 150-mile drive away. Add 4a-7a; see PX 100 Tables I-II, PA 477-80. The Hospitals' Customers: Although it is individual patients who use hospital services, these services are nearly always paid for in substantial part by "third party payers," such as employers, insurers and other health plans, or governmental entities. Under what has come to be termed "managed care," third party payers negotiate prices with hospitals, contract to obtain the best value, and create financial incentives to "steer" their subscribers to use a low cost provider. As the district court noted, the parties agree that managed care buyers play a significant role in individual consumers' selection of hospitals. Add. 7a. Defendants' economist discussed at the hearing that individual consumers "make their decisions based on those prices" that their managed care plan "creates" for them. Tr. 910 (Harris), PA 201. Individual consumers generally do not know what prices are at various hospitals, and the relevant price issue for them is their out-of-pocket costs. Id.; Tr. 42 (Miller), PA 5. Plaintiffs' expert likewise explained that "there is not a direct economic relationship between an individual consumer and the hospital" (Tr. 361 (Wu), PA 93), because of the important role of employers and health plans in negotiating prices and creating financial incentives for consumers. Tr. 361-63, 439, 442, 532 (Wu), PA 93-95, 138, 141, 152. Employers and insurance plans are not the only intermediaries in the consumer-hospital relationship. Since patients must be admitted to a hospital by a physician, where a physician holds hospital privileges also plays a role in an individual's use of a hospital. Tr. 52 (Miller), PA 11; Tr. 367 (Wu), PA 97; PX 16 ¶ 5, PA 315; PX 75 ¶ 6, PA 348. Patient preferences on such non-price matters as convenience and amenities are also a factor. Tr. 101-02 (Reynolds), PA 22-23; Tr. 248-49 (Wicklund), PA 63-64; Tr. 782 (Li), PA 189; PX 16 ¶ 7, PA 315; PX 73 ¶ 7, PA 345. In addition, the nature of the hospital care needed can be a consideration, and people may travel greater distances for more complex medical needs. Tr. 367 (Wu), PA 97; PX 52 ¶ 10, PA 330-31. Competition Between the Poplar Bluff Hospitals: Competition between Lucy Lee and DRMC has been intense, with each hospital focusing on the other as its primary competitor. Add. 8a, 23a. Tr. 752 (Till), PA 181; PX 103 at 643, PA 490; PX 107 at 90012, PA 508; PX 118 at 7, PA 535; PX 284 at 1016, PA 693; Kelly Inv. Hrg. 139-40, PA 261-62. Employers and commercial health plans view these hospitals as good substitutes for one another, and have been willing to shift their business from one to the other in order to get the best price. See, e.g., PX 14 ¶ 7, PA 312. The hospitals jockey for position, each granting discounts off their normal charges to employers and health plans who agree not to deal with the other, matching each other's discounts, and waiving co-payment charges for individual patients who would otherwise have a financial incentive to use the other hospital. Add. 8a; see Tr. 42-43, 49 (Miller), PA 5-6, 8; Tr. 168 (Thomas), PA 43; PX 22 ¶ 5, PA 321-22; PX 215, PA 582-83. As a result of this competition, purchasers have received substantial discounts off the hospitals' standard charges. Add. 8a. As an October 1997 memorandum by a Lucy Lee executive reported, the reimbursement rates for one major third party payer "varied greatly within the different [Missouri Tenet] facilities. Ours [Lucy Lee's] were very low, obviously, due to the current rates [that payer has] available at D.R.M.C." PX 108 at 110123, PA 526. In addition, competition between DRMC and Lucy Lee has prompted the hospitals to become more efficient providers of health care, and to meet the service and quality demands of their customers. Add. 23a-24a; see Tr. 750 (Till), PA 179; PX 14 ¶ 7, PA 312. The hospitals make efficient use of their facilities to serve both inpatient and outpatient needs, and profits at both hospitals have steadily increased. Add. 28a. See PX 158 at 4, PA 566; PX 325 at 1063, PA 702. The Acquisition: In April 1997, the defendants agreed that Tenet would acquire DRMC for $40.5 million. Add. 4a. In requesting authorization for the acquisition, Tenet officials advised their Board of Directors that the acquisition would "increase [managed care] contract negotiating leverage" and "protect the investment we currently have at Lucy Lee," and they projected substantial returns to the corporation's bottom line. PX 151 at 1-2, 4, 8 & Exhibit 1, PA 556-57, 559, 560. B. Proceedings Below At the preliminary injunction hearing, plaintiffs presented live testimony from five Poplar Bluff area employers, two health plans, and one former health plan official, as well as four experts, three of whom gave testimony relating to geographic market definition. Plaintiffs introduced additional testimony through deposition or sworn declarations from an additional eight employers and eight hospitals. Defendants' witnesses consisted of owners or employees of the merging hospitals, one individual from a firm with a plant in Poplar Bluff, and three other witnesses retained by defendants: the author of a consumer survey, an economist, and an accountant. A major focus of the hearing was the delineation of the geographic market. Plaintiffs' Case: Plaintiffs contended that the relevant geographic market for assessing the competitive effects of Tenet's proposed acquisition comprises Butler County and portions of the surrounding seven counties (Carter, Dunklin, Reynolds, Ripley, Stoddard, and Wayne Counties in Missouri, and Clay County in Arkansas) -- the "Poplar Bluff Region."(2) Plaintiffs adduced extensive testimony from employers in the Region -- on whom the financial burden of higher hospital prices would largely fall -- regarding the harmful effects of the proposed merger. These witnesses included both large and small employers located throughout the Region -- some in Poplar Bluff itself and some in outlying towns. Tr. 104 (Reynolds), PA 25; Tr. 169-70 (Thomas), PA 44-45; Tr. 207-08 (Clark), PA 54-55; Tr. 284-85 (Thorn), PA 72-73; PX 14 ¶¶ 1, 8-9, PA 310, 312-13; PX 16 ¶¶ 1, 8-9, PA 314-15; PX 18, PA 316-18; PX 19 ¶¶ 1, 5, 6, PA 319-20; PX 22 ¶¶ 1, 8, PA 321-22; PX 23 ¶¶ 1, 3, 16, PA 323, 326. They described the vigorous competition that has existed between Lucy Lee and DRMC, and the prospect of serious harm to the community if that competition were lost. The Director of Employee Benefits at Briggs and Stratton, for example, expressed concern that increased health care costs would raise the cost of his company's products and make them less competitive in national markets. Tr. 100 (Reynolds), PA 21. The employer testimony also showed that employers could not practicably avoid increased hospital prices in Poplar Bluff by obtaining services for their employees elsewhere. Based on their experience in administering health benefit plans, these witnesses testified that employees would not see hospitals outside the plaintiffs' proposed market as practical alternatives for the primary and secondary services offered in Poplar Bluff. Tr. 101 (Reynolds), PA 22; Tr. 169 (Thomas), PA 44; Tr. 248-49 (Wicklund), PA 63-64; Tr. 284 (Thorn), PA 72. Noting that switching to distant hospitals would create great inconvenience and could disrupt existing doctor-patient relationships, witnesses testified that employees would object to a health plan that sought to steer them away from Poplar Bluff. Employees use distant hospitals for more complex or specialized services unavailable in Poplar Bluff, but generally use local hospitals for primary and secondary care. Tr. 47, 52 (Miller), PA 7, 11; Tr. 165 (Thomas), PA 41; Tr. 567, 577 (Hall), PA 167, 171; PX 97 at 18-21, PA 369-72. Thus, while employers have successfully steered some patients between the two local Poplar Bluff hospitals, they do not believe that a plan to encourage use of hospitals outside the Region would be acceptable to employees. Tr. 101-02, 123 (Reynolds), PA 22-23, 28; Tr. 166 (Thomas), PA 42. Witnesses also explained why it was critically important to have a health plan that is popular with employees. Tr. 360-61 (Wu), PA 92-93; Stranglen Dep. 92-93, PA 279-80. First, health benefits are an important tool in attracting a qualified work force -- a significant factor given the competitive labor market in Poplar Bluff. Tr. 165-66 (Thomas), PA 41-42; Tr. 201 (Clark), PA 51; Tr. 244 (Wicklund), PA 60; Tr. 271 (Thorn), PA 69; Tr. 569 (Hall), PA 168. Second, an unpopular plan is likely to create labor relations problems. Tr. 102 (Reynolds), PA 23; Tr. 138-39 (Anderson), PA 35-36. Employers spoke of "an employee relations nightmare" (Tr. 102 (Reynolds), PA 23) and "[a] major moral issue" (Stranglen Dep. 92, PA 279). Some expressed concern that such employee discontent might cause workers to seek to unionize. Tr. 570-71 (Hall), PA 169-70; Stranglen Dep. 95, PA 281. In addition, the employers could not avoid these concerns by offering more than one health benefit plan, because it would prove costly for small employers and unwieldy for large employers seeking to offer uniform benefits to employees in different markets. Tr. 166 (Thomas), PA 42; Tr. 247 (Wicklund), PA 62; Tr. 1164-65 (Wu), PA 156-57; PX 23 ¶ 15, PA 326; Tr. 94 (Reynolds), PA 17; Tr. 202-03 (Clark), PA 52-53; Tr. 275 (Thorn), PA 70; Tr. 55-56 (Miller), PA 13-14; Tr. 139 (Anderson), PA 36. For all these reasons, these employer witnesses consistently confirmed that they could not and would not attempt to steer employees to hospitals outside the Region, even in the face of substantial price increases from the combined Tenet hospitals. See Tr. 100, 102, 123 (Reynolds), PA 21, 23, 27; Tr. 168-69 (Thomas), PA 43-44; Tr. 207 (Clark), PA 54; Tr. 283-84 (Thorn), PA 71-72; Tr. 564-65, 570 (Hall), PA 165-66, 169; see also PX 12 ¶ 6, PA 308-09; PX 14 ¶¶ 7, 8, PA 312-13; PX 16 ¶ 8, PA 315; PX 18 ¶ 4, PA 316-17; PX 19 ¶ 5, PA 319-20; PX 22 ¶ 8, PA 322; PX 23 ¶ 16, PA 326. Indeed, the sole employer witness called by defendants similarly testified that, if prices did go up, his firm would not try to use financial incentives to encourage employees to use hospitals outside the Poplar Bluff Region, because of the employee relations problem that would cause. Tr. 569-71 (Hall), PA 168-70. Health plan witnesses agreed that a benefit plan that sought to steer patients out of the Poplar Bluff Region through financial incentives or penalties would be unpopular, and expressed the view that they would not be able to market such a plan to employers. Tr. 54-55 (Miller), PA 12-13; Tr. 138-39 (Anderson), PA 35-36; Tr. 312-314, 346-47 (Ressel), PA 78-80, 81-82. Although employees in the eastern part of the Region are not as far from the hospitals in Cape Girardeau or Sikeston as those living in Poplar Bluff, third party payers testified that they would not try to steer such employees to the hospitals in those cities, because managed care rates at these hospitals are so high it would not save any money. Tr. 313-14 (Ressel), PA 79-80; Tr. 54-55 (Miller), PA 12-13. Other evidence confirmed that these hospitals would remain more expensive than those in Poplar Bluff, even if post-merger prices increased by 10 percent. Tr. 395-403 (Wu), PA 119-27; PX 392-95, PA 727-30. A representative of Blue Cross Blue Shield testified: "We would never try to steer anyone towards a higher cost facility," because it would raise the cost of the insurance plan and make it less marketable. Tr. 313-14 (Ressel), PA 79-80. The HealthLink witness agreed that it would cost more to steer patients to the facilities outside Poplar Bluff that have comparable services. Tr. 55 (Miller), PA 13. Testimony from the administrators of hospitals outside the Poplar Bluff Region likewise showed that these hospitals are not practical alternatives for consumers in the Region. These witnesses said they would not expect to gain patients after a 10 percent price increase in Poplar Bluff, nor would they take steps, such as building outpatient clinics within the region, to try to attract more patients from the Region. PX 58 ¶¶ 6, 8, PA 335; PX 73 ¶ 7, PA 345; PX 78 ¶¶ 4, 6, PA 350-51.