Brillant v. Royal

582 So.2d 512 (1991) Helen B. BRILLANT and Eugene G. Brillant v. Dr. Lorenza ROYAL and Sterling Medical Associates. Dr. Lorenza ROYAL and Sterling Medical Associates v. Helen B. BRILLANT and Eugene G. Brillant. 89-1303, 89-1392. Supreme Court of Alabama. May 31, 1991. *514 Tom Dutton of Pittman, Hooks, Marsh, Dutton & Hollis, Birmingham, for appellants/cross-appellees. Davis Carr, of Pierce, Carr & Alford, Mobile, for appellees/cross-appellants. ALMON, Justice. This is a medical malpractice action brought by Helen and Eugene Brillant against Dr. Lorenza Royal and Sterling Medical Associates (hereinafter "Sterling"). The Brillants allege that Dr. Royal's failure to diagnose Mrs. Brillant's cerebral aneurysm, manifested on March 25, 1987, by a sudden severe headache, which is a classical symptom of a "warning leak," was the proximate cause of paralysis she later suffered because of a full rupture of the aneurysm on April 10. Specifically, the Brillants contend that Dr. Royal negligently failed to perform a computerized tomography scan ("CT scan" or "catscan"), to perform a lumbar puncture test, or to consult a neurologist, and that this failure prevented an accurate diagnosis of Mrs. Brillant's condition and proximately caused her later paralysis. The Brillants also contend that Sterling is vicariously liable for Dr. Royal's negligence because, they contend, Dr. Royal was an employee or agent of Sterling. The trial court directed a verdict for Sterling on the ground that there was no evidence that Dr. Royal was anything other than an independent contractor and directed a verdict for Dr. Royal on the ground that the Brillants had not presented substantial evidence that Dr. Royal's actions proximately caused Mrs. Brillant's injuries. The evidence revealed that on March 24, 1987, while the Brillants were visiting friends in Valdosta, Georgia, Mrs. Brillant experienced the sudden onset of an extremely severe headache. When the headache would not subside, the Brillants left Valdosta to return home that night. At approximately 6:45 the next morning Mrs. Brillant visited the emergency room at Lyster Army Community Hospital. Dr. Royal wrote a medical record of her account of the sudden onset of the headache, including her statements that she had been vomiting, that the headache was radiating up her scalp and down her back, that her neck was sore on full range of motion and to palpation, and that the "pain was worse than her migraines" she had suffered years before (emphasis by Dr. Royal). He performed an examination to test for nerve function deficits, but found no such deficits. He diagnosed a spastic muscular headache, prescribed an analgesic and a muscle relaxer, and instructed Mrs. Brillant to "return as needed." On April 2, 1987, eight days after her first visit, Mrs. Brillant returned to Lyster Army Hospital for further treatment for her headaches, which, although persistent, were not as severe as her initial headache. Dr. Fidel Velez examined Mrs. Brillant, performed some routine tests, prescribed an analgesic and muscle relaxer, and recommended physical therapy. Apparently Dr. Velez did not know that Dr. Royal had examined Mrs. Brillant. On the morning of April 10, 1987, Mrs. Brillant suffered a full rupture of an aneurysm; that rupture produced a massive subarachnoid hemorrhage. Her husband took her to the emergency room at Dale County Hospital. Dr. John Wessner, an emergency room doctor, ordered a CT scan, which revealed her condition, and transferred her to Flowers Hospital in Dothan so that she could be evaluated and treated by a neurosurgeon. Further tests conducted at Flowers Hospital confirmed a subarachnoid hemorrhage with "a large amount of blood in the subarachnoid space." The following morning Dr. Bruce Woodham and Dr. Christopher Boxell performed neurosurgery on Mrs. Brillant in order to repair the ruptured blood vessel. Initially following surgery, Mrs. Brillant's prognosis was very good. However, three days later a CT scan demonstrated right frontal temporal edema, and, shortly thereafter, she suffered a vasospasm that caused paralysis of her left side. Subsequent rehabilitation has not corrected Mrs. Brillant's condition.[1] *515 The trial commenced with the Brillants presenting the testimony of Dr. Wayne Longmore as their expert witness. The Brillants also called Dr. Royal as an adverse witness. When the Brillants rested their case-in-chief, the defendants filed a motion for directed verdict. The court directed a verdict for Sterling on the basis that the Brillants had not presented substantial evidence that Dr. Royal was an employee of Sterling so as to justify imposing vicarious liability on Sterling. However, the court denied the motion as to Dr. Royal. At the conclusion of all the evidence, Dr. Royal renewed his motion for directed verdict. The court granted this motion after specifically finding that the Brillants had failed to produce substantial evidence[2] that Dr. Royal's alleged negligence proximately caused Mrs. Brillant's injuries. The Brillants now challenge the directed verdicts in favor of Dr. Royal and Sterling. They also raise as error two of the trial court's rulings on objections during the trial. Dr. Royal and Sterling have cross-appealed and raise as error the trial court's refusal to enter a summary judgment in their favor on (1) the theory of efficient, intervening cause, or (2) their assertion that the Brillants' exclusive remedy is against the United States under the Federal Tort Claims Act ("FTCA"). Because the FTCA, if applicable, forecloses any remedy other than that provided against the United States, we first address whether the trial court erred in denying Sterling and Dr. Royal's motion for summary judgment on this basis. The "Gonzales Act," codified at 10 U.S.C. § 1089, provides as follows: "The remedy against the United States provided in sections 1346(b) and 2672 of title 28 for damages for personal injury, including death, caused by the negligent or wrongful act or omission of any physician, dentist, nurse, pharmacist or paramedical or other supporting personnel... of the armed forces ... while acting within the scope of his duties or employment therein or therefor shall hereafter be exclusive of any other civil action or proceeding by reason of the same subject matter against such physician, dentist, nurse, pharmacist, or paramedical or other supporting personnel (or the estate of such person) whose act or omission gave rise to such action or proceeding." 10 U.S.C. § 1089 (1983). The immunity provided by § 1089 applies only if the doctor is serving in active military duty or is a civilian employee of the armed forces. Doctors practicing medicine at military hospitals as independent contractors bear personal liability for their negligent and/or wanton acts just as doctors practicing in the private sector bear their own liability. These independent contractors are not "employees of the government," and, thus, are not covered by the Government's cloak of immunity. Lilly v. Fieldstone, 876 F.2d 857 (10th Cir.1989); Norton v. Murphy, 661 F.2d 882 (10th Cir.1981). In Lilly, the Tenth Circuit Court of Appeals recognized that a determination of whether a physician was an individual contractor or was an employee of the Government did not depend strictly on the "control" exerted by the Government because "[i]t is uncontroverted that a physician must have discretion to care for a patient and may not surrender control over certain medical details." 