Characterizing the Rural Opioid Use Environment in Kentucky Using Google Earth: Virtual Audit

• Type of Material: Article
• Authors: Natalie Crawford, Emory University; Regine Haardoerfer, Emory University; Hannah Cooper, Emory University; Izraelle McKinnon, Emory University; Carla Jones-Harrell, Emory University; April Ballard, Emory University; Sierra Shantel von Hellens, Morehead State University; April Young, Emory University
• Language: English
• Date: 2019-10-04
• Publisher: JMIR Publications
• Publication Version: Final Published Version
• Copyright Statement: © Natalie Danielle Crawford, Regine Haardöerfer, Hannah Cooper, Izraelle McKinnon, Carla Jones-Harrell, April Ballard, Sierra Shantel von Hellens, April Young.
• License: Creative Commons Attribution 4.0 International
• Final Published Version (URL): http://dx.doi.org/10.2196/14923
• Title of Journal or Parent Work: JMIR
• ISSN: 1438-8871
• Volume: 21
• Issue: 10
• Start Page: e14923
• End Page: e14923
• Grant/Funding Information: The authors would like to acknowledge their funding sources: NIDA R21 DA042727-01 and the Emory University Prevention Science Core Center for AIDS Research (P30AI050409).
• Keywords: MORTALITY; Health Care Sciences & Services; OVERDOSE; Life Sciences & Biomedicine; PHYSICAL-ACTIVITY; STREET VIEW; DRUG-USE; opioid-related disorders; BEHAVIORS; Science & Technology; Medical Informatics; RELIABILITY; rural health; BUILT ENVIRONMENT; built environment; UNITED-STATES
• Research Categories: Health Sciences, Public Health

Abstract

Background: The opioid epidemic has ravaged rural communities in the United States. Despite extensive literature relating the physical environment to substance use in urban areas, little is known about the role of physical environment on the opioid epidemic in rural areas. Objective: This study aimed to examine the reliability of Google Earth to collect data on the physical environment related to substance use in rural areas. Methods: Systematic virtual audits were performed in 5 rural Kentucky counties using Google Earth between 2017 and 2018 to capture land use, health care facilities, entertainment venues, and businesses. In-person audits were performed for a subset of the census blocks. Results: We captured 533 features, most of which were images taken before 2015 (71.8%, 383/533). Reliability between the virtual audits and the gold standard was high for health care facilities (>83%), entertainment venues (>95%), and businesses (>61%) but was poor for land use features (>18%). Reliability between the virtual audit and in-person audit was high for health care facilities (83%) and entertainment venues (62%) but was poor for land use (0%) and businesses (12.5%). Conclusions: Poor reliability for land use features may reflect difficulty characterizing features that require judgment or natural changes in the environment that are not reflective of the Google Earth imagery because it was captured several years before the audit was performed. Virtual Google Earth audits were an efficient way to collect rich neighborhood data that are generally not available from other sources. However, these audits should use caution when the images in the observation area are dated.