C.J. Gideon, Jr. (argued), North & Gideon, Nashville, Tenn., for plaintiffs-appellants.

Joe B. Brown, U.S. Atty., Nashville, Tenn., Douglas Thoreson (argued), for defendant-appellee.

Before MILBURN and BOGGS, Circuit Judges, and ALDRICH, District Judge*.

MILBURN, Circuit Judge.

Plaintiffs Karen and Mary Keir appeal from the judgment of the district court in favor of defendant United States of America in this medical malpractice action under the Federal Tort Claims Act ("FTCA"). For the reasons that follow, we reverse the judgment of the district court and remand for further proceedings consistent with this opinion.

I.

This litigation arises out of a course of events ending in plaintiff Karen Keir's loss of the sight of her left eye. A premature, low birth weight infant, Karen was born on March 9, 1975, suffering from hyaline membrane disease. Karen underwent oxygen therapy shortly after birth, and received close ophthalmological follow-up to guard against the development of retrolental fibroplasia, an ophthalmological disorder which often follows intensive oxygen therapy. By the age of two, the risk of development of this disorder becomes nonexistent. Consequently, there remains no need for subsequent retinal examinations to rule out this disorder after a child passes beyond that age.

On September 30, 1977, when Karen was two and one-half years of age, her father, a drill sergeant in the United States Army, was stationed at Fort Campbell, Kentucky. On that date, Karen's parents took her to the military optometry clinic at Fort Campbell because Karen's left eye had begun turning in intermittently, a condition known as strabismus or accommodative esotropia. Karen was again examined at the Fort Campbell Optometry Clinic on November 28, 1977. Major Richardson, who examined Karen on this visit, noted that her left eye continued to turn inward on occasion even when she was wearing her glasses. J.A. 123.

In early 1978, Karen's father was transferred to Korea, and Karen and her mother moved to Nashville, Tennessee. Dr. Willard Tirrill, a Nashville ophthalmologist, treated Karen from January 26, 1978, until February 6, 1979. On Karen's second visit, February 27, 1978, Dr. Tirrill performed a dilated examination. He concluded that Karen suffered farsightedness and accommodative esotropia, treated her with glasses, and believed that this therapy was effective.

Although Dr. Tirrill's records reflect that he dilated Karen's eyes during the February 27, 1978, examination, his records do not reflect whether he used a direct or an indirect ophthalmoscope. The difference between the two is of crucial significance in this case. A direct ophthalmoscope is primarily utilized to examine the retina's posterior segment. An indirect ophthalmoscope is utilized to visualize a greater part of the retina. Use of an indirect ophthalmoscope may permit visualization to the equator of the eye and perhaps slightly anterior thereto, depending upon the cooperation of the patient. In any event, this examination did not result in the detection of pathology. Dr. Tirrill dilated Karen's eyes again on February 6, 1979, and once again the examination was negative for pathology.

Dr. Tirrill testified that he was unable to determine Karen's visual acuities during the period in which he treated her. The reason for this is that small children are unable to communicate effectively, and thus the ability to determine the quality of the child's vision is impeded. He testified that, when Karen left his care, he did not have any clinical basis to suspect that Karen suffered any abnormality other than accommodative esotropia.

Subsequent to February 6, 1979, the Keir family was transferred to Fort Dix, New Jersey. Mrs. Mary Keir contacted the central medical appointment desk at Fort Dix and requested an appointment with an ophthalmologist. Mary Keir was told that hospital policy required her to see an optometrist first, and that after this initial evaluation, she would be referred to an ophthalmologist.

Mary Keir scheduled an appointment for July 20, 1979. At that time, she and Karen reported to the optometry clinic at Fort Dix. Mary Keir testified that upon their arrival, "a man called us back and checked her glasses and did a little eye test on her and told me to wait in the waiting room, that we would see Dr. Channing." J.A. at 140. "Dr. Channing" is Major, now Lieutenant Colonel, Eugene Channing, an Army optometrist.

Mary Keir testified that when she and Karen went into Dr. Channing's office, she observed that it looked like the other offices in which Karen had been treated. She further testified that she believed Dr. Channing was an ophthalmologist. J.A. 142.

Dr. Channing testified that the standard of care required him to take a medical history from the patient upon her initial visit. J.A. 191. Instead of writing on Karen's chart that she developed strabismus at age two and had been under the care of an ophthalmologist since birth, he described her ocular history as negative. He elaborated on the statement by noting on the chart that she had a left esotropia. Dr. Channing explained his failure to record this medical history by noting that the information was available in other medical records. J.A. 192. However, when answering interrogatories, he stated that Mary Keir did not inform him of Karen's complete ocular history until her last visit on August 11, 1980. When confronted with this discrepancy, Dr. Channing stated that the answer to the interrogatory was incorrect. J.A. 263.

At Karen's initial visit on July 20, 1979, Dr. Channing measured Karen's glasses with a lensometer and took her visual acuities in order to determine whether her prescription was correct. He examined the external structure of Karen's eyes and noted no swelling, infection or lesions. He then utilized a direct ophthalmoscope to evaluate the internal structure of Karen's eyes. He testified that the use of a direct ophthalmoscope was consistent with the standard of practice in the state of New Jersey at the time. In the course of this undilated examination with a direct ophthalmoscope, Dr. Channing was able to visualize the fovea of the eye, the optic disc, the optic nerve, the visual axis, the macula, and the media. Dr. Channing detected no pathology at this time. J.A. 231-38.

During the initial visit, Dr. Channing learned that Karen had been under ophthalmological care, and he asked Mrs. Keir to provide him with a letter from Karen's former ophthalmologist. Dr. Tirrill sent such a letter, setting forth his general findings and evaluation. The letter also stated that Karen was to be examined in four months by an ophthalmologist in New Jersey.

At all times relevant to this dispute, a standard operating procedure ("SOP") in effect at Walston Army Hospital at Fort Dix provided that "[a]ll patients with strabismus, amblyopia and dyslexia are to be referred to the ophthalmologist with their prescribed glasses, if any. The ophthalmologist will then discuss the disorder and treatments to be prescribed." J.A. 66. Despite the existence of this SOP and Dr. Tirrill's statement that Karen should be seen by an ophthalmologist, Dr. Channing continued to treat her.

When Mary Keir gave the letter from Dr. Tirrill to Dr. Channing on September 25, 1979, Dr. Channing read it and put it in Karen's file. He indicated that Dr. Tirrill's findings with regard to Karen's farsightedness were similar to his. Because there was no mention of pathology, and only an indication of a refractive problem, Dr. Channing believed that he as an optometrist could handle Karen's condition through the use of glasses, patching, and possibly training. J.A. 241-46.

During the September 25, 1979, examination, Dr. Channing performed a retinoscopy for purposes of measuring refractive error. The retinoscope allows the examiner to determine that the area of central vision, the macula, is clear and functioning. No visual acuities were taken during that examination. J.A. 246-49.

Karen returned on October 3, 1979, for the purpose of evaluation under cycloplegia, a drug that causes the eye muscles to relax. The results of this examination indicated once again that Karen was farsighted. At that time, Karen's aided far visual acuity was 20/30? in the right eye and 20/40? in the left eye. The discrepancy in visual acuities is characteristic of amblyopia. Accordingly, Dr. Channing advised Karen's mother to pursue alternate patching in an attempt to make the left eye as strong as the right eye, and he instructed her to return with Karen in three months. J.A. 249-53.

Karen returned for a follow-up examination on February 12, 1980. During that examination, Dr. Channing noted that the patching procedure had not been effective because the patch was sliding off Karen's eye and that Karen could not see when the right eye was covered. Mary Keir informed Dr. Channing that Karen did not like the patch, and she was having a difficult time convincing Karen to keep it on her eye. Aided far visual acuities at that time were determined to be 20/30 in the right eye and 20/30 in the left eye. This represented a slight improvement over the visual acuities taken on October 3, 1979. J.A. 211, 256.

Dr. Channing changed Karen's prescription because her lenses were scratched, and he also performed a retinoscopy which showed results similar to those achieved on the previous visit. Dr. Channing observed no redness or inflammation in Karen's eye, indicated that there was no evidence of pathology, and decided to continue the patching procedure for three more months. J.A. 256-58.

