Using data from online geocoding services for the assessment of environmental obesogenic factors: a feasibility study

• Main Authors: Maximilian Präger, Christoph Kurz, Julian Böhm, Michael Laxy, Werner Maier
• Format: Article
• Language: English
• Published: BMC 2019-06-01
• Series: International Journal of Health Geographics
• Subjects: Obesogenic environment; Geocoding services; Validation; Diabetes
• Online Access: http://link.springer.com/article/10.1186/s12942-019-0177-9

Abstract Background The increasing prevalence of obesity is a major public health problem in many countries. Built environment factors are known to be associated with obesity, which is an important risk factor for type 2 diabetes. Online geocoding services could be used to identify regions with a high concentration of obesogenic factors. The aim of our study was to examine the feasibility of integrating information from online geocoding services for the assessment of obesogenic environments. Methods We identified environmental factors associated with obesity from the literature and translated these factors into variables from the online geocoding services Google Maps and OpenStreetMap (OSM). We tested whether spatial data points can be downloaded from these services and processed and visualized on maps. True- and false-positive values, false-negative values, sensitivities and positive predictive values of the processed data were determined using search engines and in-field inspections within four pilot areas in Bavaria, Germany. Results Several environmental factors could be identified from the literature that were either positively or negatively correlated with weight outcomes in previous studies. The diversity of query variables was higher in OSM compared with Google Maps. In each pilot area, query results from Google showed a higher absolute number of true-positive hits and of false-positive hits, but a lower number of false-negative hits during the validation process. The positive predictive value of database hits was higher in OSM and ranged between 81 and 100% compared with a range of 63–89% for Google Maps. In contrast, sensitivities were higher in Google Maps (between 59 and 98%) than in OSM (between 20 and 64%). Conclusions It was possible to operationalize obesogenic factors identified from the literature with data and variables available from geocoding services. The validity of Google Maps and OSM was reasonable. The assessment of environmental obesogenic factors via geocoding services could potentially be applied in diabetes surveillance.