Shooting at Virginia Tech!

[IJ]

"A Montgomery County magistrate issued a temporary detention order Dec. 13, 2005.Cho was "mentally ill and in need of hospitalization and present an imminent danger to self or others as a result of mental illness, or is so seriously mentally ill as to be substantially unable to care for self and is incapable of volunteering or unwilling to volunteer for treatment," the order states."

I'll betcha, but cannot say for sure, that this is the language lifted directly from the preprinted form or template for a temporary hold in virginia, meaning that everyone a psychiatrist places on a hold meets these criteria. They meet them because the shrink wants them held and the law mandates that only such people be held, therefore, all people they want held get these words. So, if true, what does that tell us? That Cho was put on a hold. From what we've heard, he was put on a hold for SUICIDALITY. Suicidality is common, suicides are rare, suicidality does not predict homicidality or homicide. We have as yet heard no evidence that Cho presented any evidence of being a threat to others.

"The doctor, who works for a private practice in Blacksburg, reported that Cho's "affect is flat and is depressed," but "he denies suicidal ideations. He does not acknowledge symptoms of a thought disorder. His insight and judgment are normal.""

"Somebody really messed up here. Not even an acknowledgement of his underlying disorder. How can it not be seen by this bozo? Every lay person who knew him understood his communication issues. Doesn't this doctor even read the patient's medical history, for Christ's sake???"

So what does this quote actually say?

1) Cho was depressed appearing = probably accurate for someone referred for suicidality, right?

2) He denies suicidal ideation. All reputable psychiatrists comment on what the patient says about their suicidality, and we have no indication this doctor missed a comment from Cho, or that Cho was actually suicidal then. He didn't kill himself until over a year later, so that seems pretty accurate, especially since his main purpose wasn't to kill himself, but others.

3) He does not acknolwedge symptoms of a thought disorder (eg, psychosis). I have not heard anything to suggest otherwise at any time including the time that Cho was assessed. This doesn't mean he was "thinking clearly," it is a specific term which means he wasn't hallucinating or having problems with reality testing. Again, not everyone who thinks someone is out to get them will tell a psychiatrist, and Cho might not have felt this way then.

4) His insight and judgement are normal. Insight means that Cho is aware of his psychiatric condition (that he had depression and maybe autism, altho the word more conjures up a kid who barely speaks than a college student) and that doesn't seem far fetched. Judgement means that he probably told the shrink that he would go to an ER or call for help if he felt like hurting himself; that he had a decent plan for coping with his situation.

I don't think your article shows this doc was a "bozo." You're implying that his outpatient plan of care for this kid was unreasonable, and there's no evidence to that effect from the FRAGMENT of a note that was quoted. You can't conceive that the items stressing normalcy were the headline grabbers here? If not, consider that the FACT was that Cho posed an imminent danger to HIMSELF and the headline deliberately omits this detail, rendering him a succinct, headline worthy "Imminent Danger !!!" The reader is invited to wonder what of--probably a school shooting, eh?

Well, Cho's doc was asked to assess for suicidality, and found none, and guess what--he was right. Nothing happened for a year. It's as if you sent me a patient with suspected pneumonia, and he was obese and had high blood pressure, and I did my exam and found he had bronchitis, not pneumonia, and sent him home from the hospital with an outpatient appointment--then you're blaming me for him having a heart attack a year later. Psychiatrists cannot hold people because they might go nuts in a year, because they'd have to hold way too many people forever in a violation of their rights.

If he HAD been admitted, what do you think would have happened for the 1-7 days he might have spent inpatient? They would have.... cured his autism? Even his depression? He would have gotten a few days of meds and some counseling and after, he would have the OPTION of continuing them (note: recently implicated were the most commonly used meds for treating depression in WORSENING suicidality (not suicide), although this was just proven less of a concern in the last week).