(3) Plaintiffs also introduced business documents and other evidence showing that the defendants' own competitive behavior is consistent with the view expressed by employers, health plans, and other hospitals. This evidence showed, for example, that in negotiations with employers and health plans, each Poplar Bluff hospital often sought "exclusive contracts" that excluded the other from a "preferred provider" network, but made no attempt to exclude hospitals outside the Poplar Bluff Region. Tr. 43 (Miller), PA 6; Tr. 128-29 (Anderson), PA 32-33; Rouff Dep. 113-14, PA 273-74. The two Poplar Bluff hospitals set prices to avoid losing patients to one another, not to hospitals 50 or more miles away. PX 108 at 110123, PA 526. In addition, the "non-compete" provision in Tenet's purchase agreement bars DRMC's physician owners from operating a competing hospital within 50 miles of Poplar Bluff, but places no restriction on their involvement with hospitals beyond a 50 mile radius. PX 7 ¶ 14.10(c), PA 305. Plaintiffs' economic expert, Dr. Lawrence Wu, considered the testimony of market participants, along with statistical data and other evidence, and determined that economic analysis supported the Poplar Bluff Region geographic market alleged by plaintiffs. PX 97 at 15 & Exh. 7, PA 366, 401; Tr. 374, PA 99. His overall conclusion was that the proposed market is the correct one for antitrust analysis because employers and health plans "will continue to contract with the hospitals and consumers will continue to go to the hospital in enough numbers that a price increase would be profitable." Tr. 357, PA 91. Dr. Wu emphasized that, in order to answer the forward-looking question of whether a hypothetical monopolist could sustain a price increase, it is essential to understand the dynamics of the market -- i.e., the "mechanisms" by which consumers of hospital services would make substitutions to alternate providers. Tr. 366-67, 442-43, PA 96-97, 141-42. As he explained, individual consumers' hospital choices are guided in large part by intermediaries -- not only physicians (who must admit patients to a hospital where they have privileges), but also employers and health plans, which enter into contracts with hospitals and give members economic incentives to use particular hospitals through the terms of the health plans offered. Tr. 360-62, 367, 532-33, PA 92-94, 97, 152-53. As a starting point, Dr. Wu used statistics regarding the area from which DRMC and Lucy Lee draw 90 percent of their patients. Tr. 374-76, PA 99-101. But he went on to consider available information concerning the dynamics of the market, such as the reasons for which patients seek hospital services outside the Poplar Bluff Region, as an aid to determining the likely effect of a hypothetical price increase. Tr. 377-80, 391-92, PA 102-05, 115-16. He examined various factors underlying the statements of employers and health plans that they would not try to steer consumers outside the Region to avoid a post-merger price increase and found that those statements were borne out by the statistical evidence. Tr. 443, 382-403, 405-06, 423-25, PA 142, 106-27, 128-29, 130-32. In addition, he considered the particular characteristics of health care markets and the realities of the Poplar Bluff Region, and concluded that the statements that payers would not try to steer patients outside the Region were logical, persuasive, and made economic sense. Tr. 406, 425, 430-33, 443, PA 129, 132, 133-36, 142. Dr. Wu specifically considered and rejected suggestions that distant hospitals, in St. Louis or Cape Girardeau, were practical alternatives that could defeat a Tenet price increase in Poplar Bluff. He pointed out, for example, that because of the strong evidence that Poplar Bluff patients went to St. Louis hospitals for specialty referrals or other forms of tertiary care, they should be considered entirely outside of the relevant market. Tr. 382-89, PA 106-13. Dr. Wu similarly pointed to strong evidence that the patient "outflow" from Poplar Bluff to Cape Girardeau was for reasons other than economic competition, such as the need to obtain tertiary services not available in the Poplar Bluff hospitals. Tr. 392-95, PA 116-19. While he acknowledged that a somewhat higher percentage of persons in Stoddard County (which is closer to Cape Girardeau than are other parts of the Region) used services in Cape Girardeau, he also pointed to an important reason why an increase in Poplar Bluff hospital prices would not be expected to induce more consumers to go to Cape Girardeau -- i.e., price levels would still be greatly in excess of those in Poplar Bluff, even if the latter increased by 10 percent. Tr. 395-98, 403, PA 119-22, 127. He also pointed out that employers in the Region cannot practicably offer multiple health plans and so must tailor their plans to their entire workforce, and would not "change their plan to satisfy a few people who live say in Stoddard County." Tr. 432-33, PA 135-36; see, e.g., Tr. 55-56 (Miller), PA 13-14. As Dr. Wu explained, his consideration of "how consumers and hospital customers would respond to a change in price" encompasses the general concept of "critical loss" in the sense that it seeks to ascertain whether the prospect of lost business would indeed prevent a price increase. Tr. 479-82, PA 145-48. After considering all of the available evidence, however, his conclusion was that Tenet, once it owned both Lucy Lee and DRMC, would be able (and likely) to sustain a 10 percent price increase, and that "not enough people would leave to make that unprofitable." Tr. 430, PA 133. Defendants' Case: Defendants' evidence of geographic market consisted primarily of the opinion of their economic expert, Dr. Barry Harris, who based his analysis on historic patient flow data. While Dr. Harris agreed that employers and health plans play an important role in determining options open to individuals (Tr. 910, PA 201), he did not interview any as part of his analysis -- although he had previously relied on the views of such payers in another hospital merger case, in which he testified that employers "have the best idea in my mind of what is likely to occur." Tr. 934-35, PA 206-07. Defendants called only one employer witness, who stated that he favored the acquisition but confirmed that he would not attempt to steer employees to other hospitals if Tenet raised prices. Tr. 569-71 (Hall), PA 168-70. To estimate the extent of switching likely to occur in response to a post-merger price increase, Dr. Harris relied on past patient use data. Looking at each zip code in the merging hospitals' 90 percent service area, he identified as "contestable" those zip codes in which 20 percent or more of the residents had been admitted to hospitals other than the ones in Poplar Bluff. Def. Exh. Y 27-29, DA 238-40. He then assumed that all patients in a "contestable zip code" were "at risk" of switching to hospitals outside the Region if prices in Poplar Bluff went up. Tr. 853-54, 863, DA 623-24, 633. He also asserted that limiting patient flow data to "overlapping DRGs" would focus the analysis on the services in which the Poplar Bluff hospitals actually compete. Tr. 861-63, DA 631-33.(4) He contended that his conclusions about the willingness of consumers to switch were confirmed by the results of a survey conducted for defendants. Tr. 905-07, PA 198-200; Def. Exh. Y at 41, DA 252. Having concluded that patients would switch to other hospitals, Dr. Harris calculated what he termed the "critical loss," which he defined as the number of patients who would have to switch to make a price increase unprofitable. Tr. 913, DA 642. Under his formula, a 5 percent price increase would be defeated by a loss of 8 percent of commercially-insured admissions, and a 10 percent price increase raises the critical loss figure to 15 percent. Tr. 916, DA 645. Based on his analysis, Dr. Harris concluded that the relevant geographic market includes "principally hospitals within 65 miles of Poplar Bluff, but also Barnes Jewish in St. Louis." Tr. 839, PA 195. Plaintiffs' Rebuttal Case. Plaintiffs' rebuttal of Dr. Harris's analysis rested on three central points. First, Dr. Wu refuted Dr. Harris's contention that historic patient use data show that a price increase would alter consumers' behavior, in light of: (a) the acknowledged role of employee health plans in providing the "mechanism" for making consumers sensitive to hospital prices; and (b) the persuasive (indeed, uncontradicted) statements of employers that they would not change their health benefit plans to encourage employees to use hospitals outside the Region. As a result, the key premise of Dr. Harris's model -- that patients in so-called "contestable zip codes" are likely to shift in response to a price increase -- is a wholly speculative assumption that is not supported by his data or consistent with a dynamic view of the market. See Tr. 1162-66, 1171 (Wu), PA 154-59. Second, plaintiffs' DRG expert, Bruce Alan Steinwald, explained that a variety of services is contained within a given DRG, and that a single DRG can encompass both secondary and tertiary services. Tr. 1142-43 (Steinwald), PA 239-40; PX 98 ¶¶ 4, 5, 12-16, PA 417, 419-20. As a result, focussing on overlapping DRGs, as Dr. Harris did, would not eliminate patients who were using hospitals outside Poplar Bluff for services not available at the merging hospitals. Indeed, a physician witness called by defendants acknowledged that, for a given DRG, some services