AI algorithm shows researchers link between built environment and heart disease risk
Click Here to Manage Email Alerts
Key takeaways:
- Ability to see trees and a clear sky from a residence was protective against heart disease.
- The researchers made the link using an artificial intelligence-based algorithm to analyze Google Street View images.
ATLANTA — With use of an artificial intelligence-based algorithm, researchers analyzed Google Street View images and determined that the ability to see trees and a clear sky from one’s residence was protective against heart disease.
The presence of sidewalks was also protective against heart disease, likely because sidewalks could be a surrogate for physical activity, according to the researchers.
“A lot of research has shown that environmental factors strongly affect our health. If we can find a way to stratify this risk and provide interventions before cardiovascular events happen, then we could save a lot of lives,” Zhuo Chen, PhD, a postdoctoral researcher at Case Western Reserve University and University Hospitals Health System in Cleveland, said in a press release. “Our study shows that with advanced computer vision algorithms and AI, we now have the ability to quantify the built environment more effectively and efficiently. If we can assess the individual’s risk at a granular level, we could provide more personalized interventions.”
Chen and colleagues used a machine learning algorithm trained to distinguish between trees, grass, sky, sidewalks, roads and buildings to analyze Google Street View images of the residences of 49,887 participants, mostly from northeastern Ohio. They also followed participants for a median of 26.86 months to determine major adverse CV events, defined as heart attack, stroke or death from heart disease.
“The method and the data source that we’re using here is cheap, open source and publicly available,” Chen said in the release. “It can assess anywhere there are Google Street View cars on the road and really provide a refined metric of the environment.”
During the study period, there were 2,083 incidents of MACE, the researchers found.
After adjustment for age, race, sex, social vulnerability index, median household income, particulate matter exposure, noise exposure, normalized difference vegetation index, coronary artery calcium score, hypertension, diabetes and dyslipidemia, higher tree-sky index and pavement index were associated with reduced MACE, with a HR of 0.95 (95% CI, 0.91-0.95; P = .03) for the combined presence of trees and sky and a HR of 0.91 (95% CI, 0.87-0.96; P < .001) for the presence of sidewalks, according to the researchers.
There was no association between grass and MACE risk, nor between trees alone and MACE risk.
“It doesn’t necessarily mean that if we plant more trees or build more sidewalks, we’ll reduce cardiovascular risk,” Chen said in the release. “But it still gives us preliminary suggestions and indicators that can help us become aware of ways to change behaviors or neighborhood planning in the future to [potentially] lower cardiovascular risk.”
Reference:
- Mapping heart health: AI illuminates neighborhood impact on well-being. https://www.acc.org/About-ACC/Press-Releases/2024/04/01/21/55/mapping-heart-health-ai-illuminates-neighborhood-impact-on-well-being. Published April 2, 2024. Accessed April 2, 2024.
Perspective
Back to Top
Martha Gulati, MD, MS, FACC, FAHA, FASPC
I find the study intriguing, although the abstract is missing finer details of their analysis. It is interesting to use Google Street View with AI to analyze vertical environment, ie, sidewalks/sky/grass/trees/buildings. It is sometimes not easy to simply use ZIP codes because if the area is large, it can vary more than one realizes. Manual measurement is a daunting task as well. This is public, accessible data, and might be possible to scale worldwide. The horizontal and vertical landscape likely affects a lot of health and represents the environment in which we live. Since the researchers controlled for particulate matter, it is beyond pollution effects.
The most important implications of this study are that where we live matters and influences overall cardiovascular health; that greenspace, clear skies and sidewalks are associated with cardiovascular health; and that although the findings don’t imply planting more trees, or building more sidewalks or changing the skylines of where we live will result in better cardiovascular outcomes, they are intriguing and may help future urban planning.
Google Street View may help us relate to our patients and understand the environment in which they live. It can give a perspective on what might influence overall and cardiovascular health.
I would point out this study included northeast Ohio. What I don’t know is whether this is just in urban areas or even rural areas as well. The authors will likely expand this technology to look across the U.S. and the world and see if their findings can be replicated. I wonder if it has more bearing on large cities compared with small towns and rural areas. I am excited to hear more about this study when the researchers present their work.
Collapse
Chen Z, et al. Abstract 1473-221. Presented at: American College of Cardiology Scientific Session; April 6-8, 2024; Atlanta.
Click Here to Manage Email Alerts