Psychiatry Gone Awry

June 10, 2007

Paris Hilton or Mary Winkler? Forensics Gone Awry

 paris and mary

 

I'll take Paris any day. 

So Paris goes back to jail after the behind the scenes/cover of darkness/MK-ULTRA deal she made to get out of jail early was met by the public with consternation.

As near as I can tell,  a/her private psychiatrist (his blog here-- mine's better, dammit) visited her for two hours in jail, then made a plea to the sheriff that serving her sentence in jail was psychiatrically harmful to her.  So they let her out to serve it at home.

The argument here, of course, is that this is rich-white-girl gets special treatment; and the easiest  way to do it is to use psychiatry. And people say, "see?  This is they type of abuse we can expect if psychiatry is allowed to influence legal matters."

Fair enough.  I don't know Hilton's case, whether it was a appropriate or not, I don't know Dr. Sophy;  all I can say is, yes, the potential for abuse exists, but perhaps it is balanced out by the cases in which it is helpful to society.
 

But consider the reverse situation, and read it carefully because then I'm going to punch someone:

SELMER, Tennessee (AP) -- A woman who killed her preacher husband with a shotgun blast to the back as he lay in bed was sentenced Friday to three years in prison, but she may end up serving only 60 days in a mental hospital.

Mary Winkler must serve 210 days of her sentence before she can be released on probation, but she gets credit for the five months she has already spent in jail, Judge Weber McCraw said.

That leaves only two months, and McCraw said up to 60 days of the sentence could be served in a facility where she could receive mental health treatment. That means Winkler may not serve any significant time in prison.

 

Same gripe: look how people use psychiatry to manipulate the legal system-- "only two months for killing someone?!" and while I agree that's pretty pathetic, what's worrying me is this: who the hell spends five months in jail without getting a trial? 

This probably didn't occur to you, and that's why it still happens.  If I kill my preacher husband, I have the right to a speedy trial.  If I can't get a speedy trial, I get to pay a fee to be released, and then show up in court when the government gets their act together.  But what if I don't have bail money?  How can the courts justify indefinite incarceration in the absence of a trial?

Enter psychiatry.  You get a psychiatrist to evaluate the person and determine that he is not competent to stand trial.  They recommend 60 days involuntary commitment/treatment in a psych hospital in order to "restore them to competency."  If at the end of 60 days the evaluator comes back, and if he still thinks they're not competent-- they get (re)committed again.  Etc.

Trial competency has very specific requirements: understanding how you wish to plea, the possible sentences, working with your lawyer, etc.  Being depressed or psychotic, per se, isn't the determinant-- only if it seriously impacts your ability to, say, understand the charges against you.

But in the vast majority of cases I have been involved in, the report really only reflects the presence of a mental illness, not its impact to the case.  As if it is de facto proof of incompetency.  It's not.

But here's the move: the "psych hospital" they get involuntarily committed to is actually their cell. 

Technically, they are supposed to be committed to an inpatient hospital.  Many jails have them on the premises.  But if the commitment is for 60 days, and the psychiatrist treating them (i.e. not the evaluator) thinks they are cured, then they get sent back into population (their cell).  Maybe they continue on medication; maybe they see the psychiatrist weekly for "outpatient" visits.

Or maybe, maybe, the treating psychiatrist doesn't think they need any treatment.  So they spend their commitment in exactly the place they started. 

Worse, much worse, is how many people I see that I say are competent and still wind up recommitted for two months.  Six months.  A year.  Think I'm kidding?  It is impossible to even estimate how many charts I have read that indicate no psychiatric contact-- not medication, not therapy, not psychiatrist--  for the entire duration of their commitment.  And why should there be? The treating psychiatrist doesn't see anything to treat.

You're probably thinking about murderers and rapists; but the majority of these cases are theft, assaults, drug possessions. Can anyone explain to me what possible justification exists for locking up a guy charged with possession for eight months, no trial?  And I'll pretend the guy is whacked out of his nut psychotic.  Ok?  Any justification at all?

