Skip to main content

Table 4 Effect of using a constant baseline attack rate on the precision (confidence interval) of the estimated protective efficacy of tafenoquine compared to assuming an attack rate based on results from hypothetical placebo controlled trials with different sample sizes

From: A retrospective analysis of the protective efficacy of tafenoquine and mefloquine as prophylactic anti-malarials in non-immune individuals during deployment to a malaria-endemic area

ARPl(cases/n)1

ARTq(cases/n)2

PETq

LL (PETq)3

1% (1/100)

0 (0/490)

100%

45% (44.87)

1% (10/1000)

0 (0/490)

100%

45% (44.91)

1% (constant)4

0 (0/490)

100%

44% (44.10)

2% (1/50)

0 (0/490)

100%

72% (72.43)

2% (2/100)

0 (0/490)

100%

72% (72.46)

2% (constant)4

0 (0/490)

100%

73% (72.54)

3% (3/100)

0 (0/490)

100%

82% (81.65)

3% (30/1000)

0 (0/490)

100%

82% (81.69)

3% (constant)4

0 (0/490)

100%

82% (81.70)

5% (1/20)

0 (0/490)

100%

89% (88.97)

5% (2/40)

0 (0/490)

100%

89% (88.98)

5% (5/100)

0 (0/490)

100%

89% (88.99)

5% (50/1000)

0 ( 0/490)

100%

89% (89.02)

5% (constant)4

0 (0/490)

100%

89% (89.02)

10% (1/10)

0 (0/490)

100%

94% (94.49)

10% (5/50)

0 (0/490)

100%

95% (94.50)

10% (10/100)

0 (0/490)

100%

95% (94.50)

10% (100/1000)

0 (0/490)

100%

95% (94.51)

10% (constant)4

0 (0/490)

100%

95% (94.51)

  1. 1Placebo attack rate based on hypothetical trial results (cases/n) or assumed to be a known constant value (no variability).
  2. 2Observed attack rate for tafenoquine in Study 033.
  3. 3Lower limit of the one-side 95% CI for the protective efficacy of tafenoquine [LL(PETq)]. Koopman's (score test) method was used when the placebo attack rate (ARPl) is based on hypothetical placebo results (cases/n). StatXact (vs.9.0) was used for calculations and the confidence level was set to 90% [for ARTq = 0 (0/n) this gives a lower limit of a one-sided 95% interval for PETq].
  4. 4When ARPl is assumed to be a known constant equation 11(bottom) was used to obtain LL(PETq).