In this action under § 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794, which prohibits a recipient of federal financial assistance from denying benefits to an “otherwise qualified” handicapped person solely because of his or her handicap, defendants, New York University and certain of its administrators (collectively referred to herein as NYU), appeal from an order of the Southern District of New York entered by Judge Gerard L. Goettel granting plaintiff Jane Doe preliminary injunctive relief directing NYU Medical School to readmit her as a first-year medical student and denying defendants’ motion for summary judgment dismissing her complaint. We reverse the grant of preliminary injunctive relief and affirm the denial of NYU’s motion for summary judgment.
Since the district court based its decision solely on a written record (consisting principally of affidavits, reports and medical records), without any evidentiary hearing, we are not limited, in determining whether preliminary injunctive relief was properly granted, to review of its exercise of discretion but may consider the record de novo, being “in as good a position as the district judge to read and interpret the pleadings, affidavits and depositions,” Jack Kahn Music Co., Inc. v. Baldwin Piano & Organ Co., 604 F.2d 755, 758 (2d Cir. 1979) (quoting Dopp v. Franklin National Bank, 461 F.2d 873, 879 (2d Cir. 1972)). Preliminary injunctive relief ordinarily “should not be resolved on the basis of affidavits which evince disputed issues of fact,” the proper course being first to resolve credibility issues through an evidentiary hearing, Forts v. Ward, 566 F.2d 849, 851-52 (2d Cir. 1977): Dopp v. Franklin National Bank, supra, 461 F.2d at 879, unless the facts admitted by the defendant would plainly entitle the plaintiff to such relief. In light of these principles, which apply with greater force where mandatory relief changing the status quo has been granted, we summarize the pertinent record facts.
Each year approximately 5,000 persons apply for admission as medical students to NYU Medical School, of whom those found best qualified by the Admissions Committee, some 170, are accepted. In 1975 Jane Doe was accepted after falsely representing *766in her application that she did not have any chronic or recurrent illnesses or emotional problems. In fact she, while gifted academically, had suffered for many years from serious psychiatric and mental disorders, which evidenced themselves in the form of numerous self-destructive acts and attacks upon others, followed by periodic treatments by psychologists and psychiatrists and admissions to various psychiatric hospitals for care and therapy. In the third grade she had trouble with her teacher, requiring the help of a psychologist and psychiatrist. In 1963, at 14 years of age, she tore up a report card and took five Doriden tablets (sleeping pills) in anger at her parents. Beginning in November, 1963, she was treated by a psychiatrist on 24 weekly visits. Despite her doctor’s recommendation that she submit to at least several more months of therapy she terminated this treatment in June, 1964.
There is no evidence of further symptoms of psychiatric difficulties on Doe’s part until November, 1972, when at the age of 23 she became upset by an interview she had had at the University of San Francisco Medical School and attempted suicide by drinking potassium cyanide. Over her resistance she was taken to the Stanford University Hospital emergency room. There she threatened to kill herself and had to be strapped to a stretcher. She was then moved to Santa Clara Valley Medical Center (Valley), where she stayed for 2}h days. In January, 1973, she injected herself with cytosine arabinoside, a drug for treatment of cancer which attacks certain cells and renders them susceptible to infection. In March, 1973, she carved a hole in her stomach with a kitchen knife, using a local anesthetic stolen from a laboratory. Later in the month, after a counsellor at school asked her to accept inpatient treatment at Langley Porter Neuropsychiatric Institute, she severed an artery in her elbow with a razor blade. She was then admitted to Langley Porter, where she stayed for eight days. During her stay there she broke a light bulb, scratched her wrists, attacked a doctor, tore his nameplate off his door, and left against medical advice.
In July, 1973, after seeing her psychiatrist, Dr. David N. Daniels, Doe cut a vein in her left arm, losing a litre of blood which necessitated a transfusion and was admitted to Valley where she stayed six days. During her stay there she tried to leave, pulled sutures out of her wound, wrote in her blood on the wall, took charts from a nursing station, bit a staff member in the arm, and departed against medical advice. In October, 1973, after a visit to Dr. Daniels, she used a razor blade to cut her foot, resisted police custody and was admitted against her will to Valley for a 14-day stay. In the hospital she attacked a woman doctor, scratching her in the forearm and kicking her in the pelvis. In January, 1974, she cut herself again and was admitted to the Cowell Student Health Service at Stanford University. In March, 1974, she cut her left elbow, tried to seize her health records, kicked a doctor in the groin and tried to break a window, following which she was taken to the Stanford Hospital where she falsified her identity, claiming to be “Mari-ta S. Williams.” This was followed by more instances in which she cut herself and was admitted to hospitals for short periods.
In May, 1974, Doe became angry when her psychiatrist, Dr. Daniels, did not see her immediately upon her arrival at his office, and proceeded to cut her arm and smear blood on the wall of his waiting room. After the doctor finished with the patient he had been seeing and observed Doe in the waiting room, she attacked him, first biting him on the wrist and attempting to kick him in the groin, and then charging at him with a pair of scissors in an apparent effort to stab him. The doctor escaped to his inner office, locking his door. Doe left, only to return shortly thereafter and stand outside the doctor’s window with a syringe, filled with a cyanide solution, that she had jabbed into her arm. The doctor called the police, who took her into custody. Doe was subsequently taken to Valley again, where she fought with staff members. In spite of a recommendation that she stay in the hospital for 90 days of treatment, she escaped through a window with the help of her husband six days after being admitted.