In virginia, I met a floridly psycho ex law student, very smart, who would go off his meds, slowly go nuts, and then get brought back in a month later for bizarre, assaultive, or destructive behavior. It would take weeks to get the court order for forced treatment, to stabilize him (at which point he'd write letters of apology) and once he left, he'd eventually quit the meds again. Repeat as necessary. These are the rights of the floridly schizophrenic in the USA, and this guy had done far weirder stuff than Cho (eg he tried to strangle one of my friends on the locked psych ward). Autistic kids aren't well known for violence, althouygh I did find this article on atusim and ASPD in Swedish killers:

http://jiv.sagepub.com/cgi/reprint/21/8/1081

This is a typical finding: " One or several childhood-onset neuropsychiatric disorders [autism, attention-deficit/hyperactivity disorder (AD/HD), tics and learning disability] affected the majority of adult offenders." : Nord J Psychiatry. 2005;59(4):246-52. So what does this kind of stuff tell us? First, we've lumped in tics, ADHD, and learning disability with autism, which is hardly useful--heck, 25% of the women in this country meet criteria for major depression alone; my little brother has a learning disability; a coworker his tics. Are they all indefinitely holdable?

Autism doesn't see a mention of violence in its criteria for diagnosis:
http://ani.autistics.org/dsm4-autism.html
Autism+violence got me 60 hits on pubmed (schizophrenia and violence got me >800).

Or take a look at this abstract:

Isr J Psychiatry Relat Sci. 1998;35(4):293-9. Links
Aggression and sexual offense in Asperger's syndrome.Kohn Y, Fahum T, Ratzoni G, Apter A.
Psychiatry Department, Hadassah University Hospital, Hebrew University-Hadassah School of Medicine, Jerusalem, Israel.

Asperger's Syndrome is one of the diagnostic subcategories of pervasive developmental disorders. It is characterized by a defect in reciprocal social interaction, lack of empathy for others and poor non-verbal communication. Antisocial acts, including aggression and sexual offense, are not considered to be common in this disorder. We describe an adolescent with Asperger's Syndrome whose main problems are his violence and sexual offenses. We assume that these symptoms are secondary to his diagnosis of Asperger's as a manifestation of his difficulties with the "theory of mind" of others. This atypical case report is in contrast with the low prevalence of aggression and sexual offense in Asperger's, as reported in the literature. We discuss the reasons for this low prevalence. Our conclusions are based on one case history and a literature review. We call for further research in this field.

Aspergers is a related disorder to autism--see largely overlapping criteria at: http://www.autism-pdd.net/dsm4.html

Someday, some aspergers person is going to kill someone. It doesn't tell us how to prevent the problem.

[Bill Glasheen]

Ian

Read what I wrote. I'll put money on Cho not having Asperger's.

Let me be less subtle.

He does not acknowledge symptoms of a thought disorder. His insight and judgment are normal.

• It isn't the duty of the patient to diagnose their own thought disorder, nor are they usually capable of doing it. That is the sole domain of the physician(s) diagnosing the patient.
• He had pretty significant communication issues. A creative writing professor reported that it took 20 seconds for him to respond to questions. People who lived with him reported that he wouldn't look you in the eye, and responded with single words.
• The REASON why he was forceably held in the first place was because he had suicidal ideations.
• The suicidal ideations were subsequent to a SECOND unlawful incident where a woman stalked by Cho reported him to the police.
• Let us not forget Cho's first victim - a woman. That happened 2 hours before the rampage.

Ian if you didn't take great delight in debating me and this was one of your interns, you'd be all over them like white on rice.

Yes, I call said physician a bozo. Unfortunately said health care professional is not alone in level of incompetence.

Well, Cho's doc was asked to assess for suicidality, and found none, and guess what--he was right. Nothing happened for a year.

Joke, right?

Cho was admitted because he expressed an interest in committing suicide after being intercepted by police for stalking.

Cho subsequently committed suicide 2 hours after killing a woman and minutes after he slaughtered 30 in a murder-suicide spree.

If he HAD been admitted, what do you think would have happened for the 1-7 days he might have spent inpatient?

What would have happened, or what SHOULD have happened? There is a difference.

Someone needed to untie the Gordian knot, and deal with the myriad issues. Unfortunately not every health care professional has the talent and passion of the shrink in Good Will Hunting. Speaking of which... That movie speaks to this situation quite nicely.

