640 A.2d 1021
Del.
1994
VEASEY, Chief Justice:

In this case, we address the admissibility of statistical evidence presented by medical expert witnesses testifying for defendant below-appellee Kent General Hospital (“Kent General”) concerning the percentage of patients who die or suffer brain damage following a cardiac arrest. Plaintiffs below-appellants Leroy Timblin (“Mr. Timblin”) and his wife (collectively, the “Timblins”) appeal from the Superior Court’s denial of their motion for a new trial and from the judgment in favor of Kent General. The Superior Court found no error justifying the granting of a new trial. The statistical evidence, however, had little probative value to the issues in dispute and was highly prejudicial to the Timblins because it allowed the jury to base its verdict on irrelevant statistical probabilities. We therefore REVERSE the judgment of the Superior Court and REMAND the case for proceedings consistent with this opinion.

I. FACTS

On March 14, 1988, Mr. Timblin entered Kent General complaining of chest pain. He was diagnosed as suffering from a heart attack. Mr. Timblin was admitted to the coronary intensive care unit of Kent General. The on-call internist, Dr. Christopher Giles (“Dr. Giles”), recommended a course of treatment that included thrombolytic therapy. The prescribed therapy involves the administration of anti-clotting medication, in this case, a combination of Activase and Heparin.

During the hours following his admission, Mr. Timblin continued to experience an irregular heartbeat. Lidocaine was given to correct the problem. Mr. Timblin had a severe reaction to the medication, and experienced a grand mal seizure in which he became unconscious and stopped breathing. Valium was administered to control the seizure, and Mr. Timblin subsequently went into cardiac arrest.

Dr. Giles, along with Dr. Charles Allen (“Dr. Allen”), the emergency room physician, attempted unsuccessfully to establish an artificial airway for Mr. Timblin by inserting a tube down his throat (a procedure known as “intubation”). Eventually, a nurse anesthetist was able to insert the tube and establish an artificial airway. By that time, Mr. Timb-lin had been without oxygen for approximately 25 minutes. Following the intubation, Mr. Timblin was successfully resuscitated. Nevertheless, he suffered neurological damage due to lack of oxygen to his brain.

The Timblins filed suit in Superior Court alleging that the hospital staff was not adequately trained to intubate Mr. Timblin or to resuscitate him properly. The Timblins claimed that the staff of Kent General was required to be certified by the American Heart Association in Advanced Cardiac Life Support (“A.C.L.S.”). Dr. Allen was the only A.C.L.S.-certified member of the group that treated Mr. Timblin. The Timblins contended that the failure to ventilate Mr. Timblin properly was the proximate cause of his neurological deficit.

II. PROCEEDINGS IN SUPERIOR COURT

At trial the Timblins presented expert medical testimony from Dr. Calvin Fuhrm-ann (“Dr. Fuhrmann”). Dr. Fuhrmann testified that, in his opinion, the attending physicians failed to ventilate Mr. Timblin properly while he was in cardiac arrest. His opinion was based upon the medical records which showed the attending medical staffs repeated unsuccessful attempts to intubate Mr. Timblin, as well as the resulting brain damage suffered by Mr. Timblin. Dr. Fuhrmann testified that it was a departure from the appropriate standard of care not to ventilate the patient in some other manner when the intubation was unsuccessful.

During Kent General’s presentation of its defense, Dr. Giles testified as a fact witness. On direct examination, Dr. Giles, in response to a question about the risk of death in the hours following a cardiac arrest stated that “published studies cite a short-term hospital mortality without thrombolytic therapy within the range of ten to fifteen percent.” The Timblins’ counsel immediately objected, contending that the testimony concerned matters not properly disclosed in discovery. The trial judge overruled the objection, subject to a possible further ruling, and ordered defense counsel to provide the Timblins’ counsel with a copy of the study mentioned by Dr. Giles in his testimony.

Kent General also presented expert medical testimony from Dr. James Bouzoukis (“Dr. Bouzoukis”). Dr. Bouzoukis testified that the risks involved with the alternative ventilation procedures suggested by Dr. Fuhrmann rendered the procedures inappropriate in the situation presented by Mr. Timblin’s cardiac arrest. Kent General’s counsel inquired whether the fact that Mr. Timblin suffered a neurological deficit indicated that he was negligently resuscitated. Dr. Bouzoukis responded as follows:

[Ujnfortunately, the facts are that when a patient experiences a cardiopulmonary arrest, certainly less than half — and that’s being generous — will survive. In fact, I recently reviewed two articles reviewing the results of in-hospital arrests, and they had a success rate, a survival rate, if you will, of only fourteen and sixteen percent respectively. That’s only one out of seven survives.
This is an arrest that occurred within the hospital. And we know from the pre-hospital setting that at least eighty percent will have some kind of brain damage because of the incident.

