This is an historical post from an earlier blog, Adopted Son of the Autism Family, which I had before this current blog. It is re-posted without modification (other than this introductory sentence).

First of all, I’m sorry for not writing more content for my blog recently. We’ve got something very exciting coming up at my Center, which is still unfortunately under wraps. It’s going to be public knowledge soon, but let me just say it’s incredibly exciting, and it’s kept me incredibly busy.

I’ve also been a little reticent to post on the topic of the DSM-V, for a variety of reasons, although I’ve tweeted about it and e-mailed privately with people I value highly in the autism community. This is in part because my own view, as a pragmatist, has been evolving, and in part because, the more I think about it, the question of whether the “DSM-V got it right” is really fundamentally the wrong question to ask.

Dr. Thomas Insel of NIMH wrote a post a few months ago that many of you have read. I think it raises some great points while simultaneously being somewhat disingenuous. He raises an analogy between how we diagnose complex neurodevelopmental disorders like autism and how we diagnose heart attack. I think this is as unreasonable as comparing how we diagnose cancer with how we diagnose leg fractures. Let alone that heart attack is ultimately an acute illness and autism is a developmental concern, this creates a false dichotomy in which psychiatric diseases are arbitrary and non-psychiatric medicine is black and white. I don’t mean this as disrespect to Dr. Insel — rather, I think that he (and the proponents of RDoC) are thinking radically, but they’re actually not thinking quite radically enough. I think this is true in part because RDoC is a NIMH project — a mental health project. Thus, it fails to be radical enough by failing to stop and critically ask how diagnosis occurs in complex disease states outside of of psychiatry. Even just a step as far as neurology is highly informative.

I think epilepsy is a great study, and it’s accessible to me, since I wrote my dissertation on epilepsy, and because, as a neuropsychologist, I’ve spent a large amount of time by the sides of children and adults both who have “neurological” disorders and those who have “psychiatric” disorders. Just briefly about epilepsy, it is a complex cluster of disorders that have the common feature of recurrent, unprovoked episodes of electrical abnormality in the brain (that is, epilepsy is, in essence, seizures that occur recurrently and at least sometimes in the absence of an acute medical cause such as fever). Diagnosing epilepsy is complicated — although there are some very well understood forms of epilepsy, in many complex cases, multiple diagnostics (EEG, MRI, PET/fMRI, MEG neuropsychological evaluation, etc.) can provide only partially overlapping results. Thus, some cases of epilepsy are not purely algorithmic — there is no conceptual machine whose crank can be turned and spit out a reliable diagnosis without use of clinical skill, at least not in all cases. Many cases of epilepsy are also “idiopathic,” or “cryptogenic,” meaning that, although we have possible models for how they came to be and why they occur, we don’t really understand their biological basis. The diagnosis of epilepsy is also evolving over time as new instruments and techniques come into use. Thus, epilepsy highlights the fact that complex disorders are not black and white, even outside of psychiatry. The ICD-9 has diagnostic groupings for epilepsy. The ICD-10 has diagnostic groupings for epilepsy. Note how very common conceptual classifications of epilepsy, like medial temporal lobe epilepsy, with mesial temporal sclerosis, are not in either the ICD-9 or -10. And yet the world does not end, and unlike children with autism, children with epilepsy (including children with autism and epilepsy) actually get access to proper and efficacious treatments … for their seizures. How do they do this? By way of classification schema that focus solely on the question of how we do best by people with seizures, rather than on some broader and amorphous question like “what is mental illness?”

This is where disingenuous comes in. These other complex disorders have in common that, although there are ICD diagnostic classifications (which change only slowly), there is international consensus driven by experts in that field who are mutually accountable and accountable to stakeholders in that same disorder community. It is the International League Against Epilepsy, ultimately, and not the American Academy of Neurology, that has traditionally taken the lead on epilepsy classification, and they do so with an eye to making progress in epilepsy, not worrying about whether there is some grand scheme in which the needs of people with seizures are not over-weighted against the needs of people with peripheral neuropathy. In contrast, in the DSM process, members of the autism professional community are picked by a team of psychiatrists, and then paraded in front of psychiatric review committees, and autism is put in the position of selling a diagnostic criteria onto psychiatry, for absurd reasons such as limiting the overall rate of psychiatric diagnosis. A sensible diagnostic process has to be driven by autism stakeholders, and not as part of some package deal for psychiatry, and this is not going to happen if “mental health” is held in a ghetto for “disorders” that aren’t “real.” (Which is a concept no parent of a non-verbal child with autism has ever worried about!).

The second major distinction between other complex disorders, epilepsy again as an example, and the way autism is treated today, is that it is well accepted outside of psychiatry that, in the post-modern era, knowledge can advance rapidly, and definitions will not be static. The diagnostic basis for autism should change when the data says it should change, not when there is a decision to make a new DSM. Other fields get by fine with making as-needed changes to the accepted practice for a disorder, without changing the related ICD codes. Note that this discussion does not really occur in other complex aspects of medicine — in spite of how quickly, for instance, cancer knowledge has been changing (while we are implementing, in the USA, an ICD that is already 20 years old). As far as I know, other fields dealing with complex disorders do not have or need a “bible” governing practice, but rather have many practice guidelines that are updated as needed and specific to diagnostic groups and entities.

So, my point in examining epilepsy is that, really, asking if the DSM-V got it right is the wrong question, because the DSM itself has become the answer to the wrong question. RDoC is a nice idea, but I’m not sold on it, particularly because it continues the tradition of an arbitrary ghetto that pulls psychiatric disease together and apart from other kinds of diseases (thus, for instance, emphasizing similarities and differences between autism and schizophrenia, over similarities and differences between autism and epilepsy). And, it is really not necessary, even with very complex disorders, to have a central “bible” that sets criteria in stone, in order to make progress (rather, to the point, one would have to ask why psychiatry has not far outstripped all other medicine, having had this advantage for decades).

In this way, I worry that the NIMH itself is as flawed a way to solve this problem as the APA (their APA, not ours) is. The real solution should start with the radical concept of simply considering these psychiatric disorders in the same way we consider complex disorders in other parts of medicine. I honestly do not know exactly what the answer is. I think there are almost certainly differentiable neuroendophenotypes within autism, and I think that mixtures of neuropsychological results, genetics, neurodiagnostics, and other markers will likely enable us to diagnose them correctly. I personally doubt that they happen to take the form of the DSM-IV framework. But, I think, in order to solve this riddle, we need to focus on autism and not engage in grand bargains, whether they be NIMH or APA driven.