TABLE 3.

Incidence of the primary outcome in the ginger and placebo arms

	ITT equal case scenario	ITT worst‐case scenario	PPA
Placebo event rate placebo	65/75 0.87 [0.79 to 0.94]	61/75 0.81 [0.73 to 0.90]	61/71 0.86 [0.78 to 0.94]
Ginger event rate ginger	50/75 0.67 [0.56 to 0.77]	50/75 0.67 [0.56 to 0.77]	45/70 0.64 [0.53 to 0.76]
Absolute risk reduction (ginger‐placebo)	−0.20 a [−0.33 to −0.07] P = 0.003 (Wald)	−15 a [−0.29 to −0.006] P = 0.038 (Wald)	−0.22 a [−0.35 to −0.08] P = 0.002 (Wald)
Number needed to treat	5 b [3 to 15]	7 b [4 to 167]	5 b [3 to 14]
Number of children	150	150	141

Values are proportions and 95% confidence interval (in square brackets) from binomial regression. The ITT “equal case scenario” analysis assumes the occurrence of the best outcome (no vomiting) in the 5 children lost to follow‐up in the ginger arm and of the same outcome (no vomiting) in the 4 children lost to follow‐up in the placebo arm. The ITT “worst case scenario” analysis assumes the occurrence of the worst outcome (vomiting) in the 5 children lost to follow‐up in the ginger arm and of the best outcome (no vomiting) in the 4 children lost to follow‐up in the placebo arm.

Abbreviations: ITT, intention‐to treat analysis; PPA, per‐protocol analysis.

^a95 confidence interval calculated from binomial regression.

^b95 confidence interval calculated from Bender's formula.