The effect of case management and vector-control interventions on space–time patterns of malaria incidence in Uganda

BACKGROUND: Electronic reporting of routine health facility data in Uganda began with the adoption of the District Health Information Software System version 2 (DHIS2) in 2011. This has improved health facility reporting and overall data quality. In this study, the effects of case management with artemisinin-based combination therapy (ACT) and vector control interventions on space–time patterns of disease incidence were determined using DHIS2 data reported during 2013–2016. METHODS: Bayesian spatio-temporal negative binomial models were fitted on district-aggregated monthly malaria cases, reported by two age groups, defined by a cut-off age of 5 years. The effects of interventions were adjusted for socio-economic and climatic factors. Spatial and temporal correlations were taken into account by assuming a conditional autoregressive and a first-order autoregressive AR(1) process on district and monthly specific random effects, respectively. Fourier trigonometric functions were incorporated in the models to take into account seasonal fluctuations in malaria transmission. RESULTS: The temporal variation in incidence was similar in both age groups and depicted a steady decline up to February 2014, followed by an increase from March 2015 onwards. The trends were characterized by a strong bi-annual seasonal pattern with two peaks during May–July and September-December. Average monthly incidence in children < 5 years declined from 74.7 cases (95% CI 72.4–77.1) in 2013 to 49.4 (95% CI 42.9–55.8) per 1000 in 2015 and followed by an increase in 2016 of up to 51.3 (95% CI 42.9–55.8). In individuals ≥ 5 years, a decline in incidence from 2013 to 2015 was followed by an increase in 2016. A 100% increase in insecticide-treated nets (ITN) coverage was associated with a decline in incidence by 44% (95% BCI 28–59%). Similarly, a 100% increase in ACT coverage reduces incidence by 28% (95% BCI 11–45%) and 25% (95% BCI 20–28%) in children < 5 years and individuals ≥ 5 years, respectively. The ITN effect was not statistically important in older individuals. The space–time patterns of malaria incidence in children < 5 are similar to those of parasitaemia risk predicted from the malaria indicator survey of 2014–15. CONCLUSION: The decline in malaria incidence highlights the effectiveness of vector-control interventions and case management with ACT in Uganda. This calls for optimizing and sustaining interventions to achieve universal coverage and curb reverses in malaria decline. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12936-018-2312-7) contains supplementary material, which is available to authorized users.