N.A.A.C.P. v. MEDICAL CTR., INC.

Reargued In Banc May 11, 1981.

Before ALDISERT, ADAMS, GIBBONS, HUNTER, WEIS, GARTH, HIGGINBOTHAM and SLOVITER, Circuit Judges.

OPINION OF THE COURT

WEIS, Circuit Judge.

The Wilmington Medical Center has been embroiled in litigation for the past five years because of its proposal to construct a new building in the suburbs and renovate one of its buildings in downtown Wilmington, Delaware. In this latest appeal, we hold that disparate impacts of a neutral policy may be adequate to establish discrimination under Title VI of the Civil Rights Act of 1964. Assuming, without deciding, that the plaintiffs presented a prima facie case, we conclude that the Medical Center produced adequate evidence to justify its relocation and reorganization plan. Accordingly, we will affirm the action of the district court in refusing to enjoin implementation of the proposal.

Alleging unlawful discrimination, the plaintiff organizations, representing minority, handicapped, and elderly persons, sought an injunction against the relocation and reorganization of the Medical Center. After we held that the plaintiffs had private rights of action under Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq. (1976), and § 504 of the Rehabilitation Act of 1975, 29 U.S.C. § 794 (Supp. II 1978), see NAACP v. The Medical Center, Inc., 599 F.2d 1247 (3d Cir. 1979), the district court brought the matter to trial. The City of Wilmington was added as a party plaintiff, and the complaint was amended to include allegations that the Age Discrimination Act, 42 U.S.C. §§ 6101-6107 (1976 & Supp. II 1978) had been violated. In addition, plaintiffs charged the defendant with intentional discrimination as well as conduct that had a disparate impact on the classes represented by the plaintiffs.

Following a bench trial lasting more than a month, the district court filed a comprehensive and detailed opinion, concluding that the plaintiffs had failed to prove discrimination under any of the three statutes. Judgment was accordingly entered for the defendant. NAACP v. Wilmington Medical Center, Inc., 491 F.Supp. 290 (D.Del.1980).1 The plaintiffs' appeal was heard initially by a panel and then, because of the nature of the issues, was reheard by the court in banc.

The Wilmington Medical Center (WMC) was organized in 1965 by the merger of three non-profit hospitals, General, Memorial, and Delaware, in different areas of Wilmington. WMC furnishes general medical and surgical services, as well as secondary and tertiary hospital care. It provides 1,104 of the 1,471 non-profit, acute general hospital beds in New Castle County. Other institutions in the county include St. Francis Hospital, which has approximately 290 beds, and Riverside Osteopathic Hospital, with a capacity of 100. The concentration of hospital beds in Wilmington proper is higher than is desirable under national standards, while at the same time the southwestern part of the county surrounding Newark, Delaware, is quite underserved.

WMC is the only hospital in the county with a teaching program approved by the American Medical Association. Medical students and residents are important to WMC's delivery of health care to the community. Without their assistance, current levels of care could not be maintained.

Because its physical structures are aging and are not in compliance with Delaware's licensing law, WMC has encountered serious problems. Recruitment for its residency program has been hindered by the fragmenting of its plants, as well as by a lack of conference space and adequate research facilities. The surgical residency program has been placed on probation by its accrediting body and WMC itself is also in danger of losing its certification by the Joint Commission on Accreditation of Hospitals. On two recent occasions, only "probational" accreditation was granted. Loss of accreditation could result in denial of Medicare and Medicaid reimbursements, a situation which would be disastrous to WMC financially, since it relies on these funds for more than one third of its total budget.

WMC has other monetary problems. It provides the largest amount of free care in the county — approximately $8,000,000 annually. Because Medicare and Medicaid do not reimburse it for any portion of fees attributable to subsidization of free care, WMC must depend upon its endowment and the fees assessed upon paying patients and private insurers.

The population shift to the southwestern suburbs and the possibility that another health care institution might be established in that area present another threat to WMC. If it should lose the patronage of people there, most of whom pay for services or are privately insured, the subsidization of a higher percentage of unreimbursed care would become an even more serious drain on its financial resources.

Recognizing the need for remedial action, the WMC Board canvassed the options open to it. After studying about 50 plans for relocation and consolidation, it decided upon Plan Omega. Essentially, this proposal would close the General and Memorial facilities, renovate the Delaware one, and reduce the number of downtown beds to 250. In addition, a new facility of 780 beds would be built in the suburban area 9.35 miles southwest of the Delaware plant. A division of services between the two locations was part of the arrangement.2

After the district court ordered a departmental review, HEW found discriminatory effects in the plan. To ensure that Omega would comply with Title VI and the Rehabilitation Act, WMC contracted to make a number of modifications. Because no public transportation to the southwest site is available, WMC agreed to provide shuttle bus service between the Delaware and Southwest divisions for the convenience of patients, visitors, and employees. In addition, WMC committed itself to renovate the Delaware plant, devise inpatient service plans for the two branches to prevent racial identifiability at either location, and operate the two facilities on a unitary basis.

Upon acceptance of these conditions, HEW withdrew its objections to Omega.3 Plaintiffs, however, continued their opposition, contending in the district court that the relocation would subject members of the class to inferior health care and disproportionate travel burdens. Moreover, it was alleged that there has been a misallocation of services between the two divisions.

The district court analyzed the case under alternate theories of intentional discrimination and unintended discriminatory effects. The court first determined that there was no evidence of discriminatory purpose. It then applied a disparate effect standard, but concluded after a lengthy review of the evidence that plaintiffs had failed to present a prima facie case.

Rather than ending the inquiry at that point, the court assumed arguendo that a showing of disparate impact had been made. The record was then scrutinized to determine if the defendant had successfully rebutted the plaintiffs' contentions. The court concluded that even if disparate impact had been shown, WMC had demonstrated it had bona fide needs that could not be satisfied by any less discriminatory plan. Finally, the court determined that plaintiffs did not prove that a feasible alternative to Omega was available.

Consideration of the alleged disparate impact was divided into several general categories — access, quality of care, linguistic discrimination, and racial identifiability. Initially, the court found that Plan Omega would bring about vast improvements in the quality of care for all patients, including the classes represented by the plaintiffs. The detrimental effects to minorities and the elderly were determined to be minor and insignificant. With respect to the handicapped, plaintiffs failed to show any adverse impact.

The first issue considered was the plaintiffs' contention that they will lack access to the Southwest facility and, consequently, will suffer a diminution in health care. The court found that WMC would meet its obligation under the HEW agreement to provide adequate shuttle bus and ambulance service. Furthermore, the court concluded that the increased travel time would generally not deter patients from seeking treatment for serious illness at the Southwest division.

A possible exception was a group of women in need of services at the high risk obstetrical clinics at the Southwest division. It was acknowledged that minority women have a greater incidence of high risk pregnancies and that patients seeking prenatal care are more likely to be deterred from seeking medical attention than others. However, the court found that the plaintiffs had overestimated minority usage of the high risk clinics in the Southwest facility and that utilization by whites would be slightly less proportionately.

The plaintiffs' expert erroneously included in her high risk category minority teenagers who are poor users of health care and statistically more likely to have pregnancies with complications. Omega, however, included special clinics at the Delaware division for teenagers and Hispanics. Thus, the group affected by the location of high risk clinics at Southwest division would be much smaller than plaintiffs projected. In addition, the WMC director of obstetrics testified that if a large number of high risk patients appeared at the Delaware division, a clinic would be created at that location, although some patients might have to be referred to Southwest where the most sophisticated equipment would be placed.

Plaintiffs also were concerned with the fact that because obstetrical services would be offered at the Southwest division, emergency room treatment of those cases at Delaware would be inferior. The court, however, found that the vast majority of women about to deliver and those with obstetrical problems would go directly to Southwest. In only exceptional instances would the absence of inhouse obstetricians affect emergency room treatment, because an obstetrical resident would be assigned to the clinics and obstetricians would be on call. The court opined that the cases where treatment would be impaired would be extremely rare, assuming that any at all would occur. In this context, therefore, the possibility of detrimental effects was insignificant, particularly when contrasted with the improvement in quality of care Omega would provide.

The other adverse impact that plaintiffs attributed to travel difficulties is a possible decrease in the number of minority and elderly visitors to inpatients at the Southwest facility. Plaintiffs suggested that visitors would be discouraged by the longer ride to unfamiliar surroundings. Evening visits would be further hindered because the proposed shuttle bus service would stop at 7:00 p. m.

The district judge found that elderly inpatients might have fewer visitors at the Southwest division. While this might result in some detriment to the health of elderly patients, the district court characterized the level of harm as "very minor." 491 F.Supp. at 332. The court similarly concluded that the negative impact on obstetrical patients would be "insubstantial." Id.

Plaintiffs also argued that another group, minority pediatric inpatients, would be adversely affected by a decrease in the number of visitors caused by the location of services at Southwest. Recognizing the importance of family visits to the health of a child, the court determined that steps would be taken under Omega to provide those visits. Parents would be encouraged, and in some cases required, to spend the night with their children. When parents of infants could not stay, the hospital would assign staff members to give special attention to those children.4

With respect to plaintiffs' second major contention — that treatment at the Delaware division would be inferior — the court stated, "[T]he general medical and surgical care that will be rendered at the Delaware Division under Omega will be entirely equal to that rendered at the Southwest Division and superior to that which is now rendered by WMC." 491 F.Supp. at 325.

Plaintiffs asserted that the Delaware division would be housed in an inferior building and even after the proposed renovations, the two facilities would not be comparable. Moreover, it was questioned whether WMC would be financially able to meet its commitment to rehabilitate the Delaware plant. The court described these contentions as "purely speculative and wholly unsupported on the record." 491 F.Supp. at 325. In addition to assuming an obligation under the HEW agreement to refurbish, WMC had allocated more than $12,000,000 for that purpose up to the time of trial. This amount, coupled with projected commitments and funds to be set aside under the agreement, produced a total of $18,000,000 committed to renovation. The judge concluded that the additional $4,000,000 needed to complete the work could be raised from either the operating budget or unrestricted funds.

Furthermore, the court was convinced that shuttle bus service would, in fact, be provided. The cost would be minimal in comparison with WMC's annual budget and could be absorbed with no strain on the institution's financial resources.

