Randomized clinical trials in pediatric critical care: Rarely done but desperately needed

OBJECTIVE: To review the benefits and challenges of using the randomized, controlled trial (RCT) study design to evaluate preventive and therapeutic interventions in pediatric critical care medicine. CONCLUSIONS: The RCT design is able to control for many sources of potential bias that other types of study designs cannot. The findings of RCTs often contradict the findings of less rigorous study designs. Before performing an RCT, there must exist a state of clinical equipoise, a sufficient number of eligible patients must be available, and the epidemiology of the disorder in question must be well studied. There are many challenges to performing high-quality RCTs. Studying multiple element support strategies in the critically ill patient population is more complex than studying a single drug therapy. High patient and practice variability and hazy diagnostic definitions can dilute the signal-to-noise ratio. Most interventions in critical care are expected to have a modest or small effect. This markedly increases the requisite sample size. There is a paucity of accepted clinically important measurements of the outcome of critical care, making mortality a common outcome to evaluate with a not-so-common incidence. Developmental issues, the inability to give informed consent, and the failure to perform the appropriate pharmacokinetic and safety studies are additional challenges facing pediatric investigators. Despite these limitations, a good RCT remains the best way to prove that an intervention is working or not. Indeed, RCTs are and will remain the "gold standard" method to estimate the efficacy of a therapeutic or prophylactic intervention.