*767In July, 1974, Doe applied to five medical schools through the American Colleges Admissions System, falsely representing that she did not then have and had not had any chronic or recurrent illnesses or emotional problems. In December, 1974, she applied to NYU Medical School, answering “NO” to the question, “Do you have any chronic or recurrent illnesses, emotional problems, or bodily defects?” Her application was accepted in April of 1975, on the understanding she would become a student in September of that year. In the meantime, in January, 1975, while still in California, Doe cut herself in the waiting room of another psychiatrist, Dr. Charles W. Casella, who was scheduled to see her. When she lost a large amount of blood she was admitted to Kaiser Permanente Medical Center where she remained five days despite a recommendation that she stay for an indefinite period.
After starting at NYU Medical School in September, Doe was required as a matriculating student to undergo a medical examination early that month but delayed doing so until October 30, 1975. When the doctor who examined her, Dr. Michael Ruoff, noticed scars on Doe’s arm, and inquired as to their cause she for the first time informed the University of some of her history of psychiatric problems. Dr. Ruoff recommended that Doe undergo a psychiatric examination to determine whether she was fit to stay at the school. Her reply was, “It is my body and what I do to it is my concern. If I want to go out and f___a cat I can.” She then walked out without completing the examination.
Shortly thereafter Doe met with Dr. David S. Scotch, Associate Dean of NYU Medical School, and agreed to be examined by a Student Health Service psychiatrist, Dr. Marvin Stern. In an interview with Dr. Stern on November 3, 1975, she gave a more detailed history of her psychiatric problems. Dr. Stern reached the conclusion that Doe had a “fragile personality” and sent her to Dr. Emmanuel Fisher for psychological tests. Doe underwent testing by Fisher on November 5, 1975. On the basis of these tests and his interview of her, Dr. Fisher also concluded that Doe had a serious psychiatric problem. He “noted that she had a grossly detached and alienated personality, with no effective intellectual or emotional contact with the world of things or people.”
On November 10 Doe was again examined by Dr. Stem, following which Dr. Fisher recommended to Dr. Scotch that Doe be asked to withdraw from the medical school. Dr. Scotch agreed, and informed Doe of the decision on November 12. Doe asked to discuss the matter with Dr. Ivan Bennett, the Dean of the Medical School, and did so the same day. After reviewing the matter further Scotch and Bennett agreed to permit Doe to remain at NYU on the condition that she undertake psychiatric therapy with a medical follow-up by the Student Health Service. Doe accepted these conditions and was advised that if she had further psychiatric trouble she would be expected to withdraw from the school.
Doe began psychiatric therapy with Dr. Grace Frank of Bellevue Hospital while continuing her medical school program. Apparently this therapy was unsuccessful and it was eventually terminated. Doe maintains that the reason for termination was a scheduling problem; NYU maintains that the reason was Doe’s lack of cooperation. Other than for this difficulty, the rest of November, all of December, and early January, 1976, were all uneventful.
On Friday, January 30,1976, Doe went to Dr. Scotch’s office, apparently by appointment, to discuss with him a conflict in her schedule that would occur the following Monday. Dr. Scotch was not in his office when Doe arrived, and she became distressed and angry. She left the office and attempted to calm herself, but to no avail. She returned and told Dr. Scotch that she would have to revert to her past habits in order to cope with the situation. She retreated to a bathroom, where she bled herself with a catheter. An hour later she returned to Dr. Scotch’s office and told him what she had done, explaining that it was the only way she could cope with her stress. Dr. Scotch requested her resignation.
*768Doe met with Dr. Bennett the following Monday, February 2, discussed her situation with him and made a written proposal for a leave of absence, which was granted on the understanding that she might request reinstatement, which would be considered but not guaranteed in view of the problems she presented. Her tuition and dormitory fee for the second semester were refunded. After going to the University Hospital’s psychiatric ward and then refusing to be admitted, on February 9 Doe was told by NYU to move out of her dormitory room and stop attending classes. She finally agreed to enter the Payne-Whitney Psychiatric Clinic (not affiliated with NYU) two days later, but left against medical advice on February 16. Doe apparently wandered around New York during the next three days, sleeping in public restrooms. She returned to Payne-Whitney on February 19, admitting that she was in need of help, and stayed there for 17 more days, after which her insurance ran out, but made no progress. Her condition on discharge was listed as “no improvement,” with a diagnosis by Dr. James H. Spencer of “Borderline personality . . . Personality disorders, other specified types 301.89.”
A personality disorder classified as “Borderline Personality” is a serious condition, manifesting itself by a series of five or more recognizable characteristics, according to the American Psychiatric Association’s Diagnostic and Statistical Manual (3d ed.) (known as “DSM-III”). A person suffering from it is likely to have it continue through most of his or her adult life, subject to modification only by treatment by well-trained therapists over a period of years and adoption of a lifestyle which avoids situations that subject the person to types of stress with which he or she cannot cope.
After leaving Payne-Whitney in March of 1976, Doe returned to California, where she was treated by two psychiatrists simultaneously on an outpatient basis, one (Dr. Casella) providing individual therapy, the other (Dr. Richards) family therapy in which her husband joined. Doe states that leaving NYU was “a major turning point in my life.” She maintains that through the treatment she received in California she was able to master her psychiatric problems, develop healthy ways of dealing with stress, and cease her self-destructive behavior. Apparently she sought readmittance to NYU in July of 1976; the record before us is unclear on how she did so and on how NYU responded.
Doe and her husband moved to New York in October of 1976, apparently because he had a fellowship there. She took a job with an advertising agency and continued to undergo psychiatric treatment in New York until sometime between April and June of 1977. NYU and Doe disagree about the degree to which this treatment was successful. According to NYU she was unable to handle group therapy sessions at Payne-Whitney. According to Dr. Warren Tanenbaum, with whom she began treatment in February, 1977, which ended in April or May, she was ill at ease with him, hated psychiatrists, was late for her sessions, blamed Dr. Scotch for her difficulties at NYU, had thought of sticking scissors into her ribs, and feared she could not manage medical school.
