When unfalsifiability is your business plan.

This is a follow-up to my discussion last month of the Acu-Gen Baby Gender Mentor test, prompted by the report today on Morning Edition that authorities in Illinois are investigating the company marketing the test to determine whether the claims made to consumers in marketing the test rise to the level of consumer fraud.

So, for those just tuning in, the deal is that this test promises an accurate determination (99.9% accurate, if you want the numbers) of fetal gender, as early as 5 weeks into a pregnancy, from only a few drops of the mother's blood. And, they promise a 200% refund if the test is wrong.

Set aside, for the moment, concerns about whether there's good scientific evidence that this test could be so accurate. (That was the subject of the last entry.) Cast your gaze, for a moment, on what seems to be the standard operating procedure when a consumer tries to get a refund for an inaccurate test result.

Scenario 1: Baby Gender Mentor test says boy, but the sonogram says girl. The lab does a retest. If the retest still says "boy", the consumer is told that sonograms give inaccurate gender identification 20% of the time. Still a chance Baby Gender Mentor will be right, so no refund yet.

Scenario 2: Baby Gender Mentor test says boy, but amniocentesis says girl. The lab does a retest. If the retest still says "boy", the consumer is told that the pregnancy began with a set of fraternal twins, one boy and one girl, and that the boy was a "vanishing twin". No refund, even though the boy Baby Gender Mentor detected vanished.

Scenario 3: Sonogram indicates a single fetus, after which Baby Gender Mentor test says boy and amniocentesis says girl. The lab does a retest. If the retest still says "boy", the consumer is told that there was a vanishing boy twin, and that the sonogram that indicated that it was a single fetus is no proof of anything, since sonograms give inaccurate gender identification 20% of the time. No refund, because you can't prove Baby Gender Mentor was wrong!

Do you see the pattern here?

It is true, of course, that sonograms don't always give enough information to make an accurate determination, either of gender or of how many fetuses are present. (A woman of my acquaintance discovered, in her eighth month of pregnancy, that she was expecting twins — on the fifth sonogram.) But the great part here for the folks selling Baby Gender Mentor is they've got an excuse worked out for any mismatch between their test results and what other diagnostic tools (including visual inspection of the newborn) indicate. No, they aren't producing independent evidence that there was a vanishing twin, but since it could have happened, that's enough for them to say that their test result hasn't been falsified. And that, it seems, means that they'll never have to act on their double-your-money-back pledge.

With a business plan like that, maybe it's time to branch out from biotech to religion.


Technorati tags: biotech, unsupported claims, falsifiability