Wednesday, July 13, 2011

Do diff OCD sx differ genetically? Let's ask normal ppl.

Do diff OCD sx differ genetically? Let's ask normal ppl.

http://archpsyc.ama-assn.org/cgi/content/abstract/68/6/637


A Multivariate Twin Study of Obsessive-Compulsive Symptom Dimensions.
Alessandra C. Iervolino, PhD; Fruhling V. Rijsdijk, PhD; Lynn Cherkas, DPhil; Miquel A. Fullana, PhD; David Mataix-Cols, PhD. Arch Gen Psychiatry. 2011;68(6):637-644. doi:10.1001/archgenpsychiatry.2011.54

This study was recently published. It is a good example to support my idea that we generally should be wary of "findings" published in psychiatry. Apparently, anything goes.

It also is a good example of the leading way to critique any study: evaluate the degree that the hypothesis, what the study is supposed to be about, with what was actually analyzed - how were the different constructs measured, what was the study sample, etc.

I will note a little abt this study just to show how commonplace it is for "findings" to pop up in the literature, and be accepted as knowledge, when the actual evidence is nothing like the supposed finding or knowledge.

Here is the trick: recognize that for every scientific idea, there is the general, global statement being investigated, and then there is the technical specification of that. Kind of like the difference between what a new "law" is supposed to be, and what the legalese sounds like. The health care reform bill says a few things, such as "there will be these state 'exchanges," and "kids can be on their parents' health ins until the age of 25." Now, does it take 1,000 pages to say those things? No. But the law has to be technically specified with all of that stuff about "the party of the first party" and "the party of the second party."

So: Big Picture, and Devilish Details.
We often have these questions about some psychological issue. We go get articles, and try to figure out what evidence there is. Since the technical aspects can be difficult to understand, there is the opportunity for the authors to fudge, or embellish, or interpret beyond the data.
We should always have our skeptical-lens glasses on. We should always ev aluate the degree that the devilish details line up with the big picture - the goal.

Here is an example: these researchers claimed to investigate whether the genetics differ between people with different types of OCD. We know some ppl with OCD do washing, and some do checking. Are there genetic differences?

That is my rephrasing. We can look at theirs (it is not necessarily the case that authors actually tell you the big picture - here, they are not too clear).
From  the abstract, which should hit the essentials clearly and succinctly:
"Context: Obsessive-compulsive disorder (OCD) is clinically heterogeneous, but it is unclear whether this phenotypic heterogeneity reflects distinct, or partially distinct, etiologic mechanisms."
"Objective: To clarify the structure of the genetic and environmental risk factors for the major symptom dimensions of OCD."
(Did you notice the drift? No. Neither did I.)
"Results: A common pathway model did not fit the data well, indicating that no single latent factor can explain the heterogeneity of OCD."
(Of what? Of "OCD.)
"Conclusions: Obsessive-compulsive disorder is unlikely to be an etiologically homogeneous condition."

OK: look back at my paraphrase - very close to their "conclusion:" "OCD is unlikely to be homogenous."
(did you see that drift? Is "unlikely" in there because they found support for the competing hypothesis, and they are just refraining from strong wording so as to be more speculative, as early science should be? Or, are they saying "unlikely," because their study actually has no dirent relevance to this "conclusion"? -If it has no relevance, then why conduct the study and why publish it?
Serioulsy. This is a problem. We are failing to advance our knowledge of psych problems because we are goofing around with non-issues and worthless studies. This gets us nowhere.
How does this stuff get funded and published? We have to look elsewhere, like sociology and anthropology.

OK, MedsVsTherapy, what is your beef?
Here it is: genetic connections were studied between categories of obsessive characteristics and compulsive characterostics in - get this - totally normal ppl.
As far as we can tell, there was NO ONE with OCD in the data set. No one.
How do you study the genetics of OCD when your data set has no one with OCD?

You can't.

Look back: Big Picture, Devilish Details. The big picture is abt OCD types, and the devilish details are abt normal ppl.

Those of us who do not have OCD still recognize OCD traits in ourselves.
If you gave 4,000 of us a questionnaire that asked us how much we like to hoard things, or put things in order, or be scrupulous about washing for fear of germs, some of would score higher on each of these characteristics, and some of us woould score lower.

Then, you could look at our genetic similarities and differences. Do non-clinical OC traits vary by gen similarities and differences? Yes, no, or maybe.
This is what the study did. They looked at many pairs of twins. Not ppl with OCD. Why? Because they were there.

Do they make this clear? Not really. They drift from talking about abnormal to analyzing normal with little fanfare - they move from a Big Picture abt one thing to Devilish Details abt another thing without making it clear.

Now, go back to those statements I quoted earlier, and see how they drift. Now, you see the drift.
Why not study this in lab rats, and never make it obvious that rats were used?
Now - is this legit? Depends.

It is possible that the genetic variance in non-OCD ppl matches onto OCD ppl? Are the genes in a non-OCD person who scores high on "non-clinical hoarding" also having their influence upon the OCD hoarder? Possibly. But that is way mor efar-fetched than this study assumes. This study has little problem tying the two together tightly.

Consider this: just as we non-OCD ppl vary on washing compulsiveness, within a normal range, and we vary, within a normal range, on hoarding proclivity, we vary on height, and on forehead size.
If the taller among us share genes, does that mean that gigantism is caused by those genes?
Probably not. Gigantism is probably an injection of some other factor that makes that person uniquely different. A phenomenon beyond our normal height range.

If there is shared genes among those of us with bigger foreheads, does that mean elephantitis, like the Elephant Man, is a problem of those genes? No. The Elephant Man had something distincly differnt going on beyond what is going on for those of us with larger foreheads.

Ther is something distinctly different about someone who has OCD. Those of us with OCD have gone beyond the normal range of  obsessive thoughts or compulsive behaviors. This may or may not relate to normal OCD characteristics. what we do know is that those of us without OCD do not have whatever it takes to make one have OCD - whether our genes are one way or another.

The authors could clue us in regarding how extreme some high scoring people were in this study.
But they commit a major fundamental flaw - with the assistance of the journal's editorial board: the devil is in the details - nowhere so they report the actual scores on the OCD "trait" scale: means and deviations. Nowhere do they tell us whether a score of some level or greater is in the "pretty-darn-high, possibly OCD" level. Nowhere. So, we readers cannot get a grasp of the degree that higher scorers might simply be a bit more OC than the next guy, or whether high scorers are really alarmingly OC.

Not in the noted "online supplemental" info, either.

So, there you have it. If OCD is merely a matter of being a high scorer, as prompted by a strong genetic effect that is also seen at non-clinical levels, then OCD types are genetically different. But if OCD is a matter of some qualitative difference - maybe an additional genetic or environmental issue - then this study is 1. irrelevant and 2. misleading to valuable OCD research. This is at the heart of the matter in this whole study, but is not directly addressed.

They propose to study one thing, but actually study another. With this type of strategy, the field of psychiatry can mislead us abt all kinds of stuff. Now, ppl will be going around saying, "Iervolino found that the subtypes of OCD, including hoarding, are genetically distinct." THe present study speaks nothing to this.

I can't really blame the authors - hey, I need to get published, too. I will publish any half-baked, off-track study that Arch Gen Psych is willing to publish. The journal, however - they should know better.

No comments: