This vague Newsweek story describes how scientists from Google have reported developing an algorithm that can use large data sets of retinal scans to determine a person’s risk ofdeveloping cardiovascular problems.
The story establishes the novelty of this approach and mentions current assessment tools. However, there’s no mention of costs or possible harms from this approach, and no numerical data on how effective it might be. There are also no outside sources nor alternatives mentioned, and it doesn’t indicate when, if ever, this method might be available.
Heart disease is the leading killer in the U.S., causing more than 600,000 deaths per year, according to the CDC. So if there was a simple, non-invasive method of gauging a person’s risk for developing a heart attack later on, simply by scanning a person’s eyes, that would be a substantive aid in public health. Sadly this story fails to back up that dream with any real substance, and by doing so, harms readers by leading them on. Better to have not offered the story at all.
This Newsweek story offers no information on what the cost of such retinal scan might be, nor does it mention whether health insurance would cover such screening, if it is proven to eventually work. There are a number of free cardiovascular risk assessment tools available online.
The only numerical data offered by this story is that they used “data from more than a quarter of a million patients.” There is no comparison of the success rate of this new approach with other mechanisms for determining cardiovascular risk, so readers have no way of assessing its potential value.
The story offers no mention of potential harms that might arise from using this new approach and since it would be non-invasive, it is hard to imagine many. But the potential for incorrect assessments is always present as either false-positives or false-negatives and the harms from either of those can be considerable.
The story really doesn’t offer any evidence of the success of this approach. The only statement that relates to results from this research says, “the scientists were able to predict the cardiovascular risk factors that were not previously thought to be present in retinal fundus images.” But how effective were those? How does this method compare with other techniques for determining cardiovascular risk?
The published abstract pointed out that the results were “validated on two independent datasets of 12,026 and 999 patients” suggesting that some additional data was available to the story’s author. It’s not clear that any patient has been tested using this method, even though both the headline an the lede sentence of the story emphatically state it works.
There’s no outright disease-mongering here. However, the story provides no context about how many people are at risk of a heart attack. For that reason, it doesn’t meet our criteria.
The story does point out that the researchers work for “Google’s parent company Alphabet” and for the company Verily, which used to be Google Life Sciences, so the conflict of interest is obvious. But it fails to identify any of the researchers, even though it includes two quotes from a nameless source. No independent sources are quoted.
The story mentions cardiovascular risk calculators, which use parameters such as a person’s gender, smoking status, blood pressure, and age.
“Most cardiovascular risk calculators use some combination of these parameters to identify patients at risk of experiencing either a major cardiovascular event or cardiac-related mortality within a pre-specified time period, such as 10 years,” the paper states.
“However, some of these parameters may be unavailable…We therefore explored whether additional signals for cardiovascular risk can be extracted from retinal images, which can be obtained quickly, cheaply and non-invasively in an outpatient setting.”
The story indicates that “further tests” are needed, but readers are left with no idea as to when, if ever, this new approach to determining risk might be available. There’s no mention of when, if ever, clinical trials might be done to gauge its efficacy.
The idea that a non-invasive scan of a person’s retina could yield enough information to calculate that person’s risk of a heart attack appears to be novel.
This story does not appear to be based on a news release.