Tuesday, September 23, 2014

The Ostrich

Well, here I go complaining again. Yesterday's NY Times had an article about the proliferation of genetic tests, and how, according to Dr. Kenneth Offit, chief of clinical genetics at Memorial Sloan Kettering Cancer Center in New York, genetic testing companies are “rushing headlong into this era,” and that “individuals are getting results we’re not fully educated to counsel them on.” This got my dander up.

When my wife was pregnant with our first child, we were advised to take the AFP test. Thinking very little about it, we toddled off to one of the prenatal exams, were told that the result would indicate if our child would have Down Syndrome, and were sent home with a bit of trepidation, but having no real reason to worry excessively. After having endured the wait of almost three weeks for the test results, which was standard of care at the time, it came back "positive," and we were asked to come in for genetic counseling. We were in shock - there was no history on either side for many generations.

At the time, I was a grad student, and knew enough about genetics to be dangerous. I really didn't know much about such prenatal tests, and relied on the information provided by our doctor and the genetic counselor. After 15 or 20 minutes of listening to the counselor in a daze, I finally leaned across the table and asked, "So let me get this straight: there's no causal relationship here? These results are associative?" What I lacked in understanding of genetics, I made up for in mathematical understanding, having been a math and bio undergrad major, and having just finished a year of graduate statistics. I was furious. Had we been told that the test was not based on a scientific understanding of what AFP does, but rather a "people-who-wear-brown-shoes-live-longer" level correlative test, we would have lost a lot less sleep. It turned out that my wife gained extremely little weight during her pregnancies, and that wasn't taken into account, thus skewing the result.

So I get it. Not being appropriately counseled or informed can cause stress for the patient and family. However, I also felt and feel responsible in some measure for my own ignorance. I could have easily done more research about the test, but was guided by our healthcare professionals, and candidly, was doing so a bit blindly. Does that mean we shouldn't do AFP (or now the Triple Test) because there's some uncertainty? Absolutely not. First, times are very different now. Virtually anyone can find extremely reliable information about genetic tests and their implications in seconds, and far more easily than it would have been for me back then. Second, as a society, we learn from continually pushing the envelope. Data should not be reported to patients and their docs because we don't know everything? When will we? The paternalism is infuriating. We're now willing to withhold information, but back then we were perfectly happy giving the results of a far less robust test to pregnant couples?

The article suggests that "variants of unknown significance" should not be reported because "patients are not getting closure." Really? How much clearer can we get than "unknown significance?" These data are reported so that sophisticated healthcare practitioners can use it if they want it, but with the caveat from the reporting entity, that, based on the extant scientific literature, we don't really have any clue if it's important or not. At least Mary-Claire King concedes in the article that "most experts do not agree with her about withholding uncertain findings."

The article chooses to continue to flog the go-to whipping boy, breast cancer. The subject of the article turns out to not have genetic alterations associated with breast cancer, but rather, with stomach cancer. However, it would be impossible to advise her to take drastic actions, because the prognostic value of the test has not yet been determined. But the article also points out that "When Myriad began BRCA testing, its rate of unknown variants was 40 percent. Now it is 2 percent." So we do learn more as time goes on, right?

I remain exquisitely puzzled why tests where the analyte is DNA are somehow different from others. Is it because we collectively have the incorrect view that DNA is more deterministic than other biological substrates? Should we stop doing PSAs, too, because they don't have 100% specificity and sensitivity?

No comments: