Exploring predictive frameworks for malaria in Burundi.

In Burundi, malaria infection has been increasing in the last decade despite efforts to increase access to health services, and several intervention programs. The use of heterogeneous data can help to build predictive models of malaria cases. We built predictive frameworks: the generalized linear model (GLM), and artificial neural network (ANN), to predict malaria cases in four sub-groups and the overall general population. Descriptive results showed that more than half of malaria infections are observed in pregnant women and children under 5 years, with high burden to children between 12 and 59 months. Modelling results showed that, ANN model performed better in predicting total cases compared to GLM. Both model frameworks showed that education rates and Insecticide Treated Bed Nets (ITNs) had decreasing effects on malaria cases, some other variables had an increasing effect. Thus, malaria control and prevention interventions program are encouraged to understand those variables, and take appropriate measures such as providing ITNs, sensitization in schools and the communities, starting within high dense communities, among others. Early prediction of cases can provide timely information needed to be proactive for intervention strategies, and it can help to mitigate the epidemics and reduce its impact on populations and the economy.

Performance of highly sensitive and conventional rapid diagnostic tests for clinical and subclinical Plasmodium falciparum infections, and hrp2/3 deletion status in Burundi.

Rapid diagnostic tests (RDTs) are a key tool for the diagnosis of malaria infections among clinical and subclinical individuals. Low-density infections, and deletions of the P. falciparum hrp2/3 genes (encoding the HRP2 and HRP3 proteins detected by many RDTs) present challenges for RDT-based diagnosis. The novel Rapigen Biocredit three-band Plasmodium falciparum HRP2/LDH RDT was evaluated among 444 clinical and 468 subclinical individuals in a high transmission setting in Burundi. Results were compared to the AccessBio CareStart HRP2 RDT, and qPCR with a sensitivity of <0.3 parasites/µL blood. Sensitivity compared to qPCR among clinical patients for the Biocredit RDT was 79.9% (250/313, either of HRP2/LDH positive), compared to 73.2% (229/313) for CareStart (P = 0.048). Specificity of the Biocredit was 82.4% compared to 96.2% for CareStart. Among subclinical infections, sensitivity was 72.3% (162/224) compared to 58.5% (131/224) for CareStart (P = 0.003), and reached 88.3% (53/60) in children <15 years. Specificity was 84.4% for the Biocredit and 93.4% for the CareStart RDT. No (0/362) hrp2 and 2/366 hrp3 deletions were observed. In conclusion, the novel RDT showed improved sensitivity for the diagnosis of P. falciparum.

Assessing the uncertainty around age-mixing patterns in HIV transmission inferred from phylogenetic trees.

Understanding age-mixing patterns in Human Immunodeficiency Virus (HIV) transmission networks can enhance the design and implementation of HIV prevention strategies in sub-Saharan Africa. Due to ethical consideration, it is less likely possible to conduct a benchmark study to assess which sampling strategy, and sub-optimal sampling coverage which can yield best estimates for these patterns. We conducted a simulation study, using phylogenetic trees to infer estimates of age-mixing patterns in HIV transmission, through the computation of proportions of pairings between men and women, who were phylogenetically linked across different age groups (15-24 years, 25-39 years, and 40-49 years); and the means, and standard deviations of their age difference. We investigated also the uncertainty around these estimates as a function of the sampling coverage in four sampling strategies: when missing sequence data were missing completely at random (MCAR), and missing at random (MAR) with at most 30%-50%-70% of women in different age groups being in the sample. The results suggested that age-mixing patterns in HIV transmission can be unveiled from proportions of phylogenetic pairings between men and women across age groups; and the mean, and standard deviation of their age difference. A 55% sampling coverage was sufficient to provide the best values of estimates of age-mixing patterns in HIV transmission with MCAR scenario. But we should be cautious in interpreting proportions of men phylogenetically linked to women because they may be overestimated or underestimated, even at higher sampling coverage. The findings showed that, MCAR was the best sampling strategy. This means, it is advisable not to use sequence data collected in settings where we can find a systematic imbalance of age and gender to investigate age-mixing in HIV transmission. If not possible, ensure to take into consideration the imbalance in interpreting the results.

SimpactCyan 1.0: An Open-source Simulator for Individual-Based Models in HIV Epidemiology with R and Python Interfaces.

SimpactCyan is an open-source simulator for individual-based models in HIV epidemiology. Its core algorithm is written in C++ for computational efficiency, while the R and Python interfaces aim to make the tool accessible to the fast-growing community of R and Python users. Transmission, treatment and prevention of HIV infections in dynamic sexual networks are simulated by discrete events. A generic "intervention" event allows model parameters to be changed over time, and can be used to model medical and behavioural HIV prevention programmes. First, we describe a more efficient variant of the modified Next Reaction Method that drives our continuous-time simulator. Next, we outline key built-in features and assumptions of individual-based models formulated in SimpactCyan, and provide code snippets for how to formulate, execute and analyse models in SimpactCyan through its R and Python interfaces. Lastly, we give two examples of applications in HIV epidemiology: the first demonstrates how the software can be used to estimate the impact of progressive changes to the eligibility criteria for HIV treatment on HIV incidence. The second example illustrates the use of SimpactCyan as a data-generating tool for assessing the performance of a phylodynamic inference framework.