Antibody and Antigen Prevalence as Indicators of Ongoing Transmission or Elimination of Visceral Leishmaniasis: A Modeling Study

Abstract

Background Control of visceral leishmaniasis (VL) on the Indian subcontinent has been highly successful. Control efforts such as indoor residual spraying and active case detection will be scaled down or even halted over the coming years. We explored how after scale-down, potential recurrence of VL cases may be predicted based on population-based surveys of antibody or antigenemia prevalence. Methods Using a stochastic age-structured transmission model of VL, we predicted trends in case incidence and biomarker prevalence over time after scaling down control efforts when the target of 3 successive years without VL cases has been achieved. Next, we correlated biomarker prevalence with the occurrence of new VL cases within 10 years of scale-down. Results Occurrence of at least 1 new VL case in a population of 10 000 was highly correlated with the seroprevalence and antigenemia prevalence at the moment of scale-down, or 1 or 2 years afterward. Receiver operating characteristic curves indicated that biomarker prevalence in adults provided the most predictive information, and seroprevalence was a more informative predictor of new VL cases than antigenemia prevalence. Thresholds for biomarker prevalence to predict occurrence of new VL cases with high certainty were robust to variation in precontrol endemicity. Conclusions The risk of recrudescence of VL after scaling down control efforts can be monitored and mitigated by means of population-based surveys. Our findings highlight that rapid point-of-care diagnostic tools to assess (preferably) seroprevalence or (otherwise) antigenemia in the general population could be a key ingredient of sustainable VL control.

Two-stage lot quality assurance sampling framework for monitoring and evaluation of neglected tropical diseases, allowing for imperfect diagnostics and spatial heterogeneity

“…Here, we assume that the target prevalence threshold T is defined in terms of the true prevalence π i (not as measured by some imperfect diagnostic test). We further assume that the value of T is appropriate for scaling interventions up or down, which may [ 7 ] or may not be the case [ 8 – 10 ].…”

“…Given a program decision threshold T = 50% (vertical straight line), we can now deduce the error probability of unnecessarily continuing or upscaling the intervention at a frequency or intensity that is greater than needed ( ε overtreat ) when true prevalence π i < T , and the error probability of prematurely reducing the frequency or intensity of interventions ( ε undertreat ) when true prevalence π i ≥ T . These error probabilities can be considered to be community-level analogues of one minus the negative predictive value and one minus the positive predicted value [ 5 ], as used in recent NTD modelling exercises [ 8 , 11 ]. In Fig 2B , we deduce the decision cut-off c for which ε overtreat and ε undertreat do not exceed acceptable risk levels for arbitrary choices of π i < T and π i ≥ T .…”

“…The determination of the decision cut-off c is straightforward in a 1-stage LQAS ignoring the spatial heterogeneity of infections across clusters [ 8 ], but there is no analytical solution for a 2-stage LQAS allowing for spatial heterogeneity [ 6 ]. We therefore applied a Monte Carlo (MC) simulation framework to determine the decision cut-off c .…”

Assessment of the required performance and the development of corresponding program decision rules for neglected tropical diseases diagnostic tests: Monitoring and evaluation of soil-transmitted helminthiasis control programs as a case study

“…Given a program decision threshold T of 50% (vertical straight line), we can deduce both the error related to unnecessarily selecting a PC frequency that is greater than needed ( ε overtreat ) or prematurely reducing the frequency of PC ( ε undertreat ). These errors are analogous to 1 minus the negative predictive value and 1 minus the positive predicted value, as used in recent NTD modelling studies on optimal program decision thresholds [ 25 – 27 ]. Subsequently, we can also deduce to what extent this diagnostic test allows for reliable decision making.…”

A general framework to support cost-efficient fecal egg count methods and study design choices for large-scale STH deworming programs–monitoring of therapeutic drug efficacy as a case study

“…This is because the suitability and the cost of different survey designs and diagnostic techniques will vary by epidemiological setting [17]. For instance, the probability of making correct policy decisions may strongly depend on both the performance of a particular diagnostic method and the associated decision criterion in a particular epidemiological setting [18][19][20][21][22][23]. For STH, this performance depends on the average intensity of infection in a community as well as the level of variation in egg excretion (between individuals and within individuals over time), and in the case of the evaluation of drug efficacy, variation in individual drug responses [20,23].…”