Predicting malaria epidemics in Burkina Faso with machine learning.
Ontology highlight
ABSTRACT:
SUBMITTER: Harvey D
PROVIDER: S-EPMC8213140 | biostudies-literature | 2021
REPOSITORIES: biostudies-literature
Publications
Predicting malaria epidemics in Burkina Faso with machine learning.
Harvey David D Valkenburg Wessel W Amara Amara A
PloS one 20210618 6
Accurately forecasting the case rate of malaria would enable key decision makers to intervene months before the onset of any outbreak, potentially saving lives. Until now, methods that forecast malaria have involved complicated numerical simulations that model transmission through a community. Here we present the first data-driven malaria epidemic early warning system that can predict the 13-week case rate in a primary health facility in Burkina Faso. Using the extraordinarily high-fidelity data ...[more]
Similar Datasets
Project description:Malaria is a parasitic disease, endemic in many tropical and sub-tropical countries. Malaria is a well-known disease, familiar to almost all people in endemic regions, as they or their family are regularly confronted with it; everyone in these regions has probably experienced the disease, at least once in their life. To investigate the social perceptions of malaria in Burkina Faso, including its diagnosis-driven treatment, we have conducted a survey in both urban (Saint Camille Hospital, Ouagadougou HOSCO) and rural (Boussé Hospital) areas. Fifty-six individuals, mostly representatives of the society variability, were surveyed by questionnaires and 2 focus groups were organized with traditional healers. In general, populations seem to have grasped the causes, symptoms and means of preventing the disease. However, the majority of interviewees make a marked confusion between malaria and dengue; dengue fever is considered like a severe form of malaria. The care modalities (modern and/or traditional medicine) are plural and the choice of therapeutic practice depends on both the socio-economic conditions and education level of the patient. Whereas some patients mark preferences for one type of medicine, others simultaneously recourse to both; for these, a medicine does not outperform the other and their combination multiplies the chances of a quick recovery. Whether for modern or traditional medicine, the diagnosis is considered very important for effective disease management. Modern medicine uses diagnostic tools based on light microscopy and immunochromatography (rapid diagnostic tests; RDT); traditional medicine has its own diagnostic logic but nevertheless recognizes modern medicine diagnosis to guide its therapy. 90 % of those interviewed first use modern medicine to seek an accurate diagnosis of their disease and thus to receive adequate treatment. Presumptive treatments are still widely prescribed and accepted by most patients who trust the judgment of their caregiver, not perceiving any benefit to an objective diagnosis. In front of a negative diagnosis, patient reactions are diverse, some accepting investigations for other diseases (45 %), others opting for self-medication (15 %), others resorting to traditional medicine (20 %). All are unanimous in the importance of diagnosis and are in favor of in-development diagnostic technologies, provided these obviously meet the features of reliability, ease of use, availability and, of course, economical accessibility.
| S-EPMC7797373 | biostudies-literature
Project description:BackgroundImproving the knowledge and understanding of the environmental determinants of malaria vector abundance at fine spatiotemporal scales is essential to design locally tailored vector control intervention. This work is aimed at exploring the environmental tenets of human-biting activity in the main malaria vectors (Anopheles gambiae s.s., Anopheles coluzzii and Anopheles funestus) in the health district of Diébougou, rural Burkina Faso.MethodsAnopheles human-biting activity was monitored in 27 villages during 15 months (in 2017-2018), and environmental variables (meteorological and landscape) were extracted from high-resolution satellite imagery. A two-step data-driven modeling study was then carried out. Correlation coefficients between the biting rates of each vector species and the environmental variables taken at various temporal lags and spatial distances from the biting events were first calculated. Then, multivariate machine-learning models were generated and interpreted to (i) pinpoint primary and secondary environmental drivers of variation in the biting rates of each species and (ii) identify complex associations between the environmental conditions and the biting rates.ResultsMeteorological and landscape variables were often significantly correlated with the vectors' biting rates. Many nonlinear associations and thresholds were unveiled by the multivariate models, for both meteorological and landscape variables. From these results, several aspects of the bio-ecology of the main malaria vectors were identified or hypothesized for the Diébougou area, including breeding site typologies, development and survival rates in relation to weather, flight ranges from breeding sites and dispersal related to landscape openness.ConclusionsUsing high-resolution data in an interpretable machine-learning modeling framework proved to be an efficient way to enhance the knowledge of the complex links between the environment and the malaria vectors at a local scale. More broadly, the emerging field of interpretable machine learning has significant potential to help improve our understanding of the complex processes leading to malaria transmission, and to aid in developing operational tools to support the fight against the disease (e.g. vector control intervention plans, seasonal maps of predicted biting rates, early warning systems).
