Skip to main content

Table 1 Estimated number of clusters for three sample size calculation methods used in SWTs, as a function of the ICC and outcome type (continuous, binary and rate) to obtain 80 % power

From: Sample size calculation for a stepped wedge trial

ICC Standard CRT inflation factor DE inflationfactor based onWoertman et al. Analytical power based on HH
  K=20,J=1 K=120,J=1 K=20,J=6 K=20,J=6
Continuous outcomea
0 26 5 8 9
0.1 74 55 12 12
0.2 122 105 11 11
0.3 170 155 10 10
0.4 218 205 9 9
0.5 266 256 7 7
Binary outcomeb
0 25 5 8 10
0.1 71 53 12 13
0.2 117 101 11 12
0.3 163 149 9 11
0.4 209 197 8 10
0.5 256 246 7 8
Count outcomec
0 24 4 8 8
0.1 69 51 11 11
0.2 114 98 10 10
0.3 159 145 9 9
0.4 203 192 8 8
0.5 248 238 7  
  1. aIntervention effect = −0.3785; σ y =1.55; sample size for a parallel RCT = 253 subjects per arm.
  2. bBaseline outcome probability = 0.26; OR = 0.56; sample size for a parallel RCT = 243 subjects per arm.
  3. cBaseline outcome rate = 1.5; RR = 0.8; sample size for a parallel RCT = 236 subjects per arm.
  4. Notation: K= number of subjects per cluster; J= total number of time points, including one baseline