Table 3 Results of analyses of the GetTested trial. All models included main effects only. Link function is canonical unless otherwise specified. The dash symbol - means model did not converge, except for * where the log risk ratio estimated as 541, indicating separation for one or more covariates
From: Planning a method for covariate adjustment in individually randomised trials: a practical guide
Outcome measure | Summary measure | Adjustment method | Model (variable modelled) | Treatment effect estimate (SE) |
|---|---|---|---|---|
Any test (occurred in 35%) | Risk difference | Direct | Identity-link binomial (outcome) | - |
| Â | Â | Standardisation | Logistic (outcome) | 0.260 (0.021) |
| Â | Â | IPTW | Logistic (treatment) | 0.262 (0.021) |
| Â | Log risk ratio | Direct | Poisson, robust SE (outcome) | -* |
| Â | Â | Direct | Log-link binomial (outcome) | 0.797 (0.075) |
| Â | Â | Standardisation | Logistic (outcome) | 0.796 (0.074) |
| Â | Â | IPTW | Logistic (treatment) | 0.806 (0.075) |
Any diagnosis (occurred in 1.6%) | Risk difference | Direct | Identity-link binomial (outcome) | - |
| Â | Â | Standardisation | Logistic (outcome) | 0.015 (0.006) |
| Â | Â | IPTW | Logistic (treatment) | 0.013 (0.006) |
| Â | Log risk ratio | Direct | Poisson, robust SE (outcome) | 0.972 (0.433) |
| Â | Â | Direct | Log-link binomial (outcome) | 0.915 (0.412) |
| Â | Â | Standardisation | Logistic (outcome) | 0.959 (0.412) |
| Â | Â | IPTW | Logistic (treatment) | 0.855 (0.412) |