Although Mary Keir was instructed to return with Karen after three months, she did not return until August 11, 1980. The delay was caused in substantial part by medical complications arising from Mary Keir's pregnancy. When Karen was examined on August 11, 1980, Mary Keir pointed out a white spot on Karen's eye. At that time, Dr. Channing immediately referred her to ophthalmology. He arranged to have Karen seen by Dr. Biagio Mignone at the emergency room of the Fort Dix hospital. When Dr. Mignone began reading Karen's chart, he threw it across the room and told Mary Keir that Dr. Channing had no right to be treating Karen. He then looked at Karen's eye and informed Mrs. Keir that Karen was suffering from either a dogworm infection or a tumor. Accordingly, Dr. Mignone made arrangements for Karen to be examined under anesthesia on August 12, 1980. J.A. 151, 275.

During the examination under anesthesia, Dr. Mignone discovered a large amount of inflammation in the vitreous. He also indicated that the tumor was very large, approximately twelve to fifteen disc diameters, or roughly the size of a little fingernail. He indicated that the tumor was mostly anterior to the equator. He testified that, after examining the tumor, he believed that it had been present in Karen's eye for a minimum of six to eight months. J.A. 275-77.

After the examination under anesthesia, Dr. Mignone referred Karen to Dr. Jerry Shields of the Wills Eye Hospital in Philadelphia, Pennsylvania. Dr. Mignone also informed Dr. Channing that he was displeased with the care Dr. Channing had provided for Karen Keir. Dr. Mignone dictated a memorandum requesting that Karen's chart be secured and transferred to the U.S. Army Branch at Walter Reed Hospital, but the memorandum has been lost or destroyed.

The initial examination at Wills Eye Hospital revealed a fluffy, white nodular retinal mass of the infero temporal periphery with scattered satellite lesions extending in the periphery from approximately the three o'clock meridian clockwise to the eight o'clock meridian with some finer discrete lesions extending as far superiorly in the infero nasal quadrant as the nine o'clock meridian in Karen's left eye. The central portion of the vitreous in her left eye was filled with hazy cells. There were large blood vessels entering and exiting the tumor in her left eye with some blood on the surface of the tumor. Ultrasonography of Karen's left eye confirmed the presence of retinoblastoma, the most common form of intraocular cancer in children. J.A. 69-70, 72.

On August 15, 1980, Karen underwent a second ocular examination under anesthesia. During that examination, Dr. James Augsburger and Dr. Shields confirmed the presence of the peripheral white tumor in Karen's left eye. The tumor had begun to seed. The largest mass in the eye measured thirteen by fifteen milimeters in diameter and was eight milimeters thick. J.A. 70-71. A pediatric examination was performed at the Children's Hospital Oncology Unit. No evidence of metastatic retinoblastoma was detected; in other words, at the time of the diagnosis, the retinoblastoma had not spread outside of the eye. J.A. 71.

Because of the location of the tumor and the extent of vitreous seeding, only two treatment modalities were available. Karen and her parents were told that the doctors could perform external beam irradiation or enucleation, removal of the eye. Karen made it perfectly clear that she did not want her eye removed, and, accordingly, the Keir family opted for radiation.

Karen was then referred to the Department of Radiation Oncology at Hanemann Hospital in Philadelphia, Pennsylvania. Over a course of therapy completed on October 1, 1980, Karen received between 5,076 and 5,139 rads of radiation during twenty-seven treatments over a period of thirty-eight days. J.A. 71.

Karen remained under the care of Drs. Shields and Augsburger until she and her mother moved to Nashville, Tennessee, in December 1981. At that time, she came under the care of Dr. Steven Feman of Vanderbilt University Medical Center.

Karen's cancer has been cured. However, her visual acuity in the left eye is 20/300. As a result of the radiation, the lens in her left eye is clouded with a cataract, and her left retina is detached. There is no hope for improving her vision through medication or surgery.

Plaintiffs contend that Karen's loss of vision in the left eye was caused by Dr. Channing's negligent treatment, and that his failure to diagnose her condition at an earlier date precluded resort to more conservative therapy. They argue that, had detection occurred earlier, Karen would have had a better opportunity for continued use of the left eye.

The record in this case consists of the testimony of various experts regarding the proper procedures for treatment of a child with Karen's symptoms. Several of the experts also gave testimony regarding the growth rate of retinoblastoma and the various treatment alternatives available as the disease progresses.

Dr. Channing testified that, although he knew how to use an indirect ophthalmoscope at the time he was treating Karen Keir, he did not use the instrument because the standard of practice in New Jersey indicated that use of a direct ophthalmoscope was appropriate. He testified that he recognized that eye crossing can be characteristic of retinoblastoma; however, retinoblastoma will not cause an eye to cross unless the tumor is in the macula. J.A. 201, 299, 515, 550. That area of the eye can be visualized in an undilated examination with the use of a direct ophthalmoscope. J.A. 201, 298. Because there was no tumor in that area, he was able to eliminate pathology as a cause for Karen's esotropia. In other words, although Karen suffered from retinoblastoma and esotropia, the two disorders were not related.

Dr. Mignone was highly critical of Dr. Channing's treatment of Karen Keir. He testified that the SOP in effect at Walston Army Hospital clearly required Dr. Channing to refer Karen to ophthalmology. He indicated that he and Dr. Channing had disagreed about the effectiveness of the SOP and that Dr. Channing had attempted to have the policy abolished. Nevertheless, Dr. Mignone stated that he believed the policy was appropriate and demanded that it be followed. Dr. Mignone testified that, even in the absence of such a policy, the standard of care at Fort Dix would have required such a referral. This was especially true because Mary Keir told Dr. Channing that she brought her daughter to see an ophthalmologist. J.A. 268-70.

Dr. Mignone also gave testimony regarding the alternative treatment methods for eye tumors. He explained that cryotherapy utilizes a cold probe to freeze the tumor. He stated that it has no systemic side effects and has been proven to be effective in causing the death of ocular tumors. J.A. 278. He stated that another conservative treatment for retinoblastoma is photocoagulation, which utilizes light to burn the tumor. Photocoagulation is difficult when a tumor is lodged in the periphery of the eye, because it is difficult to aim the light to a very peripheral lesion. Id. When asked to assume that Karen's tumor had not seeded at the time of discovery and to further assume that the more conservative modalities of cryotherapy and photocoagulation were still available, Dr. Mignone testified that the prognosis for vision in Karen's left eye was excellent. J.A. 279. However, on cross-examination, Mignone admitted that he had never written an article on retinoblastoma, had never treated a child for the disease, had never utilized photocoagulation or another relatively conservative treatment, cobalt plaque 60, in the treatment of retinoblastoma, and did not consider himself an expert in the treatment of retinoblastoma. J.A. 283-85. In fact, he had seen only one case other than that of Karen Keir, when a child was presented to him with a white pupil precisely like that possessed by Karen Keir at the time of diagnosis. J.A. 285.

Dr. Mignone admitted that, although he estimated that the tumor had been present in Karen's eye for six to eight months before diagnosis, he recognized that the doubling pattern for this particular type of cancer is elusive, as some grow quickly and others grow slowly. However, he stated that his impression was that the larger the tumor, the longer it had been present in the eye. J.A. 305-06.

Dr. Ralph Wesley, an ophthalmic plastic surgeon, also testified on behalf of the plaintiffs. Dr. Wesley served two years on active duty as a battalion surgeon in the United States Army. He testified that, in his experience, cases such as Karen's are routinely referred to an ophthalmologist. J.A. 315. He testified that he believes in the long-standing rule that any child who has a crossed eye suffers from retinoblastoma until that disease is ruled out. He also testified that the standard operating procedure in effect at Walston Army Hospital at Fort Dix was reasonable by acceptable standards of health care. J.A. 316-17.

Dr. Wesley further testified that the standard of care during the time Karen was treated by Dr. Channing required her to be examined with an indirect ophthalmoscope while her eyes were dilated. J.A. 319. This examination should have been performed on her initial visit. The reason for performing such an examination is that it allows a more complete view of the retina. He testified that a health care provider reasonably competent in the use of an indirect ophthalmoscope could see well past the equator, even when examining a child. J.A. 320. In his opinion, based upon a reasonable degree of medical certainty, the tumor could have been diagnosed during the February 1980 examination by someone reasonably competent with an indirect ophthalmoscope. J.A. 322.

