Taragin and coworkers [1] investigated the medical malpractice tort system through use of closed claims data from the Medical Inter-Insurance Exchange of New Jersey. They presented the conclusion that physician care in malpractice cases is usually defensible and suggested that unjustified payments are probably uncommon. Their findings are generally consistent with the American Society of Anesthesiologists' study of 1004 closed anesthesia malpractice claims [2]. Their study, however, differed from ours and others [3] in one important aspect: In their study the standard of care assessment was not associated with severity of injury. In contrast, we found that in cases of severe injury, care was more commonly assessed as substandard, whereas in cases of nondisabling injuries, care was more commonly assessed as meeting standards [2]. We subsequently found in an experimental study that knowledge of severity of injury influenced anesthesiologist reviewers' judgments of the appropriateness of care [4].

One possible explanation is the difference in sampling: Taragin and colleagues studied claims from a single state; we studied a national sample with claims from 17 insurance companies (including the source of Taragin and colleagues' claims). Taragin and colleagues studied all medical malpractice claims; we studied only anesthesia-related claims. A second explanation concerns how "defensibility" was defined and assessed. The two studies did not use identical definitions of "defensibility" [1] or "appropriateness of care" [2]. Although peer reviewers in both studies had access to complete files, only in Taragin and colleagues' study did the reviewer have access to the defendant and defense attorney. Another explanation is that reviewer bias may be more common among anesthesiologists than other specialists because of the nature of anesthesia practice. The goal of anesthesia is usually to enable or facilitate therapy, not to provide a treatment or cure in itself. The avoidance of unnecessary side effects and poor outcomes is a fundamental and highly valued objective [4]. If outcome bias in judgments about standard of care is specialty specific, then specialty-specific data are required.