In June or July of 1977, after having returned to California, Doe applied for readmission to NYU. Her application was supported by both of the psychiatrists who had treated her while she had been in California after leaving NYU, Drs. Richards and Casella, even though the first had seen her at most three times and the second at most two times during that summer. Both sent letters of support to NYU. According to NYU’s standards a student seeking readmission
“must demonstrate, that the problems that precipitated the leave are resolved, that the applicant must be able to handle all of the academic and emotional stress of attending medical school, and that the school must be satisfied that the applicant will be able to function properly after graduation as a physician.”
More specifically NYU states that Doe would be required to show
*769“(c) that she does not pose a significant risk of reexhibiting her prior disorder either; (i) when readmitted to the same Medical School environment which previously caused her such difficulty; or (ii) when she might in the future be fully licensed and authorized to practice medicine; and
“(d) that in addition to being fully cured, she possesses the additional qualifications of good judgment, personal integrity and truthfulness, and a genuine commitment to the medical profession.”
The Payne-Whitney files on Doe were obtained by NYU for purposes of considering her application and turned over to the Medical School’s Chairman of the Department of Psychiatry, Dr. Robert Cancro, for a recommendation' as to her fitness to return to the school.
Upon Dr. Cancro’s recommendation and the judgments of other NYU faculty members, made after considering the letters from Doe’s two California psychiatrists, NYU decided not to readmit Doe. On August 29, 1977, Associate Dean Jacobus Potter told Doe of that decision over the phone. Dean Potter sent Doe a letter confirming the decision on September 12, 1977. In the letter Dean Potter stated:
“Major factors entering into this decision were of course the responsibilities and obligations of the School of Medicine, but it should be emphasized that your own best interests were not ignored. While the personal interview which you inquired about was not judged to be necessary for adequate consideration of your application. Dr. Cancro has kindly indicated his willingness to meet with you should you wish to discuss aspects of your own situation that you feel may not be covered in the record.”
At approximately the same time (fall of 1977) Doe was admitted as a graduate student to the Harvard School of Public Health to undertake a course of study leading to a degree of Masters of Science in Health and Policy Management. In her Harvard University medical questionnaire she falsely stated that she had not experienced any nervousness, worry or emotional disturbance causing a loss of time from work or study.
In response to NYU’s refusal to readmit her Doe sought legal assistance, and on October 25, 1977, an attorney representing her wrote to NYU requesting that it reconsider Doe’s application and threatening to bring legal action if it did not. NYU responded by agreeing to reopen the application on condition that Doe agree to be interviewed by an NYU psychiatrist. In order to determine whether it was in her own best interest to be interviewed, Doe consulted Dr. William Stage, a psychiatrist trainee at the Harvard University Health Services Psychiatric Clinic (Holyoke Center), whom she had previously seen for the purpose of obtaining therapy. Stage diagnosed her as having borderline personality disorganization, which he viewed as having been severe during 1972-76 and moderate to severe as of November 29, 1977, and noted that she “would likely have trouble” in medical school and might have trouble functioning in an emergency room context. On the other hand, Dr. Samuel Bojar, Associate Professor of Psychiatry at the Harvard Medical School, who apparently was Stage’s superior and was also consulted by Doe on seven occasions from December, 1977 well into 1978, found that she did not manifest the characteristics of a Borderline Personality Disorder but a “chronic, neurotic depression” which he described as “a treatable condition.”
On December 2, 1977, the parties (NYU and Doe) agreed to an interview by Dr. Veva H. Zimmerman, an Associate Professor of Clinical Psychiatry at NYU Medical School.1 Dr. Zimmerman examined Doe on *770December 13, 1977, and made a report to Dr. Scotch by letter on December 27, 1977, explaining that she had met with Jane Doe for one hour and 40 minutes and that in order to avoid biasing herself she had not read any reports on Doe before the examination. Dr. Zimmerman stated that she had then read all materials in Doe’s file, including the letters of recommendation by Doe’s two California psychiatrists, and concluded that there were strong indications that Doe’s “basic personality remains essentially the same as reported by Drs. Stern and Fisher in 1976 and in the discharge summaries from Payne-Whitney Hospital.” While recognizing that some progress had been made, Dr. Zimmerman noted that Doe depended greatly on her marriage “to help her maintain psychological equilibrium” and that although Doe might be able to handle a graduate program involving “purely intellectual material,” medical school, which required successful interaction with people, was quite a different matter. She therefore recommended against readmitting Doe.2
On the basis of the Zimmerman report NYU adhered to its earlier decision and on December 28, 1977, Jane Doe brought this action, claiming that the Zimmerman report could not be validated as a predictor of success in medical school and that NYU, a recipient of federal funds, had denied her readmission in violation of her rights under § 504 of the Rehabilitation Act of 1973, 29 U.S.C. §§ 791, 794 and regulations thereunder, 45 C.F.R. Part 84. Her complaint sought declaratory and injunctive relief plus damages. She immediately sought a temporary restraining order and a preliminary injunction requiring NYU to readmit her.3 The district court, Goettel, J., denied both motions in a memorandum opinion and order on January 3, 1978. It also required Doe to exhaust her administrative remedies with the Office of Civil Rights (OCR) of the Department of Health, Education, and Welfare (HEW). Doe v. New York University, 442 F.Supp. 522 (S.D.N.Y.1978). Discovery proceeded while the parties awaited a decision by the OCR.
