Skip to main content

Table 3 Specification of the logistic regression model used to impose missingness

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

 

Log (Odds Ratio)†

Odds Ratio

Mechanism 2a

Mechanism 2b

Mechanism 2a

Mechanism 2b

Treatment

β2j

β2j

exp(β2j)

exp(β2j)

AL (reference)

0.00

0.00

1.00

1.00

ASAQ

-0.05

-0.10

0.95

0.90

DP

-0.20

-0.40

0.82

0.67

Transmission

β1k

β1k

exp(β1k)

exp(β1k)

High (reference)

0.00

0.00

1.00

1.00

Low

-0.25

-0.50

0.78

0.61

Moderate

-0.15

-0.30

0.86

0.74

  1. AL artemether-lumefantrine , ASAQ artesunate-amodiaquine , DP dihydroartemisinin-piperaquine
  2. †The intercept of the logistic regression (β0) was chosen by iteration to achieve the desired proportion of missingness conditional on recurrence status. This was -2.10, -0.75 and -0.11 for 10%, 30% and 45% respectively under missingness mechanism 2a, and -2.03, -0.68 and -0.02 respectively for 10%, 30% and 45% missingness under mechanism 2b.