Title: Spatio-temporal modelling of respiratory disease risk with changing spatial boundaries
Authors: Eilidh Jack - University of Glasgow (United Kingdom) [presenting]
Duncan Lee - University of Glasgow (United Kingdom)
Nema Dean - University of Glasgow (United Kingdom)
Abstract: Spatio-temporal patterns in population-level disease risk are often estimated from data relating to a set of irregularly shaped areal units, such as electoral wards or census tracts. One shortcoming of traditional areal unit modelling techniques is that the areal units are themselves artificial units of spatial recording and can influence the spatial pattern observed in the data. That is, if the areal units changed then so would the results. This is known as the modifiable areal unit problem (MAUP). Another common problem in areal unit data of this type is that often there are changes to boundaries that occur during the time period for which data are available. Statistically, this poses a challenge since using data from before and after this change would lead to non-comparable inference due to spatial misalignment. A statistical framework is proposed for solving these problems, by using the areal unit data to obtain inference on the spatio-temporal pattern in disease risk on a regular grid. This framework is illustrated with a study on the spatio-temporal trends in health inequalities in respiratory disease risk in Glasgow, Scotland.