Surveilling Malaria through machine learning and clustering tools in pregnancy

Informacje na temat projektu

PregMal

Identyfikator umowy o grant: 890477

DOI

10.3030/890477

Projekt został zamknięty

Data podpisania przez KE 21 Kwietnia 2020

Data rozpoczęcia 1 Kwietnia 2021

Data zakończenia 21 Grudnia 2023

Finansowanie w ramach

EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Koszt całkowity

€ 172 932,48

Wkład UE

€ 172 932,48

172 932,48

Koordynowany przez

FUNDACION PRIVADA INSTITUTO DE SALUD GLOBAL BARCELONA
Spain

Periodic Reporting for period 1 - PregMal (Surveilling Malaria through machine learning and clustering tools in pregnancy)

Okres sprawozdawczy: 2021-04-01 do 2023-03-31

Malaria still represents a humanitarian thread that has become a high priority in the World Health Organization (WHO) and European Union (EU) directives. Robust malaria surveillance is key for an optimal and equitable allocation of interventions, but estimating malaria trends from current surveillance systems remains unreliable and expensive.
Pregnant women represent a promising convenient group for sentinel surveillance. They are of easy access at antenatal care (ANC) visits, and their high attendance provide a good representation of the population. They can inform about asympomatic infections, providing a timely monitoring and a representative geographical distribution of the population. However, few studies analysed the consistency of malaria burden levels between pregnant women and the community.
The overarching goal of the project is to assess the potential of pregnant women at maternal health care clinics for providing precise and effective actionable information for malaria control and elimination.

The specific objectives are three:
1. To develop new statistical tools to better characterise spatial clustering, temporal trends of Pf transmission and genetic differentiation of the parasites.
2. To assess the potential of parasitological and serological data from pregnant women at first ANC visit as a source of reliable data to reflect temporal and spatial Pf trends in the community.
3. To compare genetic metrics in the parasite population collected from pregnant women and from the overall community that can inform about changes of Pf transmission, clustering of infections and parasite importation.

The work performed in this project can be described in five components or work packages:

WP1: Database preparation and validation
Different data sources were combined in this project, including data from cross-sectional studies, from cases passively detected at health facilities and from pregnant women at ANC visits, all them from Maputo province in Mozambique in the period of three years.

WP2: Development of tools
- Spatial analysis: Two-point correlation function statistics were developed to analyse the spatial structure of malaria infections. In addition, a new hotspot and outbreak detector algorithm was developed in order to detect fine-scale hotspots and outbreaks and to monitor them in time.
- Temporal analysis: different statistical tools were develop in order to compare the temporal trends of different indicators from the different data sources and identify potential time lags between them.
- Genomics: different approaches were explored to quantify the parasite genetic relatedness accross samples and areas and to classify individual cases as imported or local.

WP3: analysis and results
- Spatio-temporal trends of malaria transmission were compared between pregnant women at first ANC visit and children from health facilities and cross-sectional studies. Positivity rates using quantitative polymerase chain reaction (qPCR) were found to be very consistent between pregnant women and children regardless of factors such as gravidity or HIV status. An effect of gravidity was found in moderate-high transmission settings for rapid diagnostic tests, where multigravid women showed lower positivity rates. A 3-month time lag was found between temporal trends of ANC data and clinical cases. Finally, hotspots between ANC data and clinical cases were compared, finding consistent results.
- Parasite genetic populations were compared between pregnant women and children in order to assess the potential of ANC data for genomic surveillance, finding consistency across different metrics and markers.
- Genetic relatedness analysis was conducted to assess the spatial connectivity of parasite genetic populations across provinces, finding a strong south-north differentiation. These results were combined with reported travels in two low-transmission districts to identify imported cases. The 25% of reported cases were classified as imported, highlighting the strong impact of importation in these areas.

WP4: training activities
- A secondment at Sequentia Biotech S.L. to receive training on genomics.
- Course: "Gender awareness", by ISGlobal and Grupitagora
- Course: "Awareness in cybersecurity", by ES-CIBER
- Course: "GIS for exposure assessment in Environmental Health Research", by ISGlobal and University of Barcelona

WP5: Dissemination and communication
The following activities were conducted:
- Press article in Horizon Magazine
- 3 outreach presentations in schools and institutes.
- 3 poster presentations at the European Scientific Conference on Applied Infectous Disease Epidemiology (ESCAIDE) 2023 (Barcelona, Spain), and at the Annual Meetings of American Society of Tropical Medicine & Hygiene 2022 and 2023 (USA)
- Oral and Invited presentations at the Annual Meeting of American Society of Tropical Medicine & Hygiene 2021 (virtual conference) and at the GenMoz meeting 2023 (Maputo, Mozambique)

The project contributed to expand the current understanding of the potential of pregnant women attending antenatal care (ANC). The study combined different transmission settings, surveillance data sources (passive detection and cross-sectional surveys), and analytical tools to compare the spatial and temporal patterns of malaria burden in children and pregnant women, integrating as well the main factors affecting positivity rates in pregnant women (gravidity, HIV status and test sensitivity). The project also analysed the potential of the different data sources, including parasitological, serological and genetic data, to characterise the burden level and the spatio-temporal patterns of transmission. The results confirm the capacity of ANC data for surveillance. A new hotspot detector was designed that adapts to the spatial distribution of the population. This software allowed to detect fine-scale regions with higher burden than expected.

Impact:
The results of the project allow for a better understanding of malaria suirveillance using ANC data, and indicate the different factors to be taken into account. The whole scientific community can benefit from the conclusions of the study to improve surveillance analysis in any malaria endemic country. Also, pregnant women attending ANC would benefit from ANC surveillance, since they would be tested for malaria and treated if infected.
Non-academic institutions will also benefit from this project: governmental organisations working to improve public health systems; organisations with an interest in malaria elimination (e.g. WHO, Goodbye Malaria Initiative); humanitarian organisations such as Médicins Sans Frontières that require to monitor malaria control in emergency situations. Finally, this research has the potential to contribute to additional, more distal, impacts by increasing the feasibility of elimination activities that can ultimately improve health, well-being and economical sustainability in malaria endemic countries.

Spatial distribution of serological clusters of different antigens in pregnant women

Spatial map of households of pregnant women and their qPCR results

Map showing clinical cases in children and detected hotspots

Periodic Reporting for period 1 - PregMal (Surveilling Malaria through machine learning and clustering tools in pregnancy)

Pobierz Pobierz zawartość strony