It is estimated that only 86% of children worldwide receive the complete three doses of major vaccines, despite global coverage. Identifying data gaps in vaccination coverage will deepen understanding of the barriers for these left-out communities and ultimately support better vaccination access around the world. This project models vaccination recall, caregiver provided data used when a child’s vaccination card is not available. Recall is only 53% accurate, yet UNICEF vaccination coverage statistics are calculated using this unreliable data. We use machine-learning techniques to model conditions of caregivers without cards, supporting card loss prevention and more accurate vaccination coverage data.