In June, 1978, Doe successfully completed her course at Harvard, receiving a masters degree, and took a summer internship at the Department of Health, Education and Welfare in Washington. There she successfully served in various capacities from the summer of 1978 until October, 1981, when the district court required that NYU readmit her. She was promoted from intern to a research assistant in the Office of the Deputy Assistant Secretary of HEW for Planning and Evaluation, which develops policy and manages the decision-making process for major proposals to Congress. In the spring of 1979 her supervisor evaluated her performance as excellent and she, with seven of her colleagues, received an unusual cash award with a statement by the Assistant Secretary that she had worked long hours, producing high quality work under pressure. She then served as liaison with the Food and Drug Administration, attending meetings with the Secretary of HEW (later HHS) and representatives of groups dealing with departmental policy. The high quality of her services was attested to by various officials. There is no evidence that she has engaged in any self-destructive or anti-social behavior since the fall of 1977.
Pursuant to the district court’s order of January 3, 1978, Doe filed a complaint with *771OCR in January, 1978. OCR investigated her claim, and on August 14, 1979, its Regional Director, Charles Tejada, issued a letter to her attorney finding that she had been discriminated against by NYU in violation of § 504. At approximately the same time Mr. Tejada sent a detailed letter to NYU stating and explaining OCR’s findings. On September 12 NYU responded with a detailed letter disagreeing with the findings and requesting reconsideration. By this time discovery had gone forward and the case was on Judge Goettel’s active docket. On September 19, 1979, Judge Goettel removed it by order without objection from either party in order to allow the OCR to attempt to effect a conciliation, as it is required to do by its regulations. The order provided that either party could reinstate the case on the active calendar.4 On January 14, 1980, Regional Director Tejada sent a second letter to NYU reaffirming the OCR Regional Office’s earlier finding.
The matter then languished in OCR while it considered whether to cut off NYU’s federal funding or take other steps to enforce its findings. Efforts by OCR to effectuate a conciliation failed. The Department of Education, which had the power to seek enforcement of OCR’s ruling, declined to do so or to seek a cut-off of federal funding, in view of the pendency of the present action.
On October 27, 1980, Doe moved to have the case restored to the active calendar. Judge Goettel granted her motion on April 9, 1981. On August 6, 1981, Doe moved for a preliminary injunction and subsequently NYU cross-moved for summary judgment.
In support of her motion Doe argued as she does to us that she met this Circuit’s requirements for preliminary injunctive relief. She claimed that she would be irreparably harmed unless NYU was required to readmit her to medical school; that she was likely to succeed on the merits because she was a handicapped person covered by § 504, 29 U.S.C. § 794, who had recovered from her earlier psychiatric difficulties and was now fully qualified to study medicine at NYU; and that NYU had refused to admit her solely because of her past history of psychiatric difficulties. To support her claim that she had overcome her psychiatric difficulties, Doe produced affidavits from several psychiatrists (Drs. Casella, Richards, Bojar, Calkins) who had either treated her at one time or another or talked with her, stating that in their judgment she was fit to attend medical school.
NYU contended that its decision not to readmit Doe should be upheld against a § 504 challenge as long as it could show that there was a “substantial basis” for its decision. It argued that such a basis had been shown, and thus that its motion for summary judgment should be granted. It pointed to Doe’s history of psychiatric difficulties and to the diagnoses of several psychiatrists who had examined her that she had suffered and continued to suffer from a “Borderline Personality disorder,” a serious psychiatric problem that is extremely difficult to cure and that can lead to the type of self-destructive behavior that Doe had engaged in in the past. NYU relied not only on the opinions of in-house psychiatrists, such as Drs. Fisher, Stem, and Zimmerman, but also on Dr. Lawrence C. Kolb, a former President of the American Psychiatric Association and an eminent clinical psychiatrist.5 *772When Doe refused to submit to an examination by Dr. Kolb he executed an affidavit setting forth his opinion on the basis of his study of the pertinent records in the case, including extensive depositions of Doe and the numerous medical reports and interviews. Dr. Kolb stated:
“It is my opinion, which I can state with a reasonable degree of medical certainty, that Jane Doe suffers from a Personality Disorder known as Borderline Personality, and that her condition is serious.. . . None of her therapuetic efforts ... would be considered of sufficient duration or intensity by me or by most experienced therapists to bring about major personality change of such a nature that Jane Doe would not relapse in the face of future emotional stress such as that encountered in medical school.”
As the result of an order by the district court, Doe finally submitted to an examination by Dr. Kolb on August 27, 1981. He reported, “In my opinion Mrs. Jane Doe remains at high risk of recurrence of personality disorganization if exposed to situations of stress such as would occur on return to medical school.”
NYU also produced an affidavit of Dr. Zimmerman, who had re-examined Doe in September of 1978 and had again concluded at that time that she was not qualified for readmission. Dr. Zimmerman reaffirmed this conclusion, stating:
“The character of [the] stress [of medical school], Jane Doe’s prior history of failing to be able to deal with it, her diagnosis of personality disorder, and her failure to have received any appropriate degree of psychiatric treatment for it leads me to conclude that she has not made any reasonable showing that she is now qualified to re-enter medical school.”
NYU also relied on the Payne-Whitney diagnosis of Borderline Personality, which confirmed several other medical diagnoses. It pointed out that even one of Jane Doe’s own doctors, Dr. Samuel Bojar, although disagreeing with the diagnosis in Doe’s case, admitted that “a borderline person is not very treatable. It’s a fixed situation.” On the basis of these medical diagnoses, which it claimed were well supported by Doe’s history, NYU argued that Doe poses a danger to herself, her teachers, her fellow students, and her patients.
