This story looks at advances in technology that can analyze human odors and possibly serve as ways to provide earlier or more accurate diagnoses.
There’s an underlying assumption in the story that these technologies will not only work, they’ll come to market in the next few years. This is a common (and unfortunate) framing we’ve seen over the years in other news stories on breath tests. We have no idea if that will really happen. To put things in perspective, the story needed an independent source–with no vested interest in these products getting approved for sale–to comment on the outlook, particularly the importance of overcoming common screening problems, like a high false-positive or false-negative rate.
While it might be true that someday a machine might be able to sense disease through analyzing your odor, the technology is unproven.
And even if it works, then what? We have seen through a number of cancer screening technologies (PSA testing and mammography screening) that a less-than-perfect attempt to find disease early often involves more harm than benefit–diagnosing earlier isn’t always a good thing, it just leads to more treatment with the same outcomes. A better title for this article might have been: “One day a machine may smell whether you’re sick, then what?”
Given that all the sources here have a vested interest in seeing these products come to market, clearly there’s investor excitement–aka “profit potential.” When that’s the case, it’s a good idea to spend some time discussing what these products might cost, especially if they’re not any more precise than current methods, which remains unknown. Healthcare costs in the U.S. are staggering–will these products magnify this problem, or alleviate it?
There is no real discussion of the actual benefits of any of these diagnostic technologies except to say that they are being studied in clinical trials. The closest we get to a discussion of actual benefits is when the article cites a paper in ACS Nano “showing that this artificially intelligent nanoarray could distinguish among 17 different diseases with up to 86 percent accuracy.” Even with over 1,400 participants in that trial we are told “the sample sizes for each disease were quite small. And the machine was better at distinguishing among some diseases than others,” so it is unclear what one is to make of this.
We understand that this story is previewing technology before it’s getting rigorously tested, but there’s a lot of assumption being made that it will work–after all, the headline is “One Day, a Machine Will Smell Whether You’re Sick.” The story says researchers are on “the cusp of succeeding.” This could be said about almost any medical technology that’s being tested clinically, and we didn’t see any evidence that it was warranted in this case.
The downsides of medical screening or diagnostic tools were left unmentioned. This is an unfortunate lost opportunity to explain to readers that any diagnostic or screening technology, while potentially lifesaving, could potentially harm as well. What if, for example, it flags someone as sick when they’re not — leading to unnecessary anxiety or treatment? And what if it detects a disease that would never cause the patient symptoms and yet the patient gets treated anyway? What if it misses someone who is sick and provides them with false reassurance?
The story acknowledges that the trials of these technologies have not yet been done, so very little can be said about the quality of the evidence.
There is no obvious disease mongering here and the diseases that the odor sensing technology will look for include some serious forms of cancer (ovarian, lung or colon), However, the story would have been stronger if it had mentioned that early detection doesn’t always guarantee better outcomes–sometimes all it does it ensure a longer period of treatment with the same outcome.
All of the sources have a vested interest in seeing these technologies come to market. Someone without such interests might have provided a helpful brake on the story’s enthusiasm.
There are currently ways to detect colon, lung, breast and ovarian cancers, all with some major benefits and some troublesome harms. If this odor-sensing technology is to succeed, we need to know if it is able to detect early signs of the disease more accurately than current technologies that are in use. The story should have discussed that.
We’re told we’re 3-5 years away from seeing this technology on the street, so that establishes its current (lack of) availability.
But the certainty with which this estimate is passed along by a researcher who’s developing such products should have received more scrutiny. This assumes that current testing will work out as planned. Moreover, three to five years is likely only enough time to show that the detection of cancer may be earlier by using the odor technology, and not the needed outcomes that we discussed above.
5-year predictions in medicine generally have a poor track record.
The story establishes the newness of this technological development, even though, as we are told, the idea of using ones olfactory glands to detect disease has been around for thousands of years of human history.
We could not locate a news release, but there’s enough original reporting that we’re comfortable rating this Satisfactory.