Kent General’s counsel later asked Dr. Bouzoukis whether anything about Mr. Timblin’s condition suggested that the ventilation was properly performed. Dr. Bouzoukis responded:

The fact that he was resuscitated.... Statistically we know less than twenty-five percent of patients whose heart [sic] goes into asystole [cessation of electrical activity] can be resuscitated; I think it’s one out of twenty. I think it was remarkable that they were able to bring his heart back under those circumstances.

The Timblins’ counsel made no immediate objection to Dr. BouzouMs’ testimony.

Prior to the parties’ closing statements, the trial judge held a prayer conference with counsel at which the Timblins’ attorney raised the fact that Kent General had not supplied the previously ordered study. The trial judge stated at the conference that, if the Timblins could show prejudice from the admission of the statistical evidence, they could apply for a remedy. Following the conference, the parties made their closing statements. During his closing, Kent General’s counsel stressed the statistical evidence by repeatedly stating that people often die or suffer neurological damage after a cardiac arrest. The following day, the jury entered a verdict in favor of Kent General.

Two days after the jury verdict, Kent General’s counsel provided the Timblins’ counsel with a study in support of Dr. Giles’ statistical references. The Timblins moved for a new trial, contending that the study, which they assert should have been provided during discovery, did not support the testimony of either Dr. Giles or Dr. BouzouMs. The Timblins attached an affidavit from Dr. Enrico Veltri (“Dr. Veltri”) in support of their contention. The trial court denied the motion for a new trial because it found that the complained-of discovery defect did not warrant a new trial.

On the same day that the trial court denied the motion for a new trial, the Timblins submitted a response to Kent General’s answer opposing the Timblins’ motion. The response included an affidavit from Dr. Henry Halperin (“Dr. Halperin”) supporting Mr. Timblin’s claim that the testimony of Dr. Giles and Dr. BouzouMs concerning the probability of death or brain damage following a cardiac arrest was improper. The trial court treated the response as a motion for reargument, which it denied for the same reason it had articulated in its denial of the motion for a new trial. The Timblins then filed a timely appeal of the judgment of the Superior Court, based on the verdict of the jury in favor of Kent General, and the denial of their motion for a new trial.

III. THE ADMISSION OF THE STATISTICAL EVIDENCE

This Court must determine whether the Superior Court abused its discretion by permitting the introduction of the challenged evidence. See Pope v. State, Del.Supr., 632 A.2d 73, 78-79 (1993); Firestone Tire and Rubber Co. v. Adams, Del.Supr., 541 A.2d 567, 570 (1988). The Timblins present several arguments to show that the introduction of statistical evidence by Kent General was improper. ' First, they claim that the statistical evidence was irrelevant to Kent General’s treatment of Mr. Timblin and was also highly prejudicial. Second, the Timblins claim that the statistical evidence was unsupported by the authorities provided to them by Kent General after trial. Finally, they contend that no proper basis for the statistical evidence was disclosed prior to the testimony of Kent General’s medical experts. Since we find that the admission of the statistical evidence was erroneous because its probative value was substantially outweighed by its prejudicial effect, we need not address the other points raised by the Timblins on appeal.

The propriety of admitting the challenged statistical evidence must be determined under the balancing test of Delaware Rule of Evidence (“D.R.E.”) 403. Rule 403 provides:

Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues or misleading the jury, or by considerations of undue delay, waste of time or needless presentation of cumulative evidence.

Rule 403 therefore requires a court to assess the probative value of the proffered evidence and to weigh that value against the negative consequences of admitting the evidence, including the risk of unfair prejudice and jury confusion. An analysis of the statistical probability evidence introduced by Kent General reveals that such evidence had minimal probative value and created a substantial risk of confusion and prejudice.

In a medical malpractice action, the plaintiff must present expert medical testimony regarding the defendant’s alleged deviation from the applicable standard of care and the cause of the alleged personal injury. 18 Del.C. § 6853; Burkhart v. Davies, Del.Supr., 602 A.2d 56, 59 (1991), cert. denied, — U.S. -, 112 S.Ct. 1946, 118 L.Ed.2d 551 (1992). Expert medical testimony may also be introduced by a defendant to show that the applicable standard of care was met or that any departure therefrom did not cause the plaintiffs injury. The statistical evidence presented by Dr. Bouzoukis, however, is not probative on either of these issues.

