Abstract / Summary
ObjectivesIndoor residual spraying (IRS) was introduced in Côte d’Ivoire in 2020. We seek to identify potential factors that could compromise the quality of routine malaria incidence and determine how impact quantification of IRS on malaria incidence depends on data quality by emphasizing robust counterfactuals with contemporaneous covariates.MethodsUsing mixed-effect models and correlation analysis we compare data from two sources (routine surveillance and re-digitized registries). To quantify decrease in incidence post intervention, we apply Bayesian structural time series analysis with posterior feature selection to determine impact confounders.ResultsThe presence of IRS, test positivity and reporting rates are significant factors impacting data quality. For high-quality data, we estimate a 12% resp. 19% decrease in incidence rate 1 year after the intervention, while for low-quality data estimates are 30% resp. 42%. Important covariates for counterfactual modeling are control district incidence, reporting and test positivity rates.ConclusionImprovements in data quality by scrutinizing test positivity rates and health facility reporting can help avoid overly optimistic impact estimations for IRS. Incidence control districts remain inevitable in order to factor out impact confounders.