After their expert suggested that operating deficits might occur in the years following completion of construction, plaintiffs questioned whether the high cost of Omega would cause WMC to discontinue the remodeling and free care. The court found such evidence irrelevant and believed that financial feasibility of Plan Omega would be determined by bond market forces: "[T]he Court refuse[d] to construe the civil rights statutes as a license ... to act as a financial overseer to those who provide services to minorities." 491 F.Supp. at 328.

Finally, the district court rejected the claims that Omega would create linguistic discrimination or racial identifiability. Plaintiffs' fear of a shortage of interpreters for Hispanics at the Southwest division was rejected. The trial judge found no present shortage of WMC personnel capable of acting as interpreters for Hispanics and expected that none would arise under Plan Omega. Additionally, the court held that Plan Omega as drafted would not create two racially identifiable facilities but if, in practice, problems arose, remedial measures would be taken. All other arguments raised by the plaintiffs were found to be frivolous, and the court concluded that they had failed to present a prima facie case of disproportionate impact.

Recognizing that review in this lengthy and hard-fought litigation was inevitable, however, the trial judge assumed arguendo that a prima facie case had been established and discussed the defendant's burden. Concluding that the defendant was required to go forward with rebuttal evidence, the court found that WMC had met its burden of showing that it had bona fide needs, that Omega would satisfy them, and that other, less discriminatory plans would not.

The court recognized that WMC's immediate need to preserve its educational program and accreditation, as well as improve its quality of care, made it obvious that something had to be done. WMC was aware that to insure its financial stability, and at the same time care for those dependent on its services, it had to provide facilities both in the city and on the outskirts. Omega met these requirements, and the court found "Omega can be completed within WMC's means and will allow WMC to hold its costs down." 491 F.Supp. at 340.

Although the WMC Board had investigated many different plans, six alternatives were discussed. Assuming that any plan which had all or substantially more beds in Wilmington would be less discriminatory, the court found that these plans would not meet WMC's needs. Some were financially infeasible, as in the example of a single large hospital in the city. Rehabilitation of all existing structures within the city was objectionable because it would perpetuate excessive fragmentation. A more even division of services between the Delaware and Southwest facilities (450 beds at Delaware, 570 at Southwest) was rejected by the medical staff as failing to meet the goal of consolidation. Thus after reviewing the options, the court concluded, "WMC has met its burden upon rebuttal by showing that even if Omega may have some meager disparate impacts, those impacts are justified by bona fide needs which could not be accomplished by any less discriminatory plan." 491 F.Supp. at 343.

In turn, plaintiffs attempted to show that "Reverse Omega" (800 beds at Delaware and 200 at Southwest) was a feasible alternative. The court, however, found that cost estimates of reverse Omega given by plaintiffs' expert were unreliable, and concluded that this plan would be "prohibitively expensive." 491 F.Supp. at 342. The court determined, therefore, that "plaintiffs have failed to meet their burden of persuasion of showing a feasible, less discriminatory alternative." 491 F.Supp. at 345.

I

The lengthy recitation of the background makes it clear that this case turns largely on factual matters. There are, however, several discrete legal issues essential to a resolution of the dispute. The first that we shall discuss implicates the nature of the evidence necessary to show a violation of Title VI. If the plaintiff must show intent to discriminate, then our task is a simple one because the trial court found no such evidence and that holding is not contested. We are persuaded, however, that intent is not required under Title VI and proof of disparate impact or effects is sufficient. Our conclusion applies to the other two statutes that have been invoked as well.

Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d (1976), bans discrimination based on race, color, or national origin in any program receiving federal financial assistance.5 WMC concedes that Medicare and Medicaid payments made to it call Title VI into play.

In Lau v. Nichols, 414 U.S. 563, 94 S.Ct. 786, 39 L.Ed.2d 1 (1974), the Supreme Court was confronted with a racial discrimination charge growing out of a school system's decision not to provide English language instruction to students of Chinese ancestry. The Court declined to reach an equal protection argument but chose instead to rely on Title VI, interpreting it as follows:

Id. at 568, 94 S.Ct. at 789 (emphasis the Court's), quoting HEW regulation, 45 C.F.R. § 80.3(b)(2).

Lau makes it clear that discriminatory impact is enough to constitute a violation of Title VI. WMC, however, argues that Lau was overruled by Board of Education v. Harris, 444 U.S. 130, 100 S.Ct. 363, 62 L.Ed.2d 275 (1979), and Regents of the University of California v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978).6 We are not convinced, however, that either case did so.

In Bakke, the question was whether a state school could properly adopt an admissions policy clearly intended to prefer minorities. It is true, as WMC notes, that five justices expressed reservations in Bakke about the holding in Lau. In the opinion written by Justice Brennan, in which Justices White, Marshall, and Blackmun joined, it was said, "[W]e have serious doubts concerning the correctness of what appears to be the premise of [Lau]." 438 U.S. at 352, 98 S.Ct. at 2779.

The issue did not have to be resolved, however, because "even accepting Lau's implication that impact alone is in some contexts sufficient to establish a prima facie violation of Title VI, contrary to our view that Title VI's definition of racial discrimination is absolutely coextensive with the Constitution's, this would not assist the respondent in the least." 438 U.S. at 352-53, 98 S.Ct. at 2779. It did not matter, the group wrote, whether Title VI proscribed some acts, such as those at issue in Lau, that would survive constitutional scrutiny. As the group read the legislative history of the Civil Rights Act, Congress did not intend to proscribe the particular type of practice challenged by Bakke — preferences designed to remedy past discrimination. As stated in another portion of the opinion, "[A]pplied to the case before us, Title VI goes no further in prohibiting the use of race than the Equal Protection Clause of the Fourteenth Amendment itself." 438 U.S. at 325, 98 S.Ct. at 2766 (emphasis supplied).

In a separate opinion, Justice Powell used language that may be inconsistent with Lau, but he stopped short of advocating that the case be overruled. He wrote, "Title VI must be held to proscribe only those racial classifications that would violate the Equal Protection Clause or the Fifth Amendment." 438 U.S. at 287, 98 S.Ct. at 2746. He then went on to distinguish Lau, saying significantly, "[T]he `preference' approved [in Lau] did not result in the denial of the relevant benefit — `meaningful opportunity to participate in the educational program' — to anyone else." 438 U.S. at 304, 98 S.Ct. at 2755.

In determining what weight is to be given to these separate statements, it is important to recognize that the issue presented to the Court in Bakke differs substantially from that in the case at bar. It was clear in Bakke that whatever the reach of Title VI, the plaintiff had established a prima facie case by showing intentional discrimination. The question facing the Court, then, was whether some forms of intentional discrimination were nevertheless permissible. A majority of the Court concluded that those forms of intentional discrimination that would survive constitutional analysis also were exempt from Title VI. Congress, in enacting the Civil Rights Act of 1964, did not intend to prohibit those racial preferences that are permitted under the Constitution.

It does not inexorably follow, however, that Congress also intended the constitutional standard to control every allegation of discrimination. It would be consistent with Congress's expansive, remedial intent to interpret Title VI as prohibiting acts that have the effect of discrimination yet permitting patent preferences designed to remedy past discrimination.

The Powell-Brennan opinions, therefore, may be read as expressing the theory that at least when the charge is intentional discrimination in the nature of a governmental preference, Title VI incorporates the constitutional standard. The case sub judice, however, is not one of a discriminatory governmental preference but one of a neutral program with disparate impact. As we see it, it is still permissible to hold that when the charge is disparate impact, a prima facie case can be established without proof of intent.

The other case on which defendant relies, Board of Education v. Harris, supra, held that § 702(b) of the Emergency School Aid Act (ESAA) prohibits school districts from maintaining racially identifiable faculties even when the segregation is unintentional. The Court upheld the power of Congress in the exercise of its authority under the spending clause to require the recipients of federal funds to go further in eliminating discrimination than mandated by the Constitution. Lau v. Nichols was not cited.

In dissent, Justice Stewart argued that since five justices in Bakke had stated Title VI prohibited only intentional discrimination, the same premise should govern claims under the ESAA. 444 U.S. at 160, 100 S.Ct. at 379. In this argument, however, he was joined only by Justice Powell. The majority expressly disclaimed any necessity to pass on the standard applicable to Title VI. Id. at 149, 100 S.Ct. at 373.

Fullilove v. Klutznick, 448 U.S. 448, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980), is another case that considered the constitutionality of a statutory preferential program. A plurality of the Court cited with approval Lau's validation of the HEW regulation proscribing actions "which have the effect" of discriminating. 448 U.S. at 479, 100 S.Ct. at 2775 (emphasis supplied by Court). Joining in the opinion were Justices White and Powell, who in Bakke had taken the position that intent was necessary to establish a Title VI violation.

Although there is ample ground for argument that the Supreme Court has doubts about Lau's continued viability, a requiem may be premature and, in any event, should not be sung by this choir. The prerogative of overruling its cases rests with the Supreme Court, and not with us. Americans United for Separation of Church and State, Inc. v. HEW, 619 F.2d 252, 271 (1980) (Weis, J., dissenting), cert. granted, Valley Forge Christian College v. Americans United for Separation of Church and State, Inc., 450 U.S. 909, 101 S.Ct. 1345, 67 L.Ed.2d 332 (1981); United States ex rel. Gockley v. Myers, 450 F.2d 232 (3d Cir. 1971), cert. denied, 404 U.S. 1063, 92 S.Ct. 738, 30 L.Ed.2d 752 (1972).7

The question is not one of congressional power but rather of intent. Providing federal funding conditioned on an even-handed application is a positive measure to discourage all forms of discrimination, intentional or not. The use of an effects test, therefore, is consistent with the legislative aim of eliminating discrimination and is in harmony with Title VII of the same Act, and Title VIII, Resident Advisory Board v. Rizzo, 564 F.2d 126 (3d Cir. 1977), cert. denied, 435 U.S. 908, 98 S.Ct. 1457, 55 L.Ed.2d 499 (1978), as well as our previous reference to Title VI in Shannon v. United States Department of Housing & Urban Development, 436 F.2d 809, 816, 820 (3d Cir. 1970) (Title VI provides redress for discriminatory effects of local housing plans). Moreover this approach parallels regulations adopted by HEW and other departments charged under § 602 of the Civil Rights Act, 42 U.S.C. § 2000d-1, with enforcing the statute.8

With due deference to Lau v. Nichols and congressional intent as we perceive it, therefore, we conclude that plaintiffs in a Title VI case alleging discrimination in the application of federal funds in a facially neutral program need only establish disparate impact. The Rehabilitation Act and the Age Discrimination Act of 1975 provide equally strong cases for application of an impact test since both are patterned after Title VI.9 We therefore use the same standard.

II

The next inquiry is whether, applying an effects test, the plaintiffs have established a prima facie case. Before addressing this issue, it is helpful to review the provisions of the agreement between WMC and HEW. Included in the early paragraphs is a statement that the Secretary of HEW desires assurances that operation of the hospital facilities under Plan Omega will be in compliance with Title VI and the Rehabilitation Act.

The agreement obligates WMC to provide free transportation between the Delaware and Southwest divisions, to designate an ombudsman to receive and act upon complaints of discrimination, to adopt a system of inpatient utilization control, and to prevent either division from becoming racially identifiable. It is additionally required that both divisions be operated on a unitary basis, with a single Board of Directors, Executive Committee, medical staff, teaching program and accounting procedure. Any proposed expansion of services at Southwest or reduction at Delaware must be first submitted to HEW for approval. WMC agreed to set aside $2,800,000 for use exclusively in renovating the Delaware facility. WMC also agreed to recognize the need for employment by minority groups, "including in particular urban minority groups." As noted earlier, the court found that WMC would carry out its categorical obligations under this agreement.

A. THE HANDICAPPED

There is no evidence that either facility will not comply with the structural requirements of the Rehabilitation Act. Indeed, the provisions for handicapped with respect to barriers, entry, and free movement within the buildings will be an improvement over existing conditions. The alleged disparate impact upon the handicapped, therefore, rests upon the location of major portions of hospital services and jobs in the Southwest division. The plaintiffs produced no credible evidence, however, establishing the residential distribution of handicapped persons within the county. In the absence of such information, we cannot tell what effect, if any, Plan Omega will have upon disabled persons in the area, and thus agree with the district court that plaintiffs did not establish a prima facie case under § 504.

B. THE AGED AND MINORITIES

Unlike the evidence with respect to the handicapped, there was testimony that most of the elderly and minorities who would be served by the Wilmington Medical Center live closer to the Delaware than the Southwest division some nine miles away.10 Since many of the medical services would be located at the suburban building, transportation to the new facility would be required, and hence treatment would not be as convenient as if provided at Delaware. Although the trial court did find that there would be some effect upon the elderly and minorities because of the travel aspects, those impacts upon patients were described as "de minimis," "insignificant," and "minor."11 We agree with these characterizations and have serious doubts that such effects are enough to establish a prima facie case of discrimination.

The nine mile trip in an area like Wilmington does not impose a significant hardship. Changes to alleviate some problems, even though resulting in improvement, often impose other burdens or confer unequal benefits. Whatever was done here could not possibly distribute the inconveniences and benefits with precise equality, but inaction would have a profound adverse impact upon all who depend upon the medical center.

All concede that something must be done or all will suffer. To establish a prima facie case under Title VI in these circumstances, some definite, measurable disparate impact is required, otherwise needed and worthwhile efforts at improvement will be paralyzed. Reasonable accommodations must be made, but when they have been reached, new programs must be allowed to proceed. Although all of us are not completely persuaded that plaintiffs met their burden here, we will assume arguendo, as did the district court, that a prima facie case was presented.12

III

The next step, therefore, is to determine what burden is placed upon the defendant and whether it was met in this case. The district court concluded that once the plaintiffs had met their initial burden, the defendant had to go forward with evidence to "rebut [that] prima facie case." 491 F.Supp. at 315. The plaintiffs argue that the defendant's burden is a heavier one, that of persuasion.

The parties agree that the decisional law allocating the burdens of production and persuasion under Title VII is instructive in this case, but disagree as to the proper interpretation of the opinions. It is not disputed that when a prima facie Title VII case of discriminatory intent is established, the defendant must go forward with evidence of a legitimate, nondiscriminatory reason for its action. The plaintiff may rebut by showing that the stated reason is mere pretext.

The ultimate burden of persuasion on the issue of illegal discrimination always remains with the plaintiff. Whatever doubt may have existed on that score has been resolved in recent years by a series of cases in the Supreme Court and this court. Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981); Board of Trustees of Keene State College v. Sweeney, 439 U.S. 24, 99 S.Ct. 295, 58 L.Ed.2d 216 (1978); Furnco Construction Corp. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978); McNeil v. McDonough, 648 F.2d 178 (3d Cir., 1981); Smithers v. Bailar, 629 F.2d 892 (3d Cir. 1980); Kunda v. Muhlenberg College, 621 F.2d 532, 543 n.3 (3d Cir. 1980); and Whack v. Peabody & Wind Engineering Co., 595 F.2d 190 (3d Cir. 1979). See also Resident Advisory Board v. Rizzo, supra at 149 n.37 (Title VIII); Rodriguez v. Taylor, 569 F.2d 1231, 1239 (3d Cir. 1977), cert. denied, 436 U.S. 913, 98 S.Ct. 2254, 56 L.Ed.2d 414 (1978) (Age Discrimination in Employment Act).

The plaintiffs contend that there should be a difference in the defendant's burden when the charge is discriminatory impact rather than discriminatory intent. Their theory is that in countering a prima facie case of discriminatory impact, the defendant is presenting something in the nature of an affirmative defense that requires shouldering the burden of persuasion. See Kirby v. Colony Furniture Co., 613 F.2d 696, 703 n.5 (8th Cir. 1980) (opinion of one judge, others not joining).

That reasoning is not convincing. One could just as readily say in an intent case that the necessity to prove a nondiscriminatory reason is an affirmative defense carrying a burden of persuasion. Holdings of the Supreme Court and this court, however, are to the contrary.

In Furnco Construction Corp. v. Waters, supra, the Supreme Court explained its allocation of the burden of proof in intentional discrimination cases. If the plaintiff produces evidence sufficient to meet the standards of a prima facie case under McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), an inference of discrimination is raised "because we presume [the complained of] acts, if otherwise unexplained, are more likely than not based on the consideration of impermissible factors." Furnco Construction Corp. v. Waters, supra, 438 U.S. at 577, 98 S.Ct. at 2949. The effect of such a prima facie case is only to put in issue whether the employer's conduct "was based upon legitimate, nondiscriminatory reasons and therefore permissible." Id. at 576 n.8, 98 S.Ct. at 2949, n.8. A prima facie case does not necessarily constitute proof of the ultimate fact of discrimination under Title VII. Id. at 576, 98 S.Ct. at 2949.

To meet a McDonnell Douglas prima facie case, a defendant must produce evidence of an acceptable reason but is not required to show an absence of discriminatory motive. Board of Trustees of Keene State College v. Sweeney, supra, 439 U.S. at 24, 99 S.Ct. at 295. The burden of persuasion on the ultimate fact of discrimination remains with the plaintiff who may show that the proffered legitimate reason was a pretext. Furnco Construction Corp. v. Waters, supra, 438 U.S. at 578, 98 S.Ct. at 2950; McDonnell Douglas Corp. v. Green, supra 411 U.S. at 804, 93 S.Ct. at 1825.

Disproportionate impact or effect is simply an additional method of demonstrating impermissible discrimination under Title VII. Teamsters v. United States, 431 U.S. 324, 336 n.15, 97 S.Ct. 1843, 1854 n.15, 52 L.Ed.2d 396 (1977). In Albemarle Paper Co. v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975), the Court held that a prima facie case could be established under the impact theory if the plaintiff demonstrated that a facially neutral policy disproportionately affected persons protected by Title VII. If the plaintiff meets his initial burden, the defendant must show "`that any given requirement [has] ... a manifest relationship to the employment in question.'" Id. at 425, 95 S.Ct. at 2375, quoting Griggs v. Duke Power Co., supra 401 U.S. at 432, 91 S.Ct. at 854. In formulating this approach, the Court referred to the related test it had devised in Albemarle Paper Co. v. Moody, supra, 422 U.S. at 425, 95 S.Ct. at 2375, and went on to include a similar third step: "it remains open to the complaining party to show that other ... selection devices, without a similarly undesirable ... effect, would also serve the ... legitimate interest." 422 U.S. at 425, 95 S.Ct. at 2375.

In characterizing the defendant's obligation to show a manifest relationship as an affirmative defense, the plaintiffs here apparently assume that making out a prima face case of disproportionate impact is the equivalent of establishing a Title VII violation by a preponderance of the evidence. This assumption cannot stand because the Furnco analysis should control impact, as well as intent, cases.13

When the Supreme Court first held that Title VII prohibited some facially neutral practices, it described the congressional purpose as "the removal of artificial, arbitrary, and unnecessary barriers ... when the barriers operate invidiously to discriminate on the basis of racial or other impermissible classification." Griggs v. Duke Power Co., supra, 401 U.S. at 431, 91 S.Ct. at 853. A showing of disproportionate effect or impact alone may not establish a violation. "The touchstone is business necessity. If an employment practice which operates to exclude ... cannot be shown to be related to job performance, the practice is prohibited." Id. To be proscribed, then, the challenged practice must not only affect disproportionately, it must do so unnecessarily.

To establish a prima facie case the plaintiff need not show that the practice was unnecessary but may rely on inferences. If the defendant presents no evidence of business relatedness in his case, the court may assume that there was no permissible reason for the impact.14 In the event that the defendant does come forward with evidence to meet the inference of discrimination raised by the prima facie case, the plaintiff may still carry his burden of persuasion by demonstrating that a feasible, yet less onerous alternative exists. Albemarle Paper Co. v. Moody, supra, 422 U.S. at 425, 95 S.Ct. at 2375.

The contention plaintiffs make here, that business relatedness constitutes an affirmative defense, is incompatible with the third step of Albemarle. Plaintiffs would have WMC bear the burden of persuasion on this issue by showing a dearth of less objectionable alternatives. But in Albemarle, this burden was imposed on the complaining party. As the Court explained in an analogous context, if the plaintiffs were correct in their assessment of the various burdens of production and persuasion, the third step in the analysis would be rendered "entirely superfluous ..., since it would place on the [defendant] at the second stage the burden of showing that the reason ... was not a pretext, rather than requiring such proof from the [plaintiffs] as a part of the third step." Board of Trustees of Keene State College v. Sweeney, supra 439 U.S. at 24-25 n.1, 99 S.Ct. at 295-296 n.1.

The Supreme Court has not given any indication that it requires a shifting of the burden of persuasion in effects cases. To the contrary, the Court stated in New York Transit Authority v. Beazer, 440 U.S. 568, 587 n.31, 99 S.Ct. 1355, 1366 n.31, 59 L.Ed.2d 587 (1979), that the ultimate burden of proving discriminatory impact is the plaintiff's. Albemarle Paper Co. v. Moody, supra, was an impact case, yet in referring to the employer's burden to meet the plaintiff's prima facie showing, the Court cited McDonnell Douglas Corp. v. Green, supra, an intent case. As other examples of cross-references to McDonnell Douglas in effects cases, see Dothard v. Rawlinson, 433 U.S. 321, 329, 339, 97 S.Ct. 2720, 2726, 2731, 53 L.Ed.2d 786 (1977), and Nashville Gas Co. v. Satty, 434 U.S. 136, 144, 98 S.Ct. 347, 352, 54 L.Ed.2d 356 (1977). See also Teamsters v. United States, 431 U.S. 324, 336, 358, 360, 97 S.Ct. 1843, 1855, 1866, 1867, 52 L.Ed.2d 396 (1977).15

The plaintiffs have cited no current authority for their position. Precedents antedating Board of Trustees of Keene State College v. Sweeney, supra, have little persuasive effect since that case settled the confusion that surrounded this issue. Although the facts and inferences required to prove a case vary between intent and effect situations, that factor does not call for the shifting of the burdens of production and persuasion depending on the theory advanced.16

Moreover, it is illogical to impose a heavier burden on a defendant in a case where a neutral policy results in disparate impact than in one where the charge is unlawful animus. Indeed, if there is to be a difference, quite the opposite result should follow. The defendant who intentionally discriminates should not fare better than the one whose conduct may be subjectively blameless, but because of its effects may require remedial action.

As a practical matter, a procedural distinction between the impact and intent cases would cause unnecessary confusion in the trial courts, particularly so in cases like the one at hand in which both theories are advanced. See, e. g., Whack v. Peabody & Wind Engineering Co., supra. It is difficult to understand what important interests would be served by imposing two different burdens on the defendant in a case of this nature. Certainly the multiplication of procedural devices is not a desirable development in trial practice.

All things considered, uniformity in the procedural aspects of impact and intent cases is highly desirable and should not be sacrificed on the dubious theory that plaintiffs advance here. Although we need not worship at its shrine, symmetry is not always sinful. Just as we permit plaintiffs to establish discrimination through effects under both Title VI and VII, so should there be a consistent burden on defendants.

The district court determined that WMC should go forward "with evidence that Omega will `in theory and practice' serve `a legitimate bona fide interest of [WMC] ... and ... show that no alternative course of action could be adopted that would enable that interest to be served with less discriminatory impact.'" 491 F.Supp. at 315-16, quoting Resident Advisory Board v. Rizzo, supra at 149.

Following this, plaintiff was allowed to produce further evidence consistent with the third step of demonstrating pretext that the Supreme Court has mentioned in both intent and impact cases brought under Title VII.17 McDonnell Douglas Corp. v. Green, supra; Albemarle Paper Co. v. Moody, supra; Teamsters v. United States, supra.

In Resident Advisory Council v. Rizzo, supra, we held that under Title VIII of the Civil Rights Act of 1964, a plaintiff retains the burden of persuasion on the existence of less discriminatory alternatives. Requiring plaintiffs to meet that obligation by demonstrating that feasible, less discriminatory alternatives exist is neither unjust nor impractical in view of the extensive discovery material that was available in this case.

The district court's test is actually more stringent than that suggested in Jefferson v. Hackney, 406 U.S. 535, 92 S.Ct. 1724, 32 L.Ed.2d 285 (1972). There, the Supreme Court found that application of a percentage reduction factor to determine reduced needs of welfare recipients was rationally related to the purpose of the separate welfare programs and, consequently, did not violate the equal protection clause. For similar reasons, the Court also concluded that the challenged system would not contravene Title VI. The relationship of the reduction factor to the purposes of the State's welfare programs distinguished Jefferson from Griggs:

Jefferson v. Hackney, supra at 550 n.19, 92 S.Ct. at 1733 n.19. In Jefferson, the state was not required to produce evidence that alternate formulae for computing need would not have served the purposes of the program with less of a disparate impact.

In Bryan v. Koch, 627 F.2d 612 (2d Cir. 1980), the Court of Appeals for the Second Circuit was confronted with a Title VI challenge to the closing of a city hospital. The court said that Title VI did not require consideration of alternatives beyond "an assessment of all the municipal hospitals in order to select one or more for closing." Id. at 619. Since the appropriateness of the city's choice had been sufficiently demonstrated, the court's role ended. Expressing doubt about the feasibility of a more open ended judicial evaluation of alternative means of economizing, the opinion stated,

Id. The court added that its skepticism extended even to requiring courts to consider "alternative locations for placement ... of facilities." Id., citing NAACP v. The Wilmington Medical Center, Inc., 491 F.Supp. 290 (D.Del.1980).

By contrast, the district court in the case at hand did evaluate the alternatives. It required WMC to go "forward with evidence showing that it has chosen the least discriminatory alternative." 491 F.Supp. at 340. That is a stringent standard which more than adequately serves Title VI aims.18 The court discussed six possible, less discriminatory alternatives to the Omega Plan, including the plaintiffs' "Reverse Omega" proposal and found that none of the plans would serve WMC's needs. The court also said that WMC had "investigated approximately 50 different plans, all of which it rejected for bona fide reasons." Id. at 340 n.314. Indeed, the court found that "Omega is the only plan which can adequately meet WMC's needs." Id. at 340.

On this record, we conclude that the district court did not err in concluding that the defendant had carried its burden of meeting the plaintiffs' prima facie case.19

IV

The plaintiffs also argue that the district court erred in refusing to assess the financial feasibility of Plan Omega and in deferring instead to the judgment of the bond market. But as noted earlier, the court did make specific findings with respect to WMC's financial ability to complete the renovation at Delaware and pay for whatever shuttle bus service is required. When the court referred to the sanction of the bond market for the financial consequences of Plan Omega, it apparently was referring to the question whether WMC was wise in undertaking such an extensive project.

As we read the district judge's opinion, whether funds could be obtained was not a matter which he could confidently predict, but was a circumstance subject to market forces. If the bonds were not sold, Omega could not proceed. Obviously an undertaking of this magnitude involves some element of financial risk and predicting the ultimate outcome is not a field in which the courts have a special competence. The district judge's hesitancy to wander into this area of uncertainty is understandable. On the record we do not find it to be reversible error.

In fact, it would have been pure speculation for the court to accept the plaintiffs' argument. Even if it could be demonstrated that WMC was overly optimistic, there is no way of knowing with any certainty what remedial measures would be taken. It is far from clear that WMC would take the path suggested by plaintiffs and reduce free care and renovations of the Delaware division. Not only would this breach the HEW contractual obligations but it would also place WMC in jeopardy of losing its Medicare-Medicaid reimbursements. A facility already in financial difficulty is not likely to risk forfeiting federal funding that makes up 30 percent of its budget.

It must be remembered that the Omega Plan was submitted for administrative review to hospital planning organizations and to HEW. After some changes had been made, the proposal was approved by HEW as being in compliance with Title VI. We are not called upon to appraise the wisdom of Omega but are limited to reviewing the decision of the district court by appropriate appellate guidelines. From that perspective, we do not find legal error in the standards the district court utilized nor can we say that the factual findings are clearly erroneous. Accordingly, the judgment of the district court will be affirmed. The mandate will issue forthwith.

ADAMS, Circuit Judge, concurring.

I arrive at the result reached by the majority but, because I do so by a somewhat different route, I find it necessary to write separately.

The record here reveals a problem confronting American hospitals with increasing frequency: an aging physical plant, escalation of health care costs and maldistribution of services have combined to create a health care crisis.1 Deteriorating physical facilities threatened the Wilmington Medical Center with a loss of accreditation. Such a loss would further erode the quality of the hospital's medical care and its financial foundation by triggering a loss of qualification for the Center's teaching program and a termination of Medicare and Medicaid funds. In addition, without the construction of new facilities, the community would suffer a shortage of acute care beds. See NAACP v. Wilmington Medical Center, Inc., 491 F.Supp. at 290, 297-98 (D.Del.1980).

Faced with these demographic and fiscal pressures, the Center concluded that rehabilitation of some of its facilities in downtown Wilmington and construction of new facilities in the suburban area, to prevent paying patients from gravitating to a potential competitor, would most appropriately fulfill the hospital's needs. See 491 F.Supp. at 310. The Center considered approximately 50 proposals before arriving at a final plan that was denominated "Omega." Plan Omega was approved first by a state designated planning agency, which ensured that the project conformed to local needs for adequate health care, and then by the federal Department of Health, Education and Welfare (now Health and Human Services). See Wilmington United Neighborhoods v. United States Dept. of Health, Education and Welfare, 615 F.2d 112, 124-25 (3d Cir. 1980).

Under Plan Omega, the Center proposed to invest approximately $18-24 million in the rehabilitation of certain of the inner city facilities; to make a substantial investment in facilities in the suburbs; and to provide transportation for center city residents in need of specified services located at the suburban installation. See 491 F.Supp. at 325-27, 343, 319. The plaintiffs maintain that the proposed program violates Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, the Age Discrimination Act of 1975, 42 U.S.C. § 6101 et seq., and Section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794, because it adversely affects the quality of care and access to that care for the handicapped, the elderly and certain minority groups.2 The Center, however, contends that, as a whole, the physical rehabilitation, new construction and consolidation of services will result in improved care for all patients, and that the availability of a shuttle service will minimize any transportation barriers for the handicapped, elderly and minority groups.

