Quantifying the value of surveillance data for improving model predictions of lymphatic filariasis elimination

Abstract

BackgroundMathematical models are increasingly being used to evaluate strategies aiming to achieve the control or elimination of parasitic diseases. Recently, owing to growing realization that process-oriented models are useful for ecological forecasts only if the biological processes are well defined, attention has focused on data assimilation as a means to improve the predictive performance of these models.Methodology and principal findingsWe report on the development of an analytical framework to quantify the relative values of various longitudinal infection surveillance data collected in field sites undergoing mass drug administrations (MDAs) for calibrating three lymphatic filariasis (LF) models (EPIFIL, LYMFASIM, and TRANSFIL), and for improving their predictions of the required durations of drug interventions to achieve parasite elimination in endemic populations. The relative information contribution of site-specific data collected at the time points proposed by the WHO monitoring framework was evaluated using model-data updating procedures, and via calculations of the Shannon information index and weighted variances from the probability distributions of the estimated timelines to parasite extinction made by each model. Results show that data-informed models provided more precise forecasts of elimination timelines in each site compared to model-only simulations. Data streams that included year 5 post-MDA microfilariae (mf) survey data, however, reduced each model’s uncertainty most compared to data streams containing only baseline and/or post-MDA 3 or longer-term mf survey data irrespective of MDA coverage, suggesting that data up to this monitoring point may be optimal for informing the present LF models. We show that the improvements observed in the predictive performance of the best data-informed models may be a function of temporal changes in inter-parameter interactions. Such best data-informed models may also produce more accurate predictions of the durations of drug interventions required to achieve parasite elimination.SignificanceKnowledge of relative information contributions of model only versus data-informed models is valuable for improving the usefulness of LF model predictions in management decision making, learning system dynamics, and for supporting the design of parasite monitoring programmes. The present results further pinpoint the crucial need for longitudinal infection surveillance data for enhancing the precision and accuracy of model predictions of the intervention durations required to achieve parasite elimination in an endemic location.

Predicting lymphatic filariasis elimination in data-limited settings: A reconstructive computational framework for combining data generation and model discovery

“…This suggests that the observed loss of information, while statistically relevant, may not be practically meaningful. This conclusion complements previous findings that predictive performance of LF models is improved by calibrating the model to baseline and monitoring data together [18]. In , for example, we show that data-informed models provide more precise predictions than unconstrained models and that it is valuable to also include post-intervention data in addition to baseline data to the model calibration procedure [18].…”

“…This constitutes a highly useful capability for disease management applications, since the discovery of information about pre-control prevalence in settings without such data is essential to policy makers for planning necessary interventions, determining the optimal frequency of monitoring activities, evaluating program progress, and applying for formal validation of elimination from the WHO [4,5,26,53]. This work has also provided additional insights regarding the value of data for refining parasite transmission models [13,[16][17][18]22]. We found that there was a loss of information (as measured by the Shannon information index) when we calibrated the model to mid-MDA data and hindcasted the initial conditions as opposed to fitting the model directly to baseline data.…”

“…This requirement implies that these models should accurately reflect the key biological and ecological processes underpinning local parasite transmission through the reliable specification of initial conditions, parameterizations, and the functional forms describing specific transmission processes whose uncertainties are well-defined [13][14][15]. A growing theme is also how best to discover or infer these input and latent variables from external observations so that information contained within data can be reliably captured for mathematically reconstructing local transmission dynamics [13,[16][17][18]. This requirement suggests that improving model predictive performance will ultimately depend on the development and application of computational discovery systems that can construct suitable models from process information contained in empirical data [19].…”

“…lymphatic filariasis (LF), a highly debilitating disease which is targeted for global elimination by the World Health Organization (WHO) [9,10,20,21]. This work has demonstrated how, in addition to improving model accuracy, data assimilation can also significantly decrease the uncertainty associated with model predictions, highlighting a clear benefit of using these approaches for policy applications that require both accuracy and precision [13,16,18,22].…”

Data-driven modelling and spatial complexity supports heterogeneity-based integrative management for eliminating Simulium neavei-transmitted river blindness

“…The above conclusions are clearly conditioned on the ability of our modelling framework to reliably capture the dominant transmission dynamics of S. neavei -transmitted onchocerciasis in an endemic focus. While our BM-based modelling algorithm combines the advantageous features of mechanistic and statistical approaches to improve the estimation of local models for facilitating forecasts of interventions applied under a variety of field conditions, it is dependent, as for any data-driven predictive system, on the model structure employed, estimation procedure, and on the data used for facilitating model discovery 16,17 . Although our BM framework primarily focused attention on addressing parameter uncertainty with data, we note here firstly that the present model is based on previously established population models of onchocerciasis transmission 7,47 , with appropriate structural extensions made with regard to population-averaged mf uptake and larval development in the S. neavei vector host as well as the operation of different forms of host immunity in populations 48 .…”