Search

As malaria incidence decreases and more countries move towards elimination, maps of malaria risk in low-prevalence areas are increasingly needed. For low-burden areas, disaggregation regression models have been developed to estimate risk at high spatial resolution from routine surveillance reports aggregated by administrative unit polygons. However, in areas with both routine surveillance data and prevalence surveys, models that make use of the spatial information from prevalence point-surveys might make more accurate predictions. Using case studies in Indonesia, Senegal and Madagascar, we compare the out-of-sample mean absolute error for two methods for incorporating point-level, spatial information into disaggregation regression models. The first simply fits a binomial-likelihood, logit-link, Gaussian random field to prevalence point-surveys to create a new covariate. The second is a multi-likelihood model that is fitted jointly to prevalence point-surveys and polygon incidence data. We find that in most cases there is no difference in mean absolute error between models. In only one case, did the new models perform the best. More generally, our results demonstrate that combining these types of data has the potential to reduce absolute error in estimates of malaria incidence but that simpler baseline models should always be fitted as a benchmark.

Insecticide-treated nets (ITNs) are one of the most widespread and impactful malaria interventions in Africa, yet a spatially-resolved time series of ITN coverage has never been published. Using data from multiple sources, we generate high-resolution maps of ITN access, use, and nets-per-capita annually from 2000 to 2020 across the 40 highest-burden African countries. Our findings support several existing hypotheses: that use is high among those with access, that nets are discarded more quickly than official policy presumes, and that effectively distributing nets grows more difficult as coverage increases. These results can inform both policy decisions and downstream malaria analyses.

Background Anti-malarial drugs play a critical role in reducing malaria morbidity and mortality, but their role is mediated by their effectiveness. Effectiveness is defined as the probability that an anti-malarial drug will successfully treat an individual infected with malaria parasites under routine health care delivery system. Anti-malarial drug effectiveness (AmE) is influenced by drug resistance, drug quality, health system quality, and patient adherence to drug use; its influence on malaria burden varies through space and time. Methods This study uses data from 232 efficacy trials comprised of 86,776 infected individuals to estimate the artemisinin-based and non-artemisinin-based AmE for treating falciparum malaria between 1991 and 2019. Bayesian spatiotemporal models were fitted and used to predict effectiveness at the pixel-level (5 km × 5 km). The median and interquartile ranges (IQR) of AmE are presented for all malaria-endemic countries. Results The global effectiveness of artemisinin-based drugs was 67.4% (IQR: 33.3–75.8), 70.1% (43.6–76.0) and 71.8% (46.9–76.4) for the 1991–2000, 2006–2010, and 2016–2019 periods, respectively. Countries in central Africa, a few in South America, and in the Asian region faced the challenge of lower effectiveness of artemisinin-based anti-malarials. However, improvements were seen after 2016, leaving only a few hotspots in Southeast Asia where resistance to artemisinin and partner drugs is currently problematic and in the central Africa where socio-demographic challenges limit effectiveness. The use of artemisinin-based combination therapy (ACT) with a competent partner drug and having multiple ACT as first-line treatment choice sustained high levels of effectiveness. High levels of access to healthcare, human resource capacity, education, and proximity to cities were associated with increased effectiveness. Effectiveness of non-artemisinin-based drugs was much lower than that of artemisinin-based with no improvement over time: 52.3% (17.9–74.9) for 1991–2000 and 55.5% (27.1–73.4) for 2011–2015. Overall, AmE for artemisinin-based and non-artemisinin-based drugs were, respectively, 29.6 and 36% below clinical efficacy as measured in anti-malarial drug trials. Conclusions This study provides evidence that health system performance, drug quality and patient adherence influence the effectiveness of anti-malarials used in treating uncomplicated falciparum malaria. These results provide guidance to countries’ treatment practises and are critical inputs for malaria prevalence and incidence models used to estimate national level malaria burden.

Maps of disease burden are a core tool needed for the control and elimination of malaria. Reliable routine surveillance data of malaria incidence, typically aggregated to administrative units, is becoming more widely available. Disaggregation regression is an important model framework for estimating high resolution risk maps from aggregated data. However, the aggregation of incidence over large, heterogeneous areas means that these data are underpowered for estimating complex, non-linear models. In contrast, prevalence point-surveys are directly linked to local environmental conditions but are not common in many areas of the world. Here, we train multiple non-linear, machine learning models on Plasmodium falciparum prevalence point-surveys. We then ensemble the predictions from these machine learning models with a disaggregation regression model that uses aggregated malaria incidences as response data. We find that using a disaggregation regression model to combine predictions from machine learning models improves model accuracy relative to a baseline model.