Because I would affirm the district court on different grounds than the majority, I find it unnecessary to decide whether Lau v. Nichols, 414 U.S. 563, 94 S.Ct. 786, 39 L.Ed.2d 1 (1974), in which the Supreme Court construed Title VI to prohibit disparate impacts, is still the relevant governing law. Although the Supreme Court has recently suggested in Regents of the University of California v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978), and Board of Education of New York City v. Harris, 444 U.S. 130, 147 n.10, 100 S.Ct. 363, 62 L.Ed.2d 275 (1979), that Title VI might incorporate the constitutional standard of specific intent to discriminate, it should be noted that, because of the wide range of activities and conduct that Title VI covers, the concerns raised in those two cases do not parallel the issues here. As the majority explains, Bakke focused on the intentional use of racial criteria in the context of voluntary remedial actions. The Supreme Court did not deal with the type question present in this case, namely whether Title VI may impose requirements on recipients of federal funds that are broader than the Constitution demands when faced with disparate impacts resulting from facially neutral actions. Moreover, the concern voiced in Harris — that because a violation of Title VI may result in a cutoff of funds, it is likely that Congress desired this drastic sanction only when discrimination is intentional — would also appear to be inapplicable here. The plaintiff's private cause of action against the Center, seeking an injunction to prevent future discrimination, involves no immediate prospect of a fund cutoff. In fact, it is attempting to ensure against such a possibility.3

Nonetheless, whether hospital relocations and renovations such as the present one should be subjected to judicial scrutiny aimed not only at preventing intentional discrimination but also at forestalling any relocation which may occasion unintentional, adverse effects on protected groups is, from my perspective, a troublesome question. Courts may not be the most competent forums for determining the effects of hospital relocations on racial minorities. Admittedly, we earlier acknowledged the desirability of judicial review by finding a private cause of action under Title VI. See NAACP v. Wilmington Medical Center, Inc., 599 F.2d 1247, 1254 (3d Cir. 1979). And courts are often well-situated to address the concerns of relatively unorganized, politically weak consumer groups whose interests may be insufficiently recognized in administrative proceedings. But I would defer deciding until a later day, the level of scrutiny, and the extent to which courts consequently might intrude upon a better-informed regulatory process.

As the district court held, and the majority here agrees, the record is devoid of proof of intentional discrimination. But, even assuming that a disparate impact test is an appropriate one under Title VI, I believe that a fair reading of the evidence in this case, and especially of the extensive findings made by the trial court, indicates that the plaintiffs failed to make out a prima facie case under any of the statutes involved. A plethora of findings underlies the trial judge's ultimate conclusion that the plaintiffs have shown only a slight disparate increase in travel time, a modest decrease in the ability of inner city residents to visit patients at the suburban site, and consequently a minimal negative effect which the decrease in visitors may have upon the quality of care for the elderly. 491 F.Supp. at 333. The district court also determined that the possibility of a few minority high risk patients missing an appointment at the speciality clinics, and the extremely rare chance of an obstetrical emergency patient receiving inadequate treatment in the inner city division constituted such unlikely effects that they failed to establish a prima facie case under Title VI. 491 F.Supp. at 337. I cannot find that the trial court clearly erred in holding that plaintiffs "failed to meet their initial burden of proving disparate impact under the civil rights statutes invoked." See 491 F.Supp. at 339.

Moreover, these specific findings are part of a larger mosaic: the trial court's overarching finding that the level of care for all population groups will improve as a result of the benefits that greater consolidation, better-trained residents and upgraded facilities will confer. Measured against HEW regulations which define Title VI violations as actions which have "the effect of defeating or substantially impairing accomplishment of the objective of the program as respect [sic] individuals of a particular race, color, or national origin," 45 C.F.R. § 80.3(b)(2) (emphasis added), these de minimis impacts simply do not pass muster. Unless a threshold is created for prima facie cases under Title VI, questions regarding hospital relocations and similar, complex socioeconomic decisions will be open to protracted court challenge, for each significant community undertaking affects slightly differently the various protected population subgroups in our country's localities.

Because I am unable to find that the plaintiffs established a prima facie case, it is unnecessary for me to resolve whether the defendants carried a burden of production or burden of proof in rebutting an initial showing of disparate impact with evidence of legitimate medical needs.

GIBBONS, Circuit Judge, concurring and dissenting.

I join in Part I of the opinion of the court, which holds, contrary to the position advanced by the Wilmington Medical Center (WMC), that Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq. (1976), Section 504 of the Rehabilitation Act of 1975, 29 U.S.C. § 794 (Supp. II 1978) and the Age Discrimination Act, 42 U.S.C. § 6101 et seq. (1976 & Supp. II), prohibit not intentional discrimination alone, but also disparate impact upon the classes those acts protect. Only an effects standard will encourage decision-makers to consider possible discriminatory consequences of a proposal before its implementation. I also join in Part II A of the opinion of the court, holding that the plaintiffs did not establish a prima facie case of violation of Section 504 of the Rehabilitation Act. I do not join in Part II B of the opinion, which assumes arguendo that the plaintiffs established a prima facie case of disparate impact against the aged and minorities. My view is that the plaintiffs quite clearly proved a prima facie case of disparate impact against both classes in significant respects. I dissent from Part III of the opinion of the court, which is entirely inconsistent with the intention which the court correctly attributes to Congress in its discussion of the federal funding statutes in Part I.

I.

In describing the requirements for a prima facie case the majority opinion states:

Maj. op. p. 1332 (emphasis supplied). If the court ended its discussion with the italicized language, I would agree with its test. But the addition of the following sentences shows that the court is confusing the requirements of a prima facie case — definite, measurable impact on the protected class — with justification for the imposition of such an impact. Medical and financial necessities may justify the adverse effects, but do not make them any less substantial. The distinction is critical. It is illogical to proceed, as the majority does, to the issue of justification, without first identifying the specific impacts which will occur, since what amount of accommodation is reasonable depends on the size of the impact.

On this record there is no question but that, contrary to the district court's alternative holding, a definite, measurable impact on the protected classes has been shown. It need not be assumed arguendo. It is plain. The district court's contrary conclusion is the result of several fundamental legal errors, which the majority opinion ignores.

A. Factual Background

Before reviewing the lower court's ultimate findings, I summarize the background facts about Plan Omega which are common ground. The plan was the culmination of a long period of planning by WMC for the improvement of its ability to offer quality medical care. WMC is a private nonsectarian hospital which evolved from the merger in 1965 of three acute care general hospitals in the City of Wilmington, which after the merger maintained three separate physical facilities containing approximately 55 percent of the available acute care beds in the State of Delaware. Although the three separate facilities are located in different areas of the city, they are all well served by bus routes which run throughout heavily populated areas of New Castle County, as would be true of any new consolidated facility if it was located in Wilmington. The primary reasons for the merger were reduction of duplicate facilities and improvement of clinical experience for a resident program in order to attract residents who would serve the Delaware community. Thwarting this purpose is the fact that WMC's physical facilities are aging and in various states of disrepair; their inadequacies have resulted in only probational accreditation by the Joint Commission on Accreditation of Hospitals. Loss of accreditation would mean ineligibility for participation in residency programs approved by the American Medical Association. Thus plant improvement is essential.

It has always been the judgment of the WMC medical staff that achievement of the hospital's objectives could best be served by placing all medical and surgical services under one roof. However, by the time planning for a new facility commenced, Delaware, like other states, was experiencing a substantial shift in population growth from older urban to newer suburban locations. In New Castle County that shift in the growth pattern produced a rapid increase in its southwestern portion, around Newark, Delaware. That area is now served only by an emergency room, and the need for some hospital beds in the vicinity of Newark is widely recognized. The Delaware Health Planning Council, a state agency, has recommended that such beds be provided. If an institution other than WMC were to do so, the latter would be adversely affected. WMC is the largest provider of free care in the County to those unable to pay for care themselves and unable to qualify for government assistance. Partly as a result of subsidization of this free care, WMC's hospital rates are the highest in the State of Delaware. Since the population in the southwestern suburbs is generally more affluent than that in the urban northern part of the County, diversion of patients from WMC to a new suburban institution would have the effect of increasing the percentage of free care patients in WMC's patient mix, and thus of adversely affecting its financial stability.

The combination of undisputed facts outlined in the preceding paragraph limited the choices available to WMC. It could build a new hospital under a single roof in the suburbs, closing the Wilmington units, or it could build a new hospital in Wilmington, running the risk that another institution would build in the suburbs, or it could attempt, with separate facilities, to serve both areas. When the third course was decided upon it became necessary to determine the mix of facilities in the two locations. In Plan Omega WMC opted for the erection of a new 780-bed hospital, the Southwest Division, in a rural location at Stanton, the closing of two of the three Wilmington divisions, and the reduction in bed capacity of the third, Delaware Division, from 480 to 250 beds. The Stanton location is not now and will not be in the foreseeable future served by public transportation, although it is on an interstate highway. The Plan involves more than allocating beds, however, for some major hospital services will be located exclusively at the larger Southwest Division, others exclusively in the smaller Delaware Division, and some in both places.1

As a result of the order of the district court that the Department of Health, Education and Welfare perform an investigation of Plan Omega's compliance with Title VI,2 proceedings took place before that agency which resulted in a finding that, as originally conceived, Plan Omega would violate Title VI. A Supplemental Agreement between the agency and WMC was then made for the purpose of assuring that the plan would not violate the statute.3 The fact, though not the legal significance, of the Supplemental Agreement is undisputed.

The plaintiffs contend that the effect of Plan Omega is to impose on protected classes in New Castle County a disproportionate burden of decreased availability of medical services and of employment opportunities when compared to the rest of the population. That disparate impact, they contend, will result because the housing patterns in New Castle County concentrate greater numbers of the protected classes in those parts of the County nearer to the present WMC divisions, which under Plan Omega would be scaled down considerably. The travel burden imposed by the proposed Southwest Division would allegedly be exacerbated by the fact that the new facility would not be served by public transportation. The burden would, therefore, fall disproportionately on minorities, the handicapped, and the elderly, who, plaintiffs urge, have far less access to private transportation than does the rest of the population. Moreover, they predict that because the more significant inpatient medical services will be at the Southwest Division, the quality of care at the Delaware Division will in many respects be inferior, and that division will ultimately become racially identified. These impacts would violate Title VI.