could be obtained in Poplar Bluff but others would be referred to tertiary hospitals. Tr. 773-74 (Li), PA 186-87. Using other measures of case severity, Mr. Steinwald analyzed patient use data and found that, both overall and within a given DRG, patients traveling to Cape Girardeau and St. Louis for care received more complex services than were offered by the Poplar Bluff hospitals. Tr. 1129-30 (Steinwald), PA 226-27; PX 98 ¶¶ 5, 22-31, PA 417, 420a-21. Dr. Wu pointed out that the results of the Steinwald analysis are consistent with other evidence that travel outside the market for care is largely for tertiary services. Tr. 387-89 (Wu), PA 111-13. Third, plaintiffs presented evidence that the consumer survey that Dr. Harris attempted to use to verify his statistical model was wholly unreliable. Plaintiffs' survey expert, Dr. Gary Ford, testified that, as a result of numerous defects in its design and execution, the survey was "fatally flawed" and "provides no useful evidence." Tr. 1104-05 (Ford), PA 212-13; see Tr. 1105-12 (Ford), PA 213-20. C. The District Court's Decision In its preliminary injunction ruling, the district court thoroughly analyzed all of the considerations pertinent under this Court's precedents: product and geographic market definition; the likely competitive effects of the merger within the defined market; possible defenses; and the balance of equitable considerations. The court concluded that the proposed acquisition would eliminate the direct and "intense" competition that has existed between Lucy Lee and DRMC, which has lowered prices and improved hospital quality and services. Add. 23a-24a. The court also found that available alternatives would not prevent Tenet from exercising the market power it would gain. Add. 14a-16a, 21a. In addition, the district court rejected claims that DRMC is a weak competitor or that the merger would enable Tenet to offer new services and achieve cost savings. Rather, the court found that DRMC is profitable and well-positioned to respond to market changes, and that defendants' efficiency claims were highly speculative. Add. 25a-29a. The court also concluded that the public interest in preserving competition outweighed any hardships the injunction might impose on Tenet. Add. 29a-30a. The court addressed the issue of geographic market definition at length. Add. 10a-21a. The court began its analysis by recognizing, in accordance with this Court's instructions, that plaintiffs must meet the burden of "identify[ing] a credible relevant market before any preliminary injunction issues." Add. 10a (citing FTC v. Freeman Hospital, 69 F.3d 260, 268 (8th Cir. 1995)). The court further recognized that an antitrust market cannot be defined solely on the basis of current patterns of use, but must encompass "the geographic area 'to which customers can practically turn for alternative sources of the product and in which the antitrust defendants face competition.'" Id. (quoting Morgenstern v. Wilson, 29 F.3d 1291, 1296 (8th Cir. 1994), cert. denied, 513 U.S. 1150 (1995)). In addressing the pivotal question of "practical alternatives," the court acknowledged the undisputed fact that a significant number of persons in the Poplar Bluff Region "have gone" to hospitals outside the Region. Add. 12a. However, the court further recognized that, in order to address the forward-looking question posed by Freeman and Morgenstern, a court must consider how the market works and why patients go to particular hospitals for various types of services. Id. Considering "the evidence as a whole," the court found "plaintiffs' evidence more credible, logical, and persuasive." Add. 13a. The court first addressed what it referred to as "anecdotal" evidence -- i.e., the extensive direct, nonstatistical evidence from employers, health plans, and other hospitals described above. Add. 13a-16a. Consistent with this Court's emphasis on evaluating practical alternatives from the consumer's perspective, and a recognition of the important role of employers and health plans as "consumers" (Add. 11a, citing Freeman), the district court focused particularly on the testimony of employers regarding what they would do in case of a price increase in Poplar Bluff. Add. 13a-15a. The court agreed with plaintiffs that, for the reasons discussed above, such evidence showed that employers could not practicably turn to outside hospitals to defeat a post-merger price increase by Tenet. See Add. 13a-14a, 16a. The court also discussed evidence that the larger hospitals closest to the Region are more expensive than those in Poplar Bluff, and would still be more expensive even after a 10 percent price increase. Add. 15a-16a. Noting the testimony discussed above, the court stated that third party payers would have no incentive to attempt to shift any patients to these larger hospitals. Add. 15a-16a. Taking all of this evidence into account, the court found that "sufficient numbers of consumers in the Poplar Bluff region would not practicably turn to these larger hospitals for acute care services in the event of a price increase by Lucy Lee and DRMC." Add. 15a. The court also discussed the testimony of administrators of hospitals located outside the Region, who stated that they did not expect a price increase in Poplar Bluff to result in any increase in their acute care admissions, and that even those hospitals with the financial resources to do so would not seek to attract additional patients from the Region. See PX 58 ¶¶ 6, 8, PA 335; PX 73 ¶ 7, PA 345; PX 78 ¶¶ 4, 6, PA 350-351. These administrators, the court noted, "would not expect their hospitals to be affected in any way" by a 10 percent price increase in Poplar Bluff. Add. 14a-15a. Turning to the statistical evidence, the court concluded that it lent further support to plaintiffs' position. Add. 16a-19a. The court stated that Dr. Wu's analysis of patient use data was "a proper first step" in the geographic market inquiry. Add. 17a. The court then noted that Dr. Wu had also considered a variety of other evidence in an effort to answer the dynamic question whether a price increase would cause those currently using Poplar Bluff hospitals to shift to hospitals outside the proposed market. In particular, it noted Dr. Wu's ultimate conclusion that "the merged hospitals would be able to profitably raise prices, and that a 10% price increase was probable." Add. 17a. With regard to the parties' use of DRGs in evaluating the statistical evidence, the court observed that a DRG "is a somewhat imprecise approximation of care complexity." Add. 17a-18a. Accordingly, the court concluded that looking to DRGs alone could not identify patients who were leaving the proposed market to obtain more specialized or complex care not available at the Poplar Bluff hospitals. Add. 18a. The court also discussed the analysis of patients leaving the Region that plaintiffs' DRG expert performed, using various measures of care complexity. The court concluded that, "even accepting some of the limitations suggested by defendants' expert, the DRG data supported plaintiffs' argument." Add. 19a. The district court rejected the contention of defendants' expert that any attempted post-merger price increase could easily be defeated because so many consumers live in "contestable" zip codes. The court found that the "contestability" model could not answer the question whether hospitals are practical alternatives, because it rests on a fundamentally flawed assumption -- i.e., that one consumer's unexplained use of an out-of-market hospital indicates that another consumer will use that hospital in the event of a price increase. Since DRG data alone could not adequately segregate those patients who were leaving to obtain services not available in Poplar Bluff, the court noted that the contestable zip code analysis would include patients leaving the Region for tertiary services unavailable in Poplar Bluff. Add.19a-20a. The court concluded that this zip code model -- and indeed the totality of defendants' factual arguments that they would face a "critical loss" of business that would defeat any price increase -- were factually unsupported because they improperly "assume[d] the answer to the ultimate question." Add. 20a n.5. Furthermore, the court concluded that the consumer survey that Dr. Harris relied on in an attempt to support the results of his contestable zip code model suffered from flaws so fundamental that the survey results were "not probative" and "wholly unhelpful." Add. 21a. As a result, the district court determined that defendants' geographic market analysis was "unsupported by the evidence and inconsistent with the economic realities of Southeastern Missouri." Id. On the contrary, the court held that plaintiffs had met their burden of establishing the relevant market, by evidence the court found to be "more credible, logical, and persuasive." Add. 13a. SUMMARY OF ARGUMENT The central question presented by this appeal is whether the district court made a clear factual error in determining that the Poplar Bluff Region -- an area encompassing Poplar Bluff and portions of seven surrounding counties -- is the relevant geographic market within which to assess the competitive effects of Tenet's proposed acquisition. After consideration of "the evidence as a whole," the court determined that the evidence presented by plaintiffs was "more credible, logical, and persuasive." Add. 13a. Such a determination of the relevant geographic market is a finding of fact, which must be upheld in the absence of a showing by defendants that the district court's delineation of the geographic market was "clearly erroneous." Community Publishers, Inc. v. DR Partners, 139 F.3d 1180, 1184 (8th Cir. 1998). In an attempt to divert attention from the fact-driven inquiry at the heart of this appeal, defendants contrive various supposed legal errors in the lower court's ruling. Their contentions are based on mischaracterizations of the district court's analysis, and a misunderstanding of this Court's precedents. In order to address the forward-looking question of where consumers "could practicably turn" for hospital services, the district court properly considered a broad range of evidence, including both statistical analyses and direct testimony by market participants. The "anecdotal" evidence that defendants criticize is an important part of the dynamic analysis called for, because it reveals likely responses to market changes. Defendants' contentions that plaintiffs should have been required to prove their case solely by statistical evidence has no basis in law, and, indeed, is directly contrary to this Court's teachings in Freeman Hospital. Nor is there any basis whatever for defendants' assertion that the court below somehow shifted the burden of proof. (Parts I.A and I.B.) Defendants similarly err in