Skip to main content

Table 1 Statistical power (percentages) to detect a log-linear trend from various designs for 700 risk sets

From: Study design: Evaluating gene–environment interactions in the etiology of breast cancer – the WECARE study

    

Counter-matching

  

Random sampling

CM 1:2

CM 2:1

2 Gy rate ratio

Alla

NCC 3

NCC 4

PN = 100%, PP = 100%

PN = 90%, PP = 90%

PN = 100%, PP = 100%

1.4

79

61

63

70

67

63

1.5

87

69

78

84

80

78

1.6

97

86

93

95

92

92

1.7

100

95

97

99

96

97

  1. CM 1:2, counter-matched design in which one member of the triplet is registry radiation treatment negative (RRT-) and two are registry radiation treatment positive (RRT+); CM 2:1, counter-matched design in which two members of the triplet are RRT- and one is RRT+; NCC 3, nested case-control study in which each case is randomly matched to two controls; NCC 4, nested case-control study in which each case is randomly matched to three controls; PN, predictive negative value; PP, predictive positive value. aFull cohort power where radiation dose is available for all controls in the risk set and used in the analysis; this is the upper limit on any case-control analysis.