Statistical evidence: Admissible or not?

P ruitt v Lewis, Super Ct No. 08C-03-033 (Del 3/10/2011) , is interesting because it addresses the question of whether statistical evidence can be admitted to prove that a defendant did or did not breach a standard of care. The defendant offered the statistical evidence to show that there was a lack of causation and the probability that he did not breach the standard of care, because the result of his ministrations to the patient fell within expected statistical parameters.

Citing Timblin v Kent General Hospital, 640 A2d 1021 (Del 1994) , the court related the following facts. The plaintiff entered the hospital complaining of chest pains. He was diagnosed as suffering from a heart attack and was admitted to the coronary care unit. He was treated via blah, blah, blah, but ultimately suffered cardiac arrest, was deprived of oxygen for about 25 minutes, was resuscitated but suffered residual brain damage, and ultimately expired. At trial, the hospital proffered expert testimony that (1) when patients suffer cardiopulmonary arrest in a hospital, less than half will survive; (2) a review of 2 peer-reviewed articles on the subject showed that treatments of in-hospital arrests have successful outcomes of only 14% and 16%, respectively (only 1 of 7 survive); (3) at least 80% of all patients experiencing cardiopulmonary arrests will have some residual brain damage; and (4) less than 25% of patients who experience cessation of cardiac electrical activity (systole) can be resuscitated (the expert’s personal experience and opinion was that it was closer to 5%). After this testimony, the jury rendered a verdict in favor of the defendant. The case was appealed, and the Delaware Supreme Court held that the admission of statistical evidence was erroneous, because its probative value was substantially outweighed by its prejudicial effect. In other words, the court said that the statistical probability of death or brain damage after a cardiac arrest cannot be used to show that a doctor did or did not act in conformity with the applicable standard of care, merely because the ultimate results were in line with the statistical probabilities presented. The Supreme Court stated:

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 similar community. A defendant may not use evidence that a patient’s treatment ended with an expected result to infer the patient received proper care. If it were a fact that 80% of cardiac arrest victims die or suffer brain damage that fact is not relevant to the issue of whether [the defendant] deviated from the applicable standard of care in trying to resuscitate [the plaintiff].

The court noted that “statistical probability evidence creates a significant risk of jury confusion and unfair prejudice, because it might cause the jury to decide the case based on what happens normally instead of what happened in the [actual] case.” In essence, the court was trying to deal with the inappropriate association that, because a majority of patients who suffer cardiac arrest either die or suffer some neurological impairment, the plaintiff who happens to suffer a cardiac arrest is therefore expected to suffer death or brain damage. Rewording its message, the court stated: “Because such an inference is not based on the facts at hand, but rather an impermissible speculation based on inapplicable statistics, the potential for misleading the jury was quite substantial.”

Coming back to Pruitt , the court fashioned its own version of the established precedent by holding that, just as it is improper to use an unusual outcome to create an inference that the standard of care was breached, it is also improper to infer that a patient received proper care merely because the treatment outcome was what is normally seen in two thirds of all patients treated in that manner. For common behavior (results within statistical norms) to be admissible, the court noted that a special nexus must exist between that common behavior (statistically expected outcome) and the facts as established in the case at hand. Finding none, the appellate court reversed and ordered a new trial.


We are running at breakneck speed to fill our educational coffers with evidence, statistical evidence, of the efficacy of 1 modality of treatment over another or that some widget or gizmo provides superior results over a different widget or gizmo. Assume for the moment that all of our research studies are methodologically sound. Assume that the patients and researchers have been appropriately blinded. Also assume that the number of subjects studied will provide statistically valid data and that the analyses themselves have been appropriately chosen and carried out. We now have evidence-based information on how to treat our patients, right? Let’s for the moment say yes, right, okay. Can this “valid evidence” be admitted to prove that a practitioner (1) did or did not possess the requisite skill, knowledge, experience, and expertise (SKEE); or (2) applied that SKEE in a manner consistent with the applicable community standards; or (3) exercised his or her best judgment in the treatment of this patient (our hypothetical plaintiff)? Of course not. No evidence shows whether a doctor deviated from or adhered to an accepted standard of care on any particular occasion, and that, my friends, is what it’s all about.

Courts don’t care how good or how bad a doctor’s reputation is in the professional community. All they do care about is whether, on this particular occasion, this doctor did or did not conform to a standard of care, and, if he did not, whether the breach of that professional duty, as defined through expert testimony, was the direct or proximate cause of the injury claimed by the plaintiff. Disraeli said: “[t]here are three kinds of lies—lies, damned lies, and statistics.”

Developing all of this evidence-based data will do us no good if we rely merely on the numbers and do what our statistics tell us to do. We are still treating patients with individual malocclusions who respond individually to the ministrations they receive. All the statistics in the world will not help me treat this particular patient better or worse because I adhered to the statistical evidence. The standard of care requires that I do 3 things: (1) possess the requisite degree of SKEE as is possessed by the average practitioner of good standing in my community, (2) exercise that SKEE in a reasonable manner as would any member of my specialty under the same or similar circumstances, and (3) exercise reasonable judgment in the care of this particular patient given the totality of the circumstances.

Statistics won’t help any of us with any of that.

Only gold members can continue reading. Log In or Register to continue

Apr 8, 2017 | Posted by in Orthodontics | Comments Off on Statistical evidence: Admissible or not?
Premium Wordpress Themes by UFO Themes