876 F.2d at 859. Therefore, the analysis employed by that court centered on "whether other evidence manifests an intent to make the professional an employee subject to other forms of control which are permissible." Id. In making that determination, the court in Lilly analyzed the following questions: (1) Did the physician have an arrangement with the hospital whereby he was always required to see patients there? (2) did the physician bill the Army separately at his standard *516 rates or did the physician reduce his fees when treating a military patient? (3) could the physician maintain a private off-base office? (4) did the physician have exclusive control over his patients and records? (5) did the Government furnish the physician with permanent and private office space or secretarial help at the hospital? (6) did the physician work under a written contract with the Government? (7) was the physician regularly scheduled on the hospital duty roster? and (8) did the physician maintain regular or prescribed office hours as a civilian physician? In view of the Lilly analysis, we will review the relationship of Dr. Royal, Sterling, and Lyster Army Hospital. The contract between Sterling and the Government reveals that Sterling was clearly an independent contractor rather than an employee of the Government.[3] The more disputed question, however, is the relationship of Dr. Royal to Sterling and Lyster. We first look at the contract between Dr. Royal and Sterling. That agreement recites that Dr. Royal is an independent contractor. He is paid by Sterling on an hourly fee basis, and his work schedule is determined by Sterling. Sterling also provided Royal with a "claims made" malpractice insurance policy and reimbursed Royal for the premiums on a "tail" malpractice insurance policy.[4] The contract also specifically allowed Dr. Royal to be employed as a doctor elsewhere, but provided that he could not be employed by any "medical activity" engaged in the delivery of emergency medical services at Lyster Army Hospital. There was no written contract for services between Dr. Royal and Lyster Army Hospital. In reality, while Sterling and Dr. Royal argue that the Government controlled Dr. Royal's work at Lyster, so that Royal was cloaked with immunity, it appears that, as was the case with the doctor in Lilly, supra, the Government asserted no more control over the manner and methods used by Royal than a private hospital asserts over a private physician with staff privileges. Although Dr. Royal treated patients exclusively at Lyster Army Hospital in 1987, he did not enter into a written contract with the Government and was clearly free to maintain a private off-base medical practice. Dr. Royal did not bill the Army for his services and was not paid by the Government. The Government did prohibit Dr. Royal from being on duty for more than 12 consecutive hours for any one shift; however, Sterling, not the Government, controlled Dr. Royal's work schedule. "Surely, being subject to hospitals' rules as a condition of staff privileges does not remotely make a private physician an employee of that hospital." Lilly, supra. Upon a thorough review of the record and a consideration of the parties' positions, we hold that the trial court correctly denied the defendants' motion for summary judgment based on the FTCA, because there was substantial evidence that both Sterling and Dr. Royal were independent contractors. The same evidence shows that Dr. Royal's relationship to Sterling was that of an independent contractor. Sterling specified the hours that Dr. Royal worked and paid the premiums on Dr. Royal's malpractice insurance. Sterling was responsible for establishing and maintaining a quality control program to assure that the requirements of the contract between Sterling and the Government were provided as specified. This quality control plan included a plan for detailed inspections to be performed on a regular basis and a specified method of identifying and correcting deficiencies in the quality of services performed. Nevertheless, *517 the quality control plan did not give Sterling the right to tell Dr. Royal how to perform his duties or how to exercise his professional judgment. Rather, it was consistent with an owner or a general contractor's right to supervise the course of an independent contractor's performance without causing the independent contractor to be deemed an agent, so long as the contractor does not reserve or exercise the right to control the details of the independent contractor's performance. See, e.g., Alabama Power Co. v. Beam, 472 So.2d 619 (Ala.1985). Sterling also retained the right to terminate its contract with Dr. Royal if, for any reason, he was considered "unacceptable" by the Government and/or Sterling, but that right is also consistent with Dr. Royal's status as an independent contractor. Thus, the evidence did not contradict the contract's recitation that Dr. Royal was an independent contractor, and the trial court correctly directed a verdict for Sterling. We now address the Brillants' argument that the trial court erred in directing a verdict for Dr. Royal on the basis that the Brillants failed to present sufficient evidence that Dr. Royal's conduct proximately caused Mrs. Brillant's injuries. A directed verdict for the defendant is proper only when no reasonable inferences in favor of the plaintiff's claim can be drawn from the evidence or when there is no conflict in the evidence for a jury to resolve. Sasser v. Connery, 565 So.2d 50 (Ala.1990); Peden v. Ashmore, 554 So.2d 1010 (Ala.1989). "When a directed verdict motion is made, the evidence should be viewed in the light most favorable to the opposing party, and if a reasonable inference can be drawn against the moving party, then the trial court should deny the motion." Sasser, 565 So.2d at 51, quoting Peden, 554 So.2d at 1013. There was conflicting evidence as to whether Dr. Royal's failure to order a CT scan, followed, if necessary and appropriate in view of the CT scan results, by a lumbar puncture, fell below the appropriate standard of care. Dr. Longmore, the Brillants' expert, who is a board certified emergency room specialist, testified that Dr. Royal's failure to order such tests was beneath the standard of care required of an emergency room doctor such as Dr. Royal: "Q. [W]ould someone presenting with a sudden onset of severe headache associated with nausea and vomiting and neck pain; would that picture be compatible with a warning leak? "A. Yes, it would. "Q. Would you have an opinion as to whether or not an emergency room physician practicing in the United States of America in 1987 would be bound to know that by the standard of care? "A. Yes, I have an opinion. "Q. And your opinion would be what? "A. My opinion is that every emergency physician must be familiar with this diagnosis and [this] compilation of symptoms; that's why emergency physicians are there, to sort out the sick from the less sick. And this is a story you get with a sick patient, sudden severe headache, associated with nausea, vomiting, radiating to the neck; that's a classic presentation for a subarachnoid hemorrhage with a warning leak. I would expect every doctor to know that [who] practices in emergency medicine." Dr. Longmore testified that Dr. Royal adequately took Mrs. Brillant's history of the onset of the headache and adequately conducted a physical examination. The point at which Dr. Royal fell below the standard of care, according to Dr. Longmore, was in his failure to either confirm or rule out a subarachnoid hemorrhage as a possible cause of the headache: "Q. [H]ad a warning leak been included [as a possible diagnosis], would the standard of care have required Dr. Royal to do anything or have anything done, in