Here’s the gist. Theranos aims to make blood testing a lot less expensive and a lot easier. One or two drops of blood are good for dozens of tests. Part of their vision is to enable healthy people to undergo more frequent testing and in doing so, more closely monitor their health and catch problems more quickly. This makes sense intuitively.
Unfortunately, more frequent testing in a more healthy population is fraught with problems. More about that in a short bit.
Apparently, this point of view didn’t do down well with Mr. Cuban. Here’s his first tweet:
He unpacked his position more enthusiastically in the next tweet:
I did my best to better articulate the issue, but that was kind of hopeless because Twitter.
But the 140 character limit and Mr. Cuban’s disbelief in my point of view or misunderstanding of why I was taking it did cause me to sharpen my line of thinking into ten brief points.
And for that, Mark Cuban, I am grateful.
Here’s the gist (you can see it on Twitter here; I edited the text for readability given lack of the 140 character limit here):
- Testing has value only when it changes a subsequent / downstream decision.
- That value is positive when then the new decision creates net value relative to what you’d do without test.
- That value is negative when the new decision causes net loss relative to what you’d do without the test.
- Thus, whether testing creates health benefit depends on whether we know what to do with the new test information.
- What to do about “out of range” test values in otherwise healthy people being tested more frequently is not known.
- But many patients and physicians will feel the need to act in light of these “out of range” values.
- If and when that happens, people will be treating out of range test values as though they are disease.
- And that’s when the harms and costs of treatment are likely to outweigh the benefits.
- That’s when more testing overshoots, causing more harm than good. And we know it happens (e.g., PSA – prostate specific antigen testing for prostate cancer).
- Is this a problem for Theranos? Maybe, maybe not. But without a head-to-head randomized controlled trial of more testing vs. usual testing, there’s no way to know.
Overdiagnosis and resulting therapeutic mishaps are not theoretical; it happens. Whether it will happen for Theranos, and whether the net benefit of cheaper and more accessible testing will – from a public health perspective – outweigh these harms cannot be known in advance. But what we absolutely do know is that more frequent testing in otherwise healthy people does raise the chance that the technology will do more harm than good.