Policy influence
IRC, in coordination with the Uganda Ministry of Health, is scaling up the mReach application and the Fifth Child community engagement strategy in Kitgum and Lamwo districts. A family planning module will be added to the existing mobile application, to promote integrated immunisation and postpartum family planning services.
Synopsis
This study evaluated the International Rescue Committee’s (IRC’s) Fifth Child integrated community engagement strategy to improve immunisation coverage in Northern Uganda.
Context
Northern Uganda faces several challenges in meeting the health needs of its population. Maternal and child health services, including immunisation, remain weak and have stagnated, with only 73 per cent of children receiving the third and final dose of the diphtheria-pertussis-tetanus vaccination (DPT3) (Uganda Demographic Health Survey, 2011). This rate has a range of causes, including limited community involvement and poor quality data to support decision-making. In early 2015, IRC established Fifth Child: a data-informed community engagement strategy to improve the stagnating immunisation coverage in northern Uganda.
Research questions
This study addresses the following questions:
- Does the intervention lead to an increase in DPT3 and Measles containing vaccine (MCV) combined immunisation coverage in 12 to 23-month-old children;
- Is there a reduction in vaccination drop-out rates for DPT 1, 2, 3 and oral polio vaccine 1, 2, 3;
- Will the intervention lead to improvements in the timely uptake of immunisations; and
- Has the intervention increased integrated community case management of children aged 6 – 59 months?
Methodology
Intervention design
The intervention aimed to enhance community engagement through the use of better quality data for target defaulter-tracing using a mobile-health data collection platform called mReach. The mReach application aggregated and presented data on the immunisation status of children, which was used for tracking those who were due for (or had defaulted from) their scheduled immunisations. Using the application, the facility-based health workers downloaded lists of children due for immunisations as well as defaulters. These lists were shared with community health volunteers known as Village Health Teams for follow-ups during home visits. These data were also shared with community leaders, who were actively co-managing immunisation outreaches with health workers. A second component of the intervention provided logistical support and supervision to health staff in both treatment and control areas for effective delivery of immunisation services.
Theory of change
The theory of change for this intervention hypothesised that data-driven, targeted community engagement will facilitate more effective defaulter-tracing, contributing to increased immunisation coverage in underserved, vulnerable and hard-to-reach communities. The theory of change rested on the following assumptions: (1) the intervention would facilitate community co-management of following up with immunisation defaulters and planning outreaches; (2) the mReach platform would improve data quality and provide health workers and community leaders with accurate and up-to-date data on immunisation defaulters; and (3) health workers and the Village Health Teams would be adequately trained to use the mReach data to conduct effective home visits that would prompt defaulting caregivers that are due for immunisation to complete the immunisation schedule of their children on time.
Evaluation design
A cluster randomised controlled trial was conducted to evaluate whether the intervention increased coverage of DPT3 and MCV in children 12 to 23 months old. A cluster was defined as a health facility catchment area. Thirty-two of these were included; sixteen in the control arm and the remaining sixteen were intervention clusters. The sample comprised 55 children in the age group of 9-23 months per cluster, which brought the total number of children included in this study to 1,760. Cross-sectional household surveys were used at baseline and endline to measure the impact of the intervention.
Findings
Please read these findings in light of the limitations detailed below.
- The study found significant increases in DPT3 and MCV coverage rates in both intervention and control (from 68% to 77%) groups. In the intervention group, coverage increased from 65% to 76%, while in the control group coverage increased from 68% to 77%. However, the difference between intervention and control at endline was not significant.
- The mReach platform was designed to allow Village Health Teams to easily identify defaulters and perform targeted outreach. However, the results showed no difference between intervention and control groups in exposure to household visits or the actions taken by caregivers after such visits.
- The intervention appeared to have no effect on the timeliness of vaccinations, as there was no significant difference between treatment and control groups at endline for timely receipt of DPT3 and MCV.
- Use of the platform proved challenging for some healthcare workers, who did not initially find the application straightforward. Additional training and supervision from IRC staff and an mHealth specialist were required to ensure proper use of the application.
- Caregivers had positive views towards the involvement of community leaders in supporting immunisation, particularly the community leaders’ ability to disseminate information to community members and to influence behaviour.
- All health workers in both treatment and control sites received immunisation update training from IRC and district leaders, which may have increased attention to immunisation services and improved coverage at study sites. Vaccine stock-outs were found to be common across all study sites.
Limitations of the study
The study has several major limitations, largely stemming from contamination between intervention and control groups. Some control sites implemented defaulter tracing systems that were similar to that provided by the mReach platform in intervention sites (albeit paper-based). In addition, some health workers were transferred from intervention to control sites during the course of the intervention. Thus, there is limited basis for drawing inferences about the causal impact of the intervention. Also, different measures of immunisation coverage showed different results: reported coverage based on administrative records (i.e., immunisation cards) showed marked improvement while coverage based on caregiver self-report showed no increase. Thus, it is possible that the overall increase in recorded coverage resulted only from improvements in administrative tracking.