In a decision issued on September 25, 1981, Judge Goettel found that Doe was a “handicapped person” under § 504 and implementing regulations, 34 C.F.R. § 104.3(j) and 45 C.F.R. § 84.3(j). He stated that whether she was “otherwise qualified” for admission to NYU Medical School depended on whether her psychiatric symptoms would recur, preventing her from performing as a medical student and causing her to pose a danger to herself and others. In deciding this issue the judge discounted the testimony of the psychiatrists on both sides and relied upon Doe’s “actual behavior and condition over the past five years.” He also gave no weight to the findings of the OCR except for its finding that Doe was a handicapped person, since the OCR had not conducted any independent investigation, the Regional Director and Department of Education had backed away from seeking enforcement of its determination, Doe was an HEW employee at the time when the OCR recommendations were made, and there were indications of possible OCR bias in her favor, including the unauthorized disclosure to her counsel of agency action. He concluded that Doe “will more likely than not be able to complete her course of medical studies and serve creditably as a physician,” that she was therefore “otherwise qualified” under the Act, and that she had been *773denied readmission “solely because of the handicap.” Applying the approach of the Tenth Circuit in its recent decision in Pushkin v. University of Colorado, 658 F.2d 1372, 1981, he further concluded that NYU had failed to sustain its burden of going forward and proving that Doe was not an otherwise qualified handicapped person or that her application for readmission was rejected for reasons other than her handicap. No deference was given to NYU’s evaluation of Doe’s qualifications as compared with those of other qualified first-year students. Judge Goettel denied NYU’s motion for summary judgment and found that Doe was likely to prevail on the merits.
On the question of whether a preliminary injunction should issue, Judge Goettel found that an additional year’s delay in Doe’s medical studies would irreparably harm her. Because likelihood of success on the merits had already been shown, this finding would have been sufficient to satisfy the requirements for a preliminary injunction, but the judge went on to find that if Doe had a relapse, “it will manifest itself in self-destructive acts or hostility to authority figures such as professor and doctors” rather than to fellow students or patients. He conceded, however, that there was a danger to patients. He further found that NYU would suffer no other hardships if it were forced to admit Doe. Accordingly he issued a mandatory preliminary injunction directing that Doe be readmitted to NYU Medical School pending trial of the action. This appeal was expedited, although a stay pending appeal was denied; Doe entered medical school on October 6, 1981.
DISCUSSION
Preliminary Injunctive Relief
Turning first to the propriety of preliminary injunctive relief, Doe was entitled to a preliminary injunction only upon showing (a) that she would otherwise suffer irreparable injury and (b) either (1) a likelihood of success on the merits or (2) sufficiently serious questions going to the merits to make them a fair ground for litigation and a balance of hardships tipping decidedly in her favor. Jackson Dairy, Inc. v. H. P. Hood & Sons, Inc., 596 F.2d 70, 72 (2d Cir. 1979); Caulfield v. Board of Education, 583 F.2d 605 (2d Cir. 1978). Where, as here, mandatory relief is sought, as distinguished from maintenance of the status quo, a strong showing of irreparable injury must be made, since relief changing the status quo is not favored unless the facts and law clearly support the moving party. Clune v. Publishers’ Association, 214 F.Supp. 520, 531 (S.D.N.Y.), aff’d per curiam on opinion below, 314 F.2d 343 (2d Cir. 1963); Martinez v. Mathews, 544 F.2d 1233, 1243 (5th Cir. 1976).
Irreparable Injury
On this record Doe did not in our view make a sufficient showing of irreparable injury to entitle her to mandatory injunctive relief. Ordinarily a one-year delay in obtaining admission to a graduate school for the purpose of pursuing professional studies, as distinguished from interruption or termination of attendance already in progress, is insufficient to warrant an injunction in the absence of other circumstances militating in favor of such relief. See, e.g., Donnelly v. Boston College, 401 F.Supp. 1, 4 (D.Mass.1975), aff’d per curiam, 558 F.2d 634 (1st Cir.), cert. denied, 434 U.S. 987, 98 S.Ct. 618, 64 L.Ed.2d 483 (1977); Timmerman v. University of Toledo, 421 F.Supp. 464, 466-67 (N.D.Ohio 1976). No such other circumstances exist here. On the contrary, there are strong indications that Doe could have waited a year without any' serious loss. At the time when she sought relief she was gainfully employed in a responsible HEW position where she was highly valued by her supervisors. She and her husband had purchased a home near Washington, DC. If she prevailed on the merits she would be entitled to commence her studies at 33 years of age and despite her argument to the contrary would not be precluded from having a child during her matriculation.
*774It further appears that Doe could have reapplied for preliminary relief or sought trial at least a year earlier than she did. Although Judge Goettel, when he originally denied preliminary relief on January 3, 1978, directed Doe to exhaust her administrative remedies with OCR and on September 19, 1979, after OCR had issued its first letter of findings, removed the case from his active calendar to give the parties an opportunity to reach a voluntary resolution of the matter, he permitted discovery to proceed concurrently with the OCR investigation and provided that either party could apply to restore the case to his active calendar at any time. Despite OCR’s issuance on January 14, 1980, of its second letter of findings, Doe did not move to restore the case to active status until October 27, 1980 (well after the NYU Medical School year had commenced) and simultaneously moved to shift the burden of proof to NYU. After Judge Goettel restored the case to active status on April 9, 1981, denying Doe’s motion to shift the burden, and on June 16, 1981, ruled that he would not give any weight to the OCR’s findings, Doe did not move for preliminary relief until August 10, 1981.
Under these circumstances we would ordinarily reverse the grant of preliminary relief solely on the ground that Doe failed to make a sufficient showing of irreparable injury. We are confronted, however, with the district court’s issuance of mandatory injunctive relief, in reliance on which Doe has entered NYU and apparently left her position in Washington.6 It therefore becomes advisable to consider the merits of her claim, not only for the purpose of determining whether she has shown a likelihood of success but also to ascertain whether the district court properly denied summary judgment.
The Merits
Section 504 of the Rehabilitation Act of 1973 provides that “[n]o otherwise qualified handicapped individual in the United States, as defined in section 706(7) of this title, shall, solely by reason of his handicap, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving Federal financial assistance. . . . ” 29 U.S.C. § 794 (Supp. Ill 1979).
