A Bayesian shared-frailty spatial scan statistic model for time-to-event data

TL;DR

This paper introduces a Bayesian shared-frailty spatial scan statistic model for time-to-event data that accounts for intra-unit and inter-unit spatial correlations, improving cluster detection accuracy.

Contribution

It proposes a novel Cox model with shared frailty for spatial scan statistics, addressing limitations of existing models in handling spatial correlations.

Findings

Classical models fail to maintain type I error with intra-unit correlation.

Proposed model performs well with both intra- and inter-unit correlations.

Applied successfully to epidemiological data on mortality clusters.

Abstract

Spatial scan statistics are well known and widely used methods for the detection of spatial clusters of events. In the field of spatial analysis of time-to-event data, several models of scan statistics have been proposed. However, these models do not take into account the potential intra-unit spatial correlation of individuals nor a potential correlation between spatial units. To overcome this problem, we propose here a scan statistic based on a Cox model with shared frailty that takes into account the spatial correlation between spatial units. In simulation studies, we have shown that (i) classical models of spatial scan statistics for time-to-event data fail to maintain the type I error in the presence of intra-spatial unit correlation, and (ii) our model performs well in the presence of both intra-spatial unit correlation and inter-spatial unit correlation. Our method has been applied to epidemiological data and to the detection of spatial clusters of mortality in patients with end-stage renal disease in northern France.