Dr. Wesley testified that, after reviewing Karen's medical records, he did not believe that Dr. Channing's visual acuities were accurate. He criticized Dr. Channing's use of alternate patching, and stated that Dr. Channing failed to adhere to the standard of care by failing to refer Karen to an ophthalmologist and failing to see that she underwent a dilated exam with an indirect ophthalmoscope. J.A. 323-27. He indicated that a dilated exam should have occurred during the February 1980 visit at the very latest because one year had elapsed since Karen's last dilated examination. J.A. 327-28. He indicated that the presence of vitreal seeding was evidence that the tumor had progressed to a late stage, and that had the tumor been diagnosed earlier and had more conservative therapy been available, the chances for vision would have been better. J.A. 330.

Dr. Robert Estes gave extensive testimony regarding Karen's current condition. Estes has been required to diagnose retinoblastoma in the course of his practice, and he gave testimony regarding the diagnosis of the disease. Although Dr. Estes has never treated a retinoblastoma, he was permitted to give testimony regarding the various treatment alternatives available at various stages of the disease, as his practice required him to be familiar with these alternatives.

Dr. Estes testified that accurate and prompt diagnosis and treatment of retinoblastoma are absolutely mandatory. J.A. 350. He testified that photocoagulation is the treatment of choice for small retinoblastomas and that cryotherapy is also a preferred method for treating small retinoblastomas. However, photocoagulation is contraindicated when there is widespread vitreal seeding, as there was in the present case. Dr. Estes testified that, in his opinion, external beam irradiation, the treatment utilized on Karen Keir, has greater complications than cryotherapy. J.A. 350-51.

Dr. Estes stated that the SOP adopted at Fort Dix was reasonable in light of the prevailing professional practice. J.A. 352. He criticized the treatment rendered by Dr. Channing in this case because it did not involve a dilated examination and because it involved extensive patching. J.A. 354-57.

Dr. Estes testified that, if Karen had been under the care of an ophthalmologist, the tumor would have been diagnosed by February 1980. J.A. 358. However, he admitted that the determination of growth rates for retinoblastoma involves "pure conjecture." J.A. 372. He testified that if cryotherapy or photocoagulation had been available, "in the best situation" Karen could have had normal vision in the left eye. J.A. 359. However, he had never seen photocoagulation or cryotherapy used as the only source of treatment for retinoblastoma. J.A. 381.

As earlier indicated, Dr. Jerry Shields is the physician who treated Karen at the Wills Eye Center in Philadelphia. He testified that the clinic has treated over 200 cases of retinoblastoma, and indicated that, in his experience, the use of an indirect ophthalmoscope is the most important diagnostic step in the evaluation of a child suspected of suffering from retinoblastoma. J.A. 385. When discussing treatment options, Dr. Shields indicated that the most common complication of external beam irradiation is damage to the retinal blood vessels, or radiation retinopathy. J.A. 405. Other complications include cataracts, glaucoma, atrophy of the optic nerve, and facial bone asymmetry. Dr. Shields indicated that the use of cobalt plaque 60, another relatively conservative treatment, presents many of the same complications as external beam irradiation, although with the latter treatment the complications are somewhat more localized to the area of radiation. J.A. 405.

Dr. Shields further testified that photocoagulation is the treatment of choice for selected small retinoblastomas. J.A. 413. When administered properly, it has fewer complications than irradiation. J.A. 385. However, it is contraindicated in cases where there is widespread vitreal seeding. Id. Dr. Shields testified that cryotherapy is also an effective method of dealing with small retinoblastomas. However, there are potentially serious complications associated with cryotherapy, including retinal detachment, vitreous hemorrhage, and low ocular pressure. J.A. 414.

When asked to estimate the origin of the tumor, Dr. Shields testified that, assuming the tumor began at the epicenter, it would have begun at approximately the equator of the eye, J.A. 400, "in the peripheral retina at about five o'clock." J.A. 408.1 Dr. Shields testified that, by the time he first saw the tumor, cryotherapy and photocoagulation could not be utilized because seeding had begun. He further testified that use of cobalt plaque 60 was not feasible at the time of diagnosis. J.A. 389-90, 398.

Dr. Shields testified that in unilateral retinoblastoma cases, enucleation is generally recommended. However, because this tumor was detected at a relatively early stage, as compared to most cases, he believed that an attempt at radiation was justified. J.A. 400-01. He further indicated that he had seen very few unilateral peripheral retinoblastoma cases that have been treated conservatively, because it is extremely rare for the disease to be diagnosed at that early a stage. J.A. 404. He testified that by the time most such cases are recognized, they are very far advanced. Id. Dr. Steven Feman is the physician who treated Karen at Vanderbilt Hospital after she was released from the Wills Eye Center. He testified regarding her current condition and criticized the treatment rendered by Dr. Channing. Dr. Feman treats retinoblastoma at Vanderbilt Hospital and indicated that, in the course of his career, he has diagnosed between twelve and fourteen such cases.

Dr. Feman testified that, if Karen had been placed under the care of a reasonably competent ophthalmologist, the tumor should have been diagnosed at least six months before August 1980. He testified that there was no reason why the tumor was not seen at the February 1980 examination. J.A. 425. He criticized Dr. Channing's use of the patching therapy, and stated that during the time Karen was under Dr. Channing's care, the minimal accepted standard of care required that she have a dilated indirect ophthalmoscopy. J.A. 426.

Dr. Feman indicated that vitreal seeding usually occurs at a very late stage in the development of retinoblastoma and that, after seeding occurs, cryotherapy and photocoagulation are no longer viable treatment options. J.A. 433. He testified that the fact that Karen's tumor originated anterior to the equator was not something that would preclude its discovery by a reasonably competent ophthalmologist using an indirect ophthalmoscope during a dilated examination. J.A. 435.

Of the twelve to fourteen cases of retinoblastoma that Dr. Feman has diagnosed, more than half were treated by enucleation. A little less than 50 per cent have received external beam irradiation, and cryotherapy has never been used as the primary treatment modality. J.A. 443.

Dr. Ronald Holweger was called to testify on behalf of the United States. He testified that his own standard of care required the use of a dilated examination at least once a year for patients exhibiting symptoms such as Karen Keir's.2 J.A. 511. Dr. Holweger testified that the use of an indirect ophthalmoscope is difficult with children, and that in most pediatric cases it would be difficult to see to the equator of the eye. The reason for this difficulty is that a child will not hold her eye still long enough and cannot understand directions well enough to cooperate in the manner necessary for the physician to be able to see to the equator. J.A. 514-15.

After reviewing the records in this case, Dr. Holweger stated that, during the time Dr. Channing followed this child, there was no reason for Dr. Channing to perform a dilated examination out of concern for pathology. J.A. 516-18. Dr. Holweger also indicated that he could not state to a medical certainty that the tumor was present during any of Dr. Channing's examinations. J.A. 519-20.

The government's key witness was Dr. David Abramson, whose specialty is ophthalmological oncology. He has conducted over 14,000 examinations for the purpose of diagnosis of suspected retinoblastomas. He has been involved in the active management and diagnosis of approximately 800 cases of retinoblastoma and is familiar with the use of all methods of treatment for the disease. He has written the largest series based on patient experience using various treatment methods in combating retinoblastoma. He has more experience in the diagnosis and treatment of retinoblastoma than any other expert who testified in this case.

Dr. Abramson testified that approximately 80 per cent of the children he sees have the eye filled with the retinoblastoma tumor upon the first diagnosis. In approximately 95 per cent of these cases, the tumor is first detected by the mother, and not by a physician. J.A. 526.

Dr. Abramson testified that photocoagulation is not a suitable treatment modality for tumors anterior to the equator. Because this was the position of Karen Keir's tumor, photocoagulation was not a viable treatment option. The reason for this is that light travels in straight lines, and in order to reach the anterior region of the eye, the light would have to bend. J.A. 527.

Dr. Abramson testified that cryotherapy is successful in about 70 per cent of the cases in which it is used, with success defined as tumor destruction. One disadvantage of cryotherapy is that each treatment requires anesthesia, and risks are encountered when one is repeatedly placed under a general anesthetic. J.A. 533, 537.

