Abstract
The Food and Drug Administration (FDA) recommended ≥30% decrease on patient-reported outcomes for pain be considered clinically significant in clinical trials for adults with irritable bowel syndrome. This percent change approach may not be appropriate for children. We compared three alternate approaches to determining clinically significant reductions in pain among children.
80 children with functional abdominal pain participated in a study of the efficacy of amitriptyline. Endpoints included patient-reported estimates of feeling better, and pain Visual Analog Scale (VAS). The minimum clinically important difference in pain report was calculated as (i) mean change in VAS score for children reporting being 'better'; (ii) percent changes in pain (≥30% and ≥50%) on the VAS; and (iii) statistically reliable changes on the VAS for 68% and 95% confidence intervals.
There was poor agreement between the three approaches. 43.6% of the children who met the FDA ≥30% criterion for clinically significant change did not achieve a reliable level of improvement (95% confidence interval).
Children's self-reported ratings of being better may not be statistically reliable. A combined approach in which children must report improvement as better and achieve a statistically significant change may be more appropriate for outcomes in clinical trials.
