Testing a Biometric Identification System to Improve Malaria Vaccine Completion

How to cite

Maffioli, Elisa Maria; Cohen, Jessica; Guure, Chris. 2025. "Testing a Biometric Identification System to Improve Malaria Vaccine Completion." AEA RCT Registry. AEARCTR-0016987. https://doi.org/10.1257/rct.16987-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Elisa Maria Maffioli* , Jessica Cohen , Chris Guure

* Primary investigator

Registry ID
AEARCTR-0016987
DOI
https://doi.org/10.1257/rct.16987-1.0
First registered
Last updated
Status
In development
Trial dates
2025-10-13 – 2027-12-31
Countries
Ghana (Oti)
Registry record
http://www.socialscienceregistry.org/trials/16987

Abstract

Receiving all four doses of the malaria vaccine can significantly protect children against malaria illness, hospitalization, and death. However, in Ghana, only 46% of children complete the full vaccination sequence. More broadly, many children in Ghana do not receive the full set of recommended pediatric vaccinations. To address this, Simprints, in collaboration with Ghana Health Services, will implement a digital vaccination record system linked to biometrics. This system will automatically identify children who are behind on their vaccination schedule, providing health workers with information to prioritize community outreach. Additionally, it will send voice message reminders to caregivers to improve compliance. We will conduct a cluster-randomized controlled trial (c-RCT) in the Oti region to measure the impact of this innovation on the proportion of children completing malaria and routine vaccination schedules.

Intervention

A program that implements a digital vaccination record system linked to biometrics in health facilities, coupled with voice message reminders for caregivers of children who are due for or missed vaccinations. Simprints will provide tablets and train Community Health Workers (CHWs) to use facial biometric identification for children above 6 months old. It will also automatically identify children who missed a scheduled dose, enabling CHWs to follow up with caregivers to bring the child in or to help CHWs in their micro-targeting efforts. Finally, Simprints will automatically send caregivers' voice message appointment reminders to those caregivers of children who are due for or missed vaccination.

Experimental design

This research study aims to generate rigorous evidence on how digital identification systems (face biometrics) may increase malaria and other routine vaccine completion rates for young children in Ghana.

We will implement a cluster-Randomized Controlled Trial (c-RCT) in the Oti region, considering a cluster of the catchment areas of close-by health facilities. We will select a sub-sample of about 100 clusters: half of the clusters will receive the intervention (treatment group), and half will serve as a control group. All facilities within the cluster will be assigned to either treatment or control group.

In the treatment group, Simprints will provide tablets and train Community Health Workers (CHWs) to use biometric identification for children above 6 months old. It will also identify children who missed a scheduled dose enabling CHWs to follow up with caregivers to bring the child or to help CHWs in their microtargeting efforts (i.e., bringing vaccinations to the communities). Finally, Simprints will automatically send caregivers' voice message appointment reminders on a schedule to be determined during initial field testing.

In the control group, CHWs will keep operating as before. About 30 health facilities in Jakisan district are expected to be using the GHS e-tracker. In the e-tracker facilities, CHWs will continue their usual work practice in regard to digitally recording patient details and services received, which may vary in levels of fidelity and frequency. In all facilities, regardless of whether the e-tracker is being used, paper records are maintained. This includes the CHW updating the vaccine tally register, and the aggregate numbers are reported into the Digital Health Information Management System (DHIMS) monthly.

Stratification - We will stratify based on three geographical regions (groups of districts), and a proxy for volume of patients (above or below the median of total number of caregivers visiting health facilities in March, April, May 2025, bringing children for vaccinations). We will have in total 6 strata. We are excluding two districts in Northern Ghana, because of feasibility and budget constraints (Nkwanta North and Nkwanta South).

Design details

Randomization method
Stratified randomization of facilities was conducted using the Stata (v.18) “randomize” command. Two variables were used for stratification. One was an indicator for the geographical region of the facility (3 total regions), and the other was an indicator for whether the facility was above or below the median in the distribution of the total number of caregivers bringing their child to the facility for vaccination in March, April, and May 2025. The randomization (among 2000 replications) that achieved the best balance based on the total number of caregivers bringing their child to the facility for vaccination in March, April and May 2025, and the total number of malaria cases in March, April, May 2025 was selected.
Randomization unit
The catchment area of health facilities is defined as a group of health facilities grouped by geographical proximity. Health facilities include health centers and CHPS compounds.
Clusters
About 100 catchment areas, totaling about 120 health facilities
Observations
We expect on average 41 women in their last two trimesters of pregnancy or women with children under 6 months old in the catchment area of a health facility. We plan to enroll both pregnant women (in the last two trimesters) as well as women with children under 6 months in the communities around each health facility. We expect to enroll about 4100 (pregnant in the last two trimesters or with children under 6 months old) women in about 100 health facility catchment areas.
Arms
41 women X 50 clusters = 2,050 women in treatment 41 women X 50 clusters = 2,050 women in control
Primary outcomes
1. Completion of full malaria sequence (Index children) 2. Timely full malaria vaccination (Index children) 3. Completion of full routine vaccination sequence (basic antigens) (Index children) 4. Completion of full routine vaccination sequence (national schedule) (Index children)
Secondary outcomes
1. Timely full routine vaccination sequence (basic antigens) (Index children) 2. Timely full routine vaccination sequence (national schedule) (Index children) 3. Early, Late or Very Late malaria vaccination (Index children) 4. Early, Late or Very Late full routine vaccination sequence (basic antigens) (Index children) 5. Early, Late or Very Late full routine vaccination sequence (national schedule) (Index children) 6. Number of malaria vaccines taken (Index children) 7. Number of routine vaccines taken (basic antigens) (Index children) 8. Number of routine vaccines taken (full national schedule) (Index children) 9. Number of timely malaria vaccines taken (Index children) 10. Number of timely routine vaccines taken (basic antigens) (Index children) 11. Number of timely routine vaccines taken (full national schedule) (Index children) 12. Received each vaccine on time (full national schedule) (Index child), analyzed individually 13. Up to date on malaria vaccine (children under 3 years old in study households) 14. Up to date on routine vaccines (basic antigens) (Children under 3 years old in study households) 15. Up to date on routine vaccines (full schedule) (Children under 3 years old in study households) 16. Cumulative number of late days in receipt of malaria vaccination (index children) 17. Cumulative number of late days in receipt of basic antigens (index children) 18. Cumulative number of late days in receipt of routine vaccines according to national schedule (index children) 19. Mother knows when to bring the child to the clinic for the first vaccination 20. Facility vaccine data accuracy

Cited by

No citation data available for this registration yet.

Keywords