order to get a warning leak ruled in or ruled out? "A. Yes. Things like a warning leak or a brain tumor or meningitis, in order to say that those things do not exist, what the doctor is required to do is to do a catscan or lumbar puncture or consult a neurologist, essentially look for definitive *518 evidence whether the patient does or does not have that serious problem." Dr. Longmore authenticated, as being "accepted as standard and trustworthy by members of [his] profession," 17 articles offered by the Brillants on the subject of subarachnoid hemorrhages, warning leaks, and treatment of the condition. Some were in publications for emergency room specialists, while others were in publications for neurologists and neurosurgeons. They were all admitted without objection. Excerpts from some of them are attached as an appendix to this opinion. The trial judge stated that, in directing a verdict, he was not ruling that there was not a jury question on the issue of whether Dr. Royal's conduct fell beneath the appropriate standard of care. Rather, the judge expressly directed the verdict for Dr. Royal based on his holding that the Brillants had failed to present substantial competent evidence that any negligence by Dr. Royal proximately caused Mrs. Brillant's paralysis. Dr. Royal's position, accepted by the trial court, is that expert testimony by a neurologist or a neurosurgeon as to the cause of her vasospasm and resulting paralysis was necessary in order for the Brillants to meet their burden of proof, and that such testimony was lacking. The burden of proof in a medical malpractice case requires the plaintiff to show that the defendant's treatment fell beneath the standard of care and proximately caused the plaintiff's injuries. This proof must ordinarily be by expert testimony. "In a medical malpractice case, in order to find liability there must be more than a mere possibility that the alleged negligence caused the injury. Williams v. Bhoopathi, 474 So.2d 690, 691 (Ala.1985). There must be some evidence that that negligence probably caused the injury. Orange v. Shannon, 284 Ala. 202, 206, 224 So.2d 236, 239 (1969)." Howard v. Mitchell, 492 So.2d 1018, 1019 (Ala.1986); Bradford v. McGee, 534 So.2d 1076, 1079 (Ala.1988); Peden v. Ashmore, 554 So.2d 1010, 1013 (Ala.1989); Harris v. Theriault, 559 So.2d 572 (Ala.1990); Sasser v. Connery, 565 So.2d 50 (Ala.1990). Dr. Longmore gave testimony directly to the point of proximate cause: "Q. The question is this, Dr. Longmore. Do you have an opinion, based on your training, on your experience, and on your reading of the medical literature as to whether or not Helen would have suffered severe vasospasm problems had her aneurysm been diagnosed at the warning leak stage, that is, on or about March 25th, and treated appropriately, as opposed to when it was diagnosed and treated? Do you have an opinion? "[Attorney for Dr. Royal]: Renew my objection. "The Court: I overrule. "A. Yes, I have an opinion. "Q. Tell these ladies and gentlemen of the jury of your opinion, please sir. "A. My opinion is that it would be very unlikely had she been treated appropriately on the first visit that she would develop a vasospasm and have long term problems. And that opinion is really corroborated by all the medical literature we have in front of us. There are many studies that demonstrate that the patient who is treated prior to having a major bleed is going to do much better." Dr. Royal argued throughout the trial that Dr. Longmore was not competent to give this testimony because he was not a specialist in neurology or neurosurgery. To determine whether the trial court correctly held that there was a failure of competent proof of proximate cause, we must examine closely the exact theory of causation advocated by the Brillants. Proof of proximate cause in this case required evidence of the following chain of causation: (1) when Mrs. Brillant presented to Dr. Royal on March 25, 1987, she probably had suffered a warning leak; (2) a CT scan or a lumbar puncture probably would have revealed the warning leak; (3) given positive results on such tests and given her overall condition, a neurosurgeon probably would have promptly performed surgery to seal off the aneurysm and prevent a major rupture; (4) if such surgery had been performed *519 promptly, she probably would not have had a vasospasm with the resulting paralysis, or, at least, the injuries caused by her aneurysm probably would not have been as severe. The Brillants acknowledge that none of these elements can be proved with certainty, principally because of Dr. Royal's failure to order a CT scan. They contend, however, that they offered competent evidence of each of these links in the causal chain, evidence that was sufficient to create a jury question whether Dr. Royal's alleged negligence "probably caused [her] injury," Howard, supra, and cases following. The following evidence tended to prove the first link, that Mrs. Brillant suffered a warning leak. Dr. Boxell, the neurosurgeon who performed surgery on Mrs. Brillant, testified on behalf of Dr. Royal. On cross-examination the following testimony was elicited from Dr. Boxell: "Q. All right, are you saying that probably on March 25th she had either bleeding or expansion? Can you give me that? "A. I think knowing what events have occurred that you could say she did have a warning event at that time." At trial, Dr. Royal testified that he did not "have any basis to agree" that the probable cause of Mrs. Brillant's headache was a leaking aneurysm. In his deposition, however, he had admitted that, in retrospect, she probably had a leaking aneurysm when he saw her. He admitted at trial that the Brillants' attorney had correctly read his deposition testimony to that effect. Thus, in addition to the testimony of Dr. Longmore, the plaintiffs' expert, that Mrs. Brillant probably suffered a warning leak that caused the headache with which she presented herself to Dr. Royal, there was also testimony by Dr. Royal and Dr. Boxell to the same effect. As to the second link, there was conflicting evidence as to whether a CT scan or a lumbar puncture would have detected a warning leak, but the record includes sufficient evidence that such tests probably would have disclosed an existing leak. Dr. Longmore testified: "Q. Do you have an opinion as to whether or not a catscan probably would have shown blood in this case? "A. Most likely. By far [in] the majority of cases of a ruptured aneurysm or a leaking aneurysm a catscan will be positive. "Q. Can a catscan detect a small amount of blood? "A. A catscan can detect a small amount of blood[;] depending on where the blood is and what the planes of the catscan are, some small amounts of blood may not be recognized and that's the advantage of doing the lumbar puncture. If you entertain a diagnosis of sentinel blood or a leaking aneurysm, if the catscan is negative one would expect a lumbar puncture would be done to rule out the diagnosis. "Q. Now, a lumbar puncture is indicated in this type of situation if the catscan did not show blood? "A. Yes." Thus, a jury question was presented on these first two links in the chain of causation. Dr. Royal's argument, therefore, depends on an assertion that the third and fourth links in the chain of causation require evidence from a neurologist or a neurosurgeon, and that such evidence was absent. Dr. Royal cites Hanson v. Baker, 534 A.2d 665 (Me.1987), as a case squarely on point regarding the necessity of testimony by a neurologist or a neurosurgeon as to causation in brain injuries. Mr. Hanson died after suffering a head injury that led to a hematoma and a contusion. The plaintiff alleged that his death was caused by delays in treatment. The trial court rejected certain testimony by the plaintiff's expert, Dr. Bricker, who was a general practitioner. The Supreme Judicial Court of Maine summarized the controversy as follows: "By way of an offer of proof out of the presence of the jury, Dr. Bricker testified that early diagnosis and treatment of similar injuries should result in recovery eighty to ninety percent of the time and that 10 hours is near the maximum *520 time that treatment can be delayed without death or serious neurological implications. Dr. Bricker would have testified that the swelling resulting from the left epidural hematoma shifted the brain from left to right and caused the contusion on the right. The treating neurosurgeon, Dr. Irwin, agreed that early diagnosis and treatment is important and that without the contusion Hanson probably would have survived. Irwin testified further that the contusion occurred at the time of the fall and was inoperable. He stated that he could only speculate as to the effect of earlier treatment." 