Before addressing the evidence supporting these contentions, it is worthwhile noting certain provisions of the Supplemental Agreement reached by WMC and the Department of Health, Education and Welfare. The agreement recites the Secretary's desire to receive assurance that the operation of WMC under Plan Omega as implemented will be in compliance with Title VI and the Rehabilitation Act. The agreement covers main areas of future WMC operation. First, it obliges WMC to operate its own transportation system between the Delaware and Southwest Divisions. Second, it requires the adoption of a system of inpatient utilization control aimed at preventing one or the other division from becoming racially identifiable. Third, it requires that both divisions be operated on a unitary basis, with a single Board of Directors and Executive Committee, a single medical staff, consolidated teaching programs, and consolidated accounting. Fourth, it requires improvements in the plant of the Delaware Division and approval by the Regional Civil Rights Director of the Department of Health, Education and Welfare before any reductions in services at that division or material expansion of the Southwest Division occur. The agreement is, however, "subject to amendment from time to time by written instrument executed by the parties, to reflect such changes in systems for delivery of hospital care and of changing community needs."4

In determining that the plaintiffs had failed to prove a case of disparate impact, the trial court canvassed the evidence on each of the ways Plan Omega could be expected to impact upon the three protected classes. Since I agree with the majority's disposition of the claim of disparate impact on the handicapped, I address the remaining claims of the aged and minorities.

B. The Disparate Impacts Identified

The plaintiffs' evidence suggests that minorities and the elderly are concentrated in or near the City of Wilmington, that of households lacking access to automobiles, 73.8% were in that part of the County, and that such public transportation as is available in the County other than taxicabs is concentrated in that part of the County as well. Moreover, a round trip taxicab fare from the Delaware Division to the Stanton site costs approximately $15.00. 491 F.Supp. at 304-05. That evidence also suggested that minorities and the elderly have a disproportionately high incidence of need for those inpatient services which will be located exclusively at the Southwest Division.5 491 F.Supp. at 306. The plaintiffs' evidence identified three classes of services falling in this category. First, services relating to childbearing and infant health are more likely to be needed by minorities. 491 F.Supp. at 306-07. Second, minorities are more likely than whites to need services for cerebral vascular diseases and for cancer. 491 F.Supp. at 307. Third, the elderly are more likely than the general population to need services relating to cerebral vascular and cardiovascular diseases, gastroenterology, thoracic surgery, and radiation therapy. 491 F.Supp. at 308. The evidence is clear, and virtually undisputed, therefore, that certain services will, under Plan Omega, be moved over nine miles further from the population that needs them most; a population that has least access to private transportation.

The evidence also discloses important differences between the Southwest and Delaware Divisions, differences which plaintiffs predict will result in unequal quality of care. Almost all clinic (outpatient) care will be concentrated at the Delaware Division, and a majority of clinic patients are members of minority groups. Moreover, of 33 inpatient departments, only 5, psychiatry, family practice, rhinolaryngological surgery, dentistry and ophthalmology, will be exclusively at Delaware Division, while 15 will be exclusively at Southwest and only 13 at both facilities. There is evidence that the Delaware Division will be housed in a renovated facility, presently dilapidated, while the Southwest Division will be brand new. Plaintiffs also offered evidence tending to show that estimates of revenue on which Plan Omega was based, and upon which the proposed renovation of Delaware Division was dependent, were questionable, and thus that even a renovated Delaware Division may not materialize. 491 F.Supp. at 327. Plaintiffs contend that Delaware Division is destined to become a second-rate facility, racially identifiable, and rendering to the minorities and the elderly in its service area services inferior to those rendered at the Southwest Division. The inferior service will result, they contend, from unavailability of board-certified specialists for consultation at the Delaware Division in those important specialties exclusively at Southwest.

For various reasons the trial court rejected the significance of most of the evidence referred to, and concluded that no violation of either Title VI or the Age Discrimination Act was shown. The court analyzed the evidence under three categories: effects on accessibility, effects on quality of care, and racial identifiability. I will do likewise.

1. Accessibility

Patently the removal of a number of medical specialties from Wilmington to the suburbs will impose a disproportionate travel burden on minorities and the elderly who are concentrated in the inner city and have less access to private transportation. But as amended by the Supplemental Agreement, Plan Omega includes a shuttle bus component intended to satisfy transportation demand. Plaintiffs produced the testimony of Dr. Marvin Manheim, a transportation expert, to the effect that the level of shuttle bus service specified in the Supplemental Agreement will be far less than the demand and will create for passengers prohibitively long waiting periods. The court discounted the relevance of Dr. Manheim's testimony, however, construing the Supplemental Agreement as a categorical undertaking by WMC to provide adequate free transportation for patients, visitors and employees. The court concluded, moreover, that even in the worst case hypothesized by the transportation expert, the additional cost of providing needed free transportation would be minimal in comparison with WMC's annual budget. 491 F.Supp. at 319.

Since the court did not discredit Dr. Manheim's testimony on the likely inadequacy of the minimum transportation requirements set forth in the Supplemental Agreement, they must for purposes of review be accepted in determining whether plaintiffs proved a prima facie case. Rebuttal of this case hinged on the testimony of James Tyler, a defense witness, assuring that WMC would have the financial ability to meet this projected need. However, it was plaintiffs' contention throughout trial that, regardless of WMC's good intentions, its future revenue projections were seriously flawed and that, as a consequence, it would be unable to satisfy the anticipated need for shuttle service. Elsewhere in its opinion the court conceded that plaintiffs had presented "strong evidence" that revenues were overestimated. 491 F.Supp. at 328 n.230. But it deemed such evidence "irrelevant" for purposes of weighing the prima facie case. As will be discussed in more detail later,6 this attack on the credibility of WMC's revenue projections was, to the contrary, highly relevant, and the court's acceptance of Tyler's assurances, with no consideration given to their reliability, tainted its conclusion that WMC would be able to upgrade its planned shuttle service. For purposes of plaintiffs' prima facie case, we are left with undisputed evidence of a lack of public transportation from minority and residential areas to the Southwest Division, and of a planned private shuttle bus system which, as presently projected, will be inadequate.

Plaintiffs also produced evidence, and the trial court found, that even assuming adequate free shuttle service connecting the Delaware and Southwest Divisions, minority inpatients and visitors will still suffer a disproportionate increase in the amount of time spent traveling to reach WMC's facilities. 491 F.Supp. at 320. The same would appear, from the evidence, to be true of the elderly. This time the burden would fall particularly heavily on the poor for whom it might mean lost wages, and on minority mothers of small children. There is also evidence that because of the nature of the populations affected — the elderly and minorities — an increase in travel time may translate into some decrease in the use of health care. 491 F.Supp. at 321. The court discounted the significance of this evidence.

Relying on section 604 of Title VI,7 it declined to consider any transportation impact on employees. This omission was improper, for while it is true that medicaid and medicare funds are not appropriated to provide employment, and thus may fall within section 604, the impact on employees still was relevant to the claim, advanced on behalf of hospital patients, that the Southwest and Delaware Divisions will become racially identifiable. In measuring the plaintiffs' prima facie case, therefore, evidence about access to the Southwest Division by minority employees was highly relevant. The failure to consider it was error.

As to evidence of possible decrease in use of health care facilities, the court rejected it, noting that

491 F.Supp. at 322. While there was evidence from which another inference might have been drawn about the deterrent effect of distance upon minorities and the elderly seeking inpatient care at the Southwest Division, the above finding cannot be described as clearly erroneous.

The court went on to find, however, that in two respects which were treated as de minimis the added distance and travel time would have a deterrent effect on the statutorily protected classes. First, the location of all specialty obstetrical care at the Southwest Division, including specialty clinics and high-risk prenatal care, may deter a few minority women to skip appointments. Second, there will be a disproportionate decrease in the number of minority and elderly visitors to the Southwest Division. 491 F.Supp. at 322. The significance of these disparate impacts upon the protected classes is discussed in connection with quality of care.

2. Quality of Care

(i) Physical Facilities

Plaintiffs presented expert testimony that the cost of the features of Plan Omega which called for the renovation of the Delaware Division to make it physically and aesthetically equal to the Southwest Division was seriously underestimated by the defendants. Under the Supplemental Agreement WMC agrees to use its best efforts to take $2,800,000 from the financing of Plan Omega and place it in escrow, to be expended on renovating the Delaware Division. However, plaintiffs' expert predicted the cost of renovations would be $24,700,000. Plaintiffs also offered what the court described as "strong evidence" tending to show that the revenue projections upon which Plan Omega is predicated were overestimated, and thus that the improvements in the Delaware Division would not be carried out. 491 F.Supp. at 328. The court did not discredit this testimony, rather it disregarded the evidence as irrelevant, thus presenting us with a legal issue.

The evidence of underestimation of the cost of renovating the Delaware Division was held to be irrelevant because of the categorical requirements of the Supplemental Agreement. 491 F.Supp. at 325. Even leaving the Agreement aside, the court found that the needed renovations were of the type that could be met by use of money from the operating budget and unrestricted funds. 491 F.Supp. at 327. If, however, the revenue projections are underestimated as plaintiffs alleged, the operating budget will not be achieved and unrestricted funds will be invaded to meet opening deficits. Although it noted the "seriousness of the consequences," the court refused to "take a hard second look at the plan," holding instead that plaintiffs' evidence of overestimation of revenues — indeed the whole question of Plan Omega's financial feasibility — was irrelevant,

491 F.Supp. at 328.

Such reliance on the Supplemental Agreement or the bond market as a basis for declining to consider plaintiffs' evidence on cost underestimation and revenue overestimation cannot withstand analysis. If Delaware Division were not to be renovated substantially, after erection of Southwest Division went forward, the case for disparate impact upon the protected classes would be overwhelming. It is no answer to say that substantial evidence of financial infeasibility is irrelevant because WMC has made a promise to HEW. In the first place the Supplemental Agreement does not appear to me as categorical as the trial court believed it to be.8 Moreover, the agreement is subject to amendment at any time.9 But more fundamentally, the court's deference to HEW enforcement of an agreement, in place of a decision on the merits of the plaintiffs' claims, is inconsistent with the mandate of this court in the prior appeal. When the case was last before us we considered both the ruling that the federal statutes relied on did not permit a private cause of action, and the ruling that HEW did not have to afford a trial-type hearing. We did not pass upon the merits of the appeal from the latter ruling, because we held that it was to the court that plaintiffs could look for relief. Reliance upon the Supplemental Agreement worked out between WMC and HEW, an agreement which antedated both of the orders we reviewed in our earlier decision, is inconsistent with the clear purpose of our remand for a trial in the district court.10

Even less defensible from a legal standpoint is a rejection of the relevancy of plaintiffs' financial feasibility evidence on the ground that they could rely upon the certain wisdom of the bond market. That market may, for all we know, be perfectly confident that the balance between the interest rate it seeks and the security offered by WMC, even if a default occurs, is an acceptable risk. The one thing that is fairly predictable is that if three years after Southwest Division is completed WMC cannot afford to complete the renovations of Delaware Division, the bondholders will be far more interested in keeping the former rather than the latter in operation. Thus it was error to reject the relevancy of plaintiffs' evidence on financial feasibility as a part of their prima facie case. That evidence, if credited, casts serious doubt upon the likelihood of Delaware Division being a hospital plant equivalent to Southwest Division, since the renovation depends on the availability of an operating surplus. The district court did not discredit it, and we may not. If one projects the possibility that Delaware Division may not be renovated, the case for disparate impact upon the minorities and the elderly who are likely to be its patients is strong.11

(ii) Quality of Medical Service

Aside from their general prediction that for financial reasons Delaware Division would be an inferior facility, the plaintiffs also offered evidence attempting to prove that the quality of care received by minorities and the elderly under Plan Omega will be poorer than that received by the general population.

