Kimberly Graham and her husband, Jon Graham, sought a preliminary mandatory injunction requiring Blue Cross & Blue Shield Mutual of Ohio1 to preauthorize and pay for a specialized chemotherapy treatment for Mrs. Graham. The district court denied the request for injunctive relief because the Grahams failed to show a likelihood of success on the merits. We find that the Grahams have not demonstrated that Mrs. Graham will suffer "irreparable harm" by the denial of this injunction. Although the district court discussed irreparable harm as an alternative ground for denying the injunction, the weight of the district court's decision rests on the Grahams' failure to demonstrate a likelihood of success on the merits. Therefore, we affirm the district court's denial of injunctive relief on different grounds.

Mrs. Graham was diagnosed with breast cancer in October 1995. She underwent a modified radical mastectomy and sought preventive care to avoid recurrence of the cancer. At the time of the mastectomy, six of her lymph nodes were found to be positive for cancer, but the cancer had not metastasized, i.e. spread to other parts of her body.2 Blue Cross paid for the procedure and subsequent induction chemotherapy in accordance with the insurance policy it issued to the Grahams.

Although she was free of metastatic disease, Mrs. Graham was in a high-risk category for recurrence. Therefore Mrs. Graham sought aggressive preventive treatment. She contacted Cancer Centers of America and arranged to meet with Dr. Jorge Frank, a specialist in internal medicine with subspecialties in oncology and hematology. Dr. Frank recommended a treatment known as High Dose Chemotherapy with Stem Cell Rescue ("HDC/SCR"). HDC/SCR differs from standard chemotherapy in that patients take cancerfighting drugs in doses six to ten times higher than in standard chemotherapy. The dosage level is so high that the treatment damages the patient's bone marrow, where blood cells are formed. This results in an increased susceptibility to infection and bleeding. To combat this side effect, a quantity of the patient's stem cells are collected from the bone marrow or peripheral blood prior to the therapy and then reintroduced into the patient's body after the chemotherapy treatment.

In January 1996, Dr. Frank requested that Blue Cross precertify that it would pay for the HDC/SCR treatment. On February 7, 1996, this request was denied. Based on the information submitted by Dr. Frank, as well as on the recommendation of an independent medical consultant retained through CORE,3 Blue Cross found that HDC/SCR is an "experimental/investigative" treatment. Blue Cross excludes this category of treatment from coverage under the Grahams' policy. The denial letter advised Dr. Frank of the Grahams' right to appeal the decision.

On March 6, 1996, the Grahams instituted an appeal. After receiving the opinion of a second independent consultant and following the established review procedure, Blue Cross again denied the precertification. The Grahams requested a third and final appeal on April 6, 1996. Blue Cross sought the opinion of a third independent consultant, again followed the established review procedures, and again concluded that HDC/SCR was not covered under the policy. The denial of the second appeal prompted the Grahams to institute this action for a mandatory preliminary injunction to compel Blue Cross to precertify and pay for Mrs. Graham's HDC/SCR treatment.

The sole question before us on appeal is whether the district court abused its discretion in denying the Grahams' motion for a mandatory preliminary injunction.4 To obtain a preliminary injunction, the moving party must show "1) a reasonable likelihood of success on the merits, and 2) no adequate remedy at law and irreparable harm if preliminary relief is denied." Mil-Mar Shoe Co., Inc. v. Shonac Corp., 75 F.3d 1153, 1156 (7th Cir.1996); see also TMT North America, Inc. v. Magic Touch GmbH, 124 F.3d 876, 877 (7th Cir.1997); Grossbaum v. Indianapolis-Marion County Bldg. Auth., 100 F.3d 1287, 1291 (7th Cir.1996), cert. denied, --- U.S. ----, 117 S. Ct. 1822, 137 L. Ed. 2d 1030 (1997); Roth v. Lutheran Gen. Hosp., 57 F.3d 1446, 1453 (7th Cir.1995); Abbott Lab. v. Mead Johnson & Co., 971 F.2d 6, 11 (7th Cir.1992). If the movants satisfy this initial burden, the court must balance the irreparable harm to the non-moving party if the injunction is granted against the irreparable harm to the moving party if the injunction is denied. See Grossbaum, 100 F.3d at 1291; Publications Int'l, Ltd. v. Meredith Corp., 88 F.3d 473, 478 (7th Cir.1996); Abbott Lab., 971 F.2d at 11. The court must also consider the effect of the injunction on nonparties. See TMT North America, 124 F.3d 876, 877; Grossbaum, 100 F.3d at 1291-92; Erickson v. Trinity Theatre, Inc., 13 F.3d 1061, 1067 (7th Cir.1994).

The Grahams seek a mandatory preliminary injunction, that is, an injunction requiring an affirmative act by the defendant, Blue Cross. Because a mandatory injunction requires the court to command the defendant to take a particular action, "mandatory preliminary writs are ordinarily cautiously viewed and sparingly issued." Jordan v. Wolke, 593 F.2d 772, 774 (7th Cir.1978); see also W.A. Mack, Inc. v. General Motors Corp., 260 F.2d 886, 890 (7th Cir.1958) (finding that "mandatory injunctions are rarely issued and interlocutory mandatory injunctions are even more rarely issued, and neither except upon the clearest equitable grounds").

A district court's decision to grant or deny an injunction is entitled to deference by the reviewing courts; we will reverse its decision only for an abuse of discretion. See Roth, 57 F.3d at 1453; Storck USA, L.P. v. Farley Candy Co., 14 F.3d 311, 314 (7th Cir.1994). In deciding whether to grant or deny a preliminary injunction, a district court will make factual findings and conclusions of law; we review the district court's findings of fact for clear error and its conclusions of law de novo. See Aircraft Owners and Pilots Ass'n v. Hinson, 102 F.3d 1421 (7th Cir.1996); David K. v. Lane, 839 F.2d 1265, 1271 (7th Cir.1988).

