Spatiotemporal variations in exposure: Chagas disease in Colombia as a case study.

Age-stratified serosurvey data are often used to understand spatiotemporal trends in disease incidence and exposure through estimating the Force-of-Infection (FoI). Typically, median or mean FoI estimates are used as the response variable in predictive models, often overlooking the uncertainty in estimated FoI values when fitting models and evaluating their predictive ability. To assess how this uncertainty impact predictions, we compared three approaches with three levels of uncertainty integration. We propose a performance indicator to assess how predictions reflect initial uncertainty.In Colombia, 76 serosurveys (1980-2014) conducted at municipality level provided age-stratified Chagas disease prevalence data. The yearly FoI was estimated at the serosurvey level using a time-varying catalytic model. Environmental, demographic and entomological predictors were used to fit and predict the FoI at municipality level from 1980 to 2010 across Colombia.A stratified bootstrap method was used to fit the models without temporal autocorrelation at the serosurvey level. The predictive ability of each model was evaluated to select the best-fit models within urban, rural and (Amerindian) indigenous settings. Model averaging, with the 10 best-fit models identified, was used to generate predictions.Our analysis shows a risk of overconfidence in model predictions when median estimates of FoI alone are used to fit and evaluate models, failing to account for uncertainty in FoI estimates. Our proposed methodology fully propagates uncertainty in the estimated FoI onto the generated predictions, providing realistic assessments of both central tendency and current uncertainty surrounding exposure to Chagas disease.

From serological surveys to disease burden: a modelling pipeline for Chagas disease.

In 2012, the World Health Organization (WHO) set the elimination of Chagas disease intradomiciliary vectorial transmission as a goal by 2020. After a decade, some progress has been made, but the new 2021-2030 WHO roadmap has set even more ambitious targets. Innovative and robust modelling methods are required to monitor progress towards these goals. We present a modelling pipeline using local seroprevalence data to obtain national disease burden estimates by disease stage. Firstly, local seroprevalence information is used to estimate spatio-temporal trends in the Force-of-Infection (FoI). FoI estimates are then used to predict such trends across larger and fine-scale geographical areas. Finally, predicted FoI values are used to estimate disease burden based on a disease progression model. Using Colombia as a case study, we estimated that the number of infected people would reach 506 000 (95% credible interval (CrI) = 395 000-648 000) in 2020 with a 1.0% (95%CrI = 0.8-1.3%) prevalence in the general population and 2400 (95%CrI = 1900-3400) deaths (approx. 0.5% of those infected). The interplay between a decrease in infection exposure (FoI and relative proportion of acute cases) was overcompensated by a large increase in population size and gradual population ageing, leading to an increase in the absolute number of Chagas disease cases over time. This article is part of the theme issue 'Challenges and opportunities in the fight against neglected tropical diseases: a decade from the London Declaration on NTDs'.

Linear and Machine Learning modelling for spatiotemporal disease predictions: Force-of-Infection of Chagas disease.

BACKGROUND: Chagas disease is a long-lasting disease with a prolonged asymptomatic period. Cumulative indices of infection such as prevalence do not shed light on the current epidemiological situation, as they integrate infection over long periods. Instead, metrics such as the Force-of-Infection (FoI) provide information about the rate at which susceptible people become infected and permit sharper inference about temporal changes in infection rates. FoI is estimated by fitting (catalytic) models to available age-stratified serological (ground-truth) data. Predictive FoI modelling frameworks are then used to understand spatial and temporal trends indicative of heterogeneity in transmission and changes effected by control interventions. Ideally, these frameworks should be able to propagate uncertainty and handle spatiotemporal issues. METHODOLOGY/PRINCIPAL FINDINGS: We compare three methods in their ability to propagate uncertainty and provide reliable estimates of FoI for Chagas disease in Colombia as a case study: two Machine Learning (ML) methods (Boosted Regression Trees (BRT) and Random Forest (RF)), and a Linear Model (LM) framework that we had developed previously. Our analyses show consistent results between the three modelling methods under scrutiny. The predictors (explanatory variables) selected, as well as the location of the most uncertain FoI values, were coherent across frameworks. RF was faster than BRT and LM, and provided estimates with fewer extreme values when extrapolating to areas where no ground-truth data were available. However, BRT and RF were less efficient at propagating uncertainty. CONCLUSIONS/SIGNIFICANCE: The choice of FoI predictive models will depend on the objectives of the analysis. ML methods will help characterise the mean behaviour of the estimates, while LM will provide insight into the uncertainty surrounding such estimates. Our approach can be extended to the modelling of FoI patterns in other Chagas disease-endemic countries and to other infectious diseases for which serosurveys are regularly conducted for surveillance.

Heterogeneity of Trypanosoma cruzi infection rates in vectors and animal reservoirs in Colombia: a systematic review and meta-analysis.

BACKGROUND: The heterogeneity of Trypanosoma cruzi infection rates among triatomines insects and animal reservoirs has been studied in independent studies, but little information has been systematised to allow pooled and comparative estimates. Unravelling the main patterns of this heterogeneity could contribute to a further understanding of T. cruzi transmission in Colombia. METHODS: A systematic search was conducted in PubMed, Medline, LILACS, Embase, Web of Knowledge, Google Scholar and secondary sources with no filters of language or time and until April 2018. Based on selection criteria, all relevant studies reporting T. cruzi infection rates in reservoirs or triatomines were chosen. For pooled analyses, a random effects model for binomial distribution was used. Heterogeneity among studies is reported as I2. Subgroup analyses included: taxonomic classification, ecotope and diagnostic methods. Publication bias and sensitivity analyses were performed. RESULTS: Overall, 39 studies reporting infection rates in Colombia were found (22 for potential reservoirs and 28 for triatomine insects) for a total sample of 22,838 potential animals and 11,307 triatomines evaluated for T. cruzi infection. We have found evidence of 38/71 different animal species as potential T. cruzi reservoirs and 14/18 species as triatomine vectors for T. cruzi. Among animals, the species with the highest pooled prevalence were opossum (Didelphis marsupialis) with 48.0% (95% CI: 26-71%; I2 = 88%, τ2 = 0.07, P < 0.01) and domestic dog (Canis lupus familiaris) with 22.0% (95% CI: 4-48%; I2 = 96%, τ2 = 0.01, P < 0.01). Among triatomines, the highest prevalence was found for Triatoma maculata in the peridomestic ecotope (68.0%, 95% CI: 62-74%; I2 = 0%, τ2 = 0, P < 0.0001), followed by Rhodnius prolixus (62.0%, 95% CI: 38-84%; I2 = 95%, τ2 = 0.05, P < 0.01) and Rhodnius pallescens (54.0%, 95% CI: 37-71%; I2 = 86%, τ2 = 0.035, P < 0.01) in the sylvatic ecotope. CONCLUSIONS: To our knowledge, this is the first systematic and quantitative analyses of triatomine insects and potential animal reservoirs for T. cruzi infection in Colombia. The results highlight a marked heterogeneity between species and provide initial estimates of infection rates heterogeneity.