534 A.2d at 666-67. The Supreme Judicial Court held that the trial court did not err in ruling "that Bricker could not describe the time specific progression of bruises and bleeding in the brain or express an opinion that delay in diagnosis caused Hanson's death because he was not qualified to do so... because he lacked education and experience in neurology and neurosurgery." Id., at 667. The trial court here allowed Dr. Longmore to testify that vasospasms are much less likely if there is a small amount of blood in the brain than if there is a large amount, and allowed the introduction through Dr. Longmore of neurosurgical and neurological articles to the same effect and to the effect that early surgery, after a warning leak but before a full rupture, is likely to prevent vasospasms, neurological deficits, and death. Nevertheless, Dr. Royal insists that this evidence was not sufficient to meet the Brillants' burden of proof. In addition to the evidence from Dr. Longmore, however, there was also evidence from Dr. Boxell on what we have set out as the third and fourth links in the chain of causation. An examination of the third link will be made easier by first addressing the fourth, that is, whether, in general, diagnosis and treatment of an aneurysm after a warning leak will prevent vasospasm and subsequent neurological deficits and, in particular, whether they probably would have done so in Mrs. Brillant's case. The crucial aspect of this question is the evidence that, after a warning leak, a person will have a small amount of blood in the brain, whereas a full rupture of the aneurysm will spill a large amount of blood into the brain. There was evidence that the excess blood in the brain irritates blood vessels in the brain and causes them to collapse. This collapse is the vasospasm, which depletes the supply of blood and therefore oxygen to parts of the brain, which then die and cause neurological deficits such as Mrs. Brillant's paralysis. When surgery is conducted after a full rupture of the aneurysm, vasospasm is likely even if the surgery is successful because of the amount of blood that has been discharged into the brain. Dr. Longmore's testimony and the articles introduced through his authentication tended to establish these facts. Even if we accept Dr. Royal's contention that evidence of this fourth link in the causal chain requires testimony of a neurologist or a neurosurgeon, we find sufficient evidence on this question from Dr. Boxell: "Q. Vasospasm is caused by the irritating effect of blood? "A. That's correct. "Q. Does it follow sir, that the more blood that comes out of that aneurysm, the more likely vasospasm is to occur? "A. That's correct. "Q. Conversely, the less blood that comes about as a result of a vasospasm—as a result of an aneurysmal bleed, the less likely is vasospasm? "A. Overall the less likely that it will occur. ".... "Q. Assuming that at the time of ... Helen's visit to Dr. Royal, there was an aneurysmal bleed, wouldn't—not for sure, I'm not asking if you know for sure, but wouldn't the probability be that such bleeding would have been a small amount? "A. Again, I don't know any specific studies that have addressed that type of question. What you can say is the more blood that's present at the base of the brain the more likely you are to have *521 vasospasm and the more likely you are to develop an adverse neurological outcome. The more likely you are to suffer a stroke or effects similar to a stroke. "Q. Conversely— "A. The less likely—the smaller the quantity of blood generally the better the condition of the patient without absolutes in there. I mean there is never always. "Q. I understand. I understand the absolutes business. Okay, now, with that in mind and knowing there hasn't been—Also is it true—let me go about it this way for a moment. Also isn't it true that the less blood which comes out of an aneurysmal bleed, the less likely the person experiencing that small bleed is to show changes in neurological status or neurological function? "A. Yes. "Q. Now considering that too, if a lady presents as is shown on Plaintiff's Exhibit 12, the March 25th ER visit, if you assume such person had an aneurysmal bleed, again not the surety, but wouldn't the probability be that the amount of that bleeding would have been small? Wouldn't that have been your judgment about it sir? "A. If you had a person in very good condition and that they would have had a small amount of blood present on a scan? "Q. Answer that one. "A. They would in all probability, they would have a small quantity." Dr. Boxell was then questioned regarding an article that he was familiar with, "The Relation of Cerebral Vasospasm to the Extent and Location of Subarachnoid Blood Visualized by CT Scan: A Prospective Study," 33 Neurology 424-36 (1983), admitted as plaintiffs' exhibit 36 through Dr. Longmore. Dr. Boxell acknowledged that the results were that, of the "Group 3" patients with significant amounts of bleeding, such as Mrs. Brillant had on April 10, 20 out of 22 developed severe vasospasm, and 19 of the 22 "developed symptoms referable to the severe vasospasm." Of the "Group 1" and "Group 2" patients with little or no subarachnoid blood shown on a CT scan after a warning event such as Mrs. Brillant's headache, only 5 out of 20 developed severe vasospasm, and only 2 of the 20 "developed delayed ischemic neurologic symptoms." Thus, 86% of the patients with fully ruptured aneurysms developed neurological deficits, but only 10% of the patients treated after warning leaks developed neurological deficits. Thus, there was competent evidence that surgery after a warning leak is likely to prevent a vasospasm, but that a vasospasm is likely to occur after a full rupture even if surgery successfully repairs the ruptured aneurysm. The question as to which there is an arguable gap in the evidence is the third link in the chain of causation set out above, that is, would Mrs. Brillant have been a good candidate for neurosurgery when she presented to Dr. Royal on March 25? Dr. Boxell testified generally that a patient in a "good grade" neurologically— that is, alert, in good health generally, and without large amounts of blood in the brain—is a better candidate for neurosurgery than a person in a poor grade—that is, unconscious or in a coma and with large amounts of blood in the brain. Thus, there was sufficient evidence that a person in Mrs. Brillant's condition as she presented on March 25 would be a good candidate for neurosurgery.