Schneiderman and colleagues' study [1] of the effect of completing an advance directive on health care represents long overdue attention to a widely advocated but not adequately studied "therapeutic" intervention.

An advance directive, as generally legislated, takes effect only after a patient loses decision-making capacity. According to the investigators, however, most patients retained their decision-making capacity. Only three patients who lost decision-making capacity before death had their directive "activated". A sample of this size lacks the power to detect a real effect. Assessing treatment decisions and associated costs before the loss of decision-making capacity among patients with advance directives does not technically measure the effect of the directive. Rather, it may be a reflection of more general discussions patients have with their physicians. The authors' conclusions about the lack of effect of advance directives may not be warranted.

A second concern relates to the potential confounding by contact between investigators and patients in the control arm. Minimal contact with the investigators, even in the absence of a specific discussion of advance directives, may significantly influence patients' decisions. This would diminish the apparent effect of the advance directive.