Bandim Health Project

Background Malaria remains a major public health problem in Guinea-Bissau, where transmission is strongly influenced by climatic and environmental conditions. Despite progress, its incidence exhibits significant seasonal and geographic variations. This study aimed to analyze the influence of meteorological factors on malaria incidence, evaluate the ability of machine learning models for trend prediction, and conduct a comparative evaluation against conventional time-series approaches across the health regions of Guinea-Bissau. Methods The study used 924 monthly observations collected across Guinea-Bissau’s 11 health regions between January 2016 and December 2022, considering 28 candidate variables spanning geographical, demographic, epidemiological, environmental, and climatic dimensions. Three machine learning models Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Machine (SVM) were developed and benchmarked against conventional time-series models, namely SARIMA, ETS, and GLMM with Splines. Performance was evaluated using RMSE, MAE, and R², validated across temporal horizons up to 2022, and supported by residual diagnostics and spatial autocorrelation analyses (Moran's I). SHAP values were used to interpret variable importance. Results The XGBoost model achieved the best predictive performance, with an R² of 0.78 (95% CI: 0.52–0.97), alongside an RMSE of 2.27 (95% CI: 1.28–3.15) and an MAE of 1.74 (95% CI: 0.89–2.63) on the test set (2021), as well as strong validation metrics an R² of 0.91(95% CI: 0.82–0.98), an RMSE of 1.57 (95% CI: 0.87–2.24) and an MAE of 1.22 (95% CI: 0.65–1.85) on the 2022 horizon. Machine learning algorithms consistently outperformed conventional temporal models (SARIMA, ETS, and GLMM with Splines) in capturing non-linear relationships. The models maintained robust generalization without significant residual spatial autocorrelation. Conclusions Machine learning models, particularly XGBoost, significantly outperform conventional time-series approaches in predicting malaria dynamics in Guinea-Bissau. Integrating epidemiological, climatic, and environmental data improves the identification of high-risk periods, contributing to early warning systems and optimized control interventions.

One fourth of the world population is estimated to be infected with Mycobacterium tuberculosis (Mtb) thereby representing a large reservoir from which the disease can reactivate with severe illness and transmission as the fatal consequences. However, little is known about the lifetime consequences of tuberculosis infection (TBI) on general health besides the risk of reactivation. We conducted a retrospective follow-up study on a cohort of individuals tested with Purified Protein Derived(PPD) skin test in 1999–2000 in the capital of Guinea Bissau. The groups of PPD-reactive and -nonreactive were sought out to determine mortality during the follow-up period and to investigate subjects still alive in terms of reactivation/active Tuberculosis (TB) disease events, risk factors for TB, immunological reactivity towards Mtb and socioeconomic factors. A total of 384 (21

Abstract Background Recent data show a dramatic increase in non-communicable diseases in developing countries, including cardiovascular disease, obesity and diabetes. Most of these diseases may be preventable and to some extent treatable by alterations in physical activity. We investigated the correlations of physical activity levels, according to energy expenditure [measured in metabolic equivalent minutes per week (METs x min/week)] with fasting glycaemia levels, blood pressure, body compositions and anthropometric variables in participants from Bissau, Guinea-Bissau. Method The International Physical Activity Questionnaire was used to assess levels of physical activity, by calculating METs x min/week. These were subsequently divided into low, moderate or high physical activity categories in a healthy group of participants. Fasting glucose, arterial blood pressure, body mass index and body fat were also assessed. Data was analyzed using the SPSS statistical software. Results 429 participants were included from sports and health facilities around the country. Of these, 187 were highly active (> 3000 MET × min/ week or at least 1500 METs × min/week) and had a mean age of 26.8 ± 7.5 years. 76% (327/429) were male and 24% were female (102/429). The mean energy expenditure was 4866.7 ± 1241.2 METs × min/week and the mean fasting glucose was 94.0 ± 14.1 mg/dl. 4.1% (18/429) of the participants had diabetes and 12.8% (55/429) had hypertension. A significant negative correlation was found between blood glucose and total physical activity (rs = − 0.117, p = 0.015). Sedentary participants had the highest proportion (5.1%) of diabetes and higher rates of hypertension (15.8%). These participants had the highest obesity (10.1%) and overweight prevalence (23.7%). In addition, they also presented the highest percentage (5.1%) of body fat, with 55% of them showing high or very high visceral fat% content. In the sedentary group, a significant positive correlation, was found, indicating a low moderate association, between systolic blood pressure and weekly sitting time, rs(175) = 0.205, p = 0.006. Conclusion While high levels of physical activity had a significant impact on glycaemia control, physical activity had no significant impact on blood pressure. However, weekly sitting time may increase blood pressure.

Insecticide-treated nets (ITNs) are most effective against malaria if mosquitoes attempt to bite people when they are in bed. Collecting the biting behaviour data required to quantify a key metric - the proportion of bites taken when people are in bed - is challenging. We tested the reliability of a simplified mosquito sampling method against data from a published systematic review. Using the methodology in the remote, malaria-endemic Bijagos archipelago, Guinea-Bissau, we estimate this proportion is 0.95 (95% Credible intervals, CrI: 0.84, 1.00). Transmission modelling suggests that pyrethroid-only ITNs in a Bijagos-like setting could avert 86% of Plasmodium falciparum clinical cases in under 5-year-olds (95% CrI: 77, 89%), compared to without ITNs. Due to agricultural work, some islanders sleep without ITNs in temporary outdoor shelters, potentially reducing the predicted population-level impact of ITNs by 35%. Our methodology increases the feasibility of monitoring the impact of indoor malaria control interventions in resource-limited settings.

Submicroscopic Plasmodium infections can be a source of persistent malaria transmission. The aim of this study was to assess their frequency, distribution, morbidity and associated factors in a pre-elimination malaria setting in sub-Saharan Africa, Guinea-Bissau, where the Plasmodium falciparum is the predominant Plasmodium species. Dried fingerprick whole blood samples from 601 participants in the 2017 national, household-based, cross-sectional survey to estimate malaria prevalence were subjected to DNA extraction. The DNA was used in nested end-point PCR assays targeting genus- and species-specific regions of the Plasmodium 18S rRNA genes. Statistical analysis of socio-demographic, clinical and molecular data was carried out using the Statistical Package for the Social Sciences, version 29. Factors associated with submicroscopic P. falciparum infections and their magnitude were sought using Chi-square test and multiple logistic regression models, respectively. Statistically significant level was considered at P-value < 0.05. Nested PCR assays detected submicroscopic P. falciparum infections in 20.3