Travel-related spread of disease
I am co-author of the R package 'epiflows' for risk assessment of travel-related spread of disease. 'epiflows' produces estimates of the expected number of infections that could be introduced to other locations from the source of infection by integrating data on the number of cases, population movement, length of stay and incubation and infectious distributions.
Collaborators: R Epidemics Consortium (RECON)
I developed 'SpatialEpiApp', a Shiny web application that allows to visualize spatial and spatio-temporal disease data, estimate disease risk and detect clusters. The application incorporates modules for interactive visualizations such as maps supporting padding and zooming and tables that allow for filtering, and the generation of reports containing the analyses performed.
Malaria in Malawi
I am researching the effect of larval source management and house improvement on malaria transmission in southern Malawi, when implemented alone or in combination, in addition to the Malawi National Malaria Control Programme interventions.
Collaborators: University of Malawi and Wageningen University
Lymphatic filariasis in sub-Saharan Africa
I developed geostatistical models of lymphatic filariasis prevalence for sub-Saharan Africa which will enable geographical targeting of interventions.
Collaborator: Global Atlas of Helminth Infection (GAHI), London School of Hygiene & Tropical Medicine
Spatial variations in disease temporal trends
I worked on the development of a method for detecting spatial variations in disease temporal trends and its implementation in the SaTScan software. The method is based on spatial scan statistics and allows the detection of regions with markedly different disease trends. The method was applied to analyze cervical cancer in United States.
Collaborator: Martin Kulldorff, Harvard Medical School
Integration of misaligned data
I developed a Bayesian geostatistical model for the analysis of point-level and area-level data using INLA and SPDE approaches. The model was applied to predict air pollution in los Angeles and Ventura counties, United States.
Collaborator: Marcello Pagano, Harvard School of Public Health
Tuberculosis and HIV in Africa
I constructed spatial models to investigate the geographic distribution of Tuberculosis, HIV and risk factors in communities in Zambia, South Africa and Malawi. These models were useful to inform care and control strategies in high incidence settings.
Collaborators: Nicky McCreesh, London School of Hygiene and Tropical Medicine, and Peter Diggle, Lancaster University
Intrahepatic Hepatitis C Virus in humans
I developed spatio-temporal point process models to characterize the propagation patterns of Hepatitis C Virus in human livers. These models accounted for rates of hepatocyte infection, viral production, and immune responses and helped to understand the mechanics of infection in the liver.
Collaborators: Ruy Ribeiro, Los Alamos National Laboratory, and Peter Diggle, Lancaster University
Spatio-temporal modelling of cancer
I conducted studies on the spatio-temporal distribution of cancer in the province of Girona, Spain, and in Australia. These analyses provided a perspective of the variation in cancer and informed policy makers about geographic inequalities in cancer risk.
Collaborators: Catalan Institute of Oncology and Cancer Council Queensland
R package DClusterm
I participated in the Google Summer of Code 2011 program, developing the R package 'DClusterm' for model-based detection of disease clusters.
Collaborators: Virgilio Gomez-Rubio, Universidad de Castilla-La Mancha, and Barry Rowlingson, Lancaster University
Leptospirosis in Pau da Lima, Salvador, Brazil
I led the development of a spatial modelling architecture to understand the transmission dynamics of leptospirosis and identify targets for intervention in a urban slum in Brazil.
Collaborators: Peter Diggle, Lancaster University, and Jose Hagan, Yale School of Public Health and Oswaldo Cruz Foundation (Fiocruz)
Hospital admissions forecasting
I worked with Alder Hey Children's Hospital in Liverpool to help predict emergency admissions. This will enable them to better schedule elective procedures to make full use of hospital resources.
Collaborators: Peter Diggle and Barry Rowlingson, Lancaster University
Missing values in the Pneumonia & Influenza data from the 122 CMRS operated by the CDC
I worked on an imputation method to estimate missing values in the Pneumonia & Influenza data from the 122 Cities Mortality Reporting System operated by the Centers for Disease Control and Prevention (CDC).
Collaborator: Al Ozonoff, Harvard Medical School
Gaussian Component Mixture and CAR models in Bayesian disease mapping
I investigated new distributions to model correlated heterogeneity in Bayesian disease mapping. These distributions allow to borrow information of nearby regions when estimating the incidence or mortality of the disease.
Collaborator: Andrew B. Lawson, Medical University of South Carolina
Detection of disease clusters with local indicators of spatial association functions
I developed a method for the detection of clusters in case-control studies. The method is based on LISA functions of the product density function of the point pattern of cases, and highlights the cases which form part of an agglomeration zone.
Collaborator: Francisco Montes, Universidad de Valencia
Infectious disease outbreaks
I worked to enhance the use of the WHONET software, used to support surveillance of infections and antimicrobial resistance. My work was focused in the automated statistical detection of clusters suggestive of infectious disease outbreaks both in the hospital and community settings.
Collaborator: John M. Stelling, WHO Collaborating Centre for Surveillance of Antimicrobial Resistance, Brigham and Women's Hospital, Boston
Monitoring mortality in Europe
I was involved in the evaluation of the EuroMOMO project that monitors the mortality related to possible public health threats such as major epidemics, extreme temperatures or release of biological or chemical agents. I evaluated the statistical methods EuroMOMO uses to quantify the all-cause mortality in Europe and to detect periods where an excess of mortality occurs.
Collaborator: European Centre for Disease Prevention and Control (ECDC)
Optimal asset allocation with linear programming
I used linear programming methods to obtain the efficient frontier and determine optimal portfolios which maximize return for a defined level of risk. These methods were implemented in a financial tool to assist portfolio managers to choose optimal investment strategies.
Collaborator: Openfinance, Valencia