Using Geographically Weighted Poisson Regression to Examine the Association Between Socioeconomic Factors and Hysterectomy Incidence in Wallonia, Belgium

Abstract

Background: Various studies have investigated geographical variations in the incidence of hysterectomy in Western countries and analyzed socioeconomic factors to explain those variations. However, few studies have used spatial analysis to characterize them. Geographically weighted Poisson regression (GWPR) explores the spatially varying impacts of covariates across a study area and focuses attention on local variations. Given the potential of GWPR to guide decision-making, this study aimed to describe the geographical distribution of hysterectomy incidence for benign indications in women older than 15 years old (15+) at the municipal level in Wallonia (southern region of Belgium) and to analyze potential associations with socioeconomic factors (‘Education/training’, ‘Income and purchasing power’ and ‘Health and care’) influencing the use of this surgery.Methods: We carried out an ecological study on data for women aged 15 + living in one of the 262 Walloon municipalities who underwent hysterectomies for benign indications between 2012 and 2014. We linked standardized hysterectomy rates to three municipal-level socioeconomic factors (‘Education/training’, ‘Income and purchasing power’ and ‘Health and care’). Then, a generalized linear model (GLM) for Poisson regression and a GWPR were applied to study the relationships between hysterectomy incidence and socioeconomic covariates in Wallonia.Results: The hysterectomy rate varied across the region. The GLM Poisson regression revealed a positive and significant association between the hysterectomy rate and ‘Income and purchasing power’, and a negative and significant association between hysterectomies and ‘Health and care’. The same associations were seen in the GWPR model. The latter demonstrated that the association between hysterectomies and ‘Education and training’ ranged from negative to positive over the study area.Conclusions: Hysterectomy incidence was shown to have nonstationary relationships with socioeconomic factors. These results support the development of targeted interventions for a more appropriate use of this surgery.

Spatiotemporal heterogeneity of SARS-CoV-2 diffusion at the city level using geographically weighted Poisson regression model: The case of Bologna, Italy

“…However, a GWPR model is best for relaxing the fixed association assumption between the response and the explanatory variables thereby capturing spatial heterogeneity. Recent studies have shown that the applications of GWPR in a wide variety of fields, including but not limited to, analysis of the effects of demographic and socio-economic factors on the spatial variation of COVID-19 (Chen et al, 2022;Shawky et al, 2021;Zhang et al, 2021), health and other disease analysis (Bui et al, 2018 ;Chen et al, 2010;Goovaerts, 2005;Nakaya et al, 2005;Poliart et al, 2021;Yang et al, 2009), traffic crash modelling (Li et al, 2013;Zhao et al, 2004), population density and housing (Mennis and Jordan, 2005;Chen et al, 2017), poverty mapping (Benson et al, 2005;Loubert et al, 2018) and diseases mapping (Nakaya et al, 2005). In most of these studies, however, one of the challenges is the presentation and synthesis of the large number of "mappable" results generated by local GWR models.…”