Fig. 2From: Modelling longitudinal binary outcomes with outcome dependent observation times: an application to a malaria cohort studyEstimated marginal prevalence for children assumed to be born in the baseline year (2001) by age, study site and symptomatic (top row) or asymptomatic (bottom row) at the previous visit, obtained by fitting a naive model ignoring ODS (while relying on routine and clinical data) (Scenario 2 model) and fitting the joint model which uses all data and accommodate ODS (Scenario 4 model). Left to right column: Walukuba, Kihihi and NagongeraBack to article page