Factors in time to malaria treatment in the Brazilian Amazon: a survival analysis

Background: Malaria continues to be one of the most relevant infectious diseases in Brazil and globally, with over 150.000 cases in 2019 alone. Due to high P. vivax prevalence and P. falciparum elimination goals, the Brazilian Malaria Elimination Program placed a high priority on timely treatment. Cases in the country are concentrated in the Legal Amazon region, with diagnosis and treatment provided for free by the government through its public health system, but many relevant questions about heterogeneity in risk and access to treatment remain. Methods: We use data from the Malaria Epidemiological Surveillance System, maintained by the Ministry of Health, from 2007-2019 to analyze factors that affect time to treatment of malaria in three distinct periods. We use Kaplan-Meier survival functions to estimate the interval between the appearance of first symptoms and beginning of treatment for several administrative, demographic, social and health variables. Results: Despite advances in malaria control over the last two decades, the program has not achieved its timely treatment goals in any of the periods analyzed. Malaria risk was highest among working-age adults, concentrated in a few high risk municipalities, autochthonously transmitted, in agriculture and domestic services, lower education, mid to low levels of parasitaemia, brown, indigenous and black populations and males. The main parasite is P. vivax and the majority of cases is still passively detected. Time-to-treatment was lower when they were detected actively, had lower levels of schooling, were not pregnant or past the 2nd trimester, between 0 and 14 years of age, worked in agriculture, had low parasite counts, were infected with P. falciparum, were indigenous people and lived in high risk municipalities. Conclusions: Differences in access to treatment were small if compared to differences in risk of acquiring malaria in the first place, but active case detection, municipal risk and race showed significant differences and potential avenues for intervention.