Plaintiffs' first contention is that although both Divisions will offer some services in medicine and in surgery,12 the services delivered at Delaware Division will be inferior, (1) because of the unavailability of board-certified specialists in some specialties for consultation at that division, and (2) because of the likelihood that, even when a specialty service is offered at both, a shortage of board-certified physicians at the Delaware Division will result from their abandonment of it out of preference for Southwest. The impact in either event upon minorities and the elderly would be considerably greater than upon the general population. The court discredited the evidence suggesting that this scenario will occur; it reasoned that under Plan Omega specialists at Southwest would be required to be on call for consultation at Delaware, that Delaware would be a fully equipped hospital in which such subspecialists could practice,13 and that under the Plan doctors would be prohibited from limiting their practices to one division. In this instance, more than any other, the court was called upon to exercise intuition and judgment about human behavior in the future. While the district court's description of plaintiffs' evidence on this point as having "absolutely no merit" is hyperbolic, 491 F.Supp. at 330, one cannot say that the court erred in discrediting it. Its rejection is reinforced by the conclusion that in most cases patients in need of subspecialty services will be admitted directly to Southwest, and others can be transferred there. 491 F.Supp. at 329-30. But still, all depends on the assumption of financial feasibility.

Plaintiffs also attempted to prove that the quality of the resident staff at Delaware would be inferior in numbers and quality. The court concluded otherwise, because Plan Omega contemplates operating the residency program on a unified basis, with residents being rotated among floors in both facilities. They also challenged the quality of Emergency Room services, although the Plan calls for duplicate Class 1 emergency rooms with similar backup specialists. In both of these instances, as well, one is unable to say that the court erred in finding no disparate impact on the protected class, assuming financial feasibility of the entire program.

However, all of these findings favorable to WMC are tainted by the court's treatment of evidence of financial feasibility as irrelevant. All would probably have been determined otherwise if it were found that WMC could not afford to upgrade Delaware Division after the completion of Southwest Division. In that event, plaintiffs' case of disparate impact would be overwhelming.

In two respects the court found that plaintiffs had proved disparate impact in quality of care: access to a high risk prenatal clinic, and impact on visitors to inpatients.

As to prenatal care the evidence is clear that the incidence of high risk pregnancy is greater among minority women than in the female population at large. Obstetrical service will be concentrated at Southwest. Because of the distance to that division, the court found that persons in need of prenatal outpatient care, especially teenagers, will be more likely than others to miss appointments and be deterred from seeking care. 491 F.Supp. at 335. In considering plaintiffs' prima facie case this finding was discounted because, the court concluded,

491 F.Supp. at 335. If the quoted language was intended as a finding that there would be equivalent obstetrical clinic care at Southwest and Delaware, there is no support for it in the record. The evidence is that the only clinic at which an attending physician is present is the "obstetrical high risk clinic" (JA 525-6, 529), and that clinic will be at Southwest (JA 1002). Thus the finding of disparate impact on minority women with high risk pregnancies is unrebutted.

As to visitors, the court acknowledged that increased travel time will have a deterrent effect on minorities and the elderly. 491 F.Supp. at 332. Specialty services for which the elderly are particularly in need are concentrated at Southwest, for example, and the court acknowledged that a lack of visitors can have both psychic and physical deleterious effects on them. 491 F.Supp. at 332. All pediatrics inpatient services will likewise be at Southwest and the court conceded that visits from family members are extremely important for sick children. 491 F.Supp. at 338. Obstetrical inpatient services will be at Southwest, and OB inpatients were also found to have a great need for visitors. 491 F.Supp. at 334. But while conceding that elderly inpatients and minority users of WMC's pediatric and obstetrical services will be disproportionately impacted by the inhibiting effect of time and distance on visitors, the court disregarded those disparate impacts as insufficiently substantial to be taken into account in determining whether a prima facie case had been made out. The question of substantiality for purposes of Title VI and the Age Discrimination Act is a question of law. At least in the area of hospital care, where the marketplace typically affords few alternative facilities and impaired access may gravely affect health, I am reluctant to impose a very strict standard of substantiality. Therefore, I cannot agree with the court's conclusion that these disparate impacts are insubstantial.

3. Racial Identifiability

Finally, plaintiffs attack the Supplemental Agreement provision for inpatient utilization control, whereby patients needing services offered by both the Southwest and Delaware Divisions would be assigned to one or the other according to zip code. Expert testimony was presented suggesting that this method of allocation will fail to prevent the Delaware Division from becoming racially identifiable14 and will, in fact, cause overcrowding that can be alleviated only in such a way as to make the minority presence at the Delaware Division even more dominant. Noting that the assignment plan merely required the hospitals to offer beds to certain assigned patients, but did not compel the patients to take them, plaintiffs' experts testified that a multitude of factors would influence the choice of the more affluent, more mobile white population to go to the Southwest Division, while the poorer, less mobile minority population would await a bed at the Delaware Division. Further exacerbating the racial identifiability of the inner-city hospital, plaintiffs urge, would be the tendency for WMC's minority service employees to prefer working there because of the greater travel time required for them to reach the Southwest Division.

In the face of such evidence and an admission by HEW that it had not considered the overcrowding consequence, the court again chose to disregard it as "simply irrelevant" because of the Supplemental Agreement which requires WMC to remedy any ensuing racial identifiability "by whatever means necessary." 491 F.Supp. at 300. In response to plaintiffs' concerns that WMC might somehow avoid this contractual obligation, the court noted that "HEW can monitor WMC's activities, thus assuring that, if WMC does not meet its obligations, sanctions will be applied." 491 F.Supp. at 300.15 I have already indicated that such reliance on the terms of the Supplemental Agreement cannot substitute for a decision on the merits of plaintiffs' concerns. Since no weight was given to their evidence, the court's finding that the threatened overcrowding or racial identifiability can be remedied by simply transferring four inpatients per day from the Delaware to the Southwest Division, 491 F.Supp. at 339, cannot be upheld, especially when the transfer of minority and elderly inpatients risks reducing the quality of their care because of the impact it would have on their access to visitors.16

In summary, after looking at the entire record, including the evidence which the court erroneously disregarded, and considering the disparate impacts which were actually found, I conclude that the judgment appealed from cannot be affirmed on the basis of a failure by the plaintiffs to prove a prima facie case of disparate impact of Plan Omega on minorities and the elderly.

II.

As the opinion of the court correctly indicates,17 when plaintiffs in a Title VI or Age Discrimination Act case have produced evidence of disparate impact, the defendants must at least come forward with evidence in rebuttal or justification to "meet" the plaintiffs' prima facie case. Unfortunately, however, the court never gets more specific and, hence, there is a large gap in its analysis. By assuming that the plaintiffs' unrebutted evidence established a prima facie case of disparate impact, without discussing that evidence in detail, the opinion fails to focus upon the nature and extent of the several different impacts on which evidence was offered. It then passes on to the quite separate issue of burden of persuasion, without first pausing to inform just what it is the factfinder must be persuaded of. The content of the rebuttal or justification evidence cannot be determined in the abstract. It must be related to the precise impacts suggested by the plaintiffs' evidence.

If we were dealing with an effort to prove intentional race or age discrimination by inference from the circumstantial evidence of disparate impact upon minorities or elderly persons, the content of the rebuttal evidence would be sufficiently self-evident that an opinion writer might safely assume it need not be referred to explicitly; evidence of a nondiscriminatory business purpose for the challenged actions supports an inference that accomplishing that purpose was the true motive. Notwithstanding the inference of improper motive which arguably arises from disparate impact, the existence of a proper business purpose places the evidence on motive at best in equipoise, such that the party with the burden of persuasion on that issue loses. The plaintiff can still carry this burden by demonstrating that the business purpose was a pretext, feigned in order to hide the actual discriminatory animus. If such a pretext is established, the defendant loses, for he cannot ever justify action taken, regardless of pretext, for the purpose of discriminating.

In a case such as this one, however, where the plaintiffs do not rely upon an intent to discriminate, the legal standards for rebuttal or justification are much more complex. Part I of the opinion of the court is a welcome clarification of the analytical distinction between intent cases and disparate impact cases, and a proper recognition that spending power statutes such as Title VI and the Age Discrimination Act address both problems. Unfortunately, however, by ignoring the critical differences between the legal standards for rebuttal and justification in the two types of cases, the court leaves the law almost as confused as it was found.

