Identificação e descrição de grupos de risco de crianças nunca vacinadas em países de baixa e média renda por meio de árvores de decisão

Abstract

In order to “leave no one behind”, the Immunization Agenda 2030 targets reduction in vaccination inequalities as a goal. Decision trees are robust statistical tools that can be used for identifying risk groups of non-vaccination in children, or zero dose children, from low- and middle-income countries (LMIC). Classification and Regression Trees were created for 92 LMICs and for all countries combined using Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS) in order to identify zero dose risk groups among children aged 12-23 months. Risk groups were characterized in terms of wealth, mother’s education, place of residence and sex. Three risk groups were identified and the one with the highest zero dose prevalence (42%) represented 4% of all children, but they accounted for one in every four zerodose children. This group was composed of children whose mother did not received any doses of the tetanus vaccine, who did not have any antenatal care visits and whose delivery was not in a health facility. The higher the zero-dose prevalence in a group, the poorer, more rural and with a less educated mother the children in that group. National trees had a similar result. There is an opportunity for better integration between those services in order to achieve the Sustainable Development Goals. Decision trees have shown to be a promising technique for identifying risk groups in epidemiological research.