Although the Supreme Court has not ruled on the question of whether § 504 creates a private right of action in favor of a handicapped person injured by its violation, see Southeastern Community College v. Davis, 442 U.S. 397, 404 n.5, 99 S.Ct. 2361, 2366, 60 L.Ed.2d 980 (1979), we and other circuits have ruled that it does. Leary v. Crapsey, 566 F.2d 863, 865 (2d Cir. 1977); Lloyd v. Regional Transportation Authority, 548 F.2d 1277 (7th Cir. 1977); United Handicapped Federation v. Andre, 558 F.2d 413 (8th Cir. 1977). Since the statute, which is patterned after § 601 of the Civil Rights Act of 1964, 42 U.S.C. § 2000d and § 901 of the Education Amendments of 1972, 20 U.S.C. § 1681, was clearly intended for the special benefit of handicapped persons and a private right of action is both consistent with the underlying purposes of the legislative scheme and not one traditionally relegated to state law, the requirements of Cort v. Ash, 422 U.S. 66, 78, 95 S.Ct. 2080, 2088, 45 L.Ed.2d 26 (1975), for such an action are met.
In order to make out a case based on a violation of § 504 a plaintiff must prove (1) that she is a “handicapped person” under the Act, (2) that she is “otherwise qualified” for the position sought, (3) that she is being excluded from the position solely by reason of her handicap, and (4) that the position exists as part of a program or activity receiving Federal financial assist*775anee. Since the last of these elements is not questioned, the present case turns on proof of the first three.
Section 7(7)(B) of the Act, 29 U.S.C. § 706(7)(B), provides in pertinent part that a “handicapped” person is “any person who (i) has a physical or mental impairment which substantially limits one or more of such person’s major life activities, (ii) has a record of such an impairment, or (iii) is regarded as having such an impairment.” The term “major life activities” is defined in 45 C.F.R. § 84.3(j)(2)(ii):
“ ‘Major life activities’ means functions such as caring for one’s self, performing manual tasks, walking, seeing, hearing, speaking, breathing, learning, and work-mg.
In claiming to be a handicapped person under the Act, Doe is faced with her representations on applications that she did not suffer from any emotional problems and her testimony that her ability to function in major life activities has never been impaired and that she has never been unable to work or learn, as is attested to by her successful graduation from college, her receipt of a Masters Degree from Harvard and her outstanding record of employment as a member of HEW. Notwithstanding this evidence we believe that for present purposes she should be classified as a handicapped person under the Act, in view of the independent evidence of her extensive history of mental impairments requiring hospitalizations and her departure from NYU in 1976 because of her psychiatric problems, all of which indicate that she has suffered from a substantial limitation on a major life activity, the ability to handle stressful situations of the type faced in a medical training milieu. NYU’s refusal to readmit her on the ground that she poses an unacceptable risk to faculty, students, and patients makes clear that she is “regarded as having such an impairment.” § 7(7)(B)(iii). Our conclusion is reinforced by the wide scope of the definition in § 7(7)(B), which includes in subdivision (ii) a “record of such impairment,” and by its legislative history, which indicates that the definition is not to be construed in a niggardly fashion. See S.Rep. No. 93 — 1297, 93d Cong., 2d Sess. (1974), reprinted in [1974] U.S.Code Cong. & Ad.News 6373, 6388-91.
Turning to the Act’s term, “otherwise qualified handicapped individual,” it is now clear that this refers to a person who is qualified in spite of her handicap and that an institution is not required to disregard the disabilities of a handicapped applicant, provided the handicap is relevant to reasonable qualifications for acceptance, or to make substantial modifications in its reasonable standards or program to accommodate handicapped individuals but may take an applicant’s handicap into consideration, along with all other relevant factors, in determining whether she is qualified for admission. Southeastern Community College v. Davis, supra, 442 U.S. at 406, 99 S.Ct. at 2367. The institution need not dispense with reasonable precautions or requirements which it would normally impose for safe participation by students,, doctors and patients in its activities. Section 504 simply insures the institution’s even-handed treatment of a handicapped applicant who meets reasonable standards so that he or she will not be discriminated against solely because of the handicap. But if the handicap could reasonably be viewed as posing a substantial risk that the applicant would be unable to meet its reasonable standards, the institution is not obligated by the Act to alter, dilute or bend them to admit the handicapped applicant. Southeastern Community College v. Davis, supra, 442 U.S. at 413 n.12, 99 S.Ct. at 2370.
In determining whether a handicapped person is “otherwise qualified” for admission to an institution of higher education, a court must also consider other factors not normally encountered in evaluating ability to satisfy employment standards or to qualify for a job. The first of these is a court’s limited ability, as contrasted to that of experienced educational administrators and professionals, to determine an applicant’s qualifications and whether he or she would meet reasonable standards for academic and professional achievement estab*776lished by a university or a non-legal profession. “Courts are particularly ill-equipped to evaluate academic performance,” Board of Curators of University of Missouri v. Horowitz, 435 U.S. 78, 92, 98 S.Ct. 948, 956, 55 L.Ed.2d 124 (1978). For this reason, although the Act requires us rather than the institution to make the final determination of whether a handicapped individual is “otherwise qualified,” cf. New York State Ass’n for Retarded Children v. Carey, 612 F.2d 644 (2d Cir. 1979), considerable judicial deference must be paid to the evaluation made by the institution itself, absent proof that its standards and its application of them serve no purpose other than to deny an education to handicapped persons.