Background Substantial progress has been made in reducing the burden of malaria in Africa since 2000, but those gains could be jeopardised if the COVID-19 pandemic affects the availability of key malaria control interventions. The aim of this study was to evaluate plausible effects on malaria incidence and mortality under different levels of disruption to malaria control. Methods Using an established set of spatiotemporal Bayesian geostatistical models, we generated geospatial estimates across malaria-endemic African countries of the clinical case incidence and mortality of malaria, incorporating an updated database of parasite rate surveys, insecticide-treated net (ITN) coverage, and effective treatment rates. We established a baseline estimate for the anticipated malaria burden in Africa in the absence of COVID-19-related disruptions, and repeated the analysis for nine hypothetical scenarios in which effective treatment with an antimalarial drug and distribution of ITNs (both through routine channels and mass campaigns) were reduced to varying extents. Findings We estimated 215·2 (95% uncertainty interval 143·7–311·6) million cases and 386·4 (307·8–497·8) thousand deaths across malaria-endemic African countries in 2020 in our baseline scenario of undisrupted intervention coverage. With greater reductions in access to effective antimalarial drug treatment, our model predicted increasing numbers of cases and deaths: 224·1 (148·7–326·8) million cases and 487·9 (385·3–634·6) thousand deaths with a 25% reduction in antimalarial drug coverage; 233·1 (153·7–342·5) million cases and 597·4 (468·0–784·4) thousand deaths with a 50% reduction; and 242·3 (158·7–358·8) million cases and 715·2 (556·4–947·9) thousand deaths with a 75% reduction. Halting planned 2020 ITN mass distribution campaigns and reducing routine ITN distributions by 25%–75% also increased malaria burden to a total of 230·5 (151·6–343·3) million cases and 411·7 (322·8–545·5) thousand deaths with a 25% reduction; 232·8 (152·3–345·9) million cases and 415·5 (324·3–549·4) thousand deaths with a 50% reduction; and 234·0 (152·9–348·4) million cases and 417·6 (325·5–553·1) thousand deaths with a 75% reduction. When ITN coverage and antimalarial drug coverage were synchronously reduced, malaria burden increased to 240·5 (156·5–358·2) million cases and 520·9 (404·1–691·9) thousand deaths with a 25% reduction; 251·0 (162·2–377·0) million cases and 640·2 (492·0–856·7) thousand deaths with a 50% reduction; and 261·6 (167·7–396·8) million cases and 768·6 (586·1–1038·7) thousand deaths with a 75% reduction. Interpretation Under pessimistic scenarios, COVID-19-related disruption to malaria control in Africa could almost double malaria mortality in 2020, and potentially lead to even greater increases in subsequent years. To avoid a reversal of two decades of progress against malaria, averting this public health disaster must remain an integrated priority alongside the response to COVID-19. Funding Bill and Melinda Gates Foundation; Channel 7 Telethon Trust, Western Australia.

Access to healthcare is a requirement for human well-being that is constrained, in part, by the allocation of healthcare resources relative to the geographically dispersed human population(1-3). Quantifying access to care globally is challenging due to the absence of a comprehensive database of healthcare facilities. We harness major data collection efforts underway by OpenStreetMap, Google Maps and academic researchers to compile the most complete collection of facility locations to date. Leveraging the geographically variable strengths of our facility datasets, we use an established methodology(4)to characterize travel time to healthcare facilities in unprecedented detail. We produce maps of travel time with and without access to motorized transport, thus characterizing travel time to healthcare for populations distributed across the wealth spectrum. We find that just 8.9% of the global population (646 million people) cannot reach healthcare within one hour if they have access to motorized transport, and that 43.3% (3.16 billion people) cannot reach a healthcare facility by foot within one hour. Our maps highlight an additional vulnerability faced by poorer individuals in remote areas and can help to estimate whether individuals will seek healthcare when it is needed, as well as providing an evidence base for efficiently distributing limited healthcare and transportation resources to underserved populations both now and in the future. A global analysis generating high-resolution maps of travel time shows that 91.1% of the world's population can reach a hospital or clinic within an hour if they have access to motorized transportation, but only 56.7% can do so by walking, highlighting additional inequities for underserved populations accessing healthcare.

Malaria transmission is influenced by climate, land use and deliberate interventions. Recent declines have been observed in malaria transmission. Here we show that the African continent has witnessed a long-term decline in the prevalence of Plasmodium falciparum from 40% prevalence in the period 1900-1929 to 24% prevalence in the period 2010-2015, a trend that has been interrupted by periods of rapidly increasing or decreasing transmission. The cycles and trend over the past 115 years are inconsistent with explanations in terms of climate or deliberate intervention alone. Previous global initiatives have had minor impacts on malaria transmission, and a historically unprecedented decline has been observed since 2000. However, there has been little change in the high transmission belt that covers large parts of West and Central Africa. Previous efforts to model the changing patterns of P. falciparum transmission intensity in Africa have been limited to the past 15 years or have used maps drawn from historical expert opinions. We provide quantitative data, from 50,424 surveys at 36,966 geocoded locations, that covers 115 years of malaria history in sub-Saharan Africa; inferring from these data to future trends, we would expect continued reductions in malaria transmission, punctuated with resurgences.

