The effect of mass vaccination campaigns against polio on the utilization of routine immunization services: A regression discontinuity design

Background In most low and middle-income countries (LMIC), vaccines are primarily distributed by routine immunization services (RI) at health facilities. Additional opportunities for vaccination are also provided through mass vaccination campaigns, conducted periodically as part of disease-specific initiatives. It is unclear whether these campaigns are detrimental to RI services, or wether they may stimulate the utilization of RI. Methods Unobserved confounders and reverse causality have limited existing evaluations of the effects of mass vaccination campaigns on RI services. We explored the use of a regression discontinuity design (RDD) to measure these effects more precisely. This is a quasi-experimental method, which exploits random variations in birth dates to identify the causal effects of vaccination campaigns. We applied RDD to survey data on a nationwide vaccination campaign against Polio conducted in Bangladesh. Results We compared systematically the children born immediately before vs. after the vaccination campaign. These two groups had similar background characteristics, but differed by their exposure to the vaccination campaign. Contrary to previous studies, exposure to the campaign had positive effects on RI utilization. Children exposed to the campaign received between 0.296 and 0.469 additional doses of DPT vaccine by age 4 months than unexposed children. Conclusions RDD constitutes a promising tool to assess the effects of mass vaccination campaigns on RI services. It could be tested in additional settings, using larger and more precise datasets. It could also be extended to measure the effects of other disease-specific interventions on the functioning of health systems, in particular those that occur at a discrete point in time and/or include age-related eligibility criteria.