Clinical-statistical gap in evaluating outcome of stool patterns in young children with autistic spectrum disorder.

摘要

The prospective cohort study by Sandhu et al provided useful information about early stool patterns in young children with autistic spectrum disorder (ASD),with statistically higher prevalence of diarrhoea after 30 months of age as compared with healthy controls (odds ratio (OR) 1.68; 95% CI 1.03 to 2,74) (p=0.039). This statistical finding, however, raises two important issues. First, adjusted OR by logistic regression approximates the adjusted relative risk (RR) when disease outcome is rare (<10%) and overestimates the risk (built-in-bias) when outcome events are common (>10-20%). Indeed, RR is a more efficient effect size measure in prospective cohort studies for analysing common binary outcomes. Second, the clinical relevance of this finding is not clear from the individual and population perspectives.