Coronary disease & ACSCross-sectional study
Artificial intelligence–based assessment of built environment from Google Street View and coronary artery disease prevalence
- Cross-sectional study
- census tract level
- deep learning (convolutional neural networks)
- linear mixed-effects models
Authors' conclusion
In this cross-sectional study, the prevalence of CHD was associated with built environment factors derived from GSV through deep learning analysis, independent of census tract demographics.
Study context
- Population
- Census tracts in seven US cities (Cleveland, Fremont, Kansas City, Detroit, Bellevue, Brownsville, Denver)
- Participants
- 789 census tracts
- Outcome
- CHD prevalence at census tract level, predicted from GSV built environment features
Main finding
As reported in the abstract
Census tract variation in CHD prevalence predicted by GSV built environment features
Variation in coronary heart disease (CHD) prevalence predicted by Google Street View (GSV) features (deep learning)
63%
Observational data: associations, not effects.
Further findings
Activation maps
- Neighbourhood features represented by buildings and roads associated with CHD prevalence
Original publication
Artificial intelligence–based assessment of built environment from Google Street View and coronary artery disease prevalence. Eur Heart J. Published 28 March 2024. doi:10.1093/eurheartj/ehae158
Summarised from the published abstract; every number is checked against it