[5] Moreover, the Brillants went further, and asked Dr. Boxell a question that would have specifically addressed Mrs. Brillant's likelihood of surgery if a CT scan had been done on March 25 and had shown a leaking aneurysm. The trial court sustained Dr. Royal's objection when Dr. Boxell was asked the following question by the Brillants' attorney: "Doctor, I would like to pose this hypothetical. Please assume the following *522 facts. A 56-year-old woman comes into the emergency room complaining of a sudden headache with neck pain and nausea and vomiting. A CT scan was done by an emergency room physician. That the CT scan demonstrated a leaking aneurysm. If the patient was otherwise in good health and if she was not in a coma, do you have an opinion as to whether or not she would be a good candidate for surgery?" The objection was made on the grounds that the question was speculative, was not based on a proper predicate, and was based on facts not in evidence. The Brillants raise the sustaining of this objection as a ground for reversal of the judgment. Obviously, the question includes two facts that were not in evidence: that a CT scan was performed and that the CT scan showed a leaking aneurysm. However, as we have shown in discussing links one and two of the causal chain that the Brillants must prove, there was sufficient evidence from which the jury could have found that a CT scan should have been performed and that, if performed, a CT scan (or a followup lumbar puncture) probably would have disclosed a leaking aneurysm. The question, therefore, is whether the opinion evidence that created a jury question on these issues was a sufficient basis on which to predicate the "facts" of a CT scan being performed and the scan showing a leaking aneurysm. Until recently, this Court had held that a hypothetical question to an expert "must be based on facts and not upon opinions or conclusions of others." Chinevere v. Cullman County, 503 So.2d 841, 843 (Ala. 1987); Salotti v. Seaboard Coast Line R.R., 293 Ala. 1, 299 So.2d 695 (1974). That rule was modified, however, in Nash v. Cosby, 574 So.2d 700 (Ala.1990): "The rule disallowing expert testimony based on opinions or conclusions of others simply ignores the reality of current medical practice. It cannot be reasonably argued that doctors and other medical practitioners do not rely on the observations and conclusions of other medical professionals in making diagnoses. Thus, it is appropriate that those same conclusions and observations should be taken into account by any medical expert testifying as to an applicable standard of care or as to a breach thereof.[6] We believe that use of such conclusions and observations would only improve the reliability of expert testimony on medical issues." Id., at 705 (citation omitted). Nash dealt with observations and conclusions stated in the plaintiff's medical records. Here, the question is whether the conclusions of witnesses, based on Mrs. Brillant's medical records and on the witnesses' professional training and judgment, can serve as a predicate for a hypothetical question to an expert. Those conclusions were items that, as we have shown, the jury could have accepted as fact: because it was highly probable that the cause of Mrs. Brillant's headache was a warning leak, the jury could have found as a matter of fact that she did have a warning leak; the jury could have found that, in fact, Dr. Royal should have performed a CT scan and, if necessary, a lumbar puncture; and, because it was highly probable that one of those tests would have disclosed a warning leak, the jury could have found as a matter of fact that such tests would have disclosed that Mrs. Brillant had suffered a warning leak. Thus, the opinions or conclusions that formed the predicate were presented as fact questions for the jury, and we see no reason to prohibit the Brillants from asking Dr. Boxell the objected-to question. If Dr. Boxell had been allowed to answer this question, his general evidence that a person in a condition similar to that of Mrs. Brillant on March 25 would have been a good candidate for neurosurgery could have been supplemented by specific evidence as to whether she would have been a good candidate for neurosurgery. The trial court, in directing a verdict for Dr. Royal, cited Peden v. Ashmore, 554 So.2d 1010 (Ala.1989), and Sasser v. Connery, 565 So.2d 50 (Ala.1990). In Peden, a patient with a history of heart disease and *523 other health problems presented at the hospital; her doctor, Dr. Ashmore, was notified at 11:52 a.m. A physician's assistant examined her, prescribed several medications, and ordered an immediate brain scan. The brain scan was completed at 5:00 p.m.; it confirmed Dr. Ashmore's suspicions that she had an expanding intracranial lesion. She was transferred to a different hospital for neurosurgery but died shortly after arriving there. The plaintiff's theory was that a delay in treatment proximately caused the patient's death. Even the plaintiff's expert testified, however, that it would be speculation that any different treatment would have produced a different result, and that "I cannot say whether she would have lived or died." 554 So.2d at 1014. The neurosurgeon to whose care she had been transferred testified that he doubted that anything could have been done to save her life. The trial court directed a verdict because of a lack of proof of proximate cause, and this Court affirmed. In Sasser, the plaintiff alleged that Dr. Connery's failure to conduct tests on the patient caused her death from cancer, which, the plaintiff alleged, could have been treated if detected by Dr. Connery. "None of the experts could say that if Dr. Connery had conducted the tests in question and had found the cancer, then Ms. Sasser's life would have been saved or even extended." 565 So.2d at 51. The trial court denied Dr. Connery's motion for directed verdict, but the jury returned a verdict for Dr. Connery. This Court affirmed the judgment without reaching a jury charge issue raised on appeal by the plaintiff, holding that the trial court should have directed a verdict for lack of proof of proximate cause. A similar recent case involving a lack of proof of proximate cause is Harris v. Theriault, 559 So.2d 572 (Ala.1990). The plaintiff's decedent, Mr. Harris, was seen in the emergency room on December 25 complaining of headaches, nausea, and weakness on the right side of his body. Dr. Theriault made a diagnosis of vascular headache and prescribed painkillers. The next day Mrs. Harris brought her husband back in a much worsened condition. A neurosurgeon successfully performed a carotid endarectomy, but a later embolism caused his death. The plaintiff's expert stated in an affidavit that, if Dr. Theriault had ordered a CT scan, an earlier diagnosis and treatment "might have avoided the cerebral embolism, which led to his death." 559 So.2d at 573. That expert's deposition, however, established that the embolism was an unpreventable byproduct of the surgery and would not have been avoided by earlier surgery.