accusing the district court of ignoring their economic expert's "critical loss" analysis (Def. Br. 19). Plaintiffs agree that the bounds of the geographic market would have to be expanded if the evidence showed that a sufficient number of consumers could practicably leave the proposed market, thereby defeating Tenet's ability to raise prices. The district court understood this point, but concluded that "sufficient numbers" of consumers could not practicably do so. Add. 15a. The court did all that was legally required, and its rejection of defendants' "critical loss" analysis (Add. 20a n.5) did not reflect a misunderstanding of defendants' theories, but a recognition that they lacked factual support and were contradicted by strong evidence to the contrary. (Part I.C.) The factual record developed by plaintiffs consists of a broad array of evidence, including expert testimony, patient use data, and extensive direct, nonstatistical evidence from employers, health plans, and other market participants concerning their likely behavior in response to a post-merger price increase by a combined Lucy Lee/ DRMC. That record shows that employers and other third party payers in the Poplar Bluff Region -- who play a pivotal role in providing price signals to individual consumers -- could not practicably steer their members to hospitals outside the Region, because of their need to design health plans that employees will accept. Furthermore, the record shows that such payers see no economic purpose whatever in steering employees to the closest outside hospitals, since they are far more expensive than the Poplar Bluff hospitals and would remain so even with a 10 percent or larger post-merger price increase. Defendants' arguments about the impact of the "marginal consumer" do nothing to refute this showing, because of the practical realities faced by Poplar Bluff Region employers in the design of group health plans, as well as the circumstances facing individual members of those plans. The views of other hospitals in southeast Missouri -- and of the merging parties themselves -- confirm the lack of competition with hospitals outside the Region. (Part II.A.) Finally, the district court committed no clear error in crediting the opinion of plaintiffs' expert over that of defendants' expert. Plaintiffs' expert, Dr. Wu, performed his analysis in keeping with this Court's instructions that market analysis must take into account market realities and the dynamics of competition. By contrast, Dr. Harris's analysis for defendants was, as the district court found, flawed by its use of assumptions in place of factual showings regarding the key matters in controversy. The court correctly concluded that defendants' analysis was "unsupported by the evidence" and "inconsistent with the economic realities of Southeast Missouri." Add. 21a. (Part II.B.) ARGUMENT Standard of Review The decision of the district court preliminarily to enjoin Tenet's proposed acquisition of DRMC is subject to a highly deferential standard of review. See FTC v. Freeman Hospital, 69 F.3d 260, 267 (8th Cir. 1995). As this Court has recognized, district courts exercise "broad discretion" in deciding whether to grant or deny a motion for a preliminary injunction. See, e.g., United Industries Corp. v. Clorox Co., 140 F.3d 1175, 1179 (8th Cir. 1998). Appellate review is limited to ensuring that the district court did not rely on any clearly erroneous findings of fact or make an error on an issue of law. See West Publishing Co. v. Mead Data Central, Inc., 799 F.2d 1219, 1222-23 (8th Cir. 1986); St. Jude Medical, Inc. v. Carbomedics, Inc., 764 F.2d 500, 501 (8th Cir. 1985). Under the deferential standard applied by this Court, the district court "possesses the sole discretion" to decide when sufficient evidence has been presented to warrant injunctive relief. Freeman, 69 F.3d at 270 n.14. The "clearly erroneous" standard is satisfied if the court's factual findings are "plausible in light of the record viewed in its entirety,"(5) "supportable,"(6) or not "illogical or implausible."(7) Whether based on documentary evidence, factual inferences, or determinations of witness credibility, "[w]here there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Anderson v. City of Bessemer City, 470 U.S. at 574. This deferential standard "'rests upon the unique opportunity afforded the trial court judge to evaluate the credibility of witnesses and to weigh the evidence.'" United Industries, 140 F.3d at 1179 (quoting Inwood Lab., Inc. v. Ives Lab., Inc., 456 U.S. 844, 855 (1982)). Moreover, it is "axiomatic that a district court has the discretion to evaluate the credibility of expert witnesses and accept the testimony it finds most plausible." Freeman, 69 F.3d at 269 n.13. The principal issue for this Court to resolve on appeal -- whether the record supports the district court's finding that the Poplar Bluff Region is a "credible relevant market" within which to assess the competitive impact of Tenet's proposed acquisition of DRMC (Freeman, 69 F.3d at 268) -- presents a question of fact. See Community Publishers, 139 F.3d at 1184; Freeman, 69 F.3d at 270; White Industries, Inc. v. Cessna Aircraft Co., 845 F.2d 1497, 1501 (8th Cir.), cert. denied, 488 U.S. 856 (1988); Alexander v. National Farmers Organization, 687 F.2d 1173, 1192 (8th Cir. 1982); see also General Industries Corp. v. Hartz Mt. Corp., 810 F.2d 795, 805 (8th Cir. 1987) ("determination of relevant product market is a fact question," and so must be upheld "as long as there is evidence to support the jury's verdict"). In the absence of "clear error," the district court's delineation of the relevant geographic market must be upheld. I. THE DISTRICT COURT APPLIED THE CORRECT LEGAL STANDARD IN IDENTIFYING THE POPLAR BLUFF REGION AS THE RELEVANT GEOGRAPHIC MARKET The district court's carefully reasoned decision adheres rigorously to the analytical framework this Court uses to determine the relevant geographic market. See Freeman, 69 F.3d at 268-71; Bathke v. Casey's Gen. Stores, Inc., 64 F.3d 340, 345-47 (8th Cir. 1995); Morgenstern, 29 F.3d at 1296-97. The court recognized that plaintiffs bore the burden of identifying a "'credible relevant market,'" and concluded that they had done so, with evidence that was "credible, logical, and persuasive." Add. 10a (quoting Freeman, 69 F.3d at 268), 13a. Focusing on the forward-looking inquiry this Court has prescribed, the district court considered whether other hospitals would be "practical alternatives" that could defeat an effort by Tenet to extract higher prices. Add. 12a, 15a, 17a. And the court -- recognizing that no single type of evidence can provide the entire answer to this inquiry -- properly weighed the "evidence as a whole" in reaching its ultimate factual conclusion. Add. 13a. Accordingly, the only proper question in this appeal (see Part II, infra) is whether the court clearly erred in its factual findings. No doubt prompted by the deferential standard of review accorded factual findings, defendants strain to manufacture issues of law on which to base their appeal, going so far as to accuse the court below of "fundamental legal error." Def. Br. 16. These arguments are baseless, resting alternately on defendants' mischaracterization of this Court's prior rulings on the evidentiary standards for market definition and a refusal to acknowledge the character of the district court's reasoning. A. Geographic Market Definition Requires a Fact-Driven Analysis of the Commercial Realities Faced by Consumers Section 7 of the Clayton Act prohibits any merger or acquisition "where in any line of commerce in any section of the country, the effect of such acquisition may be substantially to lessen competition or to tend to create a monopoly." 15 U.S.C. § 18. The statute protects consumers from acquisitions or mergers that may create or enhance market power -- i.e., the ability of one or more firms to raise prices above competitive levels for a significant period of time. See Community Publishers, 139 F.3d at 1183; United States v. Archer-Daniels-Midland Co., 866 F.2d 242, 246 (8th Cir. 1988), cert. denied, 493 U.S. 809 (1989). To evaluate a particular transaction's effect on competition, the courts first identify the relevant market within which any post-transaction market power might be exercised. See United States v. Marine Bancorporation, Inc., 418 U.S. 602, 618-23 (1974); Freeman, 69 F.3d at 268; see also U.S. Dept. of Justice & Federal Trade Commission, Horizontal Merger Guidelines, 4 Trade Reg. Rep. (CCH) ¶ 13,104 (1997) ("Merger Guidelines"), DA 470-87.(8) While the product market defines the relevant product or service at issue,(9) the geographic market takes into account the locations of sellers of the product. It encompasses "the area to which consumers could practicably turn for alternative sources of the product and in which the antitrust defendants face competition." Morgenstern, 29 F.3d at 1296; see Freeman, 69 F.3d at 268 (question is where buyers "could practicably turn for alternative sources of the product should * * * [post-merger] prices become anticompetitive"). As described in the Merger Guidelines, it is the area in which a hypothetical monopolist could impose a "small but significant and nontransitory" price increase without the price increase being defeated by consumers switching to other firms. Merger Guidelines § 1.21, DA 475-76. In other words, market definition requires a dynamic, forward-looking inquiry that asks how consumer behavior would change in response to a post-merger price increase. Answering this dynamic question requires a "highly fact-driven analysis." Freeman, 69 F.3d at 271 n.16. Moreover, as this Court observed in Freeman (id.), the Supreme Court has emphasized that "Congress prescribed a pragmatic, factual approach to the definition of the relevant market and not a formal, legalistic one." Brown Shoe Co. v. United States, 370 U.S. 294, 336 (1962). The issue is to be addressed by focusing on the realities of the market, and not on mere theoretical possibilities. See Community Publishers, 139 F.3d at 1184 ("determinations of [relevant market] are necessarily fact intensive and must be based on evidence that describes real markets and not hypothetical ones"). There must be a factual inquiry into the "commercial realities" and "practical alternatives" faced by "consumers." See Eastman Kodak Co. v. Image Technical Servs., 504 U.S. 451, 482 (1992); Bathke, 64 F.3d at 345; Freeman, 69 F.3d at 271 n.16. B. The District Court Properly Considered a Broad Range of Evidence, Statistical and Nonstatistical, in Making its Findings on Market Definition Because the "decisive question" is "where consumers could practicably go for alternative sources of acute care inpatient hospital services" and not theoretical possibilities, historic or current patient flow data "[is] not, by itself" sufficient to delineate the relevant market. Freeman, 69 F.3d at 269 (emphasis added); see Bathke, 64 F.3d at 345-46; Morgenstern, 29 F.3d at 1296. Consistent with those rulings, the court below recognized that statistical evidence regarding where consumers currently go for hospital services cannot provide a "definitive answer" to the question at hand, and accordingly based its ruling on "the evidence as a whole," including both statistical and nonstatistical evidence. Add. 12a-13a. Defendants now attempt to bootstrap this entirely appropriate observation by the district court into a failure of proof on plaintiffs' part, arguing that the injunction should have been denied because this one type of evidence did not conclusively prove plaintiffs' case. Def. Br. 29 (citing Freeman, 69 F.3d at 268 & n.12). Defendants' notion that an antitrust market must be proved solely by statistical evidence is utterly without support in the decisions of this or any other court. In Freeman, the Court upheld a district court's factual determination that the FTC had failed to meet its burden of proof on geographic market, rejecting arguments that the district court had improperly ignored or discounted certain evidence. 