Dr. Abramson testified that the use of external beam irradiation, the treatment used in this case, is extremely rare in unilateral retinoblastoma cases because it is very unusual to be able to save the eye. Id. He indicated that fewer than 100 cases in the United States have resulted in successfully saving the eye. Id. The advantages of external beam irradiation are that the tumor responds to a relatively low dose of radiation, that it involves no anesthesia, and that the therapy is generally tolerated well by the patient. J.A. 537.

Dr. Abramson testified that it is "unusual and almost impossible," absent scleral depression, to be able to see anterior to the equator by using an indirect ophthalmoscope. J.A. 538-39. This is so because the procedure is uncomfortable and children tend to be uncooperative. J.A. 539. He made the following observation:

It [examination with an indirect ophthalmoscope] is unpleasant. You have to put a wire speculum between the lids, and even then it is really a disaster. The truth is that if you get out to the equator in a child, if we are talking about age 0 to 5, you should get a little award yourself because it is extremely unusual. Anybody who says that it is anything else is simply not telling the truth.

I will tell you a story, if I may, which I mentioned to you earlier. But I think it is pertinent. At the age of about three I thought my own son had retinoblastoma, and I have a photograph that shows him with a white pupillary reflex. And I decided that there was no one better to see the child than me. And I have a nice and a very cooperative son. And I brought him to my office and put eye drops in his eyes, which because he is nice he still says to me he loves me but don't ever do that again.

And I examined him, and I was really concerned that he had retinoblastoma. And I did not see out to the equator in each eye 360?, and there is no way that I could see anterior to the equator. He did not have retinoblastoma, but if he had retinoblastoma, I never would have seen it in the periphery.

J.A. at 539-40.

Dr. Abramson testified as to the varying growth rates for retinoblastomas. He stated that, in his practice, it was not at all unusual for him to see no sign of pathology on one visit and to see a tumor only three months later. Retinoblastomas can grow very quickly in very short periods of time. J.A. 546-47.

In this particular case, Dr. Abramson testified that there was no reason to believe that Karen's esotropia was caused by the retinoblastoma because the tumor did not involve the macula or visual axis. J.A. 550. As to the origin of the tumor, Dr. Abramson testified that, without doubt, it originated anterior to the equator. J.A. 552.3 Dr. Abramson testified that Karen's tumor was not suitable for treatment by cryotherapy, photocoagulation or cobalt plaque. Only two treatment modalities remained--irradiation and enucleation. J.A. 554.

Dr. Abramson testified that the fact that the tumor had seeded did not necessarily mean that it was in the advanced stages. He indicated that he had seen tumors that began to crumble when they were very small and that seeding was only "very, very, very loosely correlated with size." J.A. 559. He further indicated that he had never seen a small tumor in the periphery of the eye in a patient who did not have a history of retinoblastoma in her family. J.A. 531-32. The explanation for this phenomenon is that, in the absence of family history, a health care provider has no reason to look in the periphery of the eye for a tumor. Thus, the mere fact that Karen's tumor was large did not necessarily mean that the diagnosis was unreasonably delayed.

Dr. Abramson testified that, based on his review of the record, it was "extremely highly unlikely" that Karen's tumor was present in July 1979 when she was first examined by Dr. Channing. J.A. 568. Dr. Abramson further indicated that, in his opinion, the tumor was not present on September 25 or October 3, 1979. J.A. 569. Finally, he believed that it was more probable than not that the tumor was not present when Dr. Channing examined Karen on February 12, 1980. J.A. 570-71. He further testified that, in his opinion, he did not see how the tumor could have been detected on February 12, 1980. He stated that even assuming that the tumor was present on that date, which he doubted, because of its location and the absence of a family history to arouse suspicion, "to detect that would have been just about impossible." J.A. at 572.

Dr. Abramson's testimony was corroborated by that of Dr. Roger Hiatt. Dr. Hiatt also indicated that performance of a dilated examination with an indirect ophthalmoscope is extremely difficult with children, although a competent ophthalmologist with good dilation could see up to the equator and somewhat beyond using an indirect ophthalmoscope on a quiet, passive patient. J.A. 624-26; 641. Dr. Hiatt criticized Dr. Channing's treatment of Karen Keir, indicating that Dr. Channing could have done more than he did. However, he also stated that there was nothing in Karen's medical history which would necessitate the performance of an aggressive ophthalmological examination to explore the area anterior to the equator. J.A. 634.

Upon consideration of the evidence produced, the district court granted judgment for the United States. The court concluded that plaintiffs' first claim, which charged the government with negligence in failing to establish and implement certain internal safeguards, involved a discretionary function for which the government has not waived immunity. With regard to the second claim, a lack of informed consent, the court concluded that this was a battery under state law for which the government has not waived its immunity. The court further concluded that any claim for misrepresentation was within the exception from the waiver of immunity contained in 28 U.S.C. Sec . 2680(h).

The district court rejected plaintiffs' claim that Dr. Channing's violation of New Jersey statutes and the hospital SOP constituted negligence per se. Because New Jersey considers such violations to be only evidence of negligence, this evidence had to be considered in light of the other evidence of record.

Finally, the district court rejected plaintiffs' claim that Dr. Channing was guilty of common law negligence. The central basis for this finding was that, despite the fact that Dr. Channing had failed to use an indirect ophthalmoscope, he had stayed within the parameters for visual care authorized by optometric practice in New Jersey. Moreover, the district court found that Dr. Channing reasonably did not suspect the existence of retinoblastoma or other pathology, and thus the examination procedures he utilized were appropriate. It is from this judgment that plaintiffs appeal.

II.

A. Governmental Immunity

1. Application of the Discretionary Function Exception

The Federal Tort Claims Act, 28 U.S.C. Sec . 1346(b), authorizes actions against the United States for damages

for injury or loss of property, or personal injury or death caused by the negligent or wrongful act or ommission of any employee of the Government--while acting within the scope of his office or employment, under circumstances where the United States, if a private person, would be liable to the claimant in accordance with the law of the place where the act or omission occurred.

See also United States v. S.A. Empresa de Viacao Aerea Rio Grandense (Varig Airlines), 467 U.S. 797, 807-08, 104 S.Ct. 2755, 2761, 81 L.Ed.2d 660 (1984). The government's waiver of immunity is qualified by the exceptions contained in 28 U.S.C. Sec . 2680. Section 2680(a) contains what is commonly known as the discretionary function exception. It provides that the waiver of immunity contained in section 1346(b) does not apply to

[a]ny claim based upon an act or omission of an employee of the Government, exercising due care, in the execution of a statute or regulation, whether or not such statute or regulation be valid, or based upon the exercise or performance or the failure to exercise or perform a discretionary function or duty on the part of a federal agency or an employee of the Government, whether or not the discretion involved be abused.

On appeal, plaintiffs argue that the district court committed reversible error when it concluded that Dr. Channing's failure to comply with the SOP was within the discretionary function exception. They contend that the failure to comply with a duly adopted procedure is not a discretionary function within the meaning of 28 U.S.C. Sec . 2680(a). See Red Lake Band of Chippewa Indians v. United States, 800 F.2d 1187, 1196 (D.C.Cir.1986) (decisions involving discretion to make or implement policy are discretionary functions within the meaning of section 2680(a)); Mandel v. United States, 793 F.2d 964, 967-68 (8th Cir.1986) (failure to comply with previously adopted safety policy is not a discretionary function for which the government retains immunity); Carlyle v. United States, Dept. of the Army, 674 F.2d 554, 557 (6th Cir.1982) (Army's decision whether to supervise recruits was a discretionary function; if government had decided to provide supervisors and they had negligently performed this duty, judicial review under Federal Tort Claims Act could have been appropriate); Miller v. United States, 583 F.2d 857, 867 (6th Cir.1978) (governmental decision as to appropriate level of water in channel is a discretionary function; government not immune from liability for negligent raising and lowering of flood gates).