Another factor which must be taken into account is that the qualification of a handicapped person for admission to an institution turns not only on whether he or she meets .its reasonable standards but whether the individual, where a few (in this case 170) must be chosen out of thousands of applicants, is as well qualified despite the handicap as others accepted for one of the limited number of openings. In performing the difficult task, where there are more qualified applicants than places available, of making comparative judgments to determine which are the most promising candidates, the institution is not required to accept a qualified handicapped person if the handicap renders that individual less qualified than other qualified applicants.
Since an institution or employer is permitted to take into consideration an applicant’s handicap in deciding whether he or she is qualified, a § 504 action frequently does not lend itself easily to the analysis used for allocation of burdens and order of presentation of proof used in suits alleging discrimination based on impermissible factors (race, color, religion, sex or national origin) in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e, et seq., McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973); Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981). In such suits, although the plaintiff has the ultimate burden of proving by a fair preponderance of the evidence that the defendant discriminated against him on the basis of an impermissible factor, he may establish a prima facie case by proving that he applied for a position for which he was qualified and was rejected under circumstances indicating discrimination on the basis of an impermissible factor. The burden then shifts to the defendant to rebut the presumption of discrimination by coming forward with evidence that the plaintiff was rejected for a legitimate reason, whereupon the plaintiff must prove that the reason was not true but a pretext for impermissible discrimination. That procedure may be appropriate for § 504 suits in which the defendant disclaims any reliance on the plaintiff’s handicap.
On the other hand, in the more typical suit under § 504, the defendant acknowledges reliance on the plaintiff’s handicap and since the plaintiff’s handicap may be a permissible factor to be taken into account in determining whether he is qualified, the order of presentation of proof in such cases cannot be framed in terms of permissible versus impermissible factors. The pivotal issue is not whether the handicap was considered but whether under all of the circumstances it provides a reasonable basis for finding the plaintiff not to be qualified or not as well qualified as other applicants. Accordingly we hold that in a suit under § 504 the plaintiff may make out a prima facie case by showing that he is a handicapped person under the Act and that, although he is qualified apart from his handicap, he was denied admission or employment because of his handicap. The burden then shifts to the institution or employer to rebut the inference that the handicap was improperly taken into account by going forward with evidence that the handicap is relevant to qualifications for the position sought. Cf. Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977). The plaintiff must then bear the ultimate burden of showing by a preponderance of the evidence that in spite of the *777handicap he is qualified and, where the defendant claims and comes forward with some evidence that the plaintiff’s handicap renders him less qualified than other successful applicants, that he is at least as well qualified as other applicants who were accepted.7
Applying these principles, it is clear in the present case that Doe, a handicapped person, was denied readmission because of her handicap. NYU has come forward with evidence that the handicap was relevant to her qualifications for readmission according to its standards (see p. 770 supra) which appear reasonable enough and are not challenged. She therefore bears the burden of showing that despite her handicap she is qualified. Since her admission in 1975 was obtained on her false representation that she did not suffer from any recurrent illnesses or emotional problems, that initial admission does not serve to establish that she is “otherwise qualified” under the Act except to indicate that except for her personality disorder, which involved self-destructive and antisocial behavior, she was academically acceptable.8 NYU, however, was entitled, in determining whether she was qualified, to be advised of and to take into account her mental impairment, since it is directly relevant to her qualifications and bears upon her ability to function as a student and doctor, to get along with other persons, and to withstand stress of the type encountered in medical training and practice. NYU is of necessity concerned with the safety of other students, faculty and patients to whom Doe would be exposed, since this could adversely affect them as well as the success and reputation of its Medical School activities. Any harm done by her as a medical student to others, moreover, might expose it to legal liability for knowingly permitting such exposure.
The crucial question to be resolved in determining whether Doe is “otherwise qualified” under the Act is the substantiality of the risk that her mental disturbances will recur, resulting in behavior harmful to herself and others. The district court adopted as its test that she must be deemed qualified if it appeared “more likely than not” that she could complete her medical training and serve as a physician without recurrence of her self-destructive and antisocial conduct. We disagree with this standard. In our view she would not be qualified for readmission if there is a significant risk of such recurrence. It would be unreasonable to infer that Congress intended to force institutions to accept or readmit persons who pose a significant risk of harm to themselves or others, even if the chances of harm were less than 50%. Indeed, even if she presents any appreciable risk of such harm, this factor could properly be taken into account in deciding whether, among qualified applicants, it rendered her less qualified than others for the limited number of places available. In view of the seriousness of the harm inflicted in prior episodes, NYU is not required to give preference to her over other qualified applicants who do not pose any such appreciable risk at all.
The evidence in the record before us indicates that there is a significant risk that Doe will have a recurrence of her mental disorder, with resulting danger to herself and to others with whom she would be associated as a medical student. Several psychiatrists and mental institutions enjoying excellent reputations in their field, including a concededly eminent leader (Dr. Lawrence C. Kolb) who examined her on *778August 27,1981, and an Associate Professor who interviewed her on December 13, 1977 (Dr. Veva Zimmerman), have diagnosed her as suffering from a serious condition known as a “Borderline Personality” disorder which is likely to continue through most of her adult life. The seriousness of its manifestations can only be minimized by treatment by well-trained therapists over a long period of time and by the adoption of a lifestyle that avoids stress of the unusual types faced in medical training and practice. The opinions of Drs. Kolb and Zimmerman are corroborated in varying degrees by the early diagnosis of Dr. Daniels that she has a “passive-aggressive personality” disorder, the Kaiser Permanente Medical Center’s diagnosis of “Personality Disorder,” and Payne-Whitney’s diagnosis of “Borderline Personality Disorder.” The non-recurrence of Doe’s self-destructive and antisocial activity for the past four years, during which she has peacefully coexisted with others, is attributed to the fact that the types of stress to which she has been subjected at the Harvard School of Public Health and as an HEW employee do not approximate the seriousness of those which she would experience as a medical student and doctor.9 Moreover, her history indicates that although there were no manifestations of disorder for seven years after the earlier episodes in 1963-64, they recurred during the period 1972-1977, indicating that despite a period of dormancy they may recur again.
