Vaccination boosters: using 'village gossips' to improve childhood vaccination rates in India

Even before the pandemic disrupted the delivery of primary health care around the world, health leaders in many countries struggled to increase childhood immunization rates – with global coverage of the first dose of the measles vaccine topping out at only at 86 percent, well below the 95 percent needed to prevent measles outbreaks.

In India, as of 2021, despite several efforts over the years to improve immunization coverage with the world's largest universal immunization program, just over 75 percent of children in the 12- to 23-month age group are fully vaccinated against BCG, polio, DPT, Hepatitis B and measles. In the northern Indian state of Haryana, vaccination rates for routine immunization vary significantly, with parents in some districts reporting that only 40 percent of children are fully vaccinated.

To improve demand for immunization, especially in areas with low vaccination coverage, health leaders in the Haryana government collaborated with researchers affiliated with the J-PAL network on a large field experiment that, between 2016 and 2018, would test three innovative ideas. The first one was to use "village gossips," individuals identified by the local community as good at disseminating information. These “immunization ambassadors,” as well as others identified as trusted sources and some randomly selected individuals, were sent text messages asking them to remind others in their communities of the value of immunization. The second idea involved providing caregivers with small mobile phone credits of varying amounts, each time they brought their child in for a vaccination. The third idea was to send targeted caregivers direct reminders via text and voice messages, whenever a child was due for a specific vaccination. Through various groupings, 75 such unique interventions, randomly assigned at the sub-center and village levels, were tested.

The study was conducted across seven districts in Haryana that had the lowest full immunization rates, covering 140 primary health care centers and 755 health sub- centers across more than 2,000 villages. Another innovation was to collect administrative data on immunization, using village frontline health workers, who were provided with electronic tablets and trained to use a simple and user-friendly e-health platform called mhealth, developed by MIT’s SANA software development group. The platform was used to register all children, track vaccines administered to each child, and record mobile numbers of each caregiver. This data became the basis for delivering the innovative interventions, as it was used to calculate the mobile credit awarded for each vaccination and send targeted text message reminders to caregivers. Machine learning was also used to determine which policy combinations were the most effective and cost effective.

Our research found that using village gossips as immunization ambassadors was the single most effective means of improving full immunization rates. We found that the combination of using village gossips and targeted text reminders to parents was both inexpensive and cost effective. The most effective combination overall – which increased rates of full immunization by 55 percent relative to the comparison group – combined all three innovations: using village gossips, mobile phone credit incentives that increased across the immunization schedule, and targeted text messages to caregivers about next scheduled vaccinations. This combination was particularly effective in poorer areas that had low immunization levels.

It is also interesting to note what did not work. Targeted messages sent directly to parents about the next vaccination schedule, when not used in combination with other incentives, had no impact on vaccination rates. Neither did mobile phone credits, unless they increased across the immunization schedule. Village gossips were even more effective at transmitting information than more traditional conduits such as village leaders and officials, teachers, or frontline health workers.

We learned from this study that village gossips were particularly good and transmitting information about the value of immunization, and likely amplified the effect of incentives such as mobile phone credits. This supports many of the findings of the Exemplars in Global Health team, which has highlighted the impact of incentives such as conditional cash transfers in Peru and ambassadors, such as BRAC’s community health workers in Bangladesh, in improving health-seeking behavior.

The government of Haryana is currently working with researchers and J-PAL South Asia to scale the most cost-effective combination of using village gossips and targeted reminders in three districts. Frontline health workers are identifying these village gossips by asking caregivers - when they come to immunization sites - to identify who in their village is a good source for information. Simple nomination questions have been integrated into the standard intake forms on ANM Online (or ANMOL), the central government's mobile health platform.

Challenges remain to this work, including ensuring adoption of the mhealth platform by the frontline workers for real-time data entry at immunization sites. The model is currently being piloted in two districts where frontline workers have been trained on how to ask caregivers at immunization sites for nominations for village gossips. Once a list of village gossips is identified for each village, they start to receive targeted text reminders on the value of routine immunization. The government is notably also interested in testing the idea of using these ambassadors for diffusing information about the importance of COVID-19 vaccinations, especially the need to complete the vaccine's schedule with the second dose.

One key hurdle to scaling is related to developing a mechanism for identifying gossips. Two primary pathways are asking frontline health workers to survey their patients at the point of service, or deploying dedicated survey teams to villages, to conduct doorstep surveys. Obviously, incorporating questions about village gossips into an existing health care app is most cost effective. But it is also difficult – primarily because frontline health workers are already overburdened. Our experience from the research study indicates that it is important to invest time and effort in building capacity and securing buy-in of frontline workers to return reliable results.