[7] This Court affirmed a summary judgment for the defendant. Those three cases are distinguishable from this one. In each of them, there was a complete absence of proof that a different result probably would have occurred with different treatment. Here, there was substantial evidence that Dr. Royal's failure to perform the proper diagnostic techniques probably caused the vasospasm and paralysis by delaying neurosurgical treatment until after a full rupture of the aneurysm. Furthermore, there was evidence that, generally, surgery should be performed on a person who presents with a warning leak in an aneurysm but otherwise in good condition. Any failure to tie that general evidence to Mrs. Brillant's specific condition was caused by the court's sustaining an objection that under the (later announced) rule of Nash v. Cosby, should have been overruled. In their cross-appeal, Sterling and Dr. Royal contend that, even if there was sufficient evidence of negligence on the part of Dr. Royal, they were nevertheless entitled to a directed verdict on the theory that the conduct of Dr. Velez constituted an efficient, intervening cause. An intervening cause is defined as "one that occurs *524 after an act committed by a tort-feasor and which relieves him of his liability by breaking the chain of causation between his act and the resulting injury." Richter v. Uhrig, 534 So.2d 260, 262 (Ala.1988). However, this Court has stated: "Not every cause which comes into operation after a tortfeasor has acted will relieve him of liability for his wrongful act. More than the proper temporal relationship between the tortfeasor's act and the subsequent cause is required. In order to be an intervening cause, a subsequent cause also must have been unforeseeable and must have been sufficient in and of itself to have been the sole `cause in fact' of the injury." General Motors Corp. v. Edwards, 482 So.2d 1176, 1195 (Ala.1985). The expert testimony was in agreement that Mrs. Brillant did not exhibit the classic symptoms of a warning leak when she was seen by Dr. Velez. Therefore, his failure to diagnose her earlier warning leak was not an intervening cause relieving Dr. Royal of liability. For the reasons stated, we affirm the judgment based on the directed verdict in favor of Sterling Medical Associates, we reverse the judgment based on the directed verdict for Dr. Royal, and we remand this case for a new trial consistent with this opinion. AFFIRMED IN PART, REVERSED IN PART, AND REMANDED. HORNSBY, C.J., and SHORES, ADAMS, HOUSTON, KENNEDY and INGRAM, JJ., concur. MADDOX and STEAGALL, JJ., dissent. APPENDIX Plaintiff's Exhibits Exhibit 19—McMicken, "Emergency CT Head Scans in Traumatic and Atraumatic Conditions," 15 Annals of Emergency Med. 274 (1986): "Approximately 50% of patients will experience premonitory symptoms in the weeks before rupture. These symptoms and signs may consist of localized head pain, cranial nerve palsy, visual field defect, nausea, photophobia, or neck pain. These warning signs may be produced by `mini leaks' or aneurysmal expansion. ".... "Any patient suspected of SAH [subarachnoid hemorrhage] should undergo CT as the initial diagnostic study. This rapidly performed noninvasive procedure will be abnormal 90% of the time if performed within five days of the SAH. "As the hemorrhage is reabsorbed the initial high density becomes isodense, similar to adjacent normal brain. Therefore, evidence of subarachnoid, intracerebral, intraventricular or subdural bleeding will be present on CT 95% of the time within hours after the hemorrhage, decreasing to 74% on day 3. ".... "Improvement in the dismal mortality rate from subarachnoid hemorrhage is achieved only by early recognition and prompt treatment. CT has provided a safe, sensitive, and noninvasive answer." pp. 277-78. Exhibit 20—Duffy, "The `Warning Leak' in Spontaneous Subarachnoid Haemorrhage," Med. J. of Australia 514 (May 28, 1983): "The frequency and significance of the warning leak should be recognized. Such recognition would improve the outcome after the rupture of an intracranial aneurysm." p. 514. Exhibit 21—Drake, "Management of Cerebral Aneurysm," 12 Stroke, 273 (1981): "Only a small fraction of the patients are seen after the initial bleed when the greatest therapeutic reward would occur. "The challenge for the future then, will be the early recognition of the initial bleeding, the warning bleeding. It will require public education about the problem and a continuing emphasis on it for students and physicians. The potential for prevention of death or dreadful disability is large for thousands in the prime of life each year." p. 273. (Emphasis in original.) *525 Exhibit 22—Gillingham, "The Management of Ruptured Intracranial Aneurysms," 12 Scottish Med. J. 377 (1967): "Over half the patients who are admitted in coma, stupor, or who are otherwise severely disabled, from ruptured intracranial aneurysm have a clear-cut history of a minor haemorrhage a few days or weeks before the major episode. Although there is a sudden onset of severe headache or pain in the back of the neck associated with meningism, in these patients with a minor haemorrhage there are usually no other symptoms or clinical signs and the diagnosis is often missed. Its recognition is important because operative treatment for the prevention of recurrent haemorrhage during the first few days after this minor episode results in a low mortality and morbidity. The problems of management of ruptured intracranial aneurysm are discussed in the light of a personal experience of 291 consecutive cases over a period of fifteen years, 81 per cent of whom have been followed up." p. 383. Exhibit 23—Leblanc & Winfield, "The Warning Leak in Subarachnoid Hemorrhage and the Importance of Its Early Diagnosis," 131 Canadian Med. Ass'n J. 1235 (1984): "Subarachnoid hemorrhage caused by an aneurysm usually presents as an extremely severe, sudden and incapacitating headache associated with prostration or loss of consciousness, nuchal rigidity, photophobia and vomiting. Death or serious permanent disability occurs in over 60% of cases. Neurologists and neurosurgeons have come to recognize that the classic major subarachnoid hemorrhage may be preceded by a minor hemorrhage, a `warning leak,' that produces a moderately severe headache with minor or no accompanying features. The severity of the headache is often such that the patient seeks medical advice, but subarachnoid hemorrhage is seldom suspected because other classic manifestations of this condition are lacking. The patient usually remains well for up to 4 weeks, until a major rupture occurs and a diffuse subarachnoid hemorrhage, with its often devastating effects, ensues. If the warning leak is recognized, a potentially fatal hemorrhage can be prevented by treating the aneurysm before it ruptures. ".... "Other signs of a warning leak include nausea or vomiting, neck pain, unsteadiness and dizziness.... Unlike the classically described subarachnoid hemorrhage, a warning leak is only rarely associated with photophobia or signs of diffuse meningeal irritation. "The responsibility for recognizing the warning leak rests with the primary care physician, who is usually the first to see the patient. We recommend that individuals without a history of chronic headache who present with a severe, subjectively atypical headache of sudden onset, especially if it is accompanied by other warning signs, undergo a lumbar puncture, provided there are no indications of intracranial hypertension. If the findings are compatible with subarachnoid hemorrhage the patient should be immediately admitted to hospital for four-vessel angiography. If an aneurysm is seen, the patient should promptly undergo surgery. Aneurysmal subarachnoid hemorrhage usually occurs in the most productive years of life and is potentially fatal. Recognition of the warning leak can lead to appropriate treatment of the aneurysm before a frank rupture occurs, with its often devastating consequences." pp. 1235-36. Exhibit 24—King & Saba, "Forewarnings of Major Subarachnoid Hemorrhage Due to Congenital Berry Aneurysm," N.Y.St.J. Med., 638 (April 1974): "Spontaneous subarachnoid hemorrhage carries a high mortality rate