Wednesday, March 26, 2008

Genetic Testing for Bipolar: Are You Kidding Me?

Academics Dr. John Kelsoe and Kurt May have fired the warning shot: Genetic testing for mental disorders is on its way. Like much else in the mental health field, I fear that marketing may yet again trump science. Kelsoe and May's new test is out and it claims it can assess the risk for bipolar disorder (sort of) for a fee of $399.

Both Furious Seasons and Daniel Carlat have already opined wisely on the topic. The first issue is the science behind such testing -- if the science does not support the validity of the test in determining if someone actually has a mental disorder, then the test is a sham. So what does the science say? According to an article in Science, one genetic variant used in the test was associated with a tripling of risk for bipolar disorder. The catch: The variant was only found in 3% of individuals with bipolar disorder and 1% of the people without bipolar disorder. A genetic variant that is only possessed by 3% of people with bipolar can hardly be considered as widely useful. A combination of five variants in another study was found in 15% of individuals with bipolar disorder compared to 5% if those without the condition. As I understand it, the current test, as put forth by Kelsoe and May through the company Psynomics, tests for a combination of the previously mentioned variants. Again, the set of variants they are using are not very common even among people with bipolar disorder. So even if you are bipolar, the odds are high that this test would not label you as such. In the world of testing, this is called low sensitivity, which means that a test is nothing to cheer about.

Additionally, according to the Science piece, other researchers were unable to replicate Kelsoe's findings, making the test yet more questionable.

The thing about bipolar disorder is that it can be diagnosed by (drum roll please)... interviewing a patient thoroughly! That's right, a well-trained interviewer can simply ask questions to determine whether an individual has bipolar disorder. Imagine that. There is often a hullabaloo made over patients with bipolar disorder being initially misdiagnosed as depressed -- the way to solve this problem is not to perform a fairly useless genetic test, but rather to actually spend time with patients, perform a thorough assessment, and listen to them. How's that for a wild idea? If your response is: "But there's no time to actually talk with the patients," then no cookie for you! It is likely true that many people later diagnosed with bipolar were initially seen in primary care settings for a brief appointment, in which they were diagnosed as depressed (the underlying bipolar piece was missed). Again, giving a scientifically dubious test because "Gee, it's based on genetics so it has to be accurate" rather than training physicians to improve interviewing skills will only worsen the problem.

When I have more time emerges, I will post again on the topic. This idea of genetic testing for mental disorders certainly needs much more attention. When academics go into marketing, strange things can happen, as I have documented here on many occasions.


Anonymous said...


While you would be hard-pressed to find someone more skeptical over the pervasive genetic reductionism in media and popular conceptions, I do want to sound some reservations regarding your point on interviewing and bipolar.

It is true that a well-trained interviewer CAN diagnoses someone with bipolar quickly, but the evidence suggests that such diagnoses do not happen quickly (I think the average period from initial consultation with a mental health professional to diagnosis of bipolar II or bipolar spectrum is 8-10 years).

Bipolar in all but its classic forms can, IMO, be extremely difficult to diagnose in no small part because it correlates so strongly with GAD, panic disorder, and depression. Moreover, as you obviously know, thinking about bipolar has changed dramatically in the last decade or so, such that what was once thought to affect a relatively small percentage of the population quite severely (bipolar I) has now expanded -- and not without controversy -- to incorporate a spectrum of persons exhibiting some kind of rapid cycling.

In any case, I don't disagree that a well-trained interviewer can diagnose bipolar quickly, but I do want to say that interpreting a patient's signs and symptoms as bipolar is most assuredly not facile. Rather, it seems to me to be quite a difficult endeavor, at least if we're talking about bipolar II or bipolar spectrum.

This speaks, of course, to the inherently uncertain and ambiguous nature of diagnosis in general and in mental health in particular. This is even more important because of the unequal power dynamics between provider and an already-vulnerable patient, and because those with mental health impariments, illnesses, or disabilities, have traditionally been among the most stigmatized groups in Western society over the past 200 years.

All this is just to say that diagnosis matters, is not something to be taken lightly, and at least as to bipolar, can be quite difficult.

I'm not suggesting you disagree with any of the above, and I certainly agree on the dubiety of the genetic testing you mention here.

--Daniel Goldberg

CL Psych said...


Thanks very much for your terrific comment.

I believe that the current genetic test has only been validated on bipolar I samples, but I would have to check more carefully to confirm.

Bipolar is indeed often a difficult diagnosis to make, which is why I think we often need to allocate more time to appropriate assessment rather than having very brief initial assessments. The hurried pace of medicine is surely partially to blame for many misdiagnoses across the board.

And if this test was quite sensitive in identifying bipolar disorder as well as specific (i.e., not having a high rate of false positives), then I'd have far fewer issues with it. Think about if such poorly validated tests became accepted practice -- that is my primary concern. Even if such tests were much more accurate, there would still be ethical issues raised, but there would at least be a strong scientific argument for the use of such tests.

BPD in OKC said...

I find the entire idea of these tests to be ridiculous

Radagast said...

"Nonsense on stilts," as Bentham helpfully points out, on a not unrelated topic.

You know, when this sort of crap emerges, I know I'm wasting my time? This is the Dark Ages, I'm convinced of it.