## Project Details

### Description

Malaria is a potentially life-threatening mosquito-borne disease causing a wide variety of symptoms such as high fever, chills, headache, and vomiting among others [1]. Of the five-known species of Plasmodium parasites [2], P. falciparum is responsible for the majority of malaria deaths globally as the most severe and prevalent species in Sub-Saharan Africa [3]. Despite increased efforts to eradicate malaria worldwide with a reduced incidence and mortality of about 21% and 58%, respectively, between 2010 and 2015, malaria infections remain the leading cause of deaths in African children. In 2015, about 212 million malaria cases were recorded with 429,000 deaths, 92% of which took place in the WHO Africa region [3]. In order to reduce disease transmission and consequently malaria-related morbidity and mortality, infections should be detected and treated in time.

Social network analysis (SNA), an application of mathematical graph theory, could radically improve our understanding of malaria transmission dynamics in low transmission settings, and therefore enhance the effectiveness of targeted screening, treatment strategies and overall malaria elimination campaigns. Since social structures and risk behaviors that modify exposure to malaria infection, such as time spent in high-risk areas and the use of prevention measures and treatment seeking behavior, are typically socially shared, network structures are thought to be highly relevant to understand malaria transmission dynamics. Social networks are not confined to pre-determined geospatial units and the analysis thereof goes beyond accounting for individual-level information, as opposed to contact tracing, to reveal underlying measurable network structures.

A vast amount of research has been conducted so far with the aim of estimating quintessential epidemiological parameters related to the transmission of malaria such as the force of infection (i.e., the instantaneous rate at which individuals acquire malaria) and the entomological inoculation rate (i.e., the average number of infectious bites per unit of time). Both mathematical and statistical techniques have been employed to serve this purpose.

The aim of this research project is to develop new methodology for the estimation of age- and time-varying epidemiological parameters related to vector-borne diseases. In this research project, I will focus on malaria, which remains an important public health concern worldwide. Nevertheless, the aim is to develop a translational framework applicable for the analysis of data on relevant vector-borne infectious diseases, since a lot of recent outbreaks have been vector-borne (e.g., Zika virus). There are four main objectives in this proposal:

(1) Develop methodology to integrate both mathematical and statistical models to estimate age- and timevarying epidemiological parameters such as force of infection and point prevalence based on longitudinal malaria parasitaemia cohort data (WP1);

(2) Accommodate for the dependence in recurrent infection times within the same individual, and take into account phenomenon such as unobserved heterogeneity, outcome-dependent sampling and double interval censoring characterising longitudinal cohort data using appropriate statistical models (WP2);

(3) Relate heterogeneity in household conditions and individual attributes to social network data and study differences in social network structures between malaria-infected and uninfected individuals using SNA techniques (WP3).

(4) Combine the age- and time-dependent epidemiological malaria parameters, estimated in the presence of unobserved heterogeneity, outcome-dependent sampling and accounting for the doubly interval censored nature of the parasitaemia data, with the derived social network structures in a simulation model allowing for the study of elimination strategies (WP4).

References: 1. Miller, L. H., et al. (2013). Malaria biology and disease pathogenesis: insights for new treatments. Nature Medicine, 19(2): 156-167. 2. Phillips, M. A., et al. (2017). Malaria. Nature Reviews Disease Primers, 3:17050. 3. WHO (2016). World Malaria Report 2016. Geneva. Licence: CC BY-NC-SA 3.0 IGO.

Social network analysis (SNA), an application of mathematical graph theory, could radically improve our understanding of malaria transmission dynamics in low transmission settings, and therefore enhance the effectiveness of targeted screening, treatment strategies and overall malaria elimination campaigns. Since social structures and risk behaviors that modify exposure to malaria infection, such as time spent in high-risk areas and the use of prevention measures and treatment seeking behavior, are typically socially shared, network structures are thought to be highly relevant to understand malaria transmission dynamics. Social networks are not confined to pre-determined geospatial units and the analysis thereof goes beyond accounting for individual-level information, as opposed to contact tracing, to reveal underlying measurable network structures.

A vast amount of research has been conducted so far with the aim of estimating quintessential epidemiological parameters related to the transmission of malaria such as the force of infection (i.e., the instantaneous rate at which individuals acquire malaria) and the entomological inoculation rate (i.e., the average number of infectious bites per unit of time). Both mathematical and statistical techniques have been employed to serve this purpose.

The aim of this research project is to develop new methodology for the estimation of age- and time-varying epidemiological parameters related to vector-borne diseases. In this research project, I will focus on malaria, which remains an important public health concern worldwide. Nevertheless, the aim is to develop a translational framework applicable for the analysis of data on relevant vector-borne infectious diseases, since a lot of recent outbreaks have been vector-borne (e.g., Zika virus). There are four main objectives in this proposal:

(1) Develop methodology to integrate both mathematical and statistical models to estimate age- and timevarying epidemiological parameters such as force of infection and point prevalence based on longitudinal malaria parasitaemia cohort data (WP1);

(2) Accommodate for the dependence in recurrent infection times within the same individual, and take into account phenomenon such as unobserved heterogeneity, outcome-dependent sampling and double interval censoring characterising longitudinal cohort data using appropriate statistical models (WP2);

(3) Relate heterogeneity in household conditions and individual attributes to social network data and study differences in social network structures between malaria-infected and uninfected individuals using SNA techniques (WP3).

(4) Combine the age- and time-dependent epidemiological malaria parameters, estimated in the presence of unobserved heterogeneity, outcome-dependent sampling and accounting for the doubly interval censored nature of the parasitaemia data, with the derived social network structures in a simulation model allowing for the study of elimination strategies (WP4).

References: 1. Miller, L. H., et al. (2013). Malaria biology and disease pathogenesis: insights for new treatments. Nature Medicine, 19(2): 156-167. 2. Phillips, M. A., et al. (2017). Malaria. Nature Reviews Disease Primers, 3:17050. 3. WHO (2016). World Malaria Report 2016. Geneva. Licence: CC BY-NC-SA 3.0 IGO.

Status | Active |
---|---|

Effective start/end date | 3/11/21 → … |

### IWETO expertise domain

- B680-public-health
- B680-epidemiology