Skip to main content

Table 1 Marginal (individual) power obtained for each outcome, when analysing two (top), three (middle) or four (bottom) continuous outcomes using a variety of methods to control the FWER

From: Methods to adjust for multiple comparisons in the analysis and sample size calculation of randomised controlled trials with multiple primary outcomes

Pairwise correlation between outcomes

None

Bonferroni

Holm

Hochberg

Hommel

Stepdown minP

Two outcomes

 0

80.9

72.4

78.5

79.2

79.2

78.2

 0.2

80.6

71.8

77.8

78.6

78.6

77.7

 0.4

80.0

71.3

76.6

77.7

77.7

76.7

 0.6

80.0

71.0

76.0

77.4

77.4

76.7

 0.8

80.3

71.3

75.6

77.4

77.4

77.2

Three outcomes

 0

80.2

65.9

75.2

76.7

76.8

75.5

 0.2

80.5

66.4

75.0

76.6

76.7

75.3

 0.4

80.2

65.7

73.8

75.4

75.6

73.2

 0.6

80.0

65.7

73.3

75.0

75.2

73.8

 0.8

80.0

65.9

72.2

74.6

74.8

76.1

Four outcomes

 0

80.5

62.3

73.2

75.0

75.2

72.7

 0.2

80.4

62.3

72.6

74.4

74.8

72.2

 0.4

80.6

62.4

72.1

74.1

74.4

72.2

 0.6

80.3

62.0

70.7

73.1

73.5

72.3

 0.8

80.3

61.9

69.7

73.2

73.6

73.5

  1. D/AP method was not examined due to the poor performance observed when exploring FWER
  2. There was no missing data in any of the outcomes. The tables display various degrees of correlation between the outcomes, ranging from no correlation (ρ =0.0) to strong correlation (ρ =0.8)