Table 2.
Model parameters for the screening numbers per week in a multivariable generalized mixed-effects model (Poisson family with logarithmic link function)
| Predictor | β coefficient* | exp (β)* | 95% interval for exp(β)* | p value |
|---|---|---|---|---|
| Revision of inclusion criteria | 0.186 | 1.204 | (0.881–1.647) | 0.243 |
| Increase of travel expenses reimbursement | 0.149 | 1.161 | (0.859–1.570) | 0.331 |
| Investigator teleconferences | 0.382 | 1.466 | (1.018–2.112) | 0.0398 |
| Investigator meetings | −0.084 | 0.919 | (0.705–1.199) | 0.534 |
| Core partner site | 0.379 | 1.460 | (0.254–8.392) | 0.671 |
| Reaching 80% of site recruitment target | −0.357 | 0.699 | (0.542–0.903) | < 0.001 |
| Public holidays | −0.763 | 0.466 | (0.367–0.591) | < 0.001 |
| Week | −0.006 | 0.994 | (0.986–1.003) | 0.202 |
| Intercept | −4.19 | 0.015 | (0.005–0.045) |
* adjusted values. No evidence for overdispersion was found (dispersion parameter: 1.0098, p = 0.3820 [19]). Only a shared fixed intercept is added when the site is inactive (β0 = − 4.19, exp.(β0) = 0.015, 95% CI [0.005–0.045]). Random intercepts α for active clinical sites ranged from 3.42 to 4.35