The Grahams must show that Mrs. Graham will suffer irreparable harm if the injunction is denied. Irreparable harm is harm "which cannot be repaired, retrieved, put down again, atoned for.... [T]he injury must be of a particular nature, so that compensation in money cannot atone for it." Gause v. Perkins, 56 N.C. (3 Jones Eq.) 177 (1857) (quoted in A.O. Smith Corp. v. F.T.C., 530 F.2d 515, 525 (3d Cir.1976)); see also Roland Mach. Co. v. Dresser Indus., Inc., 749 F.2d 380, 386 (7th Cir.1984) (Irreparable harm is "harm that cannot be prevented or fully rectified by the final judgment after trial."). The Grahams argue that irreparable harm will occur because:

A. In the absence of preauthorization, [Mrs. Graham] has been treated frequently with regular chemotherapy, which causes such serious and grave destruction of bodily functions and tissues that it is further destroying, in each and every session, her limited opportunity to be treated by [HDC/ SCR]; and,

B. [Mrs. Graham's] prognosis, without the prompt commencement of the course of treatment prescribed, including [HDC/SCR], is that she will develop metastatic malignancy and increased lymph node involvement within 2 years; and, that once such condition exists, she will have very little chance for recovery or any extended period of disease-free life.

....

E. Plaintiffs cannot afford to pay for the prescribed treatment nor are funds available from any other source.

Complaint para. 30, Graham v. Blue Cross & Blue Shield Mutual of Ohio, No. 96 C 3217 (N.D.Ill. Oct. 30, 1996) (emphasis in original). These assertions rest on the presumption that HDC/SCR is more efficacious than standard chemotherapy in preventing the recurrence of cancer in patients like Mrs. Graham. The Blue Cross policy covers standard chemotherapy for Mrs. Graham's population, but the Grahams wish to pursue this alternative treatment program. If the Grahams cannot produce evidence that the prescribed treatment is indeed more effective than the traditional chemotherapy treatment she is undergoing at Blue Cross's expense, their claim that Mrs. Graham will suffer irreparable harm fails.

The district court found that determining the efficacy of a new medical treatment requires "Phase III" studies.5 Graham, slip op. at 10. This conclusion is supported by the testimony of all of Blue Cross's experts. Even Mrs. Graham's expert witness, Dr. Barry Meisenberg, admitted that the effectiveness of HDC/SCR could not be established without random testing. As stated by Dr. Meisenberg, "It may take a long time to demonstrate a benefit if there is one. This benefit could then only be proved by a randomized trial or randomized trials.... I think the randomized trial is the gold standard and the highest level of evidence.... [R]andomized trials are going to be required to prove this ... beyond a reasonable doubt." Meisenberg Dep. at 76.

It is undisputed that no Phase III studies have been completed for persons in Mrs. Graham's situation--Stage II with four to nine positive nodes. Although there is some evidence from Phase II studies that HDC/SCR may be helpful for Mrs. Graham's population, a conclusion regarding the relative efficacy of the treatment, as compared to standard chemotherapy, must await completion of the Phase III studies.

The record is replete with statements supporting this position. For example, Dr. John Erban, a physician employed by CORE to give an opinion on Mrs. Graham's precertification request, stated in his affidavit: "There are studies currently underway which are investigating the effect of [HDC/SCR] on patients in the adjuvant setting who are similarly situated to Mrs. Graham, but to date, no conclusions can be drawn from these studies as to whether either of these treatments is better than, worse than, or equal to conventional chemotherapy or other treatment options." Erban Aff. para. 5. In his deposition, Dr. Erban testified as follows:

Q. With respect to the subclassification of overall breast cancer patients that Kimberly Graham is a part of, to what extent have any well-designed studies been conducted and completed so as to permit anyone to draw conclusions about a positive net health outcome for the treatment that was proposed?

A. I believe it's too early to say that there are any that are completed that will answer the question for Mrs. Graham and for patients like Mrs. Graham.

Erban Dep. at 103-04. Even the plaintiffs' witnesses testified that the effects of HDC/SCR, as compared with standard chemotherapy, are uncertain at the present time. Dr. Jorge Frank, Mrs. Graham's treating physician, admitted that he could not conclude with certainty that HDC/SCR would be more helpful to Mrs. Graham than standard chemotherapy. He stated, "I think that there has not been enough data accumulated on sufficient numbers of patients with certainty, with certainty to say that this [HDC/SCR] is the treatment and that this is better than anything else." Rec. Tr. Vol. 2 at 130. Dr. Frank also acknowledged on cross-examination that the lack of information from randomized clinical trials meant that he did not know with certainty what would happen to Mrs. Graham if the treatment he proposed was administered. Rec. Tr. Vol. 2 at 125-27. In addition to these statements by Mrs. Graham's treating physician, her expert witness concluded: "We'll need the completion of phase 3 randomized trials to know what benefit, if any, and what magnitude the benefit is to this therapy." Meisenberg Dep. at 83.

In short, there is no evidence to suggest that the treatment Mrs. Graham seeks is any more effective than the standard chemotherapy currently available to her. As such, we find that the Grahams cannot establish that Mrs. Graham will suffer irreparable harm. Thus, the district court did not abuse its discretion by denying the Grahams' request for a mandatory preliminary injunction. We therefore AFFIRM the district court's denial of the injunction.