World Immunisation Week 2020 has arrived, and vaccinations are on a lot of people’s minds at the moment, with everyone wondering when/if a vaccine against COVID-19 will be available. But with health systems being strained and normal operations of all kinds being disrupted, it’s essential to maintain focus on delivering key preventive services to protect public health, including routine immunisation for children. While such immunisation is widely recognised as one of the most cost-effective interventions in global health and development, vaccination coverage remains low or stagnant in many poor countries.

As part of 3ie’s Immunisation Evidence Programme, we are creating an evidence gap map (EGM) on interventions to increase routine vaccination coverage in low- and middle-income countries (LMICs). (We are also conducting a systematic review of a subset of this literature, focusing on community-engagement strategies.) If you aren’t familiar with 3ie’s EGMs, you can read about them here. As this EGM nears completion (we are still working to identify additional studies), we wanted to share some details about our approach and what we’ve found so far.

First, a few words about the framework we created for this EGM. We wanted to be able to categorise interventions and outcomes in a very fine-grained way. We recognise that while researchers often think in abstractions, decision makers have to think in more concrete terms—i.e., their decisions often concern specific programmes or policies rather than broad intervention types. So while some systematisation is of course necessary, we wanted to map the evidence with as much specificity as we could. So our detailed framework contains 38 intervention categories and 44 outcome categories. But, in the interests of brevity, here I’ll be discussing things in a more aggregated way. Those who wish to see all the details are encouraged to keep an eye out for the full map, which will be published on this site in the coming weeks.

Our intervention categories are grouped based on whom the intervention targets: caregivers, the health system, other community members (e.g., traditional and religious leaders), or the community as a collective whole. For outcomes, we divide them into factors relating to the demand for vaccinations, the supply of vaccination services, vaccination coverage, and health (i.e., morbidity and mortality from vaccine-preventable causes).

So, what have we learned about the evidence base on routine immunisation in LMICs? Again, I hasten to stress that these findings are preliminary, but assuming the overall patterns remain consistent with what we’ve found so far, we can identify some interesting trends.

When we map the evidence onto our framework, some patterns jump out immediately. One is that by far the most-studied types of interventions are those using monetary and non-monetary incentives to motivate caregivers (27 instances of evidence) or health workers (21). There is also considerable focus on educating and informing caregivers (18 instances of evidence) and promoting community-level dialogue (17).

But by far the most striking pattern is how blank the outcome columns are aside from the “vaccination coverage” one. We’ve identified 114 instances of evidence on interventions that affect coverage of routine vaccinations in young children; the most we’ve identified in other categories is 13 (health outcomes), followed by 5 (caregivers’ attitudes). To an extent, this makes sense and is a good thing. Improving vaccination coverage is precisely the goal of most studies included in our map, so it’s logical that they focus on measuring effects on vaccination coverage. And it’s good that so much of the evidence gets at outcomes that are proximal to what we really care about, namely improving health and saving lives.

On the other hand, though, we have a serious lack of evidence about pretty much all of the intermediate outcomes along the way. Whether a child gets vaccinated depends on a host of demand- and supply-side factors: caregivers need to recognise the importance of vaccinations and desire to vaccinate their children, they need to know when and where to go to receive those vaccinations, and health centres need to be accessible, stocked with vaccines, and staffed by health workers with the knowledge and skills to administer vaccines appropriately. Given how vaccination coverage has stagnated in many poor countries, we know at least some of these factors are missing. The links in the causal chain leading up to vaccination need to be strengthened, and we need good evidence to tell us how to do that.

So a provisional message to the field from our EGM work is that impact evaluations on immunisation interventions need to pay greater attention to the intermediate steps in getting to full vaccination coverage: what barriers stand in the way of caregivers’ forming intentions to vaccinate and following through on those intentions? What works to improve health workers’ tracking of children due for vaccines and enhance their outreach efforts? How can supply chains and cold chains be strengthened to prevent vaccine stockouts? Getting evidence-backed answers to these questions is an essential component of developing effective strategies to achieve full vaccination coverage worldwide.