This project has been shortlisted for the DPH 2025 Innovation Prize – Best Data Driven Innovation
Team: Pratap Kumar, Mohamed Hussein Kahow and Rashed Shah (UCL, Health-E-Net Limited, Save the Children International)
Outline: Paper tally sheets are used to document vaccine administration during immunisation campaigns in Low and Middle-Income Countries (LMICs). Data from tally sheets are delayed (as paper forms are collected from different teams, summarised, transported and transcribed), often have errors due to multiple steps of manual transcription, and need complex efforts to ensure data quality. Data on vaccine administration are therefore not used for operational management of campaigns, and program funders have low confidence on routine data reported from the field. Replacing paper tally sheets with electronic immunisation systems requires significant investment in hardware, software and training, and can be difficult to deploy in challenging field settings like in Somalia.
A “smart” tally sheet (STS) system was piloted in Somalia to address these challenges. STS are paper based, but allow for data entered in tally sheets to be digitised by field staff by taking a photo of structured data fields at the end of vaccine administration activities on each campaign day. AI/ML technology extracts data on the mobile phone (using small, browser-based models running on any phone with a camera and browser) and presents data to users for review before submission. Extracted data are sent to cloud servers either through mobile internet or SMS. The system is intended to deliver near-real time data on field activity with minimal reliance on real-time connectivity, while eliminating the need for manual transcription and expensive data audit processes. Save the Children, through support from the Gates Foundation and in collaboration with Health-E-Net Limited, used STS between May and Oct 2024 in four supplemental immunisation campaigns in Somalia by 5, 15, 45 and 50 vaccination teams in successive campaigns. The first two campaigns were in the Banadir region (around Mogadishu) and the next two also included the more inaccessible Bay, Lower Juba and Mudug regions. Data from a total of 578 STS were submitted during the four campaigns from 50 unique users (vaccination team supervisors). Data from 554 (95.8%) were digitised on the same day as the campaign activity and data available on the campaign dashboard. Errors in the handwriting recognition algorithm were corrected by users in 1.9% of all data fields across all teams and campaigns (n=17,370) and an additional 0.5% corrected by manual review by Health-E-Net. A total of 13,740 children were vaccinated across the four campaigns, including 6,182 identified as zero-dose.
The paper-to-digital approach offers an affordable alternative to generate high-quality, real-time data from immunisation campaigns in LMIC settings. It uses simple, familiar, paper-based tools that can be rolled out with minimal change management across diverse needs and settings. The ability to generate real-time digital data from paper opens opportunities for novel tools such as digital twins of vaccine supply chains to be deployed in challenging public health settings.