I'm not saying you can't sentence him to eight months-- cane him, for all I care;  I'm saying you can't jail him for eight months without a trial.  Is anyone listening to me? 

The system is designed with simply one outcome in mind: keep the poor with high recidivism rates and minimal social resources in jail-- a sort of half-way house for the disenfranchised-- until you can't possibly justify it any longer, and then give them a quick trial, accept the guilty plea ("what guilty plea?") and sentence them to time served and probation-- where you can add further controls.

It's debatable whether keeping potential terrorists in Cuba is a good idea.  But when the State starts using pyschiatry to manage their population...

I know you think I am exaggerrating.   I'll bet you're not poor. 




Permalink | Comments (6)

Vote up Vote down Score: 7 (7 votes cast)


 

April 30, 2007

Does Media Reporting Of Suicides and Homicides Promote Copycats?

I won't give a detailed answer to this question here (it seems to be no), but there is an article making the reddit rounds now that I need to kill before it becomes another meme (like that other badly reported story about psychiatry.)

The article is from BMJ 2002, called Influences of the Media On Suicide, and it puts its conclusion right at the top:

Reporting and portrayal of suicidal behaviour in the media may have potentially negative influences and facilitate suicidal acts by people exposed to such stimuli. Recent systematic reviews by others and ourselves (unpublished) have found overwhelming evidence for such effects.1 (emphasis mine)

And it offers about 8 references in support.  And so now every nut with a microphone can proclaim it loudly: it's the media's fault.

We may want to take a pause and examine these 8 references: none of them offer anything close to "overwhelming evidence."  For example:

Reference 1-- the one directly cited for the above statements-- is indicative of the type of "overwhelming evidence" that exists. The study finds that media reporting of suicide is extensive and detailed, but not that there is a clear link to future suicides.  

In the summary, the authors use phrases like, "dearth of literature," "evidence is less reliable," "few studies permitting/demonstrating [the link]," "does not demonstrate consistency," "many studies fail to demonstrate" over 11 times in the 3 pages describing the studies. 

Despite this, they are sure the link exists-- but they don't actually show the link, they infer a link.    The authors repeat phrases, "it is fair to conclude that the evidence suggests an association [exists]" "tends to suggest," "probably reasonable to regard the association is causal"  13 times in two pages.  Under these criteria, it's reasonable to assume the Matrix is real.

Reference 3 (not even linked correctly) is a letter to the editor, describing two cases, where the method of suicide was affected by internet, but not the decision to commit suicide.  And the methods were rather weak: one guy took two pills of castor oil, and the other woman tried to drink water.  No, I'm not kidding.

Reference 5 is frequently cited in support of media's impact.  It supposedly says that a TV show with a Tylenol OD caused more Tylenol ODs: 20% of these suicidal viewers said it influenced their decision to attempt suicide in the first week post broadcast.  Maybe-- that 20% is really 6 people.  And most had attempted Tylenol OD in the past.  Oh, and the authors note that while 17% of the suicidal viewers' choice of Tylenol was influenced by the show, some of them chose not to use it because of the show.

Reference 12 is probably the most cited reference in this field.  In 1978 Vienna built a subway, which soon became a popular method of suicide.  So the government established guidelines for reporting-- specifically, that the method not be mentioned-- and subways suicides decreased by 80%.  Fantastic.  Overall suicide rates didn't change, though.  Too bad. 

So much for the "overwhelming evidence" for a soon to be media soundbite. 

The article doesn't make a good case for media influencing the decision to kill yourself, though I'll admit that it may influence the method.  And that's where it gets tricky.

It's important to make a distinction between copycat suicides and copycat homicides: more poeple die in the latter, and, let's postulate, they didn't want to die. That has to be part of the calculus in media reporting. Copying suicide by water (instead of pills) is different than copying a 30 person massacre (instead of killing, say, one person.)

