Here is an interesting post from Torts Prof Blog:
"After all the time and money spent, aren’t the results of medical malpractice litigation extremely accurate? Although perhaps better than earlier studies indicated, malpractice results are imprecise. The Harvard School of Public Health study, in assessing how often medical error occurred in a random sample of closed claims, found that, "73 percent of all claims for which determinations of merit were made had outcomes concordant with their merit." Studdert et al., 354 New Eng. J. Med. at 2028. That figure would likely be reduced somewhat further if the technical requirements of medical malpractice were applied to the concept of medical error. In sum, after all these resources are expended, more than one in four cases is decided incorrectly.
Data is mixed regarding the distribution of the errors (whether errors generally favor plaintiffs or health care providers). The Harvard School of Public Health study finds a fairly even distribution. On the other hand, Professor Philip G. Peters, Jr. analyzed all seven studies that have compared verdicts rendered in individual malpractice cases with independent evaluations of each claim by medical or legal experts. Philip G. Peters, Jr., Doctors & Juries, 105 Mich. L. Rev. 1453 (2007). According to Peters, the studies demonstrate that plaintiffs win about 10% to 20% of cases with weak evidence of negligence and 50% of the cases with strong evidence thereof. In other words, plaintiffs win between 10% and 20% of cases they should likely lose, but lose one-half of cases they should likely win. Id. at 1464. If that’s true, plaintiffs’ lawyers, who tend to support the current malpractice regime, take a big risk with each individual client with a strong case that goes before a jury."