ObjectiveTo assess the effects of prognostic estimates, perceived benefit of treatment, and practice style on decision-making in critical care.DesignRandomized assignment of subjects to either of two versions of a questionnaire designed to elicit treatment decisions for six intensive care unit cases based on actual patients. One version offered optimistic survival forecasts; the other, pessimistic forecasts.SubjectsA random sample of 120 clinicians obtained from the Canadian Critical Care Society was contacted by mail. One version of the questionnaire was randomly assigned and mailed to each. Thirty-four replies, 17 for each version (response rate, 28%), were received and analyzed.Measurements and Main ResultsA list of treatment/management options was developed for each case, in three categories: recommended, questionable, and unacceptable. Subjects were also able to list new options that they would order that were not on the list. The dependent variables were the number of actions ordered in each category and the total for each case. Perceived benefit was measured by comparing subjective estimates of the probability of survival with the optimistic/pessimistic forecast given in the case. Practice style was assessed by correlating the total number of actions ordered across all possible pairs of cases. There were no significant differences between the two questionnaires on actions ordered either by category or by amount per category. Perceived benefit did not appear to be an important factor in decision-making. However, statistically significant correlations provide evidence for practice style in intensive care unit decision-making on an interventionist/noninterventionist dimension.ConclusionsThere is no evidence that erroneous or biased prognostic estimates affect intensive care unit treatment choices. Neither the principle of maximizing expected utility nor the Rule of Rescue appear to affect these decisions systematically, but practice style does. (Crit Care Med 1999; 27:58-65)