Abstract

GIScience 2016 Short Paper Proceedings A Regional Approach for Modeling Dog Cancer Incidences with Regard to Different Reporting Practices G. Boo 1, 2 , S. Leyk 3 , S. I. Fabrikant 1 , A. Pospischil 2 Department of Geography, University of Zurich, Zurich, Switzerland Email: {gianluca.boo; sara.fabrikant}@geo.uzh.ch Collegium Helveticum, Zurich, Switzerland Email: apos@vetpath.uzh.ch Geography Department, University of Colorado, Boulder, CO, USA Email: stefan.leyk@colorado.edu Abstract Underreporting is a persistent limitation in research on environmental risk factors for dog cancer, impeding potential comparative investigations with human cancer. To address this challenge, we propose a regional modeling approach accounting for different reporting practices across the study area. In doing this, we demonstrate the need for new modeling strategies to improve statistical performance through more systematic assessments of spatial non-stationarity of statistical associations that can be linked to underreporting. 1. Introduction Humans and dogs have been sharing their habitat for millennia by being exposed to similar environmental conditions over time. An interesting aspect of this co-evolutionary process is that the development of cancer in humans and dogs might be comparable. Environmental exposures associated with cancer in dogs could thus serve to timely identify risk for humans (Pinho et al. 2012). However, to date, only few studies have addressed possible linkages between environmental risk factors and dog cancer. This gap is due to uncertainty in most existing dog cancer databases, typically because of underreporting. For this reason, in this study, we investigate underreporting of dog cancer, and propose a regional modeling approach to account for different reporting practices across the study area. We develop a case study based on the Swiss Canine Cancer Registry (SCCR), a unique dog cancer database that has been assembled for comparative investigations with humans. We model dog cancer incidences at the municipal level using demographic variables and indicators accounting for underreporting. Based on the model residuals, we decompose Switzerland into regions of similar model fit and build regional models of dog cancer incidences. We finally compare statistical performance and spatial distributions of model residuals to identify changes in the statistical associations across the study area. In doing this, we aim to demonstrate that underreporting challenges the use of statistical models of dog cancer incidences, and that a regional modeling approach improves statistical performance by mitigating spatial non- stationarity of statistical associations. 2. Data The SCCR stores diagnostic records collected in Switzerland between 1955 and 2008, and it is in the process of being updated to more recent years. Comprising more than 120,000 records, this is the largest and most durable animal cancer database, to date. The present study is based on the 3,509 cancer examinations performed in 2008, which have been enumerated within the 2,351

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.