69 F.3d at 269-72 & n.14. Nowhere, however, did the Court reject any particular type of evidence as "irrelevant" (cf. Def. Br. 30), or suggest that any single type of evidence must be conclusive by itself. On the contrary, the Court emphasized "the broad range of evidence that may be of value in determining a geographic market." 69 F.3d at 271. That holding is entirely in keeping with the Supreme Court's teaching that market definition must be based on a "pragmatic" inquiry into the "commercial realities faced by consumers." See Brown Shoe, 370 U.S. at 336; Eastman Kodak Co. v. Image Technical Servs., 504 U.S. at 482. The Merger Guidelines likewise emphasize the need for judgments based on "all relevant evidence." See §§ 0, 1.11, 1.21, DA 471, 474, 476. Defendants' repeated criticism (Def. Br. 13, 21, 29) of the district court for relying upon "impressionistic claims of market participants" or upon so-called "anecdotal evidence" (i.e., direct evidence) not only lacks any basis in this Court's rulings, but is wholly illogical in light of defendants' own recognition that the requisite analysis is dynamic and forward-looking. Def. Br. 17. Statistical analysis of historic patient flow data -- whether the analysis used by plaintiffs' expert as a starting point, or the "contestable zip code" analysis on which defendants' expert based his conclusions -- necessarily "provide[s] only a static picture of the market." Add. 17a (citing Freeman, 69 F.3d at 269). The district court properly considered such statistics "in conjunction with" other evidence -- e.g., the direct testimony of market participants, such as the Poplar Bluff Region employers and other large purchasers of hospital services, who are the very ones who will decide how they can practicably respond to post-merger price increases by Tenet. Add. 13a. This Court has recognized such testimony of market participants as not only "relevant" but also "important" to a determination of a proper geographic market. Freeman, 69 F.3d at 270; see Morgenstern, 29 F.3d at 1297. Such views as to current and future market conditions, as well as other current and forward-looking evidence, address "the commercial realities faced by consumers," and are highly pertinent to a proper, dynamic analysis. See Bathke, 64 F.3d at 345; Freeman, 69 F.3d at 270 & n.14; Morgenstern, 29 F.3d at 1297. Defendants' further assertion that the district court somehow relieved plaintiffs of their burden of proof as to market definition (Def. Br. 28-30) is utterly baseless. The district court carefully followed this Court's directives in Freeman, and kept the burden squarely on plaintiffs. Add. 10a. Moreover, the court elsewhere clearly demonstrated its understanding of the different burdens borne by plaintiffs and defendants, noting in a separate analysis that defendants "failed to meet their burden of proving either affirmative defense." Add. 25a. Defendants nonetheless insist that the court "effectively" relieved plaintiffs of their burden to prove a "credible market," simply because it took into account nonstatistical evidence and observed that defendants' theoretical model did not refute plaintiffs' broad-based factual showing. See Def. Br. 29.(10) Such an argument is nothing more than a transparent attempt to fabricate an issue of law; all that is truly at issue therefore is the district court's resolution of factual issues on a contested record. C. The District Court Was Not Required to Quantify Likely Consumer Responses to a Price Increase Defendants also insist that the district court erred by supposedly not taking into account the possibility that shifts in hospital use by a limited number of "marginal" consumers (e.g., from Poplar Bluff hospitals to those in Cape Girardeau or elsewhere) could undermine Tenet's ability to sustain a post-merger price increase. They complain that the court below "plainly misunderstood" their "critical loss" analysis, and therefore failed to answer the "relevant question." Def. Br. 23. In truth, however, it is defendants who plainly misunderstand the district court's ruling, and who seek to obscure its cogent analysis with an artificial and formulaic approach. There was no dispute in the court below -- nor is there here -- that the geographic market inquiry requires consideration of whether a post-merger price increase would cause enough patients to use alternative hospitals to render the price increase unprofitable. As shown in Part II, infra, plaintiffs made an extensive factual showing, uncontradicted by credible evidence from defendants, that it was unlikely there would be such defections. With all of the evidence before it, and aware of defendants' "critical loss" calculations,(11) the district court concluded that sufficient numbers of consumers in the Poplar Bluff region would not practicably turn to these larger [outside] hospitals for acute care services in the event of a price increase by Lucy Lee and DRMC. Add. 15a (emphasis added); cf. Tr. 430 (Wu), PA 133 (stating similar conclusion). Defendants flatly err, therefore, in supposing that the district court failed to consider the potential impact of "marginal customers." Def. Br. 28. Defendants' claim that the court "needed to have begun its analysis by determining the number of patients who would have to leave order to make a price increase unprofitable" (Def. Br. 22) is incorrect and contrary to their own expert's view. See DA 218; Def. Exh. Z, PA 297a. Defendants offer no support for the artificial notion that one must approach this issue by first calculating the hypothetical number of "lost" customers who could defeat a price increase, and then determining whether such a loss would occur.(12) Indeed, elsewhere in their brief (Def. 18), they acknowledge that the first question posed by their expert in defining a market is "could some consumers switch to alternative hospitals if prices were to increase." In the present case, the district court addressed that question. It considered the overwhelming evidence that even a 10 percent Tenet price increase would not induce consumers to switch their business, and concluded that "sufficient numbers" therefore could not practicably turn elsewhere.(13) Add. 15a. Defendants provide no authority for the proposition that the court was then required to engage in the fruitless exercise of quantifying some threshold level of shifting that would make a price increase unprofitable. Such a requirement has no support either in case law(14) or the Merger Guidelines, and is plainly contrary to the "pragmatic" approach that courts take to market definition. See Freeman, 69 F.3d at 271 n.16 (quoting Brown Shoe, 370 U.S. at 336-37); see also General Industries Corp. v. Hartz Mt. Corp., 810 F.2d at 805 (market definition is "merely an aid for determining whether market power exists"). Defendants are simply incorrect when they equate rigorous analysis with the sort of simple mathematical calculation they embrace. II. THE DISTRICT COURT DID NOT CLEARLY ERR IN FINDING THAT THE POPLAR BLUFF REGION IS THE RELEVANT GEOGRAPHIC MARKET The district court arrived at its factual findings after a careful analysis of the extensive body of evidence amassed at a five-day evidentiary hearing, including testimony from a wide range of witnesses with first-hand knowledge of commercial realities regarding hospital services in the Poplar Bluff Region, the testimony of expert witnesses on both sides, and numerous exhibits and deposition excerpts. Weighing that evidence "as a whole," the court concluded that plaintiffs' evidence was "more credible, logical, and persuasive." Add. 13a. Defendants contend that witnesses addressed the "wrong question," that the court failed to consider the "marginal consumer," and that the approach favored by their expert was not followed. Def. Br. 30-44. None of these arguments can refute the overwhelming evidence that hospitals outside the Poplar Bluff Region could not constrain a price increase, nor do they undermine the district court's conclusion that defendants' version of the facts "is inconsistent with the economic realities of Southeast Missouri." Add. 21a. A. The Evidence Amply Supports the District Court's Finding That Sufficient Numbers of Consumers Could Not Practicably Turn to Hospitals Outside the Poplar Bluff Region The district court's factual findings are based on a comprehensive range of the kinds of evidence this Court has held relevant to "the critical question of where consumers of acute care inpatient hospital services could practicably turn" in the event of a monopolistic price increase. See Freeman, 69 F.3d at 268. As set forth above (pp. 6-13), plaintiffs presented not only statistical evidence that reflected both current use patterns and reasons for which residents go to outside hospitals, but also direct evidence of the factors that market participants will necessarily take into account in making their decisions in response to changes in the Poplar Bluff hospital market. That evidence -- including testimony from the very people who will be making such decisions -- shows that employers and health plans would not attempt to steer employees to hospitals outside the Region. The district court, weighing this evidence in light of the mechanisms by which economic decisions are made in this market, had ample basis for its conclusion that "sufficient numbers" of consumers would not turn to hospitals outside the Poplar Bluff Region as practical alternatives. See Add. 15a. 1. A key commercial reality in this market is the pivotal role of employers and health plans as market