Table 4:
Results from the multivariable logistic regression analysis to identify independent characteristics associated with not obtaining consent (Created by the authors)
| Trial characteristics* | Adjusted OR (95% Confidence Interval) | p-value |
|---|---|---|
| Start year | <0.0001 | |
| 1998 – 2010 | Ref | |
| 2011 – 2012 | 1.24 (0.85 – 1.80) | 0.260 |
| 2013 – 2018 | 2.13 (1.52 – 3.00) | <0.0001 |
| Journal impact factor | 0.569 | |
| Tertile 1: [0.11 – 3.575) | Ref | |
| Tertile 2: [3.575 – 7.958) | 1.09 (0.78 – 1.53) | 0.608 |
| Tertile 3: [7.958 – 70.67] | 0.91 (0.64 – 1.30) | 0.587 |
| Design | <0.0001 | |
| Individually randomised trial | Ref | |
| CRT with no professional-level or cluster-level interventions | 2.17 (1.40 – 3.37) | 0.0005 |
| CRT with professional-level intervention, possibly other interventions, no cluster-level interventions | 5.09 (3.22 – 8.04) | <0.0001 |
| CRT with at least some cluster-level interventions | 8.09 (5.63 – 11.64) | <0.0001 |
| Country of study conduct | <0.0001 | |
| LMIC only | Ref | |
| Other | 2.31 (1.26 – 4.20) | 0.006 |
| EU only or UK only | 2.54 (1.47 – 4.37) | 0.0008 |
| Canada only or US only | 4.00 (2.46 – 6.51) | <0.0001 |
| Setting | 0.073 | |
| Clinical | Ref | |
| Public health | 0.70 (0.48 – 1.03) | 0.073 |
| Self-identified as pragmatic | <0.0001 | |
| No | Ref | |
| Yes | 2.14 (1.57 – 2.91) | <0.0001 |
| Primary purpose | <0.0001 | |
| Treatment | Ref | |
| Other | 1.35 (0.93 – 1.96) | 0.118 |
| Health services research | 3.15 (2.13 – 4.67) | <0.0001 |
*
Only pre-specified characteristics were entered into the model. Intervention type (Clinical, Dietary and behavioral, Other) was excluded as it was collinear with other variables.
CRT = Cluster randomised trial