“Accurate” medical tests can still be mostly wrong

hobson-theranos-1-rkIt sounds so simple that many Americans, especially with modern DIY  technology, may want to try: To get healthier, why not go out and get every possible diagnostic test done? Especially if all it takes is a drop or two of blood: Wouldn’t this increase early detection of disease and make for quicker treatment and better health?

Think again, please. This is a more nuanced issue with potentially significant cost implications and health risks from inevitable test foibles and inaccuracies, as some nifty reporting by the numerically oriented folks at the online site 538 have shown.

They have joined in the increasing media take-down of Theranos, a gee-whiz company that became a Wall Street favorite. The firm ─which is under regulatory fire for hyping its product─not only claims it has proprietary processes that will make blood testing faster, more convenient, and cheaper, it also says that its work-ups should become more ubiquitous. Theranos says patients who undergo more blood tests might learn about possible diseases sooner and get treated better.

Not exactly, as 538 points out:

[As} more and more research is showing, it’s not as simple as going exploring in healthy people to see what turns up, then treating what you find. “The reservoir of abnormalities is bottomless,” said Dr. Brenda Sirovich, co-director of the Outcomes Group at the VA Medical Center in White River Junction, Vt. “And we don’t know anything about what those things mean in an asymptomatic vacuum.”

What “asymptomatic vacuum” means is that if it doesn’t cause pain or some other symptom, it’s impossible to know if a blip on a blood test is meaningful.

That’s true even if a test is, say, 90 or even 95% accurate. A 90% accuracy rate means one in ten patients is getting a false alarm from a test result, and a 95% rate means one in twenty gets a false alarm. If  the disease you’re looking for occurs less often than the rate of inaccurate test results — i.e., here, less than one in ten or one in twenty of patients tested — that means the test is spinning off more false alarms — “false positives” — than true abnormal results.  Which means that 90% accurate isn’t really 90% accurate at all.

You can walk through the numbers generated by Bayesian analysis, which is best seen with visual mapping (click on the graphic, courtesy of 538, to enlarge it and see how it works).

As the site dissects well, testing can become excessive, costly, and it’s unclear its direct, positive effect on health. It’s one thing for a doctor to suspect a disease or condition, then to test to determine its absence or presence. It’s another to not only take tests, but then to subject patients to possible therapies and interventions─some painful and costly in themselves─as a consequence. It isn’t crystal clear, for example, that cancer tests, say those for prostate or breast cancer, conclusively should lead to therapies like surgery, radiation, and chemo.

I’ve written before about the frontiers that are opening with genetic tests available online. As technology advances, including as therapies are genetically targeted, more tests will become available to the public online, perhaps through apps, and otherwise. Caveat emptor, consult with your doctor, and don’t make serious medical testing a casual DIY curiosity.

Patrick Malone & Associates, P.C. listed in Best Lawyers Rated by Super Lawyers Patrick A. Malone
Washingtonian Top Lawyer 2011
Avvo Rating 10.0 Superb Top Attorney Best Lawyers Firm
Contact Information