We then used random effects negative binomial regression to assess the impact of the MTAT intervention on both monthly confirmed and total malaria case incidence at the health facility level using a difference-in-differences (DiD) approach, 40 with the primary effect estimate evaluated through a time (pre-post intervention period)—by a treatment group interaction term.
Additional covariates such as terminal (accounting for possible seasonal and regional differences), past smoking history, self reported SHS exposures during the work shift, and switching shifts were also considered as potential confounders and were kept in the models if they changed the primary effect estimate by ≥10%.
We believe that the present focus on the meta-analytic summaries of primary effect estimates without regard for the sources of variability will lead to the production of an incomplete picture of the intervention in question.
As primary summary effect estimates we calculated summary odds ratios and 95% confidence intervals according to the DerSimonian and Laird random effects model, which utilises weights that incorporate variance within and between studies.
A meta-analysis was performed using the hazard ratio as primary effect measure, which was estimated from Cox proportional hazard models or from summary statistics of Kaplan Meier product-limit estimations.
To see whether sex differences in behaviour were indirectly associated with sex differences in incident HCV infection, we entered any behavioural variable associated with sex and with incident HCV (in bivariate analysis) individually into a Cox model that contained female sex as its primary predictor and compared the effect estimate for female sex when it was the only variable in the model.
Adjusting for event dependence (and thus estimating the primary effect) also made only minor changes (HR = 0.73, 95% CI: 0.66, 0.80) in model 4 (Table 5).
Estimating the primary effect by adjusting for event dependence did not make major changes (HR = 0.79, 95% CI: 0.73, 0.87) in model 7 (Table 5).
When accounting for event dependence and a nonsusceptible fraction, we estimated the primary effect among susceptible children to be 60% (HR = 0.40, 95% CI: 0.33, 0.49) in model 7 (Table 4).
Estimates of the hazard ratio for posttreatment prophylaxis and the frailty are given in Web Tables 2 and 3. Comparison of models 1, 2, and 3 with the model estimating the primary effect by stratification on event order (model 4) allows separation of the primary effect of the intervention from its total effect.
Stratification on event order (and therefore estimating the primary effect of SMC) indicates a much greater primary protective effect of SMC than its total effect (HR = 0.15, 95% CI: 0.13, 0.18) (PE = 85%, 95% CI: 82, 87) as shown in model 4 (Table 4).
