Sample size calculations for the primary trial outcomes
| Outcome | Value in standard dose (20 mg/day) group | Expected value in low-dose groups | Non-inferiority margin | Power, significance level | Sample size per arm (with ~5% attrition) |
| Duration of diarrhoea, mean (SD) | 3.0 (2.7) days | 3.0 (2.7) days Lower dose equally effective | 0.3 days | 90%, 5% | 1460 |
| Percentage of children with diarrhoea >5 days | 16% | 16% Lower dose equally effective | +4% | 90%, 5% | 1500 |
| Total number of loose or watery stools during follow-up, mean (SD) | 10 (9) stools | 10 (9) stools Lower dose equally effective | +2 stools | 90%, 5% | 1460 |
| Percentage of children with vomiting during follow-up | 20% | 15% Lower dose safer | Not applicable because of superiority hypothesis for this outcome | 90%, 5% | 1340 |