| S-EPMC8243492 | biostudies-literature
Project description:The aim of this study was to analyze the spatio-temporal distribution and determinants of the 2017 dengue epidemic in Burkina Faso. A principal component analysis of meteorological and environmental factors was performed to reduce dimensions and avoid collinearities. An initial generalized additive model assessed the impact of the components derived from this analysis on dengue incidence. Dengue incidence increased mainly with relative humidity, precipitation, normalized difference vegetation index and minimum temperature with an 8-week lag. A Kulldoff Satscan scan was used to identify high-risk dengue clusters, and a second generalized additive model assessed the risk of a health area being at high risk according to land-use factors. The spatio-temporal distribution of dengue fever was heterogeneous and strongly correlated with meteorological factors. The rural communes of Sabaa and Koubri were the areas most at risk. This study provides useful information for planning targeted dengue control strategies in Burkina Faso.
| S-EPMC10675449 | biostudies-literature
Project description:BackgroundIn the context of widespread mosquito resistance to currently available pesticides, novel, precise genetic vector control methods aimed at population suppression or trait replacement are a potentially powerful approach that could complement existing malaria elimination interventions. Such methods require knowledge of vector population composition, dynamics, behaviour and role in transmission. Here were characterized these parameters in three representative villages, Bana, Pala and Souroukoudingan, of the Sudano-Sahelian belt of Burkina Faso, a region where bed net campaigns have recently intensified.MethodsFrom July 2012 to November 2015, adult mosquitoes were collected monthly using pyrethroid spray catches (PSC) and human landing catches (HLC) in each village. Larval habitat prospections assessed breeding sites abundance at each site. Mosquitoes collected by PSC were identified morphologically, and then by molecular technique to species where required, to reveal the seasonal dynamics of local vectors. Monthly entomological inoculation rates (EIR) that reflect malaria transmission dynamics were estimated by combining the HLC data with mosquito sporozoite infection rates (SIR) identified through ELISA-CSP. Finally, population and EIR fluctuations were fit to locally-collected rainfall data to highlight the strong seasonal determinants of mosquito abundance and malaria transmission in this region.ResultsThe principal malaria vectors found were in the Anopheles gambiae complex. Mosquito abundance peaked during the rainy season, but there was variation in vector species composition between villages. Mean survey HLC and SIR were similar across villages and ranged from 18 to 48 mosquitoes/person/night and from 3.1 to 6.6% prevalence. The resulting monthly EIRs were extremely high during the rainy season (0.91-2.35 infectious bites/person/day) but decreased substantially in the dry season (0.03-0.22). Vector and malaria transmission dynamics generally tracked seasonal rainfall variations, and the highest mosquito abundances and EIRs occurred in the rainy season. However, despite low residual mosquito populations, mosquitoes infected with malaria parasites remained present in the dry season.ConclusionThese results highlight the important vector control challenge facing countries with high EIR despite the recent campaigns of bed net distribution. As demonstrated in these villages, malaria transmission is sustained for large parts of the year by a very high vector abundance and high sporozoite prevalence, resulting in seasonal patterns of hyper and hypo-endemicity. There is, therefore, an urgent need for additional vector control tools that can target endo and exophillic mosquito populations.
| S-EPMC6444393 | biostudies-literature
Project description:BACKGROUND:Populations of the Anopheles gambiae complex are found during the rainy season throughout West Africa, even in arid areas with long dry seasons during which mosquitoes appear to be absent. Several hypotheses have been proposed to explain this apparent paradox, including aestivation, dispersal between neighbouring settlements, and long distance migration using high-altitude wind currents. METHODS:An individual-based, spatially explicit model of mosquito populations was developed for a region of West Africa centred on, and including all of, Burkina Faso. Populations associated with human settlements were linked by dispersal and the model incorporated geospatial data on the distribution of settlements, water bodies and rainfall. RESULTS:Local dispersal (at rates consistent with experimental data) was necessary to explain observed patterns of rainy season populations across all of the simulation area, but by itself failed to account for the presence of populations in the arid North (the Sahel). The presence of rare dry-season larval sites could explain these northern populations, but seems inconsistent with field surveys. Aestivation by female mosquitoes explained rainy-season populations in all but the very sparsest and driest areas of human habitation, while long-distance migration based on annual wind patterns could account for all observed populations. CONCLUSIONS:Modelling studies such as this can help assess the potential validity of different hypotheses and suggest priority areas for experimental study. In particular, the results highlight a shortage of empirical research on mosquito dispersal between neighbouring settlements, which may be critically important to the continued presence of many mosquito populations in West Africa. Further research that establishes the extent to which mosquitoes aestivate, and migrate using high altitude winds, is also much needed to understand Sahelian mosquito populations.