But you have to weigh this against the societal costs.  The solution offered in these articles is to restrict media reporting.  I think we can agree that the media are neither liberal nor conservative, but  sensationalists, their bias is titillation.  But to allow anyone, especially government, to affect the content of reporting-- literally, the information we are allowed to have-- seems exactly the wrong solution to a problem which may not actually exist.  (e.g. I know it seems prurient, but I actually want to know all the details of David Kelly's suicide.)

Not to mention that if you say the media are partly responsible, then you're saying that you're less responsible.

(More on copycat suicides here, and on university suicides/copycats here and here.)



Permalink | Comments (9)

Vote up Vote down Score: 0 (0 votes cast)


 

April 26, 2007

"Inflammable Means Flammable? What A Country!"

You're going to say that I made this up.  I am not making this up.  I wish I was, because then I could say,"stop making things up," and everything would be ok.  But it really happened, and I can't stop things from really happening.  So I drink.

This is the conversation I had with another psychiatrist.  He is wearing a Ermenegildo Zegna suit.  It fits him well.

I say, "... so if I have him [hypothetical bipolar depressed patient] on Depakote and Seroquel, once he's stable I try to reduce the dose, or even stop one of the medications."

He shakes his head. "I would never stop the mood stabilizer."

"Why-- wait, which one's the mood stabilizer?"

"The Depakote."

"But how do you know it wasn't the Seroquel?  That's the problem with starting two drugs immediately, you don't know which one worked."

"Well, you need them both.  Especially if you're adding an antidepressant."

"What antidepressant?"

"Seroquel." 

"Seroquel's an antidepressant?"

"In this case it is, you're using it for bipolar depression."

I blink my eyes, to make sure I still can.

"Then what's the Depakote?" 

"The Depakote is the mood stabilizer."

"But how do you know the mood stabilizer isn't the Seroquel?"

"The Seroquel is the antipsychotic."

A voice tells me to stab him.

"Why can't you just use the Seroquel-- the antipsychotic- slash- antidepressant-- alone as a first try?  What benefit does the Depakote give you over the Seroquel?"

He says it slowly, enunciates, because he deduces that I don't speak English.  "Because the Depakote is the mood stabilizer..." 

He looks at me. I look at him. He is wearing a yellow Bolgheri tie.  In my mind it is on fire.

"...Besides, if you use Seroquel off label like that, you'll get sued."  Blaming lawyers, the last refuge of the incompetent.  "And for what?  Why take the risk?"

It's at this point I realize he's not wearing socks.  "If you're going to use that logic, Depakote isn't indicated for mood stabilization, either." 

He looks at me incredulously, then suddenly he realizes something.  "Oh, okay, right, I see what you're saying now... but at least Depakote has FDA approval for Bipolar Disorder."  QED.  He's very happy now.   I can't find the waiter.  Why is my drink empty?

There's an uncomfortable pause.  He wants to show me he's a skeptic, too, that he's carefully pondered these issues.

"I have a theory, have you ever used meclizine (a drug for vertigo) as a mood stabilizer?" 

"No, I had never heard of that."

"I haven't tried it either, but it might make sense: meclizine stabilizes your balance, so perhaps it could stabilize your moods?"

I want to call the State Board of Medicine but realize I'm in a different state and I don't have the number in my phone.  "I doubt the insurance companies would ever cover it."

He slowly, purposefully, nods his head.  "Fucking meddling managed care."

 

I tell people all the time, don't get sick, don't ever get sick, but no one listens to me. 

 



Permalink | Comments (8)

Vote up Vote down Score: 4 (4 votes cast)


 

April 16, 2007

Atypicals for Maintenance Bipolar

In my post on the NEJM article about antidepressants in bipolar depression, some people couldn't see how I made the leap to a political movement away from SSRIs and seizure drugs, and towards atypicals.

First, I'm not against atypicals.  I have long advocated for fluctuating doses of antipsychotic instead of Depakote.  I do think they can treat depressive states.  I don't disagree with the study or the data. 

What I find perplexing is the timing.  I was trying to show how academic psychiatry has now decided to move towards atypicals.  Why now?