intermediaries Geographic market definition is determined by a factual inquiry into the "commercial realities faced by consumers." Bathke, 64 F.3d at 345. The district court carefully took such realities into account, recognizing the importance of managed care buyers as a mechanism for competition in this industry (Add. 7a), and factoring in "the economic realities of Southeastern Missouri" (id. at 21a). Defendants, on the other hand, ignore those realities. Throughout their brief to this Court -- and particularly in their emphasis on "marginal consumers" (e.g., Def. Br. 42) -- defendants discuss the market for hospital services as if it were the market for gasoline stations, in which consumers can easily compare prices displayed on outdoor signs and, unaffected by external constraints, drive a few miles to save a few cents per gallon. Cf. Bathke, 64 F.3d at 345-46. As plaintiffs' expert explained, however, the individual health care consumer faces a situation that is different from a few other industries. * * * [W]hen a person wants to switch gas stations he or she literally can drive to the next gas station. There's no need to call a third party payer. There's no need to get approval from anybody * * *. Tr. 367-68 (Wu), PA 97-98. Indeed, the expert witnesses for both sides agreed that health plans are important intermediaries in the market for hospital services. While individual patients are "the direct beneficiaries of the medical services provided by hospitals" (PX 97 at 5, PA 356), there is "not a direct economic relationship between an individual consumer and the hospital." Tr. 361 (Wu), PA 93. As Dr. Wu further explained, "[c]onsumers literally don't get up and admit themselves to a hospital." Id. On the contrary, the process by which individual patients become admitted to a particular hospital is a complex one, entailing considerations including existing physician relationships and physician admitting privileges. See Tr. 367 (Wu), PA 97; PX 16 ¶¶ 5, 7, PA 315; Freeman, 69 F.3d at 270. Economic considerations are generally focused on "what health plan [the patient] is in and what the * * * financial incentives are to go to a particular hospital." Tr. 367 (Wu), PA 97. Defendants' expert agreed with this assessment, recognizing that "consumers or patients make their decisions based on those prices" that managed care plans "create" for them, and not on actual hospital charges: They don't have any idea what the hospital price is. That's an irrelevancy from that point of view or maybe irrelevancy is too strong. What's important to them is their out-of-pocket costs. Tr. 910 (Harris), PA 201; see also Tr. 42 (Miller), PA 5. For all of these reasons, "the entities with whom hospitals interact closely on issues such as price and quality of care are primarily third party payors." PX 97 at 5 (emphasis added), PA 356. The price considerations that individual patients face regarding hospital choice are thus largely shaped by the decisions of their employers and health plans. The actions of those payers are the ones that truly drive competition in this context, and on which the district court properly focused. Add. 14a-15a; see Freeman, 69 F.3d at 270 n.14. 2. Extensive evidence demonstrates that purchasers of hospital services would not be able to avoid a monopolistic price increase by Tenet At the preliminary injunction hearing, the district court received evidence from a broad range of employers and health plans in the Poplar Bluff Region showing that they could not practicably defeat a post-merger price increase by Tenet by steering individuals to hospitals outside the Region. See pp. 6-9, supra. Indeed, defendants' sole employer witness conceded that his firm would not even attempt to steer its employees to Cape Girardeau if the Poplar Bluff hospitals increased their prices. Tr. 564-65 (Hall), PA 165-66. Rather than try to steer people outside the Region to discipline a merger-driven price increase, employers would either absorb the increased costs as a business expense, or pass at least a portion of it on to their employees in the form of higher premiums. See, e.g., Tr. 100 (Reynolds), PA 21; Tr. 169 (Thomas), PA 44; Tr. 207 (Clark), PA 54; PX 23 ¶ 15, PA 326. The record shows a number of reasons why employers would not try to steer their employees to hospitals outside the Region in order to defeat increased post-merger prices by Tenet. Employers testified, for example, that they would encounter substantial employee resistance if they tried to implement a benefits plan that would penalize their work force for using local providers. See, e.g., Tr. 102 (Reynolds), PA 23; Tr. 138-39 (Anderson), PA 35-36. Employers say that a health benefits plan is an important tool in attracting a qualified work force (e.g., Tr. 244-45 (Wicklund), PA 60-61), and are not willing to risk employee discontent by switching to a plan with features that would be unpopular with their work force. See, e.g., Tr. 102, 123 (Reynolds), PA 23, 27; Stranglen Dep. 92-93, PA 279-80. Distance and convenience are always important factors in identifying the relevant geographic market, (see Morgenstern, 29 F.3d at 1297), but they are particularly significant in the Poplar Bluff Region. As the owner of two local companies explained: You have to realize we're in a remote area. We're not around any city and we don't have any main highways and we don't have any railroads, we don't have a means of getting to the larger facilities. Tr. 251 (Wicklund), PA 65; see PX 16 ¶ 7, PA 315; PX 73 ¶ 7, PA 345. Even the CEO of one of the merging hospitals testified that he would not travel to Cape Girardeau for medical care because of the distance, even if his co-payment increased by 10 percent. Kelly Dep. 85-86, PA 255-56. Physician affiliation patterns pose yet another impediment. Because few if any physicians in Poplar Bluff have admitting privileges at hospitals outside the Region (e.g., Tr. 405 (Wu), PA 128; Tr. 780 (Li), PA 188; Tr. 249 (Wicklund), PA 64), many patients would have to switch physicians to be admitted at one of the more distant alternatives -- a requirement that would be highly unpopular with the work force. See, e.g., PX 14 ¶ 8, PA 312; PX 73 ¶ 7, PA 345; Freeman, 69 F.3d at 270 n.14 (noting that "patients may be limited in the hospitals to which they can practically turn" if their physicians lack privileges at alternative hospitals). The evidence from health plans -- the other major category of health care buyers -- portrays a similar picture of the difficulties they would encounter in attempting to steer their members to alternative hospitals outside the Region. Plan administrators testified that they could not market a plan in the Poplar Bluff Region without a Poplar Bluff hospital. E.g., Tr. 51 (Miller), PA 10; Tr. 138-39 (Anderson), PA 35-36; PX 58 ¶ 8, PA 335; see also Tennison Dep. 93, PA 283. Contrary to defendants' suggestions (e.g., Def. Br. 34), the foregoing considerations cannot be dismissed as mere "anecdotes" about the "current habits" of Poplar Bluff Region residents. The key economic decisions that will determine whether Tenet can exercise monopolistic power are the decisions that these employers and health plans will make. Those decisions, moreover, must necessarily be based on the judgment of these actors as to what their employees or members will accept. Direct evidence of what those judgments will be, and the reasoning behind them, provided the district court with a firm understanding of key realities of this market. Furthermore, the uncontradicted record testimony is that there would be no financial incentive for payers to steer employees to the closest comparable Missouri hospitals outside the Region, those in Cape Girardeau and Sikeston, because those hospitals are vastly more expensive than the Poplar Bluff hospitals. Tr. 55 (Miller), PA 13, Tr. 313 (Ressel), PA 79; Tr. 395-403, 529-30, 1166 (Wu), PA 119-27, 149-50, 158; PX 97 at 8, 24-26, PA 359, 375-77. Indeed, as compared to Tenet's Lucy Lee, prices to third party payers in Cape Girardeau are as much as 84 percent higher for surgical services and 37 percent higher for medical services, and the difference between DRMC and the Cape Girardeau prices is even higher. Tr. 396-97 (Wu), PA 120-21. As Dr. Wu discussed, the enormous cost differential between the low cost hospitals in Poplar Bluff and these alternatives means that Poplar Bluff hospitals would still be cheaper even after a 10 percent price increase. Tr. 395-403 (Wu), PA 119-27; PX 392-95, PA 727-30. While defendants attempt to gloss over this point (e.g., Def. Br. 27), it is a significant aspect of the "economic realities of Southeast Missouri" (Add. 21a) that the district court considered.(15) Therefore, defendants cannot fairly charge the district court with simply "assum[ing] that payors would not or could not change deductibles or utilize other incentives to discourage the use of the Poplar Bluff hospitals." (Def. Br. 32 n.19) On the contrary, the court based its findings on testimony that rested on sound logic, given the evidence of economic realties in the Poplar Bluff Region. 3. Defendants' arguments regarding "marginal consumers" ignore the realities of the market Defendants attempt to attack the foregoing evidence by contending that these witnesses were asked the "wrong question," and that the district court ignored the "marginal consumer." Def. Br. 30-32. Defendants suggest that employers could be expected to make different arrangements for some of their employees, and that special treatment of enough such employees could defeat Tenet's prospect of monopoly power. Id. These arguments are doubly flawed, in light of the factors just discussed. First, employers face significant administrative costs that make it impractical for them to provide such individualized treatment. Even larger employers -- who could most easily absorb the costs of retooling a benefits package -- prefer a single plan that will satisfy most employees and therefore would not be willing to bear the cost and administrative burden of making changes in a plan in response to price movement in a single market or to accommodate employees who live at the edge of the market. See Tr. 432-33, 531, 1165 (Wu), PA 135-36, 151, 157. Thus, even if employees in a particular area -- e.g., Stoddard County -- would be willing to use hospitals in Cape Girardeau, it would be impractical for employers to make such accommodations. Tr. 433 (Wu), PA 136. In other words, the economics of the group health plans in this market mean that employers provide the same plan to all their employees and thus gear their health plans to the needs and preferences of the "average consumer." Second, even if some employers were able and inclined to make different provisions for employees living in different locations, the large price differential between Poplar Bluff hospitals and Cape Girardeau hospitals means that employers could not save any money by shifting any of their employees to the higher-priced hospitals outside the Region. Defendants make much of the ability of employers to "steer" employees to particular hospitals (Def. Br. 27-28 n.14, 33), but lose sight of the obvious fact that managed care buyers engage in such steering to save money, not to send members to more expensive hospitals. Thus, the evidence presented below regarding the dynamics of the market for hospital services shows why defendants' "marginal consumer" arguments are misplaced. As the district court understood, employer-based and other health plans play a pivotal role in this market, providing the key signals regarding price to individual consumers. The evidence outlined above, moreover, shows that those plans cannot be expected to provide incentives to any of their members in the Poplar Bluff Region to switch to the nearest alternative comparable hospitals.