Prix / Prize
CRM-SSC 2002-2003


Charmaines's Picture


CHARMAINE B. DEAN
(Simon Fraser)


Le vendredi 14 novembre 2003 / Friday, November 14, 2003
15 h 30 / 3:30 p.m.

Centre de recherches mathematiques
Université de Montréal
Pavillon Andre-Aisenstadt
2920, chemin de la Tour
Salle / Room 6214


Une reception suivra la conférence au Salon Maurice l'Abbe (salle 6245).
There will be a reception after the lecture in Maurice l'Abbe (Room 6245).


"Spatio-temporal Mapping of Disease or Mortality Rates"


The spatio-temporal analysis of rates focuses on describing the spatial distribution of rates over a region and over time in order to suggest factors which may be linked to various diseases for further epidemiological study, or to produce smoothed maps for identifying spatial trends, or for disease surveillance and control. At Ministries of Health these maps facilitate the formulation of health policy decisions related to the allocation of health funding. At the local health area (small-area) level spatio-temporal analyses help to provide insight on the spatial distribution of health inequities and localized policy impact. They also make linkages between disease incidence and mortality and health status indicators including quality of life indicators and social capital, for example.

This talk discusses generalized additive mixed models for the analysis of geographic and temporal variability of mortality rates. This class of models accommodates random spatial effects and fixed and random temporal components.

Spatio-temporal models which use autoregressive local smoothing across the spatial dimension and spline smoothing over the temporal dimension are considered. The objective is the identification of temporal trends and the production of a series of smoothed maps from which spatial patterns of mortality risks can be monitored over time. Regions with
consistently high rate estimates may be followed for further investigation. The talk also considers the problem of accurate identification of high-risk areas when there are both smoothly varying spatial random effects and discrete hot-spots of high- and low-risk activity.

The methodology is illustrated by analysis of British Columbia infant mortality data.