In this medical malpractice case the plaintiff sued Portsmouth Hospital and Doctors Eugene B. Andes and Frank A. Graf for negligence in connection with their medical treatment of her husband following an automobile accident. A jury trial resulted in verdicts for all three defendants. On appeal, the plaintiff contends that the Trial Court (Gray, J.) should have shifted the burden of proof to the defendants on the issue of causation because the nature of the defendants’ negligence made proof of causation by the plaintiff difficult or impossible. We disagree, and therefore affirm.
At 9:20 p.m. on Friday, November 16, 1979, Daniel G. Pillsbury-Flood (the patient or deceased) was in a motor vehicle accident in *301which his head struck the windshield of the car he was driving. He did not receive medical attention that night, but the next day he went to the Portsmouth Hospital emergency department at 4:15 p.m., complaining of head and neck pain. Dr. Andes, the emergency department physician, performed a neurological examination to determine whether the patient suffered a head injury; the result was negative. The doctor ordered a series of X-rays of the patient’s neck and skull. Because of an abnormality in the neck X-ray, Dr. Andes consulted Dr. Graf, an orthopedist, to evaluate the patient’s neck. Dr. Graf viewed the X-rays, examined the patient and concluded that no fracture or other acute head or neck injury had occurred.
Neither doctor found signs or symptoms of a head injury. The patient left the hospital at 7:00 p.m. with written instructions to contact Dr. Graf and two therapists the following Monday morning. The instructions also stated: “Relaxing movement. No stress for your body, long trips.” The patient was not given a copy of the hospital’s head trauma sheet, which lists six signs of possible head injury: (1) excessive or persistent headache, (2) excessive or persistent dizziness or drowsiness, (3) excessive or persistent nausea or vomiting, (4) lack of coordination, (5) obvious difference in pupil size, and (6) unusual behavior. The sheet states that the patient should be awakened every two hours for the following twelve to twenty-four hours, and that if any of the above symptoms occur, a physician should be contacted or the patient should be brought to the emergency room.
That Sunday morning, the patient complained of head and neck pain. The plaintiff reminded her husband “that the doctor had told him that he would feel worse before he felt better.” The plaintiff left the house around 11:30 a.m., and a friend visited her husband for a short time after noon. When the plaintiff returned at 4:00 p.m., she found her husband dead.
According to the doctor who performed the autopsy, death was caused by an acute subdural hematoma, which “results from the tearing of small veins between the venous sinuses of the skull and brain, resulting in an [sic] very gradual collection of blood between the skull and the brain.” The autopsy report also stated that a generalized seizure caused the death, and that the symptoms of a subdural hematoma are characteristically slow in onset and may not occur for days or weeks.
The plaintiff, as administratrix of the deceased’s estate, sued the defendants, alleging that their medical negligence caused her husband’s death. At trial, the plaintiff’s expert witness testified that the medical care and follow-up instructions the patient received were totally inadequate. The expert also testified that a CAT scan would have disclosed the accumulation of blood in the patient’s subdural *302area, and enabled him to receive the medical care he needed within a few hours. The defendants’ expert testified that the doctors acted competently. He also indicated that he could not say as a matter of probability that the patient would not have died had he been given a head trauma sheet, and that death caused by a subdural hematoma can occur rapidly once the blood pressure reaches a high level. Finally, the doctor testified that the defendants’ care did not cause the patient’s death.
During trial, the plaintiff moved for partial directed verdicts, asserting that the burden of proof of causation had shifted to the defendants and that the defendants had failed to prove that their conduct did not cause the patient’s death. This motion was denied. The plaintiff again raised the burden of proof issue by requesting jury instructions consistent with her theory. The court refused to shift the burden of proof and instructed the jury that the plaintiff had to prove that the defendants’ negligence was a “substantial factor in causing the injury complained of.” After the jury rendered general verdicts for the defendants, the plaintiff moved to set aside the verdicts and requested a new trial, alleging error in the court’s failure to shift the burden of proof and to instruct the jury accordingly. These motions were denied, and the plaintiff appealed.
The plaintiff argues that the defendants’ failure to properly diagnose the head injury, as well as their failure to instruct the patient and give the head trauma sheet prescriptions, constituted negligence; and she therefore argues that the trial court should have shifted the burden of proof on the issue of causation to the defendants. In particular, the plaintiff asserts that the defendants should have had to demonstrate that it was more probable than not that (1) such negligence did not result in the patient’s failure to seek further treatment; (2) the subdural hematoma could not have been diagnosed even if further treatment had been sought; and (3) diagnosis of the subdural hematoma would not have minimized or eliminated the risk of death. The defendants argue that (1) the jury made no finding of negligence; (2) the trial court’s refusal to shift the burden of persuasion was correct; and (3) the cases relied upon by the plaintiff indicate only that when the plaintiff offers evidence of negligence, and a risk of injury follows from that breach of duty, the jury should be allowed to consider the negligence and causation issues, and not that the burden of proof shifts to the defendants.
In this appeal, the plaintiff relies on the “loss of a chance” doctrine as expressed in Hicks v. United States, 368 F.2d 626 (4th Cir. 1966). In Hicks, the plaintiff, who was the administrator of a deceased woman’s estate, alleged that the woman’s death was caused by the negligence of a doctor who failed to properly diagnose and treat the *303woman. The government contended that there was no breach of the medical standard of care owed to the deceased, and that even if there was negligence, there was no proof that the misdiagnosis and improper treatment were the proximate cause of the patient’s death. The government asserted that “it is mere speculation to say that [an alternative diagnosis and treatment] would have been successful.” Id. at 632.
