Impact and effect mechanisms of mass campaigns in resource-constrained health systems: quasi-experimental evidence from polio eradication in Nigeria

Abstract

Background: Mass campaigns are a key strategy for delivering life-saving interventions under Global Health Initiatives, especially in weak health system contexts. They are frequently designed parallel to the health system to rapidly achieve programme targets such as vaccination coverage, but we lack quantitative evidence demonstrating their impact and effect mechanisms on health system performance at sub-/national level. This longitudinal study responds to this gap through an analysis of polio eradication campaigns in Nigeria.

Methods: Using four rounds of Demographic and Health Surveys in Nigeria between October 2000 and December 2017, we created a longitudinal dataset containing 88 881 under-5 children/pregnancies. We estimated the relationships between individuals' campaign exposure and health system performance indices (full RI schedule attainment, maternal healthcare services utilisation and child survival) using multilevel, mixed-effects regression models applied nationally and stratified by the six geopolitical zones in Nigeria.

Results: Nationally, high-frequency mass campaigns had detrimental health systems effects that potentially left 3.6 million children deprived of full immunisation. The frequency of campaigns was most concentrated in regions with weak health systems, where the operations of RI were disrupted, alongside negative effects on child survival and institutional delivery. In contrast, regions with relatively strong health systems and few campaigns experienced beneficial effects on maternal healthcare service utilisation.

Conclusions: As we provide evidence that well-functioning health systems can benefit from mass campaigns under Global Health Initiatives, our work also challenges the established wisdom to intensify mass campaigns in weaker health systems to bypass service provision bottlenecks. Mass campaigns do not inherently benefit or damage a health system, but frequent campaigns in weak health system contexts can impede service provision. We call for an additional burden of proof and active efforts to integrate mass campaigns into routine health services by harmonising implementation plans and service delivery in weak health system contexts.

Keywords: health systems; immunisation; poliomyelitis.

Figure 1

Predicted probability of (non-polio) full immunisation as a function of SIAs to which a child had been exposed since birth (A) and as function of child age (B). Source: authors. Notes. prediction based on models 1 and 3 in table 1 for children aged 10–60 months (n=24 381), controlling for child, parental and household determinants of immunisation with survey round dummy and LGA random effect. Full immunisation defined here as at least one dose of BCG, three doses of DPT and one dose of measles vaccine as recalled by mother and/or reported by child’s health card. DPT, diphtheria, pertussis and tetanus; LGA, local government area; SIAs, supplementary immunisation activities; BCG, Bacillus Calmetter-Guerin.