Here's an example.  Eduard Vieta just released his hit single, Current Approaches to the Treatment of Bipolar Disorder With Atypical Antipsychotics, in Primary Psychiatry.  In it, there is only one short paragraph on Depakote, describing its one maintenance study, in which (it states correctly) Depakote didn't beat placebo.  That's it.  81 words.

Find me one other article written before 2007 that is so curt and dismissive of Depakote.

It goes on to explore the data on atypicals-- and there's quite a bit.  It rightfully concludes, "atypical antipsychotics have shown promising results in bipolar disorder maintenance therapy."

But here's the point: with two exceptions, all studies on atypicals referenced here came from 2004 and earlier.  The two exceptions were from 2005.

So it's not new data, it's old data.  Did they suddenly read the back issues?  Holy crap, atypicals might work?  That's why that NEJM piece is so important.  It marks the point where academia has decided to embrace atypicals and move away from Depakote.  If this move was really data driven, they would have done it in 2004.  Hell, they would have done it in 2001 when the Depakote maintenance study didn't beat placebo.

There's no conspiracy here, there's no exploitation of the weak for personal profit.  I'm not saying these are bad people, not at all. They are not conscious of it.  That's what makes this politics, not science.

You have an academic career, you want to do clinical research, who's going to pay?  NIH money is hard to get.  So you turn to Pharma.  You "get" to do a clinical trial of  Depakote for the treatment of bipolar.  When you're done, maybe they hire you to do another one.

You, personally, don't even get the money-- the department does, and they use it to pay your already set salary. But you get a career, an identity.   But you start to believe the prejudices of your chosen stomping grounds, and ignore the shortcomings.  You become a nationalist.  You start to believe that Depakote is first line, despite data; or that all seizure drugs will work; or that this thing you called bipolar is actually what you're treating.  That there are actually two poles. 

Then the money dries up.  But Abilify says, could you do a clinical trial of Abilify for bipolar?  And you say, sure, why not? maybe Abilify could be an add-on?  And then it's monotherapy.  And then Seroquel funds a study. 

The last part is when you don't simply move on from Depakote-- you distance yourself from it.  "You know, its data was never that great, it was really just an antimanic, and anyway, it had horrible side effects."  It's the next step of political hypocrisy: I was never really a citizen of that nation, I have really always been a citizen of the world.



Permalink

Vote up Vote down Score: 2 (2 votes cast)


 

April 4, 2007

Farewell, Depression



Write this day down: 4/4/07, it is the first day of the new psychiatry.  Everything changes, starting today.

Today, in the New England Journal Of Medicine, is an article ostensibly about the lack of additional benefit from adding an antidepressant to a mood stabilizer.  This is both surprising and not surprising: surprising, because, well, you'd think two drugs would be better than one.  Not surprising because, well, if the first drug worked, why would a second even be necessary?  (See #8). And if the first didn't work, how do you know the improvement didn't come entirely from the second drug

If this is all the article said, it would not be worthy of mention, let alone the herald of a new dynasty.

The study also found that the studied antidepressants did not induce mania.  That this should have been prima facie obvious even to a 9 year old without the benefit of eyes (what's an antidepressant?  They're not all chemically similar, so why should they all be blamed for the same side effects?) isn't the point here. 

The true importance of the study is contained in three statements.  If you blinked, you would have missed them.

The first is this:

Mood stabilizers were initially limited to lithium, valproate, the combination of lithium and valproate, or carbamazepine. In 2004, the protocol was amended to define mood stabilizers operationally as any FDA-approved antimanic agent.


The second is this, from the abstract:

Our study was designed to determine whether adjunctive antidepressant therapy reduces symptoms of bipolar depression...


And the third is this:

In summary, for the treatment of bipolar depression, we found that mood-stabilizing monotherapy provides as much benefit as treatment with mood stabilizers combined with a standard antidepressant.