when the diagnosis is made using customary criteria. Forewarnings of a major bleed, however, occur in many patients. They are such that the diagnosis can often be suspected and confirmed by lumbar puncture before a major ictus. This review of 175 patients suggests that these forewarnings were recorded for at least 60 per cent of the patients and form a constellation of complaints which are sufficient to lead to their diagnosis and management at a time when *526 the patient is still in good clinical condition. The mortality rate in such circumstances may be less than 10 per cent compared with 40 per cent after a major subarachnoid hemorrhage." Exhibit 25—Klara & George, "Warning Leaks and Sentinel Headaches Associated with Subarachnoid Hemorrhage," 147 Military Med. 660 (1982): "It is significant, however, that 50 per cent of the patients who present with SAH secondary to rupture of an aneurysm, experience a `warning leak' associated with a `sentinel headache' prior to the first clinically significant hemorrhage. Unfortunately, these patients are all too frequently seen in clinics or emergency rooms, where they are examined and found to be neurologically intact. They are then given analgesics and sent home. When their symptoms persist, they occasionally return only to receive the same treatment. It is the usual course of events that these patients will eventually suffer another hemorrhage and then develop a neurologic deficit or a stiff neck. Usually, it is only when they present with these findings that lumbar puncture is performed to evaluate whether or not SAH has occurred. Frequently, however, the surviving patients have already suffered significant effects from the recurrent hemorrhage, and present in a coma or with significant focal neurologic symptoms. "The purpose of this paper is to acquaint physicians and allied health care personnel with the early warning signs of SAH, in an attempt to increase significantly the early diagnosis of SAH prior to the first major hemorrhage. This is important since patients with `warning leaks' are, by definition, in Grade 1 [with or without mild headache, alert and oriented, with no motor or sensory deficit] status, and therapy for this group of patients carries an overall mortality rate of less than two per cent, with a similarly low morbidity rate. If the warning signs are missed and the patient suffers a second hemorrhage, he faces an immediate mortality of 40 per cent. He then must face the mortality and morbidity associated with surgical therapy. This mortality risk varies depending on the neurological condition of the patient, but exceeds 30 per cent for Grade 4 [semicomatose or comatose, with or without major lateralization findings] patients. "The `warning leak' or `sentinel headache' is, unfortunately, usually recognized in retrospect following a major hemorrhage. Such a `sentinel headache' is sufficiently distinctive so as to make it a recognizable entity. Hence, diagnosis prior to major hemorrhage should be possible. The headache is sudden in onset, occurring usually in patients without a past history of headache complaints. It is frequently described as `the worst headache ever.' If nuchal rigidity is present in such a setting, the classic presentation of subarachnoid hemorrhage has occurred nd should be recognizable immediately. Even without associated nuchal rigidity, sudden onset of headache in an otherwise healthy individual should alert the examiner to include `sentinel headache' or `warning leak' in the differential diagnosis. ".... "Over the past ten years, significant progress in the neurosurgical treatment of intracranial aneurysm has been made. The development of microsurgical techniques and training of neurosurgeons with expertise in this area have been supplemented by advances in neuroanesthesia and neuroradiology. These technical improvements have vastly bettered the prognosis for the individual patient with a SAH who is fortunate enough to survive the initial major hemorrhage. Many more lives can be saved if aneurysm patients are brought to neurosurgical treatment prior to experiencing the first major hemorrhage. Almost equally important, the number of individuals compelled to live out their lives with severe neurologic disabilities will be greatly reduced. "Presently, this could be most easily accomplished by increasing our level of suspicion for SAH and supplementing this with a vigorous, aggressive approach at making an early diagnosis.... Following such an approach, the overall morbidity and mortality *527 rate for patients with intracranial aneurysm and SAH should be considerably reduced." pp. 660, 662. Exhibit 28—Gillingham, "The Management of Ruptured Intracranial Aneurysm—Hunterian Lecture delivered at the Royal College of Surgeons of England on Jan. 24, 1957," 23 Annals Roy. Coll. Surg. Eng. 89 (1958) (this paper apparently was the seminal work in the field): "It would appear from the fresh appraisal of many case histories, that a patient's first episode of subarachnoid bleeding is commonly of minor severity; a mere leak of blood from the sac. There is sudden severe pain in the neck which rapidly radiates upwards over the vertex of the head. It is then followed by generalized headache which clears up in a day or two. A diagnosis of fibrositis, influenza or stiff neck may be made by the patient, his relatives or his medical attendant. Fortunately the presence of neck stiffness, even of minor degree, and other signs of meningism, usually indicate the true diagnosis. The accurate clinical assessment of such an attack is of great importance, in view of the probability and gravity of a second haemorrhage within the next few days; commonly between the second and the twenty-first day following the first haemorrhage. This second episode of bleeding, as we have seen, is often more severe than the first. The patient is unconscious for a period and is usually left with neurological deficits such as hemiplegia, visual field defects, language difficulties, apathy and negativism, dependent to some extent on the presence of an intracerebral haematoma. However, some patients die within minutes or hours of raised intracranial pressure from a massive intra-cerebral and intra-ventricular haemorrhage, and in a few patients this disastrous episode seems to represent the initial and only attack of bleeding, as opposed to the second or third.... ".... "There is now good evidence to support the view that patients who have suffered from their first, and often minor, episode of subarachnoid bleeding should be referred as a matter of urgency for neurosurgical investigation and treatment." pp. 90-93, 116. EXHIBIT 29—Okawara, "Warning Signs Prior to Rupture of an Intracranial Aneurysm," 38 J. Neurosurg., 575 (1973): "Most neurosurgeons agree that the treatment of an aneurysm before massive hemorrhage provides a more favorable outcome. If it were possible to recognize warning signs of an impending aneurysmal hemorrhage, the surgeon would have a distinct advantage. It is my impression that such warning signs do exist, erroneously suggesting influenza, sinusitis, `stiff neck,' migraine, or other diseases, but that their true nature is only appreciated after catastrophic major hemorrhage. "The purpose of this paper is to investigate the significance of warning signs that herald a major hemorrhage, thus leading to an early diagnosis and better results from treatment of intracranial aneurysms." p. 575. Exhibit 35—Waga, Ohtsubo & Handa, "Warning Signs in Intracranial Aneurysms," 3 Surg. Neurol. 