Objective: To determine whether randomized, controlled trials (RCTS) of potentially life-sustaining therapies in critically ill infants and children cause an ethical conflict for physician investigators and if ethical conflicts affect protocol implementation.
Design: Descriptive survey.
Subjects: A convenience sample of 1,050 physicians from a national pediatric critical care meeting mailing list.
Interventions: None.
Measurements and main results: The survey return rate was 41% (n = 415). Of the returned surveys, 81% (n = 331) were answered by pediatric intensivists and fellows. The remaining 19% (n = 84) were completed by other physician subspecialists. Overall, 74% had experience with RCTs involving a potentially life-saving therapy (25% had experience with three or more trials, and 26% had never participated in this type of study). The vast majority of the respondents (96%) indicated that they believe RCTs of potentially life-sustaining therapies are ethical; however, only 10% stated that they never experienced an ethical conflict with these types of studies. Most respondents (84%) indicated that published data from uncontrolled trials may bias them toward an investigational therapy. Furthermore, only 35% of the respondents indicated that they always maintain strict protocol adherence when the condition of a control patient deteriorates and parents request the experimental treatment. There was a significant association between physicians who experienced an ethical conflict and the likelihood that they would do the following if the condition of a control patient deteriorated: fail to maintain strict protocol adherence (p = .05); alter the protocol in response to parental requests for the experimental treatment (p < .01); or seek compassionate use of the experimental treatment (p < .01).
Conclusions: Although physicians consider RCTs of potentially life-sustaining therapies ethical, they acknowledge that this type of study sometimes creates an ethical conflict. Published results of uncontrolled trials lead to investigator bias in randomized trials and preclude equipoise. Our results indicate that RCTs involving life-sustaining therapies may be biased, lack consistent protocol implementation, and raise concern that data from these studies are potentially flawed.