Abstract
Home-based records (HBRs) are often considered the "gold-standard" for determining vaccination status - especially for vaccination coverage surveys (VCS). However, data on the quality and usability of HBRs when collected at the source (household) as opposed to photos for recording at a central level is scarce. This study explores the usability of HBR photographs taken during the 2022 VCS of the Democratic Republic of the Congo (DRC). It compares results from analyses of HBR transcribed in the field and those entered from corresponding photos sent to a central server in the assessment of vaccine coverage. This analysis included a random sample of 2768 children aged 12-23 months who had both HBR data entered at the time of the survey and a corresponding photo. Using the card photo, a team of 6 independent researchers transcribed the cards and assessed HBR characteristics based on a predefined set of criteria related to data quality, entry and HBR quality. Data were summarized using proportions; agreement was assessed using concordance and Kappa-Cohen statistics. Over 90 % of selected children had photos considered adequate. Most HBRs met the set quality criteria in terms of physical condition and legibility. The majority (84.6 %) included the Ministry of Health logo and listed all the vaccines in the routine immunization schedule. Concordance and the Kappa-Cohen statistic showed high levels of variability of agreement by vaccine between the two sources of data. This study illustrates that entering vaccine data using photos of HBR at a central level versus at the time of survey is feasible and can provide accurate results with moderate agreement. Further, this approach can also reduce the time per survey in the field, if interviewers are only required to take a photograph. Additionally, having an HBR photo available can be useful for secondary analyses and further training.