Skip to main content
. 2020 Jul 6;21:613. doi: 10.1186/s13063-020-04545-2

Table 2.

Summary of outcome measures and characteristics of the first primary outcome (N = 124)

Characteristics n (%)
Number of primary outcome(s)
 1 99 (80)
 2 16 (13)
 3 6 (5)
 > 3 3 (3)
Types of data reported for the primary outcomea
 Continuous outcome 101 (81)
 Categorical outcome 17 (14)
 Time-to-event outcome 4 (3)
 Others 2 (1)
Specific metric for the primary outcome
 Value at a time-point 74 (60)
 Time-to-event 4 (3)
 Change from the period-baseline 31 (25)
 Within individual difference between values at the end of each period 5 (5)
 Within individual difference between changes from baseline 1 (1)
 Others 8 (6)
Method of aggregation for the primary outcome
 Number—count of participants 21 (17)
 Mean 50 (40)
 Median 4 (3)
 Least squares mean 43 (35)
 Number—proportion/percent 6 (5)
Time frame for the primary outcome includes all randomized periods
 Yes 97 (78)
 No 27 (22)
Measured values presented by period
 Yes 4 (3)
 No 120 (97)
Provides statistical analysis for primary outcome
 Yes 78 (63)
 No 46 (37)
Statistical methods for analyzing primary outcome (n = 78)
 ANCOVA 9 (11)
 ANOVA 14 (18)
 Cochran-Mantel-Haenszel 2 (3)
 Generalized estimating equations 2 (3)
 Generalized linear mixed model 3 (4)
 Linear mixed effects model 6 (8)
 Mixed model analysis 20 (26)
 Mixed effects ANOVA crossover model 3 (4)
 Non-inferiority/equivalence testb 2 (3)
 Prescott’s test 1 (1)
t test 3 (4)
 Two-sided signed-rank test 1 (1)
 Wilcoxon (Mann-Whitney) 5b (6)
 Wilcoxon signed-rank 1 (1)
 Other 6 (7)

aTime-to-event outcome reported as Continuous Data

bFor trials reporting a statistical method and a non-inferiority/equivalence hypothesis, the method was recorded

cTrial NCT01132118 reported two statistical analyses: “Wilcoxon (Mann-Whitney)” and “Regression, Linear” for the first primary outcome