The statistical probability of death or brain damage following a cardiac arrest cannot be used to show that Kent General acted in conformity with the applicable standard of care. It is settled law that the standard of care in a medical malpractice action is established by evidence of the degree of care and competence ordinarily exercised by physicians in the same or a similar community. E.g., Riggins v. Mauriello, Del.Supr., 603 A.2d 827, 830 (1992); Register v. Wilmington Medical Ctr.,Inc., Del.Supr., 377 A.2d 8, 9-10 (1977). It is also well established that a plaintiff cannot use evidence that a medical procedure had an unusual outcome to create an inference that the proper standard of care was not exercised. See, e.g., Thomas v. St. Francis Hosp., Inc., Del.Supr., 447 A.2d 435, 438 (1982) (holding that an inference of negligent medical treatment is not warranted simply because the treatment of a patient ended with highly unusual results); DiFilippo v. Preston, Del.Supr., 173 A.2d 333, 338 (1961) (recognizing that “[t]he doctrine of res ipsa loquitur ... is not applicable in malpractice actions in which the only proof is the fact that the treatment of the patient terminated with poor results”). It therefore follows that a defendant may not use evidence that a patient’s treatment ended with an expected result to infer that the patient received proper care. If it is a fact that 80 percent of cardiac arrest victims die or suffer brain damage, that fact is not relevant to the issue of whether Kent General deviated from the applicable standard of care in trying to resuscitate Mr. Timblin. 1 That was the only issue properly before the trial court.

A plaintiff in a medical malpractice action “must prove that the doctor’s inadequate treatment was the proximate cause of the patient’s injury or death.” Shively v. Klein, Del.Supr., 551 A.2d 41, 44 (1988). Proximate cause is defined as “that direct cause without which the [injury] would not have occurred.” Chudnofsky v. Edwards, Del.Supr., 208 A.2d 516, 518 (1965); Culver v. Bennett, Del.Supr., 588 A.2d 1094, 1097 (1991). In order to prove proximate cause, the plaintiff must show that the physician’s negligence was the probable cause of the injury, i.e., the likelihood was greater than 50 percent that the negligence caused the injury. Shively, 551 A.2d at 43. Statistical evidence may be relevant in appropriate circumstances to rebut the nexus between the defendant’s conduct and the plaintiffs injury by showing that the injury was likely to occur regardless of the defendant’s alleged negligence. Even in such circumstances (which are not present here), the evidence must be carefully scrutinized to ensure its relevance and to guard against unfair prejudice. Celotex Corp. v. Wilson, Del.Supr., 607 A.2d 1223, 1229 (1992).

In his closing argument, Kent General’s counsel emphasized to the jury that individuals who experience a cardiac arrest frequently die, and most suffer a neurological injury. A close analysis of the facts of this case refutes any suggestion that the statistical evidence presented at trial by Kent General is relevant to causation by demonstrating that Mr. Timblin’s brain damage was a natural consequence of his cardiac arrest and not any alleged negligence.

There is no dispute that the damage to Mr. Timblin’s brain was caused by the approximate 25-minute period during which it was deprived of oxygen. It is also uncontested that Mr. Timblin was deprived of oxygen because an airway could not be established during this time period (i.e., he could not be intubated). The only issue of causation, therefore, is whether Kent General’s alleged negligence was responsible for the delay in establishing an airway. If Kent General could not have ventilated Mr. Timblin despite complying with the applicable standard of care, then his brain damage was not caused by Kent General’s conduct. See W. Keeton, Prosser and Keeton on the Law of Torts 266 (6th ed. 1984) (observing that a “defendant’s conduct is not a cause of [an] event, if the event would have occurred without it”); Cul-ver, 588 A.2d at 1097 (quoting Prosser). Conversely, if Mr. Timblin could have been intubated successfully, but was not because of a deviation from the appropriate standard of care, then the Timblins could meet their burden of proving causation, While we express no opinion on whether the applicable standard of care was satisfied by Kent General or whether any deviation therefrom caused Mr. Timblin’s injury, it is evident that the causation issue in this case is whether the inability to create an airway was caused by Kent General’s alleged negligence, not whether his brain damage was an inevitable result of his cardiac arrest. Accordingly, the statistical evidence had little, if any, relevance to the issue of causation in this case.

In contrast to the minimal (at best) probative value of the statistical evidence, its potential for misleading the jury was quite substantial. Delaware courts have recognized that evidence of statistical probability creates a significant risk of jury confusion and unfair prejudice because such evidence may lead a jury to decide a case based on what happens normally instead of what happened in the ease before it. In Lee v. A.C. & S. Co., Del.Super., 542 A.2d 352 (1987), the Superior Court reviewed the use of statistical evidence to prove that the decedent’s exposure to asbestos was the proximate cause of his lung cancer and death. The plaintiff introduced the testimony of an epidemiologist concerning the probability of contracting lung cancer from asbestos exposure. The court ruled that while the epidemiologist, who was not a physician, was qualified to testify regarding statistical data on asbestos and illness, he was not qualified to offer an expert medical opinion on the cause of the decedent’s fatal illness. In so holding, Judge Taylor expressed his concern that the statistical evidence would mislead the jury and be used for an improper purpose in arriving at a verdict:

Thus, admitting the epidemiologist’s opinion of causation as to the decedent would tend to cause the jury to draw a layman’s conclusion about the cause of decedent’s disease and death by treating statistical prevalence of a disease as though it was the same as evidence of the cause of this decedent’s disease and death — a subject which must rest on medical opinion.