(16) Finally, any further contention that individual consumers would make such a switch in response to a 5 or 10 percent increase in Poplar Bluff prices -- even in the absence of changes in their benefits plans -- is, like defendants' other arguments, unsupported by the evidence and inconsistent with commercial realities. Several conditions would have to be met for such a hypothesis to work, including the following. First, in order to calculate their out-of-pocket costs, such hypothetical "marginal" consumers would have to be not only knowledgeable about relative hospital prices, but also able to predict in advance what services they would require. Second, these consumers would have to be so cost-sensitive that they would switch hospitals (despite physician loyalty and affiliation or similar considerations) because of a 5 to 10 percent increase in hospital charges -- for patients with a typical 10 to 20 percent co-payment, a $20 to $80 increase in out-of-pocket costs. See Tr. 1167 (Wu), PA 158a. Third, the same hypothetical consumers who are so price conscious that they would respond to a $20 to $80 increase in their hospital bill would also have to be willing to pay far more to go to the closest comparable alternative hospitals in Cape Girardeau and Sikeston. Such a scenario, implausible on its face and unsupported by the record, cannot undermine the district court's fact-based determination. As this Court has noted, market definition "must be based on evidence that describes real markets, not hypothetical ones." Community Publishers, 139 F.3d at 1184. 4. The views of other market participants, including the merging hospitals, confirm that the Poplar Bluff Region is the relevant geographic market The views of other market participants -- which present a similar picture of the market -- are also highly relevant. See Morgenstern, 29 F.3d at 1297; Freeman, 69 F.3d at 270. As the district court observed (Add. 14a-15a), hospital administrators outside the Region do not expect an increase in their acute care patient admissions in the event of a 10 percent post-merger price increase by the Poplar Bluff hospitals. See PX 58 ¶ 8, PA 335; PX 73 ¶ 7, PA 345; PX 78 ¶ 4, PA 350. This evidence is fully consistent with a Poplar Bluff Region geographic market. The record shows that, for most purposes, Lucy Lee and DRMC view each other as their only significant competitors. While Lucy Lee and DRMC each have tried to set their prices to avoid losing patients to the other, the record does not reflect a similar concern with the prices at hospitals 50 or more miles away. See p. 10, supra. In fact, as part of its agreement to purchase DRMC, Tenet prohibited DRMC's physician-owners from operating a competing hospital within a 50-mile radius of Poplar Bluff. PX 7 ¶ 14.10(c), PA 305. Lucy Lee's and DRMC's negotiating strategies with health plans reflect a similar focus. Most significantly, while each hospital has tried to secure "exclusive contracts" that exclude the other from the plan's "preferred provider" network, there have not been similar attempts to exclude hospitals outside the Region. See, e.g., Rouff Dep. 113-14, PA 273-74. These economic assessments and practices of the parties are strong additional evidence of the market's definition. See, e.g., FTC v. Staples, Inc., 970 F. Supp. 1066, 1079-80 (D.D.C. 1997) (enjoining merger where parties considered each other their primary competitors in the relevant market). Together with the other evidence discussed above, they provide a solid basis for the district court's factual findings. B. The District Court Properly Credited the Opinion of Plaintiffs' Economic Expert Over That of the Defendants In another attempt to find reversible error, defendants criticize the district court's reliance on plaintiffs' expert witness, and attack the court's failure to credit the opinion of defendants' expert. Def. Br. 41-44. Defendants bear an especially heavy burden in asking this Court to overturn the district court on such grounds, in light of the Court's recognition of the broad discretion of a trial court to evaluate expert evidence and accept "the testimony it finds most plausible." Freeman, 69 F.3d at 269 n.13. On the present record, the district court acted well within its discretion accepting plaintiffs' expert evidence, and in recognizing the many critical flaws in defendants', rejecting the latter as "inconsistent with the economic realities of Southeast Missouri." Add. 21a. Defendants attack the analysis of plaintiffs' expert, Dr. Wu, by contending that he used a "static" analysis of a type rejected by this Court in Freeman. That argument is both factually and legally wrong. First, as both Dr. Wu and the district court recognized, his statistical analysis of patient flow data was simply a "proper first step" in the analysis. Add 17a; see Tr. 374-76 (Wu), PA 99-101. As shown above, moreover, Freeman did not "reject" any type of evidence, but simply held that evidence going beyond such a starting point is necessary. 69 F.3d at 270-71. Dr. Wu followed this approach precisely, by going on to consider additional direct evidence (ignored by defendants) regarding the courses of action practicably open to actual market participants. He also considered a variety of statistical evidence that was consistent with the testimony that Poplar Bluff Region consumers who use hospitals outside the Region generally do so to obtain more complex, tertiary care, unavailable in Poplar Bluff. See Add. 15a-17a. With respect to their own expert, defendants contend that the district court abused its discretion, committed factual error, and simply misunderstood his analysis. Def. Br. 14, 23, 41. In fact, however, the court rejected his analysis because it recognized that Dr. Harris's statistical model is "problematic" and that the model's conclusions are "unsupported by the evidence" and "inconsistent with the economic realities of Southeast Missouri." Add 19a, 21a. The record shows that these assessments were well within the district court's discretion. Throughout their brief, defendants place great emphasis on calculations that purport to show that the switch of a "very small number" of patients would defeat a post-merger price increase. It is useful to place defendants' "small numbers" in context. Defendants state that a 5 percent price increase would be defeated by "only" a switch of two patients every five days. E.g., Def. Br. 22, 31. Even on defendants' own terms, however, the magnitude of the assumed shift is large in proportional terms; defendants assume that fully 8 percent -- one out of every 12 commercially-insured patients of Lucy Lee and DRMC -- would decide to go to a distant, more expensive hospital because of a 5 percent price increase. For reasons previously discussed, however, even this figure greatly understates the number of defections that would be necessary to defeat the far more likely 10 percent price increase. See pp. 13, 32 n.13, supra. As defendants' own expert acknowledged, a 10 percent price increase would remain profitable unless 15 percent of Lucy Lee and DRMC's commercially-insured patients -- nearly one in six -- changed to distant hospitals outside the Poplar Bluff Region. Tr. 916 (Harris), DA 645. Defendants also gloss over the mechanism by which such a market shift would have to occur. As shown above, both sides' experts agreed that individual health care users are guided by the terms of their health plans. Because the vast majority of individuals are not admitted to a hospital in any given year, the shift of "two every five days" would in fact require a shift in the health coverage of a far greater number of covered individuals. As shown above, the uniform testimony of the employers and health plans who would be making such decisions shows that such a change could not practicably be effected. Ignoring the acknowledged mechanism by which individuals make health care decisions, defendants' expert relied upon a telephone survey of residents of the Region -- a matter that defendants omit from their detailed description of his methodology. See Def. Br. at 5-7. But the survey was "wholly unhelpful" and "not probative" because, in addition to being riddled with technical errors, it provided no insight into the likely reaction of patients to a price increase, failed to account for the influence of third-party payers on hospital services purchased, and failed to control for other variables that may influence a consumer's choice of hospitals. Add. 20a-21a.(17) Lacking any evidence to make their statistical analysis of patient flow data useful under this Court's geographic market inquiry (Freeman, 69 F.3d at 270-71), defendants are left with untested and unsupported assumptions about likely consumer reaction to a price increase in the Poplar Bluff hospitals. Specifically, defendants rely on patient flow data to show that some of the patients in the hospitals' service area (i.e., those living in so-called "contestable zip codes") have gone somewhere other than Lucy Lee or DRMC for inpatient care. According to defendants' expert, if some patients are already going outside the Region for health care, their neighbors would also switch to hospitals outside the Region in response to a price increase in Poplar Bluff. See Def. Br. 7. The record shows, however, that "patients are hospitalized at far away hospitals for a variety [of] reasons." Tr. 378 (Wu), PA 103. For example, emergency admissions, transfers between hospitals, and admissions for services not available locally (e.g., tertiary care) are all examples of patient flow that "one would not expect to change if prices were to increase in the Poplar Bluff region." Tr. 379 (Wu), PA 104; see also id. at 378-79, PA 103-04. Indeed, the record shows and the district court found that the current outflow from Poplar Bluff is predominantly driven by a need to obtain higher level care not available in Poplar Bluff. Add. 18a-19a; see PX 14 ¶ 8, PA 312; PX19 ¶ 5, PA 319-20; PX 78 ¶ 4, PA 350; PX 97 at 21, PA 372; Tr. 1129-43 (Steinwald), PA 226-40; PX 98, PA 417-70. This conclusion is supported, for example, by evidence that patients from the Region who go to hospitals in Cape Girardeau and St. Louis are disproportionately admitted by specialists.