The Hicks court disagreed, citing testimony indicating that the woman would have survived if she had been operated on. The court stated:
“When a defendant’s negligent action or inaction has effectively terminated a person’s chance of survival, it does not lie in the defendant’s mouth to raise conjectures as to the measure of the chances that he has put beyond the possibility of realization. If there was any substantial possibility of survival and the defendant has destroyed it, he is answerable. Rarely is it possible to demonstrate to an absolute certainty what would have happened in circumstances that the wrongdoer did not allow to come to pass. The law does not in the existing circumstances require the plaintiff to show to a certainty that the patient would have lived had she been hospitalized and operated on promptly.”
Id. The court in Hicks analogized the facts of that case to an admiralty case involving the duty to rescue a man overboard. See Gardner v. National Bulk Carriers, Inc., 310 F.2d 284 (4th Cir. 1962). In Gardner, the court stated that the ship’s master had a duty to attempt a rescue if a reasonable.possibility of success existed, and that proximate cause is proved if the master’s omission destroyed the reasonable possibility of rescue. Id. at 287. The Hicks court concluded that since expert testimony indicated prompt surgery would have saved the patient, the “doctor’s negligence nullified whatever chance of recovery [the deceased] might have had and was the proximate cause of the death.” 368 F.2d at 633.
Judicial response to the “loss of a chance” doctrine has varied. Some courts have relied on Hicks to relax the burden of proof on the issue of causation, and have submitted the negligence and causation issues to the jury on similar facts. See, e.g., Thompson v. Sun City Community Hosp., Inc., 688 P.2d 605, 616 (Ariz. 1984) (employing section 323 of the Restatement (Second) of Torts, which states that a doctor may be subject to liability if his failure to exercise reasonable care increases the risk of harm to the patient); Hershovits v. Group Health Cooperative, 99 Wash. 2d 609, 664 P.2d 474 (1983); Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978). Other courts have *304refused to lessen the burden of proof, and have granted defendant’s motions for directed verdicts when plaintiffs were unable to establish that the alleged negligence probably caused the injury. See, e.g., Gooding v. University Hosp. Bldg., Inc., 445 So. 2d 1015, 1020 (Fla. 1984); Cooper v. Sisters of Charity of Cincinnati, Inc., 27 Ohio St. 2d 242, 272 N.E.2d 97 (1971). Our review discloses no cases which shift the burden of proof on the issue of causation to the defendant.
In New Hampshire, the plaintiff in a medical malpractice action must prove that the defendant’s negligence caused the patient’s injury or loss. See Carrigan v. Sacred Heart Hospital, 104 N.H. 73, 80, 178 A.2d 502, 506-07 (1962). “[Negligent conduct is a proximate or legal cause of harm, if the actor’s conduct is a ‘substantial factor in bringing about the harm.’” Maxfield v. Maxfield, 102 N.H. 101, 105, 151 A.2d 226, 230 (1959) (quoting Restatement OF Torts § 431(a) (1934)). This issue is normally one for the trier of fact once a prima facie case is established. See id.; W. Keeton, D. Dobbs, R. Keeton & D. Owen, Prosser and Keeton on the Law OF Torts § 41, at 269 (5th ed. 1984) (hereinafter cited as PROSSER & Keeton).
The plaintiff in this case asserts that causation can only be established by showing that the deceased would have returned to the hospital for further treatment that Sunday morning had the doctor given him proper precautionary instructions, and that further treatment (i.e., a CAT scan and surgery) would have saved his life. The plaintiff reasons that such proof is difficult or impossible to adduce because the patient died, and thus the trial court should have shifted the burden of proof of causation to the defendants.
The plaintiff’s position, however, is contrary to the fundamental tenet of tort law that the plaintiff retains the ultimate burden of persuasion in negligence actions. See Jutras v. J. Scanlon Co., 106 N.H. 130, 133, 206 A.2d 615, 617 (1965); Spilene v. Company, 79 N.H. 326, 328, 108 A. 808, 809 (1920); see also RESTATEMENT (Second) of Torts § 433B, comment a (1965); Prosser & Keeton supra. This rule ensures that defendants will not have to defend against improbable claims, and that defendants will not have to disprove the element of causation. Proving a negative poses practical and theoretical problems that should not be placed on defendants. Moreover, this rule ensures that juries will not have to engage in speculation and conjecture in determining whether liability exists. We hold, therefore, that the trial court properly refused to shift the burden of proof on the issue of causation to the defendants.
Additionally, we conclude that the Hicks rule — which allows *305the trial court to relax the causation requirement and submit the issues to the jury if the defendant’s negligence increased the risk of harm — is ill-advised. Causation is a matter of probability, not possibility. See, e.g., Gooding, 445 So. 2d at 1020; Prosser & Keeton, supra § 41. We endorse the position taken by the Florida Supreme Court, which stated in a similar context:
“Relaxing the causation requirement might correct a perceived unfairness to some plaintiffs who could prove the possibility that the medical malpractice caused an injury but could not prove the probability of causation, but at the same time could create an injustice. Health care providers could find themselves defending cases simply because a patient fails to improve or where serious disease processes are not arrested because another course of action could possibly bring a better result.”
Gooding, 445 So. 2d at 1019-20.
Finally, we note that the trial court permitted the jury in this case to determine the negligence and causation issues, based on the plaintiff’s theory that the deceased may have had a “chance for better results.” Carrigan, 104 N.H. at 80, 178 A.2d at 507. The plaintiff has no cause to complain. Accordingly, we affirm.
Affirmed.
All concurred.