Skip to main content

Table 4 Power to detect a difference between RT and nRT groups in Randomisation B

From: Protocol, rationale and design of SELPHI: a randomised controlled trial assessing whether offering free HIV self-testing kits via the internet increases the rate of HIV diagnosis

HIV Incidence (per 100 PY)

Median time to diagnosis (years) in nRT group

Power (%)

Estimated HIV infections per group

Median number HIV diagnoses (nRT/RT)

Median hazard ratio (RT versus nRT)

1.5

1.0

56

53

24/40

1.73

1.5

72

53

21/40

1.91

2.0

82

53

19/40

2.08

2.5

87

53

18/40

2.21

2.0

1.0

75

71

30/53

1.70

1.5

83

71

28/52

1.94

2.0

91

71

26/52

2.08

2.5

93

71

24/51

2.20

2.5

1.0

83

89

38/66

1.73

1.5

93

89

34/65

1.92

2.0

96

89

32/65

2.09

2.5

98

89

30/64

2.25

3.0

1.0

89

107

46/78

1.74

1.5

96

107

41/77

1.92

2.0

98

107

38/77

2.09

2.5

99

107

36/77

2.23

  1. Power to detect a difference at 2α = 0.05 by log-rank test