Skip to main content

Table 4 Outline of the imputation and missingness models

From: Dealing with indeterminate outcomes in antimalarial drug efficacy trials: a comparison between complete case analysis, multiple imputation and inverse probability weighting

ModelResponsePredictors
Imputation Model\( Y=\Big\{{\displaystyle \begin{array}{l}1\kern1.25em if\ recrudescence\\ {}0\kern1.25em if\ new\ infection\end{array}}\operatorname{} \)• age (years)
• mg/kg dose of partner drug
• transmission intensity a
• treatment regimen
• time of recurrence
• parasitaemia (log)
• study sites b
• parasite density (log) on the day of recurrence
Missingness model\( {Y}_{obs}=\Big\{{\displaystyle \begin{array}{l}1\kern1.25em if\ outcome\ is\ observed\\ {}0\kern1em if\ outcome\ is\ indeterminate\end{array}}\operatorname{} \)• age (years)
• mg/kg dose of partner drug
• transmission intensity a
• treatment regimen
• time of recurrence
• recurrence status (yes/no) c
  1. a Transmission settings were categorised as low if Malaria Atlas Project estimate were less than or equal to 0.10, moderate if >0.10 and ≤ 0.40, and high if greater than 0.40
  2. b Study site was not added in the missingness model as it led to convergence issues
  3. c Excluded in the IPW-E approach