Methods for Spatial Epidemiology
Researchers: Jarno Vanhatalo, Aki Vehtari, Jouni Hartikainen, Aki Havulinna and Jouko Lampinen
The research is part of New Analysis Methods for Healthcare Process Management project. We have developed a theme map software for representing regional healthcare key figures (e.g. mortality, diagnose). Key figures are shown in standardized form using color codes and variation in time is shown as sequential images or movies. Quick adaptive binned kernel method for handling very large number of grid cells has been implemented in 2005 and software is in pilot use. Current development concentrates on implementing Bayesian spatial methods. The pilot data are mortality due to most common diseases and life expectancy in Finland.
Spatial epidemiology concerns both describing and understanding the spatial variation in the disease risk in geographically referenced health data. One of the main classes of spatial epidemiological studies is disease mapping, where the aim is to describe the overall disease distribution on a map and, for example, highlight areas of elevated or lowered mortality or morbidity risk. The spatially referenced health data in this project are aggregated from point-referenced data into lattices of various grid cell sizes. The data are geographically more accurate than areal level data where the subregions are defined by governmental districts. The models used for the spatial epidemiology in the research are based on Gaussian processes.
Gaussian processes are an attractive way to describe the spatial correlations between areas, since the correlations are included in an explicit and a natural way into the model via a covariance function. They also provide a flexible way of describing the form of the spatial prior with a combination of different covariance functions. See also, the methodological research on Gaussian processes in Bayesian modelling of complex systems.
Examples
Refereed articles
-
Jarno Vanhatalo, Pia
Mäkelä and Aki Vehtari (2010). Alkoholikuolleisuuden
alueelliset erot Suomessa 2000-luvun alussa.
Yhteiskuntapolitiikka, 75(3):265-273. (Available
online
in Finnish) (English
translation) (Online maps in
Finnish)
- Jarno Vanhatalo,
Ville Pietiläinen and Aki Vehtari (2010).
Approximate inference for disease mapping with sparse Gaussian
processes. Statistics in Medicine, 29(15):1580-1607.
(Available
online).
- Jarno Vanhatalo and Aki Vehtari (2007). Sparse Log Gaussian Processes via MCMC for Spatial Epidemiology. JMLR Workshop and Conference Proceedings, 1:73-89. (Gaussian Processes in Practice) (PDF)
Other publications
-
Jarno Vanhatalo, Pia Mäkelä, ja Aki Vehtari (2010). Regional
differences in alcohol mortality in Finland in the early 2000s.
Report A20, Department of Biomedical and Computational Science
Publications, Helsinki University of Technology.
(PDF)
- Ville Pietiläinen (2010).
Approximations for Integration Over the Hyperparameters
in Gaussian Processes, M.Sc. Thesis, Aalto University. (PDF)
- Jarno Vanhatalo and Aki Vehtari (2006). Sparse Log Gaussian Process in Spatial Epidemiology. Presentation in the international conference of the Royal Statistical Society, September 2006. (Slides in PDF).
Markus Siivola (2006). A GIS tool for visualization and spatial analysis of georeferenced health data. M.Sc. Thesis, Department of Electrical and Communications Engineering, Helsinki University of Technology. (PDF)