Plaintiffs' argument reflects a misunderstanding of the scope of the district court's decision. The district court merely held that the government's failure to adopt and implement internal safeguards was a discretionary function. Plaintiffs apparently take no issue with this holding on appeal. See Plaintiffs' Brief at 20. To the extent that plaintiffs argue that Dr. Channing's failure to comply with a duly adopted procedure is not a discretionary function, the government admits that this is a correct statement of the law. Government's Brief at 43 n. 9. Accordingly, we conclude that the discretionary function exception does not bar plaintiffs' suit to the extent that it is based upon Dr. Channing's alleged negligence in failing to follow the standard operating procedure.

2. Application of the Battery Exception

The district court concluded that it had no jurisdiction to consider plaintiffs' claims based upon the lack of consent to have Karen treated by an optometrist. The district court observed:

Under New Jersey law, informed consent is a negligence concept arising out of the duty of a health care provider to fairly disclose the appropriate risks to a patient who has consented to treatment. Ordinarily, the plaintiff must prove that a failure to disclose was negligent and that a reasonably prudent person would not have submitted to the procedure had the risks been adequately disclosed. Perna v. Pirozzi, [92 N.J. 446], 457 A.2d 431, 438 (N.J.1983). However, New Jersey makes a distinction between the concept of informed consent and the concept of a total absence of consent, which constitutes a battery.

J.A. 52. The district court concluded that the gravamen of plaintiffs' claim was not a lack of adequate disclosure of risks, but a total lack of consent for an optometrist to treat Karen. Under these circumstances, the district court concluded that plaintiffs' claim was one for battery, for which the United States has retained its immunity.

The district court correctly noted that 28 U.S.C. Sec . 2680(h) provides that the United States retains its immunity for suits in battery. However, the district court failed to recognize that 10 U.S.C. Sec . 1089(e) qualifies this provision in malpractice actions against military doctors.4 Section 1089 was enacted to provide protection for military health care providers against malpractice actions, and it provides that in such cases, the exclusive remedy for the injured party is an action against the United States under the Federal Tort Claims Act. Section 1089(e) provides:

For purposes of this section, the provisions of section 2680(h) of title 28 shall not apply to any cause of action arising out of a negligent or wrongful act or omission in the performance of medical, dental, or related health care functions (including clinical studies and investigations).

The legislative history of this section clearly indicates that it was intended to include a waiver of immunity for malpractice actions even though, under state law, they might technically be characterized as a battery.

Subsection (e) would nullify a provision of the Federal Tort Claims Act which would otherwise exclude any action for assault and battery from the coverage of the Federal Tort Claims Act. In some jurisdictions it might be possible for a claimant to characterize negligence or a wrongful act as a tort of assault and battery. In this way, the claimant could sue the medical personnel in his individual capacity notwithstanding subsection (a) simply as a result of how he pleaded his case. In short, subsection (e) makes the Federal Tort Claims Act the exclusive remedy for any action, including assault and battery, that could be characterized as malpractice.

S.Rep. No. 94-1264, 94th Cong., 2d Sess. 9-10, reprinted in 1976 U.S.Code Cong. & Admin.News 4443, 4451. "The malpractice immunity statute, 10 U.S.C. Sec . 1089(a) ... immunizes military physicians and makes tort claims against the government the exclusive remedy available in a malpractice action." Baker v. Barber, 673 F.2d 147, 148-49 (6th Cir.1982) (per curiam); see generally Lojuk v. Quandt, 706 F.2d 1456 (7th Cir.1983) (general discussion of section 1089(e)).

As the district court correctly indicated, New Jersey recognizes a distinction between claims based upon the total absence of consent and claims based upon lack of informed consent.

Informed consent is a negligence concept predicated on the duty of a physician to disclose to a patient information that will enable him to "evaluate knowledgeably the options available and the risks attendant upon each" before subjecting that patient to a course of treatment. Under the doctrine, the patient who consents to an operation is given the opportunity to show that the surgeon withheld information concerning "the inherent and potential hazards of the proposed treatment, the alternatives to that treatment, if any, and the results likely if the patient remains untreated."

Perna v. Pirozzi, 92 N.J. 446, 459-60, 457 A.2d 431, 438 (1983) (citations omitted). However, when the claim is based upon a total absence of consent, as when an operation is performed by a physician other than the one to whom consent is given, a battery occurs, and the plaintiff is entitled to nominal damages even if no injury other than that foreseeably flowing from the operation occurs. Id. at 460-61, 457 A.2d at 438.

The district court in the present case characterized plaintiffs' claim as one for battery under New Jersey law. We disagree with this characterization. Our reading of plaintiffs' complaint indicates that it is grounded in the concept of negligence. In the pretrial order, the claim is identified as one based upon the absence of informed consent. J.A. 81. Dr. Channing's alleged failure to clarify his status as an optometrist is more closely akin to the failure to provide the patient with sufficient information to enable him to evaluate treatment risks. Under New Jersey law, such a claim represents a claim for negligence.5

3. Application of the Misrepresentation Exception

In the proceedings below, plaintiffs alleged that they were induced to believe that Dr. Channing was an ophthalmologist, and that this misrepresentation formed the basis for their consent to Dr. Channing's treatment. They argued that, had they been informed that Dr. Channing was an optometrist, they never would have consented to treatment. They also argued that Dr. Channing violated N.J.Rev.Stat. Sec. 45:12-26 by using the title "Dr." without qualification by use of the term "optometrist." The district court concluded that it had no jurisdiction to consider these claims because they were barred by the misrepresentation exception to the government's waiver of sovereign immunity that is contained in 28 U.S.C. Sec . 2680(h). We believe that this conclusion was erroneous.

In Ramirez v. United States, 567 F.2d 854 (9th Cir.1977) (en banc), the Ninth Circuit concluded that the misrepresentation exception to the government's waiver of immunity is inapplicable in malpractice cases. The Ramirez court concluded that, in enacting section 2680(h), "Congress had in mind the 'traditional and commonly understood' torts of negligent misrepresentation and common law deceit." Id. at 856 (quoting United States v. Neustadt, 366 U.S. 696, 706-07, 711 n. 26, 81 S.Ct. 1294, 1300-01, 1302 n. 26, 6 L.Ed.2d 614 (1961)). As a general rule, these claims have " 'been confined "very largely to the invasion of interests of a financial or commercial character, in the course of business dealings." ' " Id. (quoting Neustadt, 366 U.S. at 711 n. 26, 81 S.Ct. at 1302 n. 26).

The court further noted that the legislative history of the Federal Tort Claims Act reveals that Congress rejected an amendment that would have retained government immunity in malpractice cases. The refusal to adopt such an amendment clearly supports the conclusion that Congress intended to provide a remedy against the United States for malpractice claims. Application of the misrepresentation exception in such a case would defeat this purpose.

The creation of strained distinctions to encompass aspects of ordinary medical malpractice within the misrepresentation exception of section 2680(h) is not justified by the language of the statute, by its history, or by Neustadt. The patient who suffers harm from [a physician's misrepresentation] is the victim of negligent conduct, not of an esoteric form of misrepresentation.

Ramirez, 567 F.2d at 857. Other courts have concluded that, when misrepresentation is a collateral aspect of a plaintiff's claim for medical malpractice, the claim is not barred by the misrepresentation exception contained in section 2680(h). See, e.g., Phillips v. United States, 508 F.Supp. 544 (D.S.C.1981). We believe that this approach to the application of the misrepresentation exception in malpractice cases is analytically sound. Thus, we conclude that the district court erred in holding that to the extent plaintiffs' claims were based upon allegations of misrepresentation, they were barred by the Federal Tort Claims Act.6

In sum, we conclude that the district court erred in its application of the Federal Tort Claims Act and its various exceptions to the facts of the present case. We conclude that Dr. Channing's alleged failure to comply with the SOP was not a discretionary function within the meaning of the Act. We further conclude that plaintiffs' claims regarding the absence of informed consent are properly characterized as negligence claims under New Jersey law. Thus, they are not encompassed by the battery exception to the waiver of immunity contained in 28 U.S.C. Sec . 2680(h). Even if the claim is one for battery under New Jersey law, it is within the jurisdiction of the district court by virtue of the application of 10 U.S.C. Sec . 1089(e). Finally, we conclude that the misrepresentation exception to the waiver of immunity contained in 28 U.S.C. Sec . 2680(h) does not preclude consideration of plaintiffs' claim of negligence.