15 (1975): "Warning signs preceding major hemorrhage were analyzed in 192 patients with intracranial aneurysms. One hundred and thirteen (59%) had warning signs. The average interval between the last warning signs and the major attacks was 16.9 days with anterior communicating and anterior cerebral aneurysms, 6 days with middle cerebral aneurysms, and 7.3 days with intradural internal carotid aneurysms. The incidence was much higher in patients with intradural internal carotid aneurysm (69%) and with multiple aneurysms (67%). The warning signs varied chiefly according to the location of the aneurysms. Early recognition of warning signs may protect patients from the devasting effects of major hemmorrhage. ".... "Subarachnoid hemorrhage due to saccular aneurysm has a tendency to be recurrent. This may give physicians the opportunity to diagnose and to treat the patients before a catastrophic situation develops. *528 Unfortunately, the `sentinel headache' was more often recognized only in retrospect. This presents an important clinical paradox, which may not be generally understood—the less affected the patient is by the initial hemorrhage the more urgent is the need for investigation and treatment. Drake wrote that in spite of increasing success in obliterating aneurysm in reasonably well patients, nearly one in ten is lost during the delay awaiting a safe period for operation. Prevention of the devastating effects of a major hemorrhage from an aneurysm must be the goal. At present the only way is for the family physician to recognize the warning signs of a minor initial leak, which occurs in over half of the patients, and to refer them for surgical attention before the brain is torn asunder." pp. 15, 19. Exhibit 37—Mizukami, Takemae, et al., "Value of Computer Tomography in the Prediction of Cerebral Vasospasm After Aneurysm Rupture," 7 Neurosurgery, 583 (1980): "The diagnostic value of computer tomography (CT) in cases of ruptured aneurysm has already been suggested. However, except for our earlier report ... and that of Fisher et al. ..., there has been little relationship noted between cerebral vasospasm and findings on the CT scan. The purpose of this study is to clarify the relationship between high density (HD) on the CT scan, which indicates a collection of blood in the subarachnoid space, and cerebral vasospasm after the rupture of an aneurysm, after first defining the occurrence of HD from the acute stage to the chronic stage. ".... "The additional findings obtained from CT scan within 4 days after SAH [subarachnoid hemorrhage] may be a useful adjunct to predict the development of cerebral vasospasm and, thereby, improve the mortality and morbidity rates after SAH. We recommend early operation in patients with a clinical grade of I, II, or III (Hunt and Kosnik) ... and, when HD has not been detected in the basal and/or insular cisterns, there is little danger of postoperative cerebral vasospasm. If HD is found in these regions, early operation for removal of the subarachnoid hematoma (within 2 or 3 days after the hemorrhage), as suggested by Sano and Saito and Suzuki et al., may be a reasonable alternative...." pp. 583, 585. Exhibit 38—Fisher, Kistler & Davis, "Relation of Cerebral Vasospasm to Subarachnoid Hemorrhage Visualized by Computerized Tomographic Scanning," 6 Neurosurgery, 1 (1980): "Cranial computerized tomography (CT) allows accurate assessment of the presence of subarachnoid hemorrhage secondary to rupture of an intracranial aneurysm. We have attempted to determine whether the amount and distribution of subarachnoid blood detected by CT early after aneurysmal rupture correlate with the later development of cerebral vasospasm visualized angiographically. We analyzed 47 cases from the Massachusetts General Hospital and found a remarkably positive correlation: When there was no subarachnoid blood or it was diffusely distributed, severe vasospasm was almost never encountered, whereas in the presence of blood clots and thick layers of blood, severe vasospasm followed almost invariably." p. 1. MADDOX, Justice (dissenting). "In Alabama, a medical malpractice plaintiff must establish through expert testimony that the defendant/physician breached the standard of care imposed upon him ... and that this breach was the proximate cause of the injury or death." Peden v. Ashmore, 554 So.2d 1010, 1013 (Ala.1989). A plaintiff in a medical negligence case bears the heavy burden of presenting evidence "that the alleged negligence probably cause the injury or death." Id. (Emphasis original.) As this Court held in Howard v. Mitchell, 492 So.2d 1018 (Ala.1986): "In a medical malpractice case, in order to find liability there must be more than a mere possibility that the alleged negligence caused the injury.... There must *529 be some evidence that that negligence probably caused the injury." 492 So.2d at 1019. The trial judge, in granting Dr. Royal's motion for a directed verdict, cited Peden and further noted that Peden applied the "scintilla" evidence rule rather than the "substantial" evidence rule that is applicable here. The trial judge specifically found: "[T]here was insufficient evidence, when taken as a whole, to constitute substantial evidence that the physician's alleged negligence probably proximately caused the injuries to Helen B. Brillant." I think that the trial judge was right. To find liability here, a jury would have to stack an inference on an inference and engage in speculation. STEAGALL, J., concurs. NOTES [1] The trial was bifurcated on the issues of liability and damages and, thus, because of the directed verdicts as to liability, no evidence was presented as to the severity of Mrs. Brillant's disability. [2] This cause of action accrued before the effective date of the Medical Liability Act of 1987, see Ala.Code 1975, § 6-5-552, so it was not tried under that Act. Because the action was not "pending in the courts of this state on June 11, 1987," however, it was tried under the substantial evidence rule of § 12-21-12. [3] The contract between Sterling and the Government specifically provided that Sterling would perform as an independent contractor and not as an employee of the Government. It also provided that the Government would assume no responsibility for the negligent acts of Sterling or Sterling's employees and specifically allowed physicians employed under the contract to engage in private practice. Further, it provided that Sterling was responsible for providing malpractice insurance and for quality control. [4] The payment of premiums for liability insurance by the alleged principal has been held to create a question of fact as to whether an agency or independent contractor relationship existed. Moore-Handley Hardware Co. v. Williams, 238 Ala. 189, 189 So. 757 (1939). [5] We recognize that Dr. Boxell stated prior to trial, "I do not personally think that the delay in establishment of the diagnosis of aneurysm had any bearing on the development of her vasospasm." Dr. Boxell acknowledged this statement at trial and otherwise gave testimony favorable to Dr. Royal. Elsewhere in his testimony, however, he stated facts or expressed opinions that, as we have shown, supported the Brillants' burden of proof. [6] Or, as in this case, as to proximate causation of the plaintiff's injury. [7] There is no evidence in the record before us that Mrs. Brillant's temporal edema (a swelling produced by blood spilled into the brain), which apparently caused the vasospasm, was an unavoidable consequence of surgery, whether the surgery was done after the warning leak or after the full rupture. On the contrary, there was evidence, as we have shown, that the blood that caused the vasospasm was the blood produced by the rupture.