In support of her claim Doe, on the other hand, has offered the opinions of various psychiatrists (Drs. Charles W. Casella, Elizabeth M. Richards, Samuel Bojar and David R. Calkins) to the effect that she is fit to pursue a medical career. In affidavits Dr. Casella expressed the view that she has “substantially recovered from her psychiatric illness,” and Drs. Richards and Bojar gave as their opinion that she does not now suffer from a Borderline Personality disorder. However, none of these psychiatrists rules out the risk that she may suffer a recurrence. Moreover, in their depositions Dr. Casella described Doe as having a “Passive-Aggressive personality” (DSM — II, 301.-81) which is a type of personality disorder and Dr. Richards, while preferring to describe her as suffering from a “Depressive Neurosis” (DSM-II, 300.4), agreed that she might fall in the category of “Passive-Aggressive personality.” Dr. Bojar, while not agreeing that Doe had a Borderline Personality disorder, described his diagnosis as follows:
“Q. Did you make a diagnosis of Jane Doe at that time?
“A. Yes. My diagnosis was that of a chronic neurotic depression with anxiety precipitated when her particular needs could not be met. If I may elaborate on my formulation — I prefer ‘formulations’ rather than diagnostic labels. I see her as a perfectionist who is an overachiever, who imposes very high demands on herself and on others; and when these can’t be met, she undergoes a great deal of frustration.
“Her security is based on her doing well, and when she cannot utilize this defense of really achieving at a high level she becomes anxious and, well, quite depressed. When she was making demands on others, when she needed others and people could not come through, she felt the same kind of frustration that she would feel from herself if she could not do well; and it was in these settings that the hostility would arise and some of these cutting episodes occurred, which I *779saw as really expression of anger and hostility against the others.”
Dr. Calkins’ summary affidavit could hardly be given much weight, since he is not a specialist in psychiatry or behavioral sciences and his training in this area was limited to a one month clerkship at medical school, from which he graduated in 1975, and occasional experience in psychiatry as a resident, principally among outpatients.
For the reasons stated by the district court, it acted properly in refusing to give any weight to the findings of the OCR and, if the action is tried, they should not be accepted as evidence. On the other hand, the court erred in ruling that it would disregard for the most part Doe’s prior psychiatric history and treat the expert testimony as being of lesser reliability than her recent behavior. In light of the type of behavioral disorder presented, which could result in a recurrence after a dormant period, expert opinion was entitled to greater weight in reaching a decision as to whether Doe was qualified. The court is required to perform the disagreeable task of considering Doe’s entire psychiatric history and weighing the expert testimony on each side.
In our view, Doe, in addition to her failure to show threatened irreparable injury, has not on this record established any likelihood of success in proving that despite her handicap she is qualified for acceptance as a medical student or to engage in the practice of medicine. Moreover, while there may possibly be questions going to the merits that are sufficiently serious to make them a fair ground for litigation, the evidence indicates that there is a significant risk of recurrence of her self-destructive and harmful conduct, which NYU should not be required to bear pending trial and that a substantial basis exists for upholding NYU’s decision to deny her readmission.10 See Kampmeier v. Nyquist, 553 F.2d 296 (2d Cir. 1977). Against this, pending a full decision on the merits, plaintiff can show only the hardship of delay of another year in admission to medical school, which as indicated above, she has voluntarily chosen to do in the past. Thus the balance of hardships tips in NYU's favor, not in Doe's. Accordingly, the grant of a mandatory preliminary injunction must be reversed.
Summary Judgment
NYU would be entitled to summary judgment only upon showing that there is no genuine issue as to material facts which would mandate a judgment in its favor as a matter of law. Fed.R.Civ.P. 56, United States v. Bosurgi, 530 F.2d 1105, 1110 (2d Cir. 1976); Heyman v. Commerce & Industry Insurance Co., 524 F.2d 1317, 1320 (2d Cir. 1975). Although on the record before us Doe has not shown a likelihood of success on the merits and, on the contrary, it appears quite likely that NYU would prevail at trial, some material facts relied upon by NYU are not attested to on personal knowledge, as required by F.R.Civ.P. 56(e), but are summaries by counsel of information contained in depositions, medical records and reports. These records and reports would probably be admissible as business records at trial upon proper authentication. Moreover, Doe’s Rule 9(g) statement, while generally challenging NYU’s statement of facts points to only a few relatively insignificant episodes (not relied upon by us) alleged on personal knowledge to be untrue and devotes itself mainly to urging acceptance of the opinions of the OCR and her medical experts rather than those of doctors relied upon by NYU, which are set forth in the form of affidavits based on personal knowledge. It is conceivable, although on this record not probable, that substantial additional evidence might be offered at trial as to the significance of the risk that she will suffer a recurrence of her mental disorder or that the testimony of the experts offered by affidavit upon the appli*780cation for preliminary relief might be sharpened or weakened by examination and cross-examination on trial. Thus we cannot say unqualifiedly that no issue exists as to the determinative material fact, which is whether Doe could adduce evidence entitling the trier of the fact to infer that despite her handicap she is at least as qualified as other applicants accepted by NYU as medical students, after following the procedures and applying the standards we have outlined. However, even if the risk of recurrence were shown to be only minimal, NYU would not be precluded from adducing evidence that it was sufficient to render her less qualified than another applicant who would be accepted in her place.
Accordingly, the district court’s denial of NYU’s motion for summary judgment is affirmed and its grant of mandatory preliminary injunctive relief is reversed. In order to insure that the parties will be restored as promptly as possible to the positions occupied by them before the district court’s decision, the mandate shall issue forthwith. Each side will bear its own costs.