SummaryThe objective of this study was to describe the association between self-reported compliance with last malaria treatment (CMT) and occurrence of malaria during follow-up, controlling for current risk factors. We conducted a prospective open cohort study in Leonislâ ndia, a rural area of Peixoto de Azevedo City, in the Amazon region of Mato Grosso, Brazil. A total of 414 individuals were interviewed at baseline regarding CMT and followed-up for either 8 or 4 months to assess malaria incidence.The associations between CMT and occurrence of malaria were examined through multiple linear regression (when the outcome was malaria episode frequency) or Cox regression (when the outcome was time to malaria onset). Poor CMT (prior to baseline) was identified as an important predictor of the occurrence of subsequent malaria episodes during follow-up among individuals with an indication of being less immune -those whose first malaria episode was relatively recent or those who had an increased number of malaria episodes during the last 2 years. Moreover, surprisingly, it seems that for individuals who are probably more immune (individuals who had experienced their first malaria episode more than 4.5 years previously or those with few or no malaria episodes during the last 2 years), CMT was found to be a poor predictor of increased risk of subsequent malaria. These findings provide compelling evidence for the need to further study CMT and its effect on malaria outcomes.