The speculation about Donald Trump’s mental health that was doing the rounds earlier in the year seems to have died down a bit.  That’s to be expected; like it or not, his Presidency is now part of normal life.  But I’ve been lagging in my blogging here, and so it’s only now that I’ve got a moment to mention in passing an op-ed article about Trump in the New Scientist that appeared just after I posted last on the topic.  (February.  I know, I know.)

It’s by Allen Frances, and it takes issue with what he calls “armchair diagnosis” of the president.  He’s right to say that there’s something disquieting about armchair diagnosis: “psychiatric diagnosis is already done far too casually and inaccurately in medical and mental health practice.  Armchair diagnosis further cheapens its currency.”  However, I do wonder whether we ought to pay some attention to whose armchair it is.  Often, it’s an armchair occupied by the genuinely ignorant, or the spiteful.  That’s the internet for you.  Accusing someone of being mentally ill or having a personality disorder on this account may be simply mistaken; or it may be intended as a jibe, the subtext of which is that there’s something shameful about having a mental health problem.  But not every armchair is the same: as Frances’ article admits, a letter with 35 signatories who work within the mental health field appeared in the New York Times.  That letter may be misguided, or ill-motivated.  But it is by people who, presumably, know a thing or two about the topic.  Their armchair is not my armchair.

But there’s something else about the piece that’s just nagging away at me.  I don’t know a heck of a lot about mental health, but (and maybe that’s why) there’s a passage in the article that strikes me as being just strange:

But the main [reason for opposing armchair psychiatry] is the inaccuracy of the narcissistic personality disorder (NPD) diagnosis: Trump may be a world-class narcissist, but this doesn’t make him mentally ill.

I wrote the criteria for NPD for the Diagnostic and Statistical Manual of Mental Disorders, which guides mental health diagnosis in the US and beyond. These require not only that the personality features be present, but also that they cause clinically significant distress and impairment. Trump appears to cause severe distress in others (rather than experiencing it himself) and has been richly rewarded (rather than punished) for his self-promoting and self-absorbed behaviours.

[…] We must avoid the frequent mistake of confusing mental illness with bad behaviour. Most people who lie, cheat and exploit others are not mentally ill, and most mentally ill people do not commit dishonourable acts.

There’s a few things that are a bit odd about this.

In the first place, let’s accept that a criterion for Narcissistic Personality Disorder is that it should cause distress and impairment in the putative sufferer.  Fine.  But it seems to me that the people who’ve been diagnosing Trump have looked at him and his behaviour, and decided that something is amiss, and reached for the NPD diagnosis as the one that comes closest to capturing the things that worry them.  Maybe, on a literal reading, he doesn’t have NPD.  It doesn’t follow from that that he doesn’t have some kind of disorder, and maybe one that’s pretty closely related.

Now, this raises a question about who should suffer from a disorder in order for a person to count as having it.  Why should the existence or otherwise of a disorder depend on whether the putative sufferer suffers from it?  I’m reminded here of “Cupid’s Disease”, one of the stories in The Man who Mistook his Wife for a Hat, concerning Natasha K.  Aged 89, she suddenly started to feel “too well”.  Syphilis, contracted decades before, had sprung out of latency, and was – to put it bluntly – making her frisky.  It was also making her happy.  So here’s an example of a disease that doesn’t cause suffering in the person who has it.  Yet we’d still want to say that the infection, and the associated cerebral change, was there.  Now, this example caused no distress to anyone else, either – but it wouldn’t be too hard to come up with a scenario in which others, and only others, were distressed.

Once we’ve accepted that there can be all-things-considered pathologies that nevertheless cause no distress to the person suffering – that now seems like the wrong word to use, but it’ll do – from them, we would have to ask why that couldn’t apply to mental illness.  Trump may not be suffering; but it doesn’t follow from that that there’s nothing amiss.

As for the impairment: well, mightn’t it be that one of the things impaired is Trump’s ability to recognise his condition?

Whether or not it’s causing distress and impairment will feed into our reasons to treat an illness – in Mrs K’s case, there was a reason to ensure that the infection didn’t progress any further, but there was no reason to be worried about the damage that it had done: quite the opposite, in fact.  An illness’ causing distress to others may also give us a reason to treat it – though this will be trickier, because there’re going to be worries about violating the autonomy and physical integrity of the putative sufferer.  We might decide never to treat Jones on Smith’s behalf.  But this comes nowhere near telling us that Jones doesn’t have whatever the medical condition is in the first place.  Neither does it tell us that we don’t have a reason to treat; it’s just that in most cases it’ll be trumped (geddit?) by the reasons not to.

There’s another concern, too – though admittedly this might come down to phrasing in a short article.

Obviously, one wouldn’t want to second-guess the work that goes into a condition making its way into the DSM, but the phrasing “I wrote the criteria” does sound a little like an insistence that a condition exists iff it is defined into existence.  That seems strange.  “Trump can’t have this personality disorder because that’s not how I defined it” may work formally, but it risks severing the link between diagnosis and reality.  There’s any number of medical conditions that are defined in one way to begin with, but where the understanding changes over time.  For example, I understand that there are some researchers who are beginning to draw links between Alzheimer’s and diabetes.  It would be very strange for neurologists to insist that the former can’t be related to the latter because that wouldn’t fit the accepted definition.

Presumably, when a condition makes it into the DSM, it does so as a hypothesis – a way of explaining observed phenomena.  If we want to explain Alzheimer’s we start of with a hypothesis about what’s going on, and expand, contract, or maybe even abandon that hypothesis as necessary.  And sometimes that’ll mean that we can rule out a diagnosis.  Suppose that someone has a tentative diagnosis of Alzheimer’s disease, but then we look more closely at his medical history and decide that there’s something else going on.  That seems fairly unremarkable.  On occasion, we might even decide that he’s actually not ill at all (though that would, I’d guess, be very rare and utterly remarkable.)

What we wouldn’t do, though, is decide that there’s no illness simply because the phenomena that led to the tentative diagnosis don’t match the diagnostic criteria exactly.  (By analogy, it would have been absurd if, in the early 1980s, when HIV/ AIDS was beginning to emerge, we’d decided that there was no disease because what we observed didn’t fit our current framework.)  That does seem to be what Frances risks doing, though.  It’s one thing to admit that observed behaviours don’t match the diagnostic criteria as they stand; quite another to insist that you’re happy with the definitions you wrote, and that reality ought to bend around them.

Is Donald Trump mentally ill?  Does he have a personality disorder?  I don’t know.  What should we do if he is or does?  I don’t know.  Frances is certainly correct to say that “we must avoid the frequent mistake of confusing mental illness with bad behaviour [because m]ost people who lie, cheat and exploit others are not mentally ill, and most mentally ill people do not commit dishonourable acts”.  But “This isn’t a psychiatric problem and I should know because I wrote the definition” strikes me as unsatisfactory.