Can Google Street View data provide heart health insights?

Google Street View

An intriguing new study has suggested that there could be a correlation between the built environment of an area – its buildings, streets, and green spaces – and the risk of inhabitants developing coronary heart disease (CHD).

The scientists behind the work applied artificial intelligence to analyse Google Street View images to glean data about the built environment of areas in seven US cities and cross-referenced those findings with census data on the prevalence of cardiovascular diseases.

Their findings, published in the European Heart Journal, suggest that built environment factors can predict 63% of the variation in the risk of CHD from one area to another.

The analysis – carried out across the cities of Cleveland, Fremont, Kansas City, Detroit, Bellevue, Brownsville, and Denver – could not only help to understand risk factors caused by the areas in which people live, but also help town planners design healthier neighbourhoods, according to the researchers.

“Here in America, they say that the zip code is a better predictor of heart disease than even personal measures of health,” said Professor Sanjay Rajagopalan from University Hospitals Harrington Heart & Vascular Institute and Case Western Reserve University, one of the leaders of the research team.

Investigating those differences scientifically is a challenge, however, so the team turned to a machine vision-based approach that looked at more than half a million images of ‘census tract’ areas of the cities, each with around 4,000 inhabitants, to look for signals.

Using a convolutional neural network to carry out attention mapping – which highlights important areas of an image – the researchers were able to interpret data on thousands of features that may have a bearing on CHD.

They found that features like green space and walkable roads were associated with lower risk, while other things, such as poorly paved roads, were associated with higher risk.

“With upcoming challenges including climate change and a shifting demographic, where close to 70% of the world’s population will live in urban environments, there is a compelling need to understand urban environments at scale,” said Prof Rajagopalan.

The authors acknowledge that it is a very early assessment of the potential of this type of approach and note in the paper that the correlations they found do not establish causality – a view echoed by Kevin McConway, a statistician at The Open University in the UK.

He said the study is an “interesting start” with a “heroic scale of data analysis,” but “tells us little new, if anything, about what actually causes people to develop CHD, or about what we might do to lessen the risks.”

“There will be many differences between people that live in neighbourhoods that look different on Google Street View, apart from what can be seen visually in those neighbourhoods,” he added. “These other differences might be the actual cause of differences in CHD risk, and not the different appearance of the neighbourhoods at all.”

In an editorial accompanying the study, Dr Rohan Khera from Yale University School of Medicine also notes that, to date, the ability to classify environmental risk factors for health has relied on population surveys that track factors like wealth, pollution, and community resources.

The new study “presents a novel and more comprehensive evaluation of the built environment,” he adds. “It is critical to use this information wisely, both in defining strategic priorities for identifying vulnerable communities and in redoubling efforts to improve cardiovascular health in communities that need it most.”

Image by petterijokela from Pixabay