| S-EPMC5879775 | biostudies-literature
Project description:BackgroundEffective and timely treatment is an essential aspect of malaria control, but remains a challenge in many parts of sub-Saharan Africa. The objective of this study was to describe young children's access to malaria treatment in Nouna Health District, Burkina Faso.MethodsIn February/March 2006, a survey was conducted in a representative sample of 1,052 households.ResultsOverall 149/1052 (14%) households reported the current possession of anti-malarial medicine, which was significantly associated with urban area, literacy of household head, having young children, and high socio-economic status. Out of a total of 802 children under five years, at least one malaria episode was reported for 239 (30%) within the last month. Overall 95% of children received treatment, either modern (72%), traditional (18%) or mixed (5%). Most of the medicines were provided as home treatment by the caregiver and half of children received some type of modern treatment within 24 hours of the occurrence of first symptoms. Despite a recent policy change to artemisinin-based combination therapy, modern anti-malarials consisted mainly of chloroquine (93%). Modern drugs were obtained more often from a health facility in localities with a health facility compared to those without (60% vs. 25.6%, p < 0.001). In contrast, beside informal providers, volunteer community health workers (CHW) were the main source of modern medicine in localities without a health centre (28% vs. 3%, p < 0.001).ConclusionAccess to modern health services providing quality controlled effective combination therapies against malaria needs to be strengthened in rural Africa, which should include a re-investigation of the role of CHW 30 years after Alma Ata.
| S-EPMC2790466 | biostudies-literature
Project description:BACKGROUND:Given the scarcity of resources in developing countries, malaria treatment requires new strategies that target specific populations, time periods and geographical areas. While the spatial pattern of malaria transmission is known to vary depending on local conditions, its temporal evolution has yet to be evaluated. The aim of this study was to determine the spatio-temporal dynamic of malaria in the central region of Burkina Faso, taking into account meteorological factors. METHODS:Drawing on national databases, 101 health areas were studied from 2011 to 2015, together with weekly meteorological data (temperature, number of rain events, rainfall, humidity, wind speed). Meteorological factors were investigated using a principal component analysis (PCA) to reduce dimensions and avoid collinearities. The Box-Jenkins ARIMA model was used to test the stationarity of the time series. The impact of meteorological factors on malaria incidence was measured with a general additive model. A change-point analysis was performed to detect malaria transmission periods. For each transmission period, malaria incidence was mapped and hotspots were identified using spatial cluster detection. RESULTS:Malaria incidence never went below 13.7 cases/10,000 person-weeks. The first and second PCA components (constituted by rain/humidity and temperatures, respectively) were correlated with malaria incidence with a lag of 2 weeks. The impact of temperature was significantly non-linear: malaria incidence increased with temperature but declined sharply with high temperature. A significant positive linear trend was found for the entire time period. Three transmission periods were detected: low (16.8-29.9 cases/10,000 person-weeks), high (51.7-84.8 cases/10,000 person-weeks), and intermediate (26.7-32.2 cases/10,000 person-weeks). The location of clusters identified as high risk varied little across transmission periods. CONCLUSION:This study highlighted the spatial variability and relative temporal stability of malaria incidence around the capital Ouagadougou, in the central region of Burkina Faso. Despite increasing efforts in fighting the disease, malaria incidence remained high and increased over the period of study. Hotspots, particularly those detected for low transmission periods, should be investigated further to uncover the local environmental and behavioural factors of transmission, and hence to allow for the development of better targeted control strategies.
