Table 1.
2×2 descriptive tables for S. mansoni infection as determined by CCA and microscopy over time; sensitivity and specificity of CCA were estimated assuming that the combination of 6 KK measurements over 3 days is 100% sensitive and 100% specific.
Table 2.
LMM estimated sensitivities and specificities over time (with 95% confidence intervals) with no gold standard assumed.
Table 3.
LMM estimated transition probability matrices.
Figure 1.
Latent Markov modelling path diagram.
The variables in boxes represent the four observed categorical indicators of the latent categorical variables C at each time point t. The three arrows between the circled variables indicate the regression model for the latent categorical variable at time point t on the latent categorical variable at time point t-1.
Figure 2.
Posterior distributions of latent variables conditioned on the responses of the CCA test for children (n = 167).
The latent variable represents a continuum of the infection severity characterized from the left to the right as to low or no infection up to high infection. The obtained scores in the horizontal axis are linked to ranking and not necessarily to the absolute displayed values.
Figure 3.
Posterior distributions of latent variables conditioned on the responses of the CCA test for adolescents and adults (n = 273).
The latent variable represents a continuum of the infection severity characterized from the left to the right as to low or no infection up to high infection. The obtained scores in the horizontal axis are linked to ranking and not necessarily to the absolute displayed values.
Figure 4.
LMM estimated “true” S. mansoni prevalence for both age groups (ages 7–16 and 17–76) with 95% confidence intervals over time.
Table 4.
Estimated S. mansoni prevalences based on 6 KK measurements and based on the LMM.