Quality of reporting of basic sample size elements from the CONSORT 2010 statement and the Cluster 2012 extension to the CONSORT statement
| All studies N=60 | 1987–2012 N=28 | 2013–2014 N=32 | Absolute difference (95% CI) | p Value | |
|---|---|---|---|---|---|
| Sample size justification | |||||
| Reported | 45 (75.0) | 18 (64.3) | 27 (84.4) | 20.1 (−1.7 to 41.8) | 0.073 |
| Item 1 | |||||
| Level of significance | 39 (65.0) | 16 (57.1) | 23 (71.8) | 14.7 (−9.3 to 38.8) | 0.233 |
| Item 2 | |||||
| Power | 45 (75.0) | 18 (64.3) | 27 (84.4) | 20.1 (−1.7 to 41.8) | 0.073 |
| Item 3 | |||||
| Treatment effect† | 33 (55.0) | 15 (53.6) | 18 (56.3) | 2.7 (−22.6 to 27.9) | 0.835 |
| Item 4 | |||||
| Consistency with primary outcome | 38 (63.3) | 14 (50.0) | 24 (75.0) | 25.0 (1.2 to 48.8) | 0.045 |
| Item 5 | |||||
| Allowance for attrition | 18 (30.0) | 7 (25.0) | 11 (34.4) | 9.4 (−13.6 to 32.4) | 0.429 |
| Item 6 | |||||
| Number of clusters | 58 (96.7) | 27 (96.4) | 31 (96.9) | 0.4 (−8.7 to 9.6) | 0.923 |
| Median cluster size | 39 (65.0) | 15 (53.6) | 24 (75.0) | 21.4 (−2.4 to 45.2) | 0.083 |
| Item 7 | |||||
| Variation in cluster size* | 6 (10.0) | 1 (3.6) | 5 (15.6) | 12.1 (−2.3 to 26.4) | 0.201 |
| Item 8 | |||||
| Variation in outcome across clusters (ie, ICC) | 33 (55.0) | 11 (39.3) | 22 (68.8) | 29.5 (5.3 to 53.7) | 0.022 |
| Item 9 | |||||
| Uncertainty of ICC (or equivalent)* | 8 (13.3) | 3 (10.7) | 5 (15.6) | 4.9 (−12.1 to 21.9) | 0.712 |
| All ItemS | |||||
| Number items reported median (IQR) | 5.0 (2.5–6.0) | 4.0 (1.0–6.0) | 6.0 (5.0–6.0) | 1.22 (0.07 to 2.36) | 0.067 |
| Reporting all nine items | 0 (0) | 0 (0) | 0 (0) | ||
Values are numbers (percentages) unless stated. p Value is for the comparison of 1987–2012 publications and 2013–2014 publications using a χ2 test for proportions (categorical outcomes) or Mann-Whitney U test (where medians are reported), or (*) using Fisher's exact test.
†A sufficient reporting of the treatment effect consists of either a standardised effect size; a mean difference and SD; means in both arms and SD; proportions in both arms; proportion in one arm and a difference.
ICC, intracluster correlation.