| S-EPMC5879937 | biostudies-literature
Project description:BackgroundEvery year 40,000 people die of malaria in Burkina Faso. In 2010, the Burkinabè authorities implemented a national malaria control program that provides for the distribution of mosquito nets and the home-based treatment of children with fever by community health workers. The objective of this study was to measure the implementation fidelity of this program.MethodsWe conducted a case study in two comparable districts (Kaya and Zorgho). Data were collected one year after the program's implementation through field observations (10 weeks), documentary analysis, and individual interviews with stakeholders (n = 48) working at different levels of the program. The analysis framework looked at the fidelity of (i) the intervention's content, (ii) its coverage, and (iii) its schedule.ResultsThe program's implementation was relatively faithful to what was originally planned and was comparable in the two districts. It encountered certain obstacles in terms of the provision of supplies. Coverage fidelity was better in Kaya than in Zorgho, where many community health workers (CHW) experienced problems with the restocking of artemisinin-based combination therapy and with remuneration for periods of training. In both districts, the community was rarely involved in the process of selecting CHWs. The components affected by scheduling all experienced successive implementation delays that pushed nets distribution and the initial provision of artemisinin-based combination therapies to the CHWs past the 2010 malaria season.ConclusionsThe activities intended by the program were mostly implemented with good fidelity. However, the implementation was plagued by delays that probably postponed the expected beneficial effects.
| S-EPMC3724672 | biostudies-literature
Project description:BackgroundMalaria in pregnancy can result in placental infection with fetal implications. This study aimed at assessing placental malaria (PM) prevalence and its associated factors in a cohort of pregnant women with peripheral malaria and their offspring.MethodThe data were collected in the framework of a clinical trial on treatments for malaria in pregnant women . Placental malaria (PM) was diagnosed by histopathological detection of parasites and/or malaria pigment on placenta biopsies taken at delivery. Factors associated with PM were assessed using logistic regression.ResultsOut of 745 biopsies examined, PM was diagnosed in 86.8 % of women. Acute, chronic and past PM were retrieved in 11 (1.5 %), 170 (22.8 %), and 466 (62.6 %) women, respectively. A modifying effect was observed in the association of gravidity or anemia at the study start with pooled PM (presence of parasites and/or malaria pigment). In women under 30, gravidity ≤ 2 was associated with an increased prevalence of pooled PM but in women aged 30 years or more, gravidity was no more associated with pooled PM (OR 6.81, 95 % CI 3.18 - 14.60; and OR 0.52, 95 % CI 0.10 - 2.76, respectively). Anemia was associated with pooled PM in women under 30 (OR 1.96, 95 % CI 1.03 - 3.72) but not in women aged 30 years or more (OR 0.68, 95 % CI 0.31 - 1.49). Similarly, the association of gravidity with past-chronic PM depended also on age. A higher prevalence of active PM was observed in women under 30 presenting with symptomatic malaria (OR 3.79, 95 % CI 1.55 - 9.27), while there was no significant increase in the prevalence of active PM (presence of parasites only) in women with symptomatic malaria when aged 30 years or more (OR 0.42, 95 % CI 0.10 - 1.75). In women with chronic PM, the prevalence of low birth weight or prematurity was the highest (31.2 %) as compared with past PM or no PM.ConclusionDespite the rapid diagnosis and efficacious treatment of peripheral infection, the prevalence of placental malaria remained high in women with P. falciparum peripheral infection in Nanoro, especially in younger women This underlines the importance of preventive measures in this specific group.
| S-EPMC8951713 | biostudies-literature
Project description:Previous research on determinants of malaria in Burkina Faso has largely focused on individual risk factors. Malaria risk, however, is also shaped by community, health system, and climatic/environmental characteristics. The aims of this study were: i) to identify such individual, household, community, and climatic/environmental risk factors for malaria in children under five years of age, and ii) to produce a parasitaemia risk map of Burkina Faso.The 2010 Demographic and Health Survey (DHS) was the first in Burkina Faso that tested children for malaria parasitaemia. Multilevel and geo-statistical models were used to explore determinants of malaria using this nationally representative database.Parasitaemia was collected from 6,102 children, of which 66.0% (95% confidence interval (CI): 64.0-68.0%) were positive for Plasmodium spp. Older children (>23 months) were more likely to be parasitaemic than younger ones, while children from wealthier households and whose mother had higher education were at a lower risk. At the community level, living in a district with a rate of attendance to health facilities lower than 2 visits per year was significantly associated with greater odds of being infected. Malaria prevalence was also associated with higher normalized difference vegetation index, lower average monthly rainfall, and lower population densities. Predicted malaria parasitaemia was spatially variable with locations falling within an 11%-92% prevalence range. The number of parasitaemic children was nonetheless concentrated in areas of high population density, albeit malaria risk was notably higher in the sparsely populated rural areas.Malaria prevalence in Burkina Faso is considerably higher than in neighbouring countries. Our spatially-explicit population-based estimates of malaria risk and infected number of children could be used by local decision-makers to identify priority areas where control efforts should be enhanced.
| S-EPMC4262087 | biostudies-other