B. Negligence

Our review of the record reveals that the fundamental basis for plaintiffs' complaint is that Dr. Channing was negligent in his treatment of Karen Keir, and that this negligence resulted in a delayed diagnosis, which in turn increased the risk that Karen would be required to resort to radical treatment methods to insure against metastasis. The use of external beam irradiation instead of more conservative treatment deprived Karen, in plaintiffs' view, of the use of her left eye. Although the negligence claim encompasses many components, including allegations of misrepresentation and allegations that the failure to comply with the SOP and certain New Jersey statutes constituted negligence per se, it is essentially a claim that Dr. Channing failed to adhere to the standard of care and that Karen Keir was injured as a result.

The standard of review of the district court's determination of negligence is narrow. "[A] finding of negligence or the absence thereof will not be set aside unless the District Court's determination is 'clearly erroneous,' under [Fed.R.Civ.P.] 52." Downs v. United States, 522 F.2d 990, 999 (6th Cir.1975); see also Ward v. United States, 838 F.2d 182 (6th Cir.1988).

"If the district court's account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently." Anderson v. City of Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985). However, when " 'the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed,' " the findings of the district court will be overturned. Id. at 573, 105 S.Ct. at 1511 (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)). Guided by this standard of review, we will consider each of the district court's findings on the issue of negligence.

1. Negligence Per Se

In the proceedings below, plaintiffs alleged that Dr. Channing was guilty of negligence per se because he unlawfully engaged in the practice of medicine in violation of N.J.Rev.Stat. Sec. 45:9-5.1 and engaged in unlawful optometric practices in violation of N.J.Rev.Stat. Secs. 45:12-1 and 45:12-26. Plaintiffs further contended that the violation of the SOP constituted negligence per se. In order to determine whether plaintiffs have stated a cognizable claim, we must look to the substantive law of New Jersey. Ward, 838 F.2d at 184; Raymer v. United States, 660 F.2d 1136 (6th Cir.1981), cert. denied, 456 U.S. 944 , 102 S.Ct. 2009, 72 L.Ed.2d 466 (1982).

We agree with the district court's conclusion that plaintiffs cannot recover under a theory of negligence per se. Under New Jersey law, the violation of a statute does not constitute negligence per se but only evidence of negligence. See, e.g., Horbal v. McNeil, 66 N.J. 99, 328 A.2d 604 (1974); Schomp v. Wilkens, 206 N.J.Super. 95, 501 A.2d 1036 (1985); Kenney v. Scientific, Inc., 204 N.J.Super. 228, 497 A.2d 1310 (1985). Assuming, arguendo, that the standard operating procedure has the force of law, violation of the SOP could have no greater implication than violation of a statute. Accordingly, violation of the SOP could not support a finding of negligence per se under New Jersey law.

2. Common Law Negligence

In order to establish negligence in a medical malpractice case governed by New Jersey law, the plaintiff must establish that he was injured by the physician's failure to utilize "the degree of care, knowledge and skill ordinarily possessed and exercised in similar situations by the average member of the profession practicing in his field." Schueler v. Strelinger, 43 N.J. 330, 344, 204 A.2d 577, 584 (1964). On appeal, plaintiffs contend that Dr. Channing breached the duty of care owed to his patient by failing to refer her to an ophthalmologist and by failing to conduct a dilated examination with an indirect ophthalmoscope.

a. Duty of Referral

The district court concluded that Dr. Channing was not negligent in failing to refer Karen Keir to an ophthalmologist, and concluded that the SOP did not create an actionable duty because under New Jersey law tort obligations cannot be created by private action.7 To support this conclusion, the district court placed heavy reliance on Coyle v. Englander's, 199 N.J.Super. 212, 488 A.2d 1083 (1985). In our view, the district court's reliance on Coyle is misplaced.

In Coyle, the issue presented for disposition by the court was "whether a disappointed promisee may recover traditional tort damages for his personal injuries caused by a breaching promisor's failure to perform a contractual term." Id. at 214, 488 A.2d at 1083. The facts indicated that plaintiff sustained injuries when moving band equipment onto a truck after defendants breached their contractual obligation to provide assistance in this endeavor. The court construed plaintiff's claim "solely as a claim that breach of a contractual term caused personal injuries to him." Id. at 217, 488 A.2d at 1085. The court concluded that the failure to perform a contractual duty could not serve as the basis for imposition of damages for personal injury "where no breach of any common-law duty is implicated in the slightest...." Id. at 276, 488 A.2d at 1090. Accordingly, recovery was denied.

In our view, Coyle does not control disposition of the present case. The duty to refer in the present case need not have been created by the SOP. Rather, the duty of due care existed by virtue of the health care provider/patient relationship. The critical question that must be answered upon a review of the record focuses on the nature of that duty. Thus, the mere fact that the SOP did not create a duty of referral does not mean that the SOP becomes irrelevant. Instead, the SOP is useful in determining the scope of the duty owed by Dr. Channing to Karen Keir.

This court has recognized that consideration of a standard operating procedure is appropriate in considering the scope of duty flowing from the defendant to a plaintiff in a negligence action. In Downs, 522 F.2d at 990, two hostages were killed during the course of the FBI's attempt to stop a hijacking at a Florida airport. Their survivors brought an action against the United States, alleging that the FBI agent was negligent in handling the situation. We concluded that, although the FBI internal hostage guidelines did not in themselves create a duty to the hostages or their survivors, the guidelines were relevant to determination of the scope of the duty. Moreover, the agent's failure to comply with those guidelines was relevant on the issue of whether he had acted as a reasonably prudent FBI agent under the circumstances.

Similarly, many other courts have held that hospital internal regulations are relevant in considering the scope of the duty of care owed by the hospital to the patient. See, e.g., Marks v. Mandel, 477 So.2d 1036 (Fla.Dist.Ct.App.1985) (per curiam) (hospital emergency room manual relevant on issue of whether the hospital violated the applicable standard of care); Darling v. Charleston Community Memorial Hosp., 33 Ill.2d 326, 211 N.E.2d 253 (1965) (state hospital regulations, standards for accreditation, and by-laws of hospital did not conclusively determine standard of care but were relevant to this determination), cert. denied, 383 U.S. 946 , 86 S.Ct. 1204, 16 L.Ed.2d 209 (1966); Taylor v. City of Beardstown, 142 Ill.App.3d 584, 96 Ill.Dec. 524, 491 N.E.2d 803 (1986) (hospital regulations, standards, and by-laws admissible to determine appropriate standard of care).

In a recent decision, a New Jersey appellate court has indicated that regulatory guidelines may be relevant in determining the appropriate standard of care. In Sanna v. Nat'l Sponge Co., 209 N.J.Super. 60, 506 A.2d 1258 (1986), plaintiff alleged that he was injured as a result of the use of improperly constructed scaffolding. The court noted that OSHA regulations contain precise guidelines for the construction of scaffolding, and such regulations could be established as objective safety standards generally prevailing in the community if expert testimony indicated that they were accepted in such. Id. at 69, 506 A.2d at 1263. Thus, Sanna supports the conclusion that the standard operating procedure is relevant in determining whether or not Dr. Channing was negligent, but that it does not necessarily create or define the precise scope of the duty running from Dr. Channing to Karen Keir.8

Our review of the record compels the conclusion that the district court was clearly erroneous in determining that Dr. Channing was not negligent in failing to refer Karen Keir. As an initial matter, it appears that Dr. Channing's failure to comply with the SOP occurred because he disagreed with Dr. Mignone about its usefulness. Dr. Channing and Dr. Mignone had argued about the viability of the SOP after Dr. Mignone became Chief of Ophthalmology at Walston Army Hospital, and Dr. Mignone insisted that the SOP would remain in place. Nevertheless, Dr. Channing believed that the SOP was not "livable," and thus he and the other optometrists utilized their own discretion in determining whether to refer patients to ophthalmology. J.A. 178-79, 185.

The overwhelming evidence presented in the record indicates that Dr. Channing should have referred Karen Keir to ophthalmology upon her first visit. Dr. Mignone stated that, aside from the SOP, the standard of care at Fort Dix required an optometrist to refer a child like Karen to ophthalmology. J.A. at 269. Dr. Ralph Wesley, who had served considerable time in the Army as an ophthalmologist, stated that in his experience cases such as Karen's were routinely referred to ophthalmology. J.A. at 315.