When a member of a class protected by a funding statute's nondiscriminatory clause has produced evidence that a defendant's actions will impose a disparate impact on that class, there are two analytically separate kinds of evidence which may be offered. First, the defendant can rebut the evidence offered by plaintiffs, by evidence that the impacts complained of will not occur. For example, a defendant can show that steps will be or have been taken to effectively prevent their occurrence. Alternatively, a defendant can justify those impacts by showing that they must necessarily occur, if certain important objectives are to be accomplished, even though a defendant has chosen what it believes to be the feasible alternative having the least discriminatory impact. This distinction between rebuttal and justification is critical. Rebuttal is in essence a factual matter. Justification, on the other hand, while it involves factual matters which may be in dispute, also involves a legal standard, which the opinion of the court never supplies. It never decides the legal question presented by this case: what must defendants do before they may justify a redistribution of federal benefits away from a protected class. In my view, a plan imposing such an impact can only be justified by a showing that defendant has valid needs essential to its business or service, that the proposed plan will feasibly meet those needs, and that other plans with lesser impact on the protected classes will not.18

In the trial court WMC offered evidence both in rebuttal and in justification. The trial court considered both, and decided in WMC's favor. If the court had weighed all the relevant evidence regarding potential impacts and the steps taken to remove or mitigate them, and concluded that the alleged impacts would not occur, it might be proper to affirm on the basis that such a finding was not clearly erroneous. Alternatively, if the court had identified specific impacts which will occur, but concluded after weighing all the relevant evidence that these impacts are inevitable, because Plan Omega is the only feasible method of meeting WMC's compelling needs, that finding might likewise be affirmed as not clearly erroneous. But neither of these courses is properly open to this court on the present record because of the errors referred to in Part I above. The district court's entire analysis is flawed by its interrelated holdings that evidence of financial infeasibility was irrelevant and that the Supplemental Agreement would overcome all defects in Plan Omega bearing upon the vital issues of transportation, racial identifiability, and quality of service. By assuming favorable revenue projections and complete compliance with the Agreement, the court relieved WMC of the need to rebut those major impacts suggested by plaintiffs' experts.19 It also thereby narrowed the number and severity of the impacts which might have to be justified. Problems which would be major if WMC lacked the financial resources to make adjustments were made de minimis by the simple stroke of disregarding plaintiffs' financial evidence. Thus although the district court sought to apply the correct legal standards, both with respect to rebuttal and with respect to justification, it did so against a background of "meager" disparate impacts created by its own assumed factual matrix which cannot be relied upon. And because the majority opinion of this court addresses neither plaintiffs' impact evidence nor the appropriate legal standard for justification, the errors below go uncorrected.

The majority does attempt to put the trial court's ruling on financial feasibility in a favorable light, but in vain.20 Evidence that Plan Omega is fiscally unsound goes to the heart of the case, for it not only raises the threat of a racially identifiable facility and other adverse impacts, but also indicates that this burden placed on minorities and the elderly is unjustified. It is one thing to demonstrate, as defendants have done, that Plan Omega is designed to serve compelling objectives; whether it will accomplish them, however, is another, equally important question left open on this record. In light of the trial court's refusal to consider "strong evidence" that Plan Omega seriously overestimates future revenues and may therefore pose a drain on WMC's resources, I cannot find the consolation others have in the district court's "overarching" finding that the level of care for all population groups will improve. As suggested above, I concede that a substantial burden on the access of minorities and the elderly to federal benefits is nonetheless justified if the relocation is the only feasible alternative capable of meeting defendant's needs. Conversely, however, an infeasible plan is insufficient justification as a matter of law. Thus, to dismiss the issue of financial feasibility as completely irrelevant subverts the Congressional policy in favor of nondiscriminatory funding.21

The discussion so far has concerned the legal standards by which a defendant's actions should be measured, and not the matter of who bears the burden of persuasion. The point is that, regardless of where that burden is placed, the errors discussed in Part I above completely tainted the trial court's substantive analysis. That alone requires a reversal and remand.

III.

Equally troubling from a precedential standpoint, however, is the majority's decision equating the defendant's burden in a case of disparate impact with that borne in a case of intentional discrimination. Today's holding achieves an artificial symmetry, but at considerable cost to the prospects of eliminating all forms of discrimination which, as the opinion of the court confirms, was the impetus behind Title VI and the Age Discrimination Act. I have no serious difficulty with leaving the burden of persuasion on the plaintiffs when what is brought forward by defendants is rebuttal — evidence that there will be no impact — rather than justification. However, since in many cases the means of proof of justification will not be within the reach of the protected class and allocation of the burden of persuasion will be dispositive, the only allocation consistent with the Congressional intention to protect the disadvantaged class from actions having discriminatory effects is to assign the burden of justifying the impact to the defendant who receives federal funds. The policy decision was made by Congress, and in filling in the gaps in the statute we should allocate the burden consistent with that policy.

The majority reaches its result without considering the interest in evenhanded access to the federal benefits Congress provides and without acknowledging the problems of proof. Instead, it looks to the one line of discrimination cases where burdens have been clearly spelled out, involving intentional employment discrimination, and argues that such rules must be uniformly applied in all cases brought under nondiscrimination statutes. The plea for symmetry has only the attraction of relieving courts from analyzing the substantive differences between two quite distinct statutory prohibitions. The defendant accused of discriminating intentionally stands in a very different position than one accused of actions which have unintended adverse effects. As noted before, the plaintiffs in this case are not asking that the factfinder draw from circumstantial evidence of disparate impact, an inference of discriminatory intent. Rather, the gravamen of the complaint is that the benefits of a federal program are being redistributed away from protected classes unnecessarily, and that racially identifiable facilities are being spawned with the help of federal funds. The object of the litigation is not so much to punish a wrongdoer, as to prevent indifference and correct for inadvertence. The presumption in a case which has proceeded to the point in litigation where definite impact is established is that defendant is subjectively willing to mitigate the impacts, but is constrained from further accommodation by other pressing needs and the lack of reasonable alternatives. Something akin to an affirmative defense makes sense logically in such a context; on the other hand, I reject the majority's proposition that "[o]ne could just as readily say in an intent case that the necessity to prove a nondiscriminatory reason is an affirmative defense...." Maj. op. p. 1333. To the contrary, such reasons are offered to disprove the existence of discriminatory intent. They can never justify it.

The logic of requiring recipients of federal aid to justify the disparate impact caused by their actions was recognized by the Supreme Court in Board of Education v. Harris, 444 U.S. 130, 100 S.Ct. 363, 62 L.Ed.2d 275 (1979). Statistical proof of discriminatory effects in that case had established a prima facie violation of the Emergency School Aid Act. That inference could be rebutted, said the Court, adding

444 U.S. at 151, 100 S.Ct. at 375; cf., Southeastern Community College v. Davis, 442 U.S. 397, 407, 99 S.Ct. 2361, 2367, 60 L.Ed.2d 980 (1979) (§ 504). However much ambiguity surrounds the nature of the burden on a nonfederally-funded employer in Title VII impact cases, Harris represents a clear signal that burden-shifting on justification is called for where principles of nondiscrimination in funding are to be enforced.

Even within the context of Title VII, on which the majority focuses, appellate courts faced with having to clarify the nature of the employer's burden on the issue of business necessity have described it as "heavy" or labeled it as one of persuasion. See, e. g., Donnell v. General Motors Corp., 576 F.2d 1292, 1298 (8th Cir. 1978); Smith v. Olin Chemical Corp., 555 F.2d 1283, 1286 (5th Cir. 1977); Vulcan Society of N.Y. v. Civil Service Comm'n, 490 F.2d 387, 393 (2d Cir. 1973). The majority's suggestion that the Supreme Court has since settled the issue of burden in such impact cases is belied by the Court's most recent pronouncement in Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981). That opinion implies that the allocation of burdens in a disparate treatment case does not necessarily govern disparate impact cases.22 450 U.S. at 252 n.5, 101 S.Ct. at 1093 n.5.

Citing intent and impact cases indiscriminately and recommending they be approached in the same way threatens to confuse the litigation process and lead to unfortunate misunderstandings such as may have occurred in the decision of the district court. It held that WMC had met its burden of producing some evidence, and that the plaintiffs failed to prove that the hospital's justification for Plan Omega was pretextual. 491 F.Supp. at 345. This misconceives what it was the court had to decide. The contention was not that the reasons advanced were not genuine, but that there were feasible alternatives with less disparate impact. That was the heart of the issue: whether Plan Omega with its disproportionately adverse effects was unnecessary.23

The majority's observation that a completed Title VI violation involves not disparate impacts per se, but only those that are unnecessary, may arguably make terminology such as "affirmative defense" unsuitable; but it does not answer the question of where the risk of non-persuasion on the issue of justification should be placed as a matter of public policy and from the viewpoint of who has the evidence. Indeed, the problems of developing evidence and assembling it in admissible form cannot be separated from the fulfillment of statutory policy flowing from Congressional intent.24

Bringing such considerations to bear in this matter, I am concerned that allocating the risk of non-persuasion on justification in an impact case by reference to precedent established in cases of intentional discrimination ignores important distinctions between access to proof in the two types of cases. In intent cases, the plaintiff-victim is often in a better position than the defendant to develop and uncover the particular acts upon which the charge of discrimination is based. In some instances, such evidence is probably more accessible to the victim, particularly when the discriminatory acts are committed by subordinates, but must be defended by their superiors. Evidence of the presence or absence of impact may also be equally available. But assuming some impact, the proof of justification is always going to be in the possession of the defendant and rarely also in the possession of plaintiff members of the protected class, since the latter are unlikely to be privy to the decision-making process or the competing considerations which ultimately form the defense of justification.

One need look no further than the instant appeal to appreciate the crucial importance which the risk of non-persuasion assumes in a case like this. All parties seem reconciled to the need for placing some facility in the Stanton area to foreclose competition from another institution, and all parties agree that WMC needs a new large hospital. The dispute has to a large extent been whether some variety of "Reverse Omega" could mitigate the disparate impacts of a relocation and still feasibly accomplish the hospital's pressing needs. By relying on intent cases and placing the burden of demonstrating feasible alternatives on the plaintiffs, the court saddled them and their expert with the task of filling in the gaps in WMC's own consideration of alternatives,25 and relieved WMC, which had far more ready access to the information, of the need to persuade that the alternatives were not financially or medically feasible. Moreover, the risk of non-persuasion was made critical by the abundance of conflicting evidence as to the projected cost differential between Omega and Reverse Omega, and it is noteworthy that the court did not accept at face value WMC's cost estimates, but made its own. 491 F.Supp. at 345. In short, the burden placed on plaintiffs was impractical and the error severe, not harmless. Thus, even if the errors discussed in Parts I and II did not require reversal, the erroneous allocation of burden of persuasion does.

IV.

I do not minimize the complexity of the trial judge's task in this case, requiring as it does consideration of future consequences. However, since the effective delivery of federally-supported health care is at stake, the federal interest in preventing creation of a racially identifiable facility and the possible curtailment of hospital services is clear. The dockets of this court and others bear testimony to the difficulties of desegregating institutions after they have become segregated. Regulations affecting site selection, designed to forestall segregation in health care delivery, intentional or otherwise, from occurring in the first place, simply reflect the wisdom — acknowledged, ironically enough, in the medical profession — that an ounce of prevention is worth a pound of cure.