Table 5. Sample size determination and target difference justifications for all trials and those with primary or coprimary PROs.
| Characteristic | Frequency (%) | |
|---|---|---|
| Trials with PRO as primary/coprimary | All trials | |
|
Was a sample size or power calculation provided? Yes, based on primary or coprimary outcome Yes, based on other outcome No |
N = 144 122 (84.7%) 4 (2.8%) 18 (12.5%) |
N = 415 358 (86.3%) 15 (3.6%) 42 (10.1%) |
|
When a sample size or power calculation was provided for the primary outcome, was the target difference justified?b Yes Anchor Distribution Standardized effect size Patient/stakeholder consultation (“opinion seeking”) Review of the evidence base Pilot study Authors specified difference was important No Target difference not justified Calculated power based on available sample size Target difference not specified |
N = 117 a | N = 358 |
| 71 (60.7%) 6 (5.1%) 5 (4.3%) 20 (17.1%) 8 (6.8%) 16 (13.7%) 31 (26.5%) 1 (0.9%) 46 (39.3%) 39 (33.3%) 4 (3.4%) 3 (2.6%) |
187 (52.2%) 6 (1.7%) 5 (1.4%) 25 (7.0%) 18 (5.0%) 49 (13.7%) 100 (27.9%) 7 (2.0%) 171 (47.8%) 151 (42.2%) 15 (4.2%) 5 (1.4%) |
|
| When a sample size or power calculation was provided for the primary outcome, what type of target difference was used? Important Realistic Both important and realistic Not specified or unclear |
N = 117a | N = 358 |
| 18 (15.4%) 37 (31.6%) 12 (10.3%) 50 (42.7%) |
39 (10.9%) 132 (36.9%) 20 (5.6%) 167 (46.6%) |
|
| When a sample size or power calculation was provided for the primary outcome, was it adjusted for attrition? Yes No Other or unclear |
N = 117a | N = 358 |
| 72 (61.5%) 44 (37.6%) 1 (0.9%) |
157 (43.9%) 197 (55.0%) 4 (1.1%) |
|
aTrials with coprimary clinical and PROs but where the clinical outcome was used to determine the sample size (n = 5) and trials presenting sample size calculation on an outcome not clearly identified as primary (n = 4) were excluded.
bMore than one method may have been used within one study.
PRO, patient-reported outcome.