Objective: To explore what design would be reasonable, acceptable, and feasible for a pediatric trial investigating the effect of low tidal volume ventilation.
Design: A two-round modified Delphi approach among pediatric intensivists with a visible special interest in mechanical ventilation.
Settings: None.
Subjects: Pediatric intensivists.
Interventions: None.
Measurements and results: In the first questionnaire "experts" classified 64 items as "important," "not so important" or "not important" (June 2010). The second questionnaire included the 29 items having been classified by more than 50% of the experts as being "important" in the first questionnaire that needed to be ranked in order of importance (August 2010). Twenty-nine of 50 (64%) experts responded to the first questionnaire, and 28 of the 29 initial responders (96.6%) to the second questionnaire. Actual expert opinion favored the following: age of study population 0 yrs to 2 yrs, expected duration of ventilation ≥ 48 hrs, and stratification by the severity of hypoxemia (cutoff PaO(2)/FIO(2) < 200). The two study arms should compare the effect of 6 mL/kg vs. 10 mL/kg on mortality. However, these views of the experts face two major issues. First, 10 mL/kg does not represent standard of care. Second, numerous uncertainties render mortality unsuitable as primary measure of outcome including a large sample size (>1500).
Conclusions: Actual expert opinion favored investigating the effect of 6 mL/kg vs.10 mL/kg on mortality. Such a design suffers from various serious criticisms. Therefore, and in our opinion, a pediatric Acute Respiratory Distress Syndrome Network trial is not realistic.