(18) Tr. 384, 394 (Wu), PA 108, 118; PX 97 at 22, PA 373. Thus, as the district court held, it was "improper" for defendants simply to assume that patients who need primary and secondary care would respond to a price increase in Poplar Bluff by traveling to distant hospitals outside the Region. Add 20a. See Tr. 1162 (Wu), PA 154 ("That's something that [Dr. Harris] assumes. It's not something that you could actually draw from the data.") Contrary to defendants' contention (Def. Br. 43), the district court's rejection of Dr. Harris's contestability analysis for its failure to eliminate patients who were leaving the market for services that are unavailable in Poplar Bluff was not premised on a "key factual error." The court found, based on the factual record, that two hospitals may treat patients that are classified in the same DRG even though they do not provide the same services to those patients. Add. 17a-18a. Thus, the district court concluded correctly that merely excluding (as Dr. Harris did) "patients who were traveling to different hospitals for treatment in DRGs that [are] not available in the Poplar Bluff hospitals" (id.) does not eliminate from the analysis those patients who use hospitals outside the area for services that are unavailable in Poplar Bluff. Thus, the court below was eminently justified in rejecting defendants' theoretical model as unsupported by evidence. See Eastman Kodak Co., 504 U.S. at 468 (a party is not entitled to prevail merely because it "enunciates any economic theory supporting its behavior, regardless of its accuracy in reflecting the actual market"); cf. St. Louis Convention & Visitors Comm'n v. National Football League, 1998-2 Trade Cas. (CCH) ¶ 72,258, at 82,655 (8th Cir. Sept. 3, 1998) (unsupported economic theory cannot prove causation). Defendants' factual presentation accordingly did nothing to refute plaintiffs' showing -- based on a proper understanding of the mechanisms by which the market works and the economic incentives faced by all participants -- that a price increase by the combined Tenet hospitals in Poplar Bluff would lead to little or no increase in patient outflow, and certainly not enough to defeat Tenet's ability to sustain such a post-merger price increase. See Tr. 430-33 (Wu), PA 133-36; Add. 15a. The district court correctly ruled that the Poplar Bluff Region constitutes a credible market for primary and secondary hospital services, in which the proposed acquisition threatens to replace vigorous competition with a virtual monopoly. CONCLUSION For all the foregoing reasons, the judgment of the district court should be affirmed. Respectfully submitted, JOHN D. GRAUBERT JOHN F. DALY LESLIE RICE MELMAN Federal Trade Commission 1. Citations to the district court's opinion use the page numbers used in the Addendum to the brief for Defendants-Appellants (Add.). Record materials contained in the Appendix filed by Defendants-Appellants are cited "DA ___"; materials set forth in the separate Appendix for the Plaintiffs-Appellees filed herewith are cited "PA ___." "PX ___" refers to plaintiffs' exhibits in the court below. "Def. Exh. __" refers to defendants' hearing exhibits. 2. A map showing the bounds of the market as proposed by plaintiffs and found by the district court is included as an addendum to this brief. 3. The four small hospitals located within the Region also stated that a postmerger price increase would not cause them to gain additional patients. PX 65 ¶ 10, PA 339; PX 69 ¶ 8, PA 341; Koppeis Dep. 78-79, PA 297b-297c; PX 70 ¶ 6, PA 342-43; PX 75 ¶ 6, PA 348. The district court addressed the evidence regarding these hospitals in its analysis of competitive effects, and concluded that they could not constrain the merged entity from raising prices. Add. 23a. 4. The Diagnosis Related Group ("DRG") system assigns a code to inpatient diagnoses and is used by hospitals and payers to calculate reimbursement. Add. 17a. 5. Anderson v. City of Bessemer City, 470 U.S. 564, 574 (1985). 6. West Publishing Co. v. Mead Data Central, Inc., 799 F.2d at 1227, 1229. 7. Anderson, 470 U.S. at 577; St. Jude Medical, 764 F.2d at 501. 8. The Merger Guidelines are intended to provide guidance to businesses and the general public on how the antitrust enforcement agencies evaluate horizontal mergers. They are not binding on the courts, but many courts have relied on or used them in assessing the legality of a proposed transaction. See, e.g., Archer-Daniels-Midland, 866 F.2d at 244; Community Publishers, Inc. v. Donrey Corp., 892 F. Supp. 1146, 1153 n.6 (W.D. Ark. 1995), aff'd, 139 F.3d 1180 (8th Cir. 1998). 9. The relevant product market in this case is not in dispute. It encompasses general acute care inpatient hospital services (sometimes also described as primary and secondary inpatient services), but not higher level, tertiary care such as open heart surgery. Add. 11a; see, e.g., Freeman, 69 F.3d at 268. Tertiary services are provided in a broader geographic market and are not affected by the elimination of competition between Lucy Lee and DRMC. 10. Defendants' supposition that plaintiffs are contesting the Freeman standard (see Def. Br. 29-30, citing 69 F.3d at 268 n.12) is entirely misplaced. In the present case, plaintiffs have sought to make the showing of a "credible geographic market" that Freeman requires. The court below concluded that plaintiffs have successfully made such a showing. Add. 10a, 13a. 11. Plaintiffs did not dispute the mathematical computation of defendants' expert regarding the number of patient defections that would be enough to deter hypothetical price increases at various levels. Add. 20a n.5. Those calculations themselves, however, are purely hypothetical; as defendants concede (Def. Br. 23.), they "tell[] the fact finder nothing about whether, in fact, the actual number [of patients] who would leave in the event of a price increase would be above or below the critical loss threshold." Moreover, they are based entirely on historic revenue and cost data (Tr. 914-16 (Harris), DA 643-45), and thus say nothing about the dynamics of the market. 12. Though defendants concede that their approach is not the only way to determine a geographic market (Def. Br. 17 n.7), their argument is presented as if the district court's refusal to use their economic expert's analysis were legal error. 13. Defendants also err in suggesting that the only pertinent hypothetical price increase for this purpose is one of 5 percent. Def. Br. 22. The Merger Guidelines posit that figure as a starting point for analysis in many contexts, but make clear that a "larger or smaller" figure may be more appropriate, depending upon the industry. Guidelines § 1.11, DA 475. In the present case, analysis in terms of a 10 percent price increase is more appropriate, in light of Dr. Wu's testimony that a post-merger increase of at least 10 percent was both likely and sustainable. Tr. 357-58, 430-31 (Wu), PA 91-91a, 133-34; PX 97 at 2, PA 353. The district court, accordingly, referred consistently to the prospect of a price increase at such a level. Add. 14a, 15a, 17a. Indeed, Tenet's ability to sustain a such a high price increase in the Poplar Bluff Region indicates all the more clearly the great degree of market power this acquisition would permit it to wield. 14. The only case the defendants cite is the vacated district court opinion in United States v. Mercy Health Services, 902 F. Supp. 968 (N.D. Iowa 1995), vacated as moot, 107 F.3d 632 (8th Cir. 1997). Def. Br. 19. Although that opinion refers to this sort of calculation (902 F. Supp. at 981), it nowhere holds that it is required for market analysis. In any event, the geographic market analysis in Mercy has no persuasive force in this case, because it turned on facts not present here, including testimony by employers and payers that they could shift people in response to a price increase (id. at 983) and evidence that: (1) health plans in the market had in the past successfully shifted people to distant hospitals in response to a price increase (id. at 982) and (2) outreach efforts by distant regional hospitals had altered hospitalization patterns (id. at 979). As this Court has recognized, geographic market definition is "highly fact-driven and therefore different in each case." Freeman, 69 F.3d at 271 n.16. 15. Defendants' reliance (Def. Br. 27 & n.13) on United States v. Archer-Daniels-Midland Co., supra, for the notion that no significance attaches to the vast price differential between the Poplar Bluff hospitals and more distant hospitals in Cape Girardeau and Sikeston is misplaced. Although the Court observed there that even a "substantial" price differential may not limit the bounds of a product market, it went on to conclude that the "large" differential of 10 to 30 percent between sugar and high fructose corn syrup (HFCS) -- resulting there from government price supports -- was an important factor in concluding that the availability of sugar would not constrain the ability of an HFCS monopolist to raise prices. 866 F.2d at 246. That price differential pales next to those shown in the present case. Moreover, the significance of the differential here is confirmed by the direct testimony of market participants that they would in fact decline to shift business to the more expensive hospitals, even if Tenet were to increase its prices in Poplar Bluff by 10 percent or more. 16. Defendants' reliance on various out-of-context excerpts of employer and payer testimony (Def. Br. 32-33) is misplaced. The fact that some area employers have steered employees to one of the hospitals in Poplar Bluff (e.g., Tr. 171 (Thomas), DA 512; Tr. 117-18 (Reynolds), DA 501-02), does not suggest that they could or would steer their employees from DRMC and Lucy Lee to hospitals outside the Poplar Bluff Region. In fact, the overwhelming testimony of employers and payers was that they would not attempt such steering. See pp. 6-9, supra. 17. For example, one of the key survey questions on which defendants relied asked consumers how they would react to a price change of $200, despite the fact that this represented 5 percent of the entire hospital bill, rather than what the individual would actually pay. Add. 20a-21a. 18. DRMC's referral relationship with Barnes-Jewish-Christian ("BJC") Health System is responsible for many of the admissions to St. Louis hospitals. PX 97 at 19 & Exh. 10, PA 370, 405; PX 52 ¶ 10, PA 330-31. |