Comparison of data reporting including missing data and its handling in 2006 and 2016 RCTs
| Characteristics | All (N=76) | Study year | P value | |
| 2006 (N=34) | 2016 (N=42) | |||
| Total patients | 0.913 | |||
| Median (IQR) | 162 (74–326) | 163 (53–367) | 159 (79–311) | |
| Range | 21–1404 | 21–1093 | 43–1404 | |
| Patient per cent completing study* | 0.089 | |||
| Median (IQR) | 86.4 (79.5–91.6) | 84.5 (72.8–91.8) | 87.9 (81.7–91.6) | |
| Range | 38.5–100 | 40.9–95.8 | 38-5–100 | |
| Patient per cent analysed for primary outcome† | 0.962 | |||
| Median (IQR) | 99.7 (97.1–100) | 99.5 (97.3–100) | 99.7 (97–100) | |
| Range | 38.5–100 | 85.8–100 | 38.5–100 | |
| Patient in experiment arm(s) | 0.81 | |||
| Median (IQR) | 100 (40–217) | 99 (29–252) | 103 (40–177) | |
| Range | 11–1286 | 11–751 | 19–1286 | |
| Patient per cent completing study in experimental arm(s)‡ | 0.405 | |||
| Median (IQR) | 87.7 (81.3–93.1) | 86.4 (81.7–93) | 88.6 (80.7–94.4) | |
| Range | 37.5–100 | 45.2–100 | 37.5–100 | |
| Patient per cent analysed for primary outcome in experimental arm(s)§ | 0.721 | |||
| Median (IQR) | 100 (99.6–100) | 100 (99.3–100) | 100 (99.6–100) | |
| Range | 37.5–100 | 88.6–100 | 37.5–100 | |
| Patient in comparator arm¶ | 0.983 | |||
| Median (IQR) | 56 (30–132) | 68 (25–141) | 53 (31–121) | |
| Range | 10–532 | 10–531 | 15–532 | |
| Patient per cent completing study in comparator arm** | 0.1 | |||
| Median (IQR) | 84.4 (73–91.5) | 81.9 (64.6–90.6) | 84.9 (78.1–93.5) | |
| Range | 23.9–100 | 30–93.7 | 23.9–100 | |
| Patient per cent analysed for primary outcome in comparator arm | 0.507 | |||
| Median (IQR) | 100 (97.2–100) | 100 (96.1–100) | 100 (97.6–100) | |
| Range | 39.5–100 | 84.5–100 | 39.5–100 | |
| Adequate sample size calculation description†† | 43 (66.2) | 19 (61.3) | 24 (70.6) | 0.429 |
| Sample size inflation anticipating follow-up loss†† | 13 (20.0) | 4 (12.9) | 9 (26.5) | 0.167 |
| Adequate follow-up description | 64 (84.2) | 28 (82.4) | 36 (85.7) | 0.689 |
| Flow diagram for patient follow-up | 0.016 | |||
| Adequate | 31 (40.8) | 17 (50.0) | 14 (33.3) | |
| Inadequate | 18 (23.7) | 3 (8.8) | 15 (35.7) | |
| Not reported | 27 (35.5) | 14 (41.2) | 13 (31.0) | |
| Amount of missing outcome data* | 0.042 | |||
| <5% | 10 (13.9) | 1 (3.1) | 9 (22.5) | |
| 5.1%–10% | 13 (18.1) | 8 (25.0) | 5 (12.5) | |
| 10.1%–20% | 31 (43.1) | 13 (40.6) | 18 (45.0) | |
| >20% | 18 (25) | 10 (31.3) | 8 (20.0) | |
| Missing data handling method given‡‡ | 0.75 | |||
| N (%) | 48 (65.8) | 23 (67.6) | 25 (64.1) | |
| Self-reported analysis term used | 0.05 | |||
| ITT | 33 (43.4) | 20 (58.8) | 13 (43.4) | |
| Modified ITT | 6 (7.9) | 2 (5.9) | 4 (9.5) | |
| None | 37 (48.7) | 12 (35.3) | 25 (59.5) | |
| Actual analysis performed | 0.266 | |||
| ITT | 29 (38.2) | 12 (35.3) | 17 (40.5) | |
| Modified ITT | 20 (26.3) | 11 (32.4) | 9 (21.4) | |
| Completer/inadequate | 15 (19.7) | 4 (11.8) | 11 (26.2) | |
| Unclear | 12 (15.8) | 7 (20.6) | 5 (11.9) | |
| Top methods to impute missing data | ||||
| LOCF, N (%) | 29 (38.2) | 17 (50.0) | 12 (28.6) | 0.056 |
| NRI, N (%) | 22 (28.9) | 10 (29.4) | 12 (28.6) | 0.936 |
| Simple imputation, N (%) | 11 (14.5) | 6 (17.6) | 9 (21.4) | 0.68 |
| Performed sensitivity analysis | 0.452 | |||
| N (%) | 14 (18.4) | 5 (14.7) | 9 (21.4) | |
*N=72, number of patients completing the trial unclear for four RCTs.
†N=73, number of patients included in primary outcome analysis unclear for three RCTs.
‡N=69, number of patients in experimental arm(s) completing the trial unclear for seven RCTs.
§N=70, number of patients in experimental arm(s) included in primary outcome analysis unclear for six RCTs.
¶N=75, one RCT had three arms with different doses of experimental intervention, no comparator group.
**N=68, number of patients in comparator arm completing the trial unclear for seven RCTs and one RCT did not have a comparator group.
††N=65, excluded 11 phase 2 RCTs as they have different sample size calculation considerations.
‡‡N=73, all enrolled patients reported to complete three RCTs with no missing data.
ITT, intention-to-treat; LOCF, last observation carried forward; N, number; NRI, non-response imputation; RCTs, randomised controlled trials.