The short-term impact of a malaria elimination initiative in Southern Mozambique: Application of the synthetic control method to routine surveillance data

Abstract

In public health epidemiology, quasi-experimental methods are widely used to estimate the causal impacts of interventions. In this paper, we demonstrate the contribution the synthetic control method (SCM) can make in evaluating public health interventions, when routine surveillance data are available and the validity of other quasi-experimental approaches may be in question. In our application, we evaluate the short-term effects of a large-scale Mass Drug Administration (MDA) based malaria elimination initiative in Southern Mozambique. We apply the SCM to district level weekly malaria incidence data and compare the observed reduction in age group specific malaria incidence. Between August 2015 and April 2017, a total of 13,322 (78%) cases of malaria were averted relative to the synthetic control. During the peak malaria seasons, the elimination initiative resulted in an 87% reduction in Year 1 (December 2015–April 2016), and 79% reduction in Year 2 (December 2016–April 2017). Comparison with an interrupted time series approach shows the SCM accounts for pre-intervention trends in the data and post-intervention weather events influencing malaria cases. We conclude MDA brought about a drastic reduction in malaria burden and can be a useful addition to existing (or new) vector control strategies and tools in accelerating towards elimination.