Id. at 355. The Superior Court therefore excluded the testimony based on D.R.E. 403. Accord Money v. Manville Corp. Asbestos Comp. Fund, Del.Supr., 596 A.2d 1372, 1376-77 (1991).

This Court has likewise recognized the prejudicial impact of statistical probability evidence. In Wheat v. State, Del.Supr., 527 A.2d 269 (1987), this Court held that the use of statistical evidence to prove the credibility of a witness was improper. In Wheat, the State presented an expert witness who provided statistical evidence about the incidence of recantation and subsequent reversals of recantations by child sexual abuse victims. The Court held that such statistical evidence invaded the province of the trier of fact by evaluating the credibility of a witness in “lie detector fashion.” Wheat, 527 A.2d at 275. In so holding, the Court noted:

[E]xpert testimony ... is not, as some current practice suggests, a mechanism for having someone of elevated education or station engage in a laying on of hands, placing an imprimatur, upon the justice of one’s cause_ Experts are not, in theory, called to tell the jury who should win. They are called, instead, to provide knowledge to the jury to permit the jury rationally to decide the case before it.

Wheat, 527 A.2d at 275 n. 5 (quoting State v. Moran, 151 Ariz. 378, 385 n. 7, 728 P.2d 248, 255 n. 7 (1986)). Thus, the Court’s primary concern with the statistical evidence in Wheat was that it would be unduly prejudicial if used by the jury to evaluate the wit ness’s credibility in terms of statistical probabilities. See also Powell v. State, Del.Supr., 527 A.2d 276 (1987) (companion case holding that admitting expert’s testimony that alleged victims tell the truth in 99 percent of child sex abuse cases amounted to plain error).

The testimony of Dr. Bouzoukis regarding the probability of death or brain damage following a cardiac arrest created a significant danger of jury confusion and unfair prejudice. The jury was permitted to consider this evidence to assess whether Kent General’s alleged negligence caused Mr. Timb-lin’s injuries. The record reflects that the jury was not instructed concerning any other purpose for which the testimony was presented. As noted, Kent General’s counsel repeatedly mentioned the statistics during his closing statement, thereby stressing their significance. “[Gjiven the strong emphasis which defense counsel placed on this [evidence], we may assume that the effect on the jury was not insubstantial.” Coe v. Schneider, De l.Supr., 424 A.2d 1, 2-3 (1980).

While a jury may draw inferences from the facts of a case, those inferences may not be based upon speculation. McGuire v. McCollum, Del .Super., 116 A.2d 897, 900 (1955). Here, the statistical evidence provided the jury with an improper basis for speculation about the cause of Mr. Timblin’s ultimate condition. The statistics invited an inference that, because the majority of patients who experience a cardiac arrest die or suffer brain damage, Mr. Timblin was expected to suffer brain damage. Such an inference is not based upon the facts of the case at hand, but rather on impermissible speculation based on inapplicable statistics. See Lee, 542 A.2d at 355-356.

This Court has held that unless a “special nexus” is shown between the evidence of common behavior and the facts of the case, the use of such common behavior evidence is highly prejudicial. Wheat, 527 A.2d at 274. In this case, the statistical evidence of what “normally happens” was irrelevant to the jury’s determination of liability. The effect of the evidence, therefore, was unfairly prejudicial to Mr. Timblin. The evidence allowed the jury to speculate as to the cause of Mr. Timblin’s injuries based on statistics which had little probative value to the issues of the case. Therefore, this evidence should have been excluded under D.R.E. 403. 2

Accordingly, the judgment of the Superior Court is REVERSED and REMANDED for proceedings consistent with this opinion.

Notes

1

. Kent General argues that Dr. Bouzoukis’ testimony was in response to the testimony of Dr. Fuhrmann, the Timblins’ expert witness. According to Kent General, Dr. Fuhrmann testified that Mr. Timblin's neurological injury indicates a negligent resuscitation and that Dr. Bouzoukis simply rebutted this assertion. Even assuming arguendo that Dr. Fuhrmann’s testimony was improper, Kent General's remedy was to object to such testimony, not to introduce equally improper testimony.

2

. Moreover, the statistical evidence was improper expert testimony because it did not provide any factual basis for the jury to decide the issue of causation, and did not add to the jury’s understanding of the facts or issues. See D.R.E. 702.