Dr. Estes agreed that the standard operating procedure represented reasonable standards of health care practice for treating patients like Karen Keir. J.A. at 352. In his letter to Dr. Channing, Dr. Tirrill stated that Karen should be seen by an ophthalmologist. At trial, Dr. Tirrill testified that he made this recommendation because of Karen's long history of ocular difficulties. Even Dr. Hiatt, one of the government's witnesses, criticized Dr. Channing's failure to refer, stating that "he is not a substitute for an ophthalmologist." Joint Appendix at 627.

There is some indication that even if Karen had not been referred initially, she should have been referred on February 12, 1980. Dr. Mignone testified that, assuming Karen's vision had not improved after the initial course of patching, referral should have been made at that time. J.A. 281-82. The assumption upon which the response was based is not entirely accurate, as there was some improvement in Karen's far visual acuity. J.A. 211, 256. Dr. Wesley testified that, given Karen's difficulty in adapting to the patching therapy, referral should have been made on February 12, 1980. J.A. 327. Dr. Feman indicated that, because the course of treatment utilized by Dr. Channing did not result in significant improvement, further evaluation should have been made through a repeated dilated examination or referral to another physician. J.A. 434.

On this state of the record, we are left with "a definite and firm conviction" that the district court's finding regarding the absence of negligence in the failure to refer is erroneous. Anderson, 470 U.S. at 573-74, 105 S.Ct. at 1511, Downs, 522 F.2d at 990. Accordingly, the district court's finding cannot be upheld.

b. Failure to Perform an Indirect Dilated Examination

The district court also rejected plaintiffs' argument that Dr. Channing failed to comply with the standard of care because he did not perform a dilated examination with an indirect ophthalmoscope. The overwhelming medical testimony presented at the trial indicated that such an examination should have been part of Karen Keir's course of treatment. Dr. Mignone testified that a dilated examination would have been part of the initial work-up if a referral to ophthalmology had been made. J.A. 272. Dr. Wesley also testified that the standard of care required a dilated indirect ophthalmoscopy to be performed on the initial visit. J.A. 317-19. Dr. Estes stated that Karen should have had a dilated retinal examination during her first visit. J.A. 354. Even Dr. Abramson indicated that it would be appropriate to do an indirect ophthalmoscopy using dilating drops during the first visit of a child with a history similar to Karen's. J.A. 595. Finally, Dr. Hiatt testified that a dilated examination should have been performed and that, if the child would cooperate, an indirect ophthalmoscope should have been utilized. J.A. 623-24.

There is also considerable support for plaintiffs' contention that a dilated examination should have been performed on February 12, 1980. Dr. Wesley testified that, because Karen did not adapt to the patching procedure, a dilated indirect ophthalmoscopy should have been performed on that date. J.A. 328. Dr. Estes agreed with this assessment. J.A. 357. Dr. Feman stated that a dilated examination should be performed as a routine matter once a year. J.A. 450. Dr. Tirrill and Dr. Holweger concurred in this assessment. J.A. 471, 511. Because the last dilated examination was performed by Dr. Tirrill in February 1979, J.A. 464, the procedure should have been repeated in February 1980.

The foregoing discussion illustrates that there was indeed overwhelming evidence to support the conclusion that the standard of care required Karen Keir to be given a dilated indirect ophthalmoscopy during the time she was under Dr. Channing's care. The district court's contrary finding is clearly erroneous.

C. Causation

The conclusion that Dr. Channing failed to adhere to the appropriate standard of care is not in itself sufficient to impose liability. The plaintiff must establish that a causal link exists between the breach of duty and the resulting injury.

The difficulty of proving causation in medical malpractice cases has caused many courts to adopt a lesser standard of proof on the issue of causation. See generally, King, Causation, Valuation and Chance in Personal Injury Torts Involving Pre-existing Conditions and Future Consequences, 90 Yale L.J. 1353 (1981). The leading case in New Jersey is Evers v. Dollinger, 95 N.J. 399, 471 A.2d 405 (1984). In Evers, the plaintiff visited her gynecologist in March 1977 because she was concerned about a lump in her right breast. Her doctor instructed her "to stop worrying and go home and relax." Id. at 402, 471 A.2d at 407. Because plaintiff continued to experience pain, she returned to her physician's office in October 1977 and was examined by his partner. At that time, a culture was taken, and the results proved negative for infection. Subsequently, plaintiff consulted a physician associated with Memorial Sloan Kettering Cancer Center in New York, and as a result of that examination, plaintiff was diagnosed as having breast cancer. She underwent a right extended mastectomy, but subsequent laboratory studies indicated that the cancer had not metastacized.

Plaintiff then commenced an action against her physician and his medical group. At the time of her trial, plaintiff had not experienced any recurrence of the cancer. However, while the case was pending on appeal, the cancer had spread into plaintiff's lungs, and her physicians concluded that she was faced with a terminal illness. The court concluded that, on the state of the record, it was improper for the trial court to grant judgment for the defendant at the conclusion of plaintiff's case. The court also gave considerable attention to the appropriate standard of causation to be applied on retrial.

The court recognized that, because proof of causation in medical malpractice cases is often extremely difficult, "the application of a standard of causation that is more flexible than that used in conventional tort claims" is warranted. 95 N.J. at 413, 471 A.2d at 413. Accordingly, the court concluded that "when there is evidence that a defendant's negligent act or omission increased the risk of harm to one in the plaintiff's position and that harm was in fact sustained," the finder of fact is entitled to determine " 'whether or not that increased risk was a substantial factor in producing the harm.' " Id. at 414-15, 471 A.2d at 413 (quoting Hamil v. Bashline, 481 Pa. 256, 269, 392 A.2d 1280, 1286 (1978)). Application of this principle to the facts before it required that

[O]n remand plaintiff should be permitted to demonstrate, within a reasonable degree of medical probability, that the seven months delay resulting from defendant's failure to have made an accurate diagnosis and to have rendered proper treatment increased the risk of recurrence or of distant spread of plaintiff's cancer, and that such increased risk was a substantial factor in producing the condition from which plaintiff currently suffers.

Id. at 417, 471 A.2d at 415.

Application of the standard adopted in Evers to the facts of the present case requires that, in order for plaintiff to establish a prima facie case, she must demonstrate that Dr. Channing was negligent and that this negligence increased the risk that she would be denied the opportunity to avail herself of conservative treatment modalities. See Evers, 95 N.J. at 414-15, 471 A.2d at 413. The trier of fact must then decide whether the increased risk was a substantial factor in producing the resulting harm. Id. at 415, 471 A.2d at 415.

In assessing the issue of causation, the district court did not engage in the analysis mandated by New Jersey law and articulated in Evers. Indeed, the district court's discussion of the issue comprises little more than the cursory observation that, according to Dr. Abramson, the chance for earlier detection was "remote." J.A. 60. The district court's treatment of the issue can perhaps be explained by its conclusion that Dr. Channing was not negligent. Because we have concluded that the district court's decision in that regard is clearly erroneous, this case must be remanded for a consideration of the question of causation under the standards established by New Jersey law. In order to facilitate the district court's consideration of the issue, a summary of the relevant precedent is provided below.

As earlier indicated, a plaintiff proceeding under a "lost chance" theory governed by New Jersey law establishes a prima facie case by demonstrating that the defendant's negligence increased the risk that the plaintiff would sustain the harm she ultimately suffered. Evers, 95 N.J. at 413, 471 A.2d at 413. In the present case, plaintiffs' medical experts testified that the tumor was present during the February 12, 1980, examination and that it should have been discovered at that time. Dr. Mignone, J.A. 305-06; Dr. Wesley, J.A. 322; Dr. Estes, J.A. 358; Dr. Feman, J.A. 425. Moreover, expert testimony supported plaintiffs' contention that vitreal seeding occurs at a late stage in the progression of the disease. Dr. Wesley, J.A. 330; Dr. Feman, J.A. 433. The presence of seeding was one of the critical factors precluding resort to more conservative therapy. Dr. Shields, J.A. 389-90; Dr. Feman, J.A. 433; Dr. Estes, J.A. 374. Thus, there was substantial evidence to support the proposition that the failure to diagnose more quickly increased the risk that Karen would be required to undergo radical treatment. See Dr. Estes, J.A. 380.

