Table 1.
Model parameters.
Table 2.
TRANSFIL model event definitions with the stochastic simulation framework.
Where necessary, the subscript i indexes individual human hosts in the population.
Table 3.
MAP estimates of parameters, with 95% credible intervals in brackets, and estimated R0 (R0 is defined as the average number of female parasites produced by a single female parasite over the course of an average lifespan, that themselves survive to have offspring, in the absence of density dependent effects).
Fig 1.
Mf prevalence time-series for (a) Malindi (Kenya) with 7 annual MDA rounds (b) Ngahmbule (PNG) with 8 annual MDA rounds and (c) Yauatong (PNG) with 9 annual MDA rounds. The first round of MDA occurs at year 1 in each setting. The blue lines are the median of the distribution, the shaded regions represent 95% of the distribution. Red dashed line represents 1% Mf stopping threshold.
Fig 2.
Aggregation of simulated Mf counts after each round of MDA, estimated using maximum likelihood estimation (MLE).
Markers represent the mean MLE from 1000 realisations. Round 0 corresponds to the aggregation estimated at baseline.
Fig 3.
Snapshots (made using kernel density estimation) of the Mf prevalence distributions (green- realisations that eliminate, blue–realisations that bounce-back), at different times for fit to Malindi, Kenya data a) at baseline (pre-intervention) b) 6 months post intervention c) 5 years post intervention d) 40 years post MDA cessation.
Fig 4.
Snapshots (made using kernel density estimation) of the Mf prevalence distributions (green- realisations that eliminate, blue–realisations that bounce-back), at different times for fit to Ngahmbule, PNG data a) at baseline (pre-intervention) b) 6 months post intervention c) 5 years post intervention d) 40 years post MDA cessation.
Fig 5.
Negative predictive values (for predicting bounce-back, given the measurement is above the threshold) and positive predictive values (for predicting elimination 40 years post MDA cessation, given the measurement is below the threshold), at different antigenemia (6–7 year olds) and filaremia monitoring thresholds, and at different measurement times (shown in varying colours).
Parameter setting—Malindi, Kenya. Bars show 95% confidence intervals. To aid visualization, the bars have been offset in the x-axis -0.1% to 0.1% in steps of 0.04%.
Fig 6.
Negative predictive values (for predicting bounce-back, given the measurement is above the threshold) and positive predictive values (for predicting elimination 40 years post MDA cessation, given the measurement is below the threshold), at different antigenemia (6–7 year olds) and filaremia monitoring thresholds, and at different measurement times (shown in varying colours).
Parameter setting—Ngahmbule, PNG. Bars show 95% confidence intervals calculated using estimates of the standard error, which depends on the number of realisations above or below the threshold. To aid visualization, the bars have been offset in the x-axis -0.1% to 0.1% in steps of 0.04%.
Fig 7.
Time-series of Mf prevalence after a single round of MDA, illustrating how the mean lifespan of adult worms affects the speed of bounce-back.
Fig 8.
Predictive values (left: NPVs for predicting bounce-back, right: PPVs for predicting elimation), for TAS prevalence measurement performed 2 years post MDA cessation, Malindi setting, with mean worm lifespans ranging from 5 years (dashed line) - 9 years (solid line).
Fig 9.
Predictive values (left: NPVs for predicting bounce-back, right: PPVs for predicting elimination), for TAS prevalence measurement performed 2 years post MDA cessation, Ngamhbule setting, with mean worm lifespans ranging from 5 years (dashed line) - 9 years (solid line).