SummaryAs malaria incidence decreases and more countries move towards elimination, maps of malaria risk in low prevalence areas are increasingly needed. For low burden areas, disaggregation regression models have been developed to estimate risk at high spatial resolution from routine surveillance reports aggregated by administrative unit polygons. However, in areas with both routine surveillance data and prevalence surveys, models that make use of the spatial information from prevalence point-surveys have great potential. Using case studies in Indonesia, Senegal and Madagascar, we compare two methods for incorporating point-level, spatial information into disaggregation regression models. The first simply fits a Gaussian random field to prevalence point-surveys to create a new covariate. The second is a multi-likelihood model that is fitted jointly to prevalence point-surveys and polygon incidence data. We find that the simple model generally performs better than a baseline disaggregation model while the joint model performance was mixed. More generally, our results demonstrate that combining these types of data improves estimates of malaria incidence.

Additional file 1. Supplementary information on methods, data and findings.

Malaria transmission in Madagascar is highly heterogeneous, exhibiting spatial, seasonal and long-term trends. Previous efforts to map malaria risk in Madagascar used prevalence data from Malaria Indicator Surveys. These cross-sectional surveys, conducted during the high transmission season most recently in 2013 and 2016, provide nationally representative prevalence data but cover relatively short time frames. Conversely, monthly case data are collected at health facilities but suffer from biases, including incomplete reporting and low rates of treatment seeking. We combined survey and case data to make monthly maps of prevalence between 2013 and 2016. Health facility catchment populations were estimated to produce incidence rates from the case data. Smoothed incidence surfaces, environmental and socioeconomic covariates, and survey data informed a Bayesian prevalence model, in which a flexible incidence-to-prevalence relationship was learned. Modelled spatial trends were consistent over time, with highest prevalence in the coastal regions and low prevalence in the highlands and desert south. Prevalence was lowest in 2014 and peaked in 2015 and seasonality was widely observed, including in some lower transmission regions. These trends highlight the utility of monthly prevalence estimates over the four year period. By combining survey and case data using this two-step modelling approach, we were able to take advantage of the relative strengths of each metric while accounting for potential bias in the case data. Similar modelling approaches combining large datasets of different malaria metrics may be applicable across sub-Saharan Africa.

Background There is a need for comprehensive evaluations of the underlying local factors that contribute to residual malaria in sub-Saharan Africa. However, it is difficult to compare the wide array of demographic, socio-economic, and environmental variables associated with malaria transmission using standard statistical approaches while accounting for seasonal differences and nonlinear relationships. This article uses a Bayesian model averaging (BMA) approach for identifying and comparing potential risk and protective factors associated with residual malaria. Results The relative influence of a comprehensive set of demographic, socio-economic, environmental, and malaria intervention variables on malaria prevalence were modelled using BMA for variable selection. Data were collected in Bunkpurugu-Yunyoo, a rural district in northeast Ghana that experiences holoendemic seasonal malaria transmission, over six biannual surveys from 2010 to 2013. A total of 10,022 children between the ages 6 to 59 months were used in the analysis. Multiple models were developed to identify important risk and protective factors, accounting for seasonal patterns and nonlinear relationships. These models revealed pronounced nonlinear associations between malaria risk and distance from the nearest urban centre and health facility. Furthermore, the association between malaria risk and age and some ethnic groups was significantly different in the rainy and dry seasons. BMA outperformed other commonly used regression approaches in out-of-sample predictive ability using a season-to-season validation approach. Conclusions This modelling framework offers an alternative approach to disease risk factor analysis that generates interpretable models, can reveal complex, nonlinear relationships, incorporates uncertainty in model selection, and produces accurate predictions. Certain modelling applications, such as designing targeted local interventions, require more sophisticated statistical methods which are capable of handling a wide range of relevant data while maintaining interpretability and predictive performance, and directly characterize uncertainty. To this end, BMA represents a valuable tool for constructing more informative models for understanding risk factors for malaria, as well as other vector-borne and environmentally mediated diseases. Electronic supplementary material The online version of this article (10.1186/s12936-018-2491-2) contains supplementary material, which is available to authorized users.

Additional file 1. Contains descriptive statistics on covariates, code for running the Gibbs sampler, and additional model outputs.

