Characteristics of the study RCTs according to the actual analysis type
| Characteristics | ITT (N=81) | mITT (N=42) | Non-ITT (N=53) | P value |
| Funding source | <0.001 | |||
| Industry, full or partial | 59 (72.8) | 31 (73.8) | 14 (26.4) | |
| Non-profit or unspecified | 22 (27.2) | 11 (26.2) | 39 (73.6) | |
| Experimental agent | 0.001 | |||
| Traditional DMARD | 13 (16.0) | 5 (11.9) | 12 (22.6) | |
| Biological DMARD | 30 (37) | 27 (64.3) | 16 (30.2) | |
| Small molecule | 8 (9.9) | 2 (4.8) | 0 (0.0) | |
| Others | 30 (37.0) | 8 (19.0) | 25 (47.2) | |
| Study phase | 0.086 | |||
| Phase 2 | 20 (24.7) | 8 (19.0) | 5 (9.4) | |
| Non-phase 2/unspecified | 59 (75.3) | 34 (81.0) | 48 (90.6) | |
| Study centres, multiple | 63 (77.8) | 42 (100) | 19 (35.8) | <0.001 |
| Study duration, months | 6 (3–12) | 9 (3–12) | 6 (3–12) | 0.481 |
| Pre-study sample size calculation reported* | 42 (68.9) | 29 (85.3) | 24 (50) | 0.003 |
| Participant flow diagram | 0.001 | |||
| Adequate | 24 (29.6) | 24 (57.1) | 11 (20.8) | |
| Inadequate | 26 (32.1) | 9 (21.4) | 12 (22.6) | |
| Not reported | 31 (38.3) | 9 (21.4) | 30 (56.6) | |
| Adequate follow-up description | 63 (77.8) | 40 (95.2) | 33 (62.3) | <0.001 |
| Number of patients enrolled | 165 (70–339) | 317 (160–549) | 70 (41–160) | <0.001 |
| Percentage of patients completing RCT† | 85.8 (70.4–89.3) | 82.9 (76–88.9) | 89 (72.8–92.3) | 0.232 |
| Percentage of missing data† | 0.341 | |||
| 0–5 | 8 (10.5) | 3 (7.1) | 6 (13.6) | |
| >5–10 | 10 (13.2) | 5 (11.9) | 12 (27.3) | |
| >10–20 | 32 (42.1) | 20 (47.5) | 13 (29.5) | |
| >20 | 26 (34.2) | 14 (33.3) | 13 (29.5) | |
| Percentage of patients analysed for the primary outcome‡ | 100 (100–100) | 98.7 (97.1–99.6) | 90.8 (86.3–97.7) | <0.001 |
| Missing data handling method described§ | 64 (84.2) | 38 (90.5) | 11 (20.8) | <0.001 |
| Efficacy, positive¶ | 54 (69.2) | 35 (87.5) | 29 (63.0) | 0.032 |
Values represent number (%) for categorical variables and median (25th–75th percentile) for the numerical variables.
*N=143 (33 phase 2 RCTs excluded as they have different sample size calculation considerations).
†N=162 (five RCTs performing ITT and nine RCTs performing non-ITT analysis did not clearly report number of patients completing the trial).
‡N=163 (13 RCTs performing non-ITT analysis did not clearly report number of patients analysed for the primary outcome).
§N=169 (all enrolled patients reported to complete seven RCTs with no missing data).
¶Positive efficacy defined as statistically significant outcome for the primary outcome favouring the experimental intervention for superiority trials or meeting the threshold for equivalence in the non-inferiority trials. N=164, 12 RCTs were strategy trials with no intervention designated as experimental.
DMARD, disease-modifying anti-rheumatic drug; ITT, intention-to-treat; mITT, modified ITT; N, number; RCT, randomised controlled trial.