Privacy Protection Versus Cluster Detection in Spatial Epidemiology

Abstract

The spatial cluster detection algorithm performed better when addresses were analyzed as exact locations than when they were analyzed as center points of zip code or census tracts, particularly when the clustered points crossed administrative boundaries. Use of precise addresses offers improved performance, but this practice must be weighed against privacy concerns in the establishment of public health data exchange policies.

Prevalencia de accidentes de tránsito no fatales en México: resultados de la ENSANUT 2006

“…La aproximación de un problema específico a través de la epidemiología espacial se ha utilizado en otros problemas de salud y ha demostrado ser de gran valía para la prevención y el diseño de políticas y servicios de salud. 27,[33][34][35][36][37] Sin embargo, el enfoque no debe partir sólo desde el sector salud, ya que la naturaleza multifactorial de los AT requiere que las acciones de promoción y prevención incluyan a otras instituciones incluidas en los AT. En México, por ejemplo, existe una reglamentación que obliga a usar cinturón de seguridad, casco en motociclistas y asientos para infantes, además de que sanciona la utilización de medios de comunicación mientras se conduce y manejar bajo el efecto del alcohol y otras sustancias tóxicas.…”

Wider Dissemination of Household Travel Survey Data Using Geographical Perturbation Methods

“…A second approach is to examine whether the masked data preserves the distances and directions to relevant geographic features that were initially observed in the original point data (Armstrong et al, 1999). Another set of methods commonly used to examine the preservation of data utility includes point-pattern analyses such as the kernel estimation of density surfaces, cross-K functions, or more informal visualizations of clustering patterns (Kwan et al, 2004;Olson et al, 2006). The location of these clusters is often more important than the mere existence of spatial clustering (Armstrong et al, 1999).…”

New developments in the spatial scan statistic

“…Accuracy, for example, may relate to the quality of geocoding, which can have a significant effect on the SSS [7]. Resolution may be related to confidentiality issues, such as when data are spatially aggregated to provide anonymity; this also can reduce the power of the SSS [8]. Consistency could, to give an epidemiological example, relate to differences in diagnosis criteria between hospitals.…”