Our review of the relevant precedent indicates that plaintiffs clearly presented sufficient evidence to support a prima facie case on the issue of causation. See Roberson v. Counselman, 235 Kan. 1006, 686 P.2d 149 (1984) (evidence indicating that chiropractor's failure to refer patient to physician decreased chance of survival was sufficient to create question of fact on issue of causation); Evers, 95 N.J. at 399, 471 A.2d at 405 (evidence that late diagnosis of breast cancer increased risk of metastasis which actually occurred was sufficient to create jury question on issue of causation); Jones v. Montefiore Hosp., 494 Pa. 410, 431 A.2d 920 (1981) (evidence tending to establish that physician's conduct increased risk of necessity of mastectomy or of metastasis sufficient to create jury question and issue of causation); Gradel v. Inouye, 491 Pa. 534, 421 A.2d 674 (1980) (evidence that failure to perform prompt biopsy increased risk of amputation created jury question on causation). Cf. Hake v. Manchester Township, 98 N.J. 302, 486 A.2d 836 (1985) (plaintiffs entitled to present evidence to support theory that defendants' negligence increased the risk that their son's attempted suicide would be successful).

Once a prima facie case has been established, the trier of fact must determine whether the defendant's conduct was a substantial factor in bringing about the injury of which plaintiff complains. See Thompson v. Sun City Community Hosp., Inc., 141 Ariz. 597, 688 P.2d 605 (1984) (en banc); Roberson, 235 Kan. at 1006, 686 P.2d at 149; Hastings v. Baton Rouge Gen. Hosp., 498 So.2d 713 (La.1986); Aasheim v. Humberger, 695 P.2d 824 (Mont.1985); McKellips v. St. Francis Hosp., Inc., 741 P.2d 467 (Okl.1987); Jones, 494 Pa. at 410, 431 A.2d at 920. Although the question of whether defendant's conduct was a substantial factor in producing the injury cannot be analyzed by resort to a mathematical formula, two factors should generally be considered.

[T]here are at least two important factors relevant to the issue of causation in cases involving negligent treatment of a potentially fatal condition: first, the patient's chances of survival if properly treated according to medical procedures generally recognized as appropriate under the circumstances; and second, the extent to which the patient's chances have been reduced by improper departure from these established procedures.

Roberson, 235 Kan. at 1019, 686 P.2d at 159 (quoting Daniels, 566 F.2d at 757) (emphasis supplied in original).

The district court's opinion did not reflect appropriate consideration of these factors, and thus its finding that the chance for earlier detection was "remote" represents an inadequate determination on the issue of causation. Although factual determinations are subject to review under the clearly erroneous standard, see United States v. City of Birmingham, 727 F.2d 560 (6th Cir.), cert. denied, 469 U.S. 821 , 105 S.Ct. 95, 83 L.Ed.2d 41 (1984); Daniels v. Hadley Memorial Hosp., 566 F.2d 749 (D.C.Cir.1977), and trial judges are given broad discretion in the assessment of expert testimony, see Waffen v. United States Dep't of Health & Human Services, 799 F.2d 911 (4th Cir.1986); TWM Mfg. Co. v. Dura Corp., 722 F.2d 1261, 1267 (6th Cir.1983), factual findings resulting from misapplication of controlling law will not be upheld. See, e.g., Weisberg v. United States Dep't. of Justice, 745 F.2d 1476, 1495-98 (D.C.Cir.1984); Harris v. Birmingham Bd. of Educ., 712 F.2d 1377 (11th Cir.1983). Cf. Senter v. General Motors Corp., 532 F.2d 511, 526 (6th Cir.) (appellate court not restricted to clearly erroneous review "where the contention is that the district court applied erroneous legal principles."), cert. denied, 429 U.S. 870 , 97 S.Ct. 182, 50 L.Ed.2d 150 (1976).

Accordingly, the district court will be given another opportunity to consider whether Dr. Channing's negligence was a "substantial factor" in producing the harm suffered by Karen Keir. See Weisberg, 745 F.2d at 1498. In making this determination, the district court must consider the evidence of record in light of the New Jersey Supreme Court's pronouncement in Evers, supra.

III.

The judgment of the district court is REVERSED, and this case is REMANDED for further proceedings consistent with this opinion. However, we intimate no opinion as to the appropriate disposition on remand.

BOGGS, Circuit Judge, concurring in part and dissenting in part.

I concur in all of the thorough and well-written opinion for the court except for the final determination on causation. The opinion is correct that the district court erred in finding no waiver of government immunity and a lack of negligence in the actions of Eugene Channing in this case. However, the district court also made clear findings on the question of causation. The question of causation was hotly disputed and considerable evidence was taken on the issue. As the court's opinion properly points out, the testimony of Dr. Abramson on this point was extensive and persuasive (supra, pp. 406-408). The crucial question is: had Karen Keir been given non-negligent care, is there any significant chance that the tumor which caused her injuries could have been detected earlier than it was (thus perhaps allowing more efficacious treatment)? Plaintiffs presented several witnesses who supported this proposition, but Dr. Abramson's testimony to the contrary was clear and unequivocal.

In these circumstances, the district court was well within the boundaries of his role as finder of fact to accept Dr. Abramson's testimony. This result is particularly justified by the fact that Dr. Abramson is clearly the leading American expert in the treatment of exactly the condition at issue here, while all of the other witnesses, on both sides, have at best modest experience.

It thus appears to me to be quibbling to hold that the district court did not appropriately consider the standard of "lost chance" causation which is established in New Jersey. The opinion well sets out the standard required: the negligent act must have increased the risk of harm and the increased risk must have been a substantial factor in producing the plaintiff's actual condition.

Here, the government's evidence, though Dr. Abramson's testimony, was fully credible that the errors of Dr. Channing, grievous though they were, did not increase the risk of harm because the tumor would not have been discovered even with appropriate care.

The opinion faults the district court for simply using the word "remote" in describing this chance of discovering the tumor. It appears to me that the district court's opinion, taken as a whole, indicates that "remote" indeed means a chance so small that it cannot properly be called an increased risk of harm. I believe it will be an exercise in redundancy to remand to the district court to have it tell us what is clearly implied, if not actually patent, in its opinion. I therefore respectfully dissent from the remand, while remaining in agreement with all of the principles of law set forth in the court's opinion.

* Honorable Ann Aldrich, Judge, United States District Court for the Northern District of Ohio, sitting by designation

1 Dr. Channing testified that, with the use of an indirect ophthalmoscope, he could see the five o'clock meridian "depending ... on how far out they are at." J.A. 262

2 This testimony appears to be of limited value, as Dr. Holweger did not even begin his residency in ophthalmology until one year after the diagnosis of the retinoblastoma in this case

3 This testimony is not inconsistent with that of Dr. Shields, who testified that the tumor originated at the five o'clock meridian in the peripheral retina at approximately the equator. J.A. 408

4 Although plaintiffs did not specifically cite section 1089(e) to the district court, we believe the issues presented below were presented in a manner comprehensive enough to permit our review of this contention

5 We emphasize that, regardless of the characterization of the claim, it is not barred by the Federal Tort Claims Act

6 Our conclusion renders it unnecessary for us to consider plaintiffs' argument that 10 U.S.C. Sec . 1089(e) nullifies the misrepresentation exception in malpractice cases against the United States Army

7 Related to this holding was the district court's conclusion that the "Good Samaritan doctrine" does not provide a basis for liability in this case. The district court concluded that, because under New Jersey law the SOP did not create an actionable duty, there was no "undertaking"' within the meaning of the Good Samaritan doctrine. On appeal, plaintiffs do not contend that the Good Samaritan doctrine establishes a basis for liability, conceding that it only applies "in the absence of a relationship which imposed a duty of reasonable care." Plaintiffs' Brief at 32 n. 7

8 Counsel for plaintiffs agreed at oral argument that this was the appropriate focus of the analysis. Specifically, counsel stated that he was not arguing that the SOP created a duty on behalf of Dr. Channing, but merely was relevant to the determination of the scope of that duty