Background: Understanding the distribution of anopheline vectors of malaria is an important prelude to the design of national malaria control and elimination programmes. A single, geo-coded continental inventory of anophelines using all available published and unpublished data has not been undertaken since the 1960s. Methods: We have searched African, European and World Health Organization archives to identify unpublished reports on anopheline surveys in 48 sub-Saharan Africa countries. This search was supplemented by identification of reports that formed part of post-graduate theses, conference abstracts, regional insecticide resistance databases and more traditional bibliographic searches of peer-reviewed literature. Finally, a check was made against two recent repositories of dominant malaria vector species locations (circa 2,500). Each report was used to extract information on the survey dates, village locations (geo-coded to provide a longitude and latitude), sampling methods, species identification methods and all anopheline species found present during the survey. Survey records were collapsed to a single site over time. Results: The search strategy took years and resulted in 13,331 unique, geo-coded survey locations of anopheline vector occurrence between 1898 and 2016. A total of 12,204 (92%) sites reported the presence of 10 dominant vector species/sibling species; 4,473 (37%) of these sites were sampled since 2005. 4,442 (33%) sites reported at least one of 13 possible secondary vector species; 1,107 (25%) of these sites were sampled since 2005. Distributions of dominant and secondary vectors conform to previous descriptions of the ecological ranges of these vectors. Conclusion: We have assembled the largest ever geo-coded database of anophelines in Africa, representing a legacy dataset for future updating and identification of knowledge gaps at national levels. The geo-coded database is available on Harvard Dataverse as a reference source for African national malaria control programmes planning their future control and elimination strategies.

Additional file 1. Description of statistical models.

Background Logistic regression is a statistical model widely used in cross-sectional and cohort studies to identify and quantify the effects of potential disease risk factors. However, the impact of imperfect tests on adjusted odds ratios (and thus on the identification of risk factors) is under-appreciated. The purpose of this article is to draw attention to the problem associated with modelling imperfect diagnostic tests, and propose simple Bayesian models to adequately address this issue. Methods A systematic literature review was conducted to determine the proportion of malaria studies that appropriately accounted for false-negatives/false-positives in a logistic regression setting. Inference from the standard logistic regression was also compared with that from three proposed Bayesian models using simulations and malaria data from the western Brazilian Amazon. Results A systematic literature review suggests that malaria epidemiologists are largely unaware of the problem of using logistic regression to model imperfect diagnostic test results. Simulation results reveal that statistical inference can be substantially improved when using the proposed Bayesian models versus the standard logistic regression. Finally, analysis of original malaria data with one of the proposed Bayesian models reveals that microscopy sensitivity is strongly influenced by how long people have lived in the study region, and an important risk factor (i.e., participation in forest extractivism) is identified that would have been missed by standard logistic regression. Conclusion Given the numerous diagnostic methods employed by malaria researchers and the ubiquitous use of logistic regression to model the results of these diagnostic tests, this paper provides critical guidelines to improve data analysis practice in the presence of misclassification error. Easy-to-use code that can be readily adapted to WinBUGS is provided, enabling straightforward implementation of the proposed Bayesian models. Electronic supplementary material The online version of this article (doi:10.1186/s12936-015-0966-y) contains supplementary material, which is available to authorized users.

Additional file 1: Code for the Bayesian models used in the main manuscript.

The transmission of malaria across the Arabian Peninsula is governed by the diversity of dominant vectors and extreme aridity. It is likely that where malaria transmission was historically possible it was intense and led to a high disease burden. Here, we review the speed of elimination, approaches taken, define the shrinking map of risk since 1960 and discuss the threats posed to a malaria-free Arabian Peninsula using the archive material, case data and published works. From as early as the 1940s, attempts were made to eliminate malaria on the peninsula but were met with varying degrees of success through to the 1970s; however, these did result in a shrinking of the margins of malaria transmission across the peninsula. Epidemics in the 1990s galvanised national malaria control programmes to reinvigorate control efforts. Before the launch of the recent global ambition for malaria eradication, countries on the Arabian Peninsula launched a collaborative malaria-free initiative in 2005. This initiative led a further shrinking of the malaria risk map and today locally acquired clinical cases of malaria are reported only in Saudi Arabia and Yemen, with the latter contributing to over 98% of the clinical burden.

Additional file 3. Sensitivity analysis to comparing age standardized local maps.

Additional file 1. Spatial covariates maps for study region.

Additional file 3. Sensitivity analysis to comparing age standardized local maps.

Additional file 2. Variable selection, posterior estimates and model comparisons.

Additional file 1. Spatial covariates maps for study region.

Additional file 2. Variable selection, posterior estimates and model comparisons.