Waisel and Truog [1] offer some good advice regarding futility policies for do-not-resuscitate orders and correctly identify several problems in defining "futility." Nonetheless, we believe that their defense of physiologic futility is problematic.

The reasons usually given to support a physiologic conception of futility are that it is somehow bad for physicians to make value judgments and that physiologic futility is value free, unlike other construals of futility. This line of reasoning is incorrect. One cannot practice medicine without making value judgments. The adoption of a physiologic criterion for futility entails a value judgment that many physicians would find flawed and indeed offensive, that is, that physicians are properly concerned with whether resuscitation restores circulatory and respiratory functions but not with whether the patient improves as a result.

Waisel and Truog also set up a false comparison when they criticize the "last hundred cases" type of policy. They note correctly that, statistically, this amounts to accepting a 95% confidence level that therapy would be successful no more than 3 of 100 times. They therefore implicitly suggest that far too much uncertainty exists in the "last hundred cases" criterion but none exists in their physiologic criterion. This suggestion is incorrect because one can never be 100% certain that a treatment will not have a certain physiologic effect; if one were to calculate the statistics, the prediction of physiologic futility might have at least the same level of uncertainty as any other policy proposal.

We believe that although we have not yet arrived at a perfect policy formulation for understanding medical futility, a physiologic conception is too narrow.