Psychiatry is not about science, it is about language, politics.  What's happened here is that "mood stabilizer" now includes atypical antipsychotics; and-- compare what the study was designed to show and what they spun it to show-- we've gone from "polypharmacy is not better" to "monotherapy with mood stabilizers [read: antipsychotics] is just as good as two drugs at once." 

There's a subtlety there, and that subtlety is magnificent.

Note the authors: Sachs, Bowden, Calabrese, Thase, etc-- the same people who pushed psychiatry into flowchart polypharmacy; where Depakote was always first line for all phases of bipolar disorder, and any exacerbations that developed were treated with the addition of a second medication.

What the article is saying is that academic psychiatrists are no longer behind antidepressants and antiepileptics.  SSRI and SNRI use will decline from here, as will Depakote.  They're behind antispychotics.  And antipsychotic use is positioned to explode. 

It goes without saying: only the antipsychotics are still branded.

But without academics pushing SSRIs, their use will wane--and, importantly, so will their support of the diagnosis "Major Depression."  This is going to sound controversial, inane, but it will happen.

Look for upcoming articles finding that "Depression" is overdiagnosed, that it is really just-- life.  Look for articles that now find SSRIs aren't that effective after all, that the old "10% better than placebo" is a statistical trick with little clinical utility.  That they are way overused in kids.

You might say, wait, isn't the decline of polypharmacy  a good thing; that SSRIs are overused in kids; that they aren't that great; and that depression is overdiagnosed?  All of this is true, but this isn't psychiatry finally coming to its senses; this is psychiatry entering the manic phase.  Sure, it's less SSRIs for kids; but it's more antipsychotics.

Because simultaneously there will be articles pushing the idea that recurrent unipolar depression is really bipolar depression; that there are common genetic or heritability patterns; that the epidemiology and course is similar, etc. The move will be to squeeze out MDD into "life" and bipolar.  This done, antipsychotics become first line agents.  Oh, and look for antipsychotics to get FDA approvals for kids.

I wish I could make this clearer, but I'm still recovering from my recent bout with death.  There's no science here, only a tinkering with language and loyalties, with staggering results.  Don't blame Pharma quite yet-- this is a NIMH study.

I am not against antipsychotics, and I have long tried to tell anyone who would listen that the data clearly show they are superior to antiepileptics.  But this isn't psychiatry suddenly waking from a coma, aha! it turns out the existing data do show antipsychotics are mood stabilizers!    Instead of using them to replace antiepileptics, they will use them to replace everything: SSRIs, benzos, antiepileptics, stimulants, etc. 

And polypharmacy will only be reincarnated-- in the form of multiple simultaneous antipsychotics (Abiliquel, anyone?), with preposterous pharmacologic justifications ("this one acts on serotonin, so it's the antidepressant, and this one on dopamine, so it's the antimanic.")   If anyone says that to you, stab them.

You don't get many changes like this, maybe once every ten years-- the last was the beginning of the Depakote era, and before that was the advent of SSRIs, each with it's own erroneous semantics ("kindling model;" "serotonin model of depression.")

I wish all the patients in the world good luck, you'll need it.  Not because of the antipsychotics themselves, which will work or not, oblivious to doctor and diagnosis; but because of the doctors, who take little interest in examining the evidence behind their practice, and even less interest in reevaluating its core principles; and who lack the courage to even treat what they see, instead resorting to artificial, and wrong, paradigms and algorithms.  There's not even pseudoscience here.  Psychiatry is being lead by the siren call of semiotics, and it is saying, follow me, I am made of words...



Permalink | Comments (17)

Vote up Vote down Score: 2 (2 votes cast)


 

February 28, 2007

This Is How I Know Society Is Collapsing

From Psychiatry Vol 4. No. 2, Feb 2007 p.42:

 

"Second, paternal postpartum depression might be related to lower levels of estrogen."

 

I didn't read any further because both my eyeballs exploded. 

 



Permalink | Comments (6)

Vote up Vote down Score: 3 (3 votes cast)


 

 

For more articles check out the Archives Web page ››