Skip to main content

Table 6 Sample sizes required to detect minimum clinically important differences

From: Performance of the SarQoL quality of life tool in a UK population of older people with probable sarcopenia and implications for use in clinical trials: findings from the SarcNet registry

  

MCID

SD

Sample size

80% power, unadjusted

80% power, adjusted

90% power, unadjusted

90% power, adjusted

SarQoL– change in fitness

Slight improvement

21

19

26

24

36

34

Any improvement

17

17

32

30

44

42

Slight worsening

1

16

8038

7476

10,760

10,008

SarQoL – change in QoL

Slight improvement

5

5

32

30

44

42

Any improvement

9

16

100

94

134

126

Slight worsening

5

12

182

170

244

228

SARC-F – change in fitness

Slight improvement

1.0

2.1

140

86

186

114

Any improvement

1.4

1.8

52

32

70

44

Slight worsening

0.3

2.0

1396

852

1868

1140

SARC-F – change in QoL

Slight improvement

2.3

1.5

14

10

18

12

Any improvement

0.8

2.5

308

188

412

252

Slight worsening

0.3

2.2

1690

1032

2262

1380

  1. MCID Minimum clinically important difference. Sample size is total sample size for a two arm trial. Calculations assume 1:1 randomisation, alpha 0.05. Adjusted: Sample size multiplied by (1 – r2) where r = 0.27 for correlation between baseline and follow-up SarQoL scores, and r = 0.63 for correlation between baseline and follow-up SARC-F scores. Where calculations deliver odd numbers, sample size is rounded up to nearest even number
Back to article page

BMC Geriatrics

ISSN: 1471-2318

Contact us