[Bill Glasheen]

Sean: Hey, Gerry, In the 1960's there was a young man that graduated from the University of Michigan. Did some brilliant work in mathematics. Specifically bounded harmonic functions. Then he went on to Berkeley. He was assistant professor. Showed amazing potential. Then he moved to Montana, and blew the competition away.

Lambeau: Yeah, so who was he?

Sean: Ted Kaczynski.

Lambeau: Haven't heard of him.

Sean: Hey, Timmy!

Tim: Yo.

Sean: Who's Ted Kaczynski?

Tim: Unabomber.

Lambeau: Yeah, you were smarter than me then, and you're smarter than me now. So, don't blame me for how your life turned out.

Sean: I don't blame you! It's not about you, you mathematical dick! It's about the boy! He's a good kid! And I won't see you f*** him up like you're trying to f*** up me right now! I won't let you make him feel like a failure too!

[IJ]

"Read what I wrote. I'll put money on Cho not having Asperger's."

Read what I wrote; I didn't say he had it.

"It isn't the duty of the patient to diagnose their own thought disorder, nor are they usually capable of doing it. That is the sole domain of the physician(s) diagnosing the patient."

True, and this diagnosis is made with BOTH observation of, and interview with, the patient. Large parts of psych diagnoses, especially assessments for suicidality, depend on what the patient says. You can like this, or not, but it won't change.

"He had pretty significant communication issues."

So? The physician quote doesn't address communication ability. He might have mentioned it in the next sentance. Either way, I haven't heard that poor communication skills make you suicidal or homicidal, certainly not to an extent that justifies a pysch hold.

"The REASON why he was forceably held in the first place was because he had suicidal ideations. The suicidal ideations were subsequent to a SECOND unlawful incident where a woman stalked by Cho reported him to the police."

And...? Lots of people feign suicidality or have suicidality when the law catches up with them or they get reported. Does that mean.... we should hold all of them? Only if they have a real liklihood of suicide. That determination isn't made by finding out why they were suicidal, but investigating their story more. And I'm not saying the guy was prince charming or that harassing a woman is excusable. Not in the slightest. But you would agree with me that thousands and thousands of men have harassed women, right? And they don't all have shooting sprees, so the psych doctor can't/shouldn't hold him as a danger to others because he stalked someone. Under Tarasoff laws, if there is a credible threat to others, he HAS to report it to the potential victim, the police, as needed--but evidently there wasn't an imminent threat to this woman. He may have killed a woman a year later... but we can't work backwards. Otherwise you'd have to claim that everyone who ever committed rape or murder or assault "SHOULD" have been forcibly committed to psych when they saw any medical or legal authority in the years before. These events are sadly nowhere near that predictable.

The other item you're missing here is that harassing someone is not a psych problem. It's a legal problem. My other half sees people all the time in the ER that some doc wants psych treatment for because they're violent or whatnot, and half the time, the diagnosis is "thug" or "criminal" and they don't need a doctor, they need jail. For Mr. Cho, we've yet to establish that his stalking someone was a prominent feature of a mental disorder or that his stalking means he should be treated any different than any other stalker--reported to the police, being subject to a restraining order, etc. We do not hospitalize stalkers.

"Let us not forget Cho's first victim - a woman."

Exactly how was the psychiatrist supposed to forsee this event one year before it happened with any kind of sensitivity or specificity?

"Ian if you didn't take great delight in debating me and this was one of your interns, you'd be all over them like white on rice."

No, I would not, because I haven't spoken to this doctor or seen the whole asessment. You're happily accusing this person of bozohood on partial information, which is precisely how we get mistakes like the Duke nonrape scandal. More to the point, I am not a psych attending. I am, however, 4 years into dating a psychiatrist who has sent home dozens of people as disturbed as Cho apparently was (based on data thusfar presented) at the time of his eval. And that has been with the express approval of the attending psychiatrists on service--about a dozen of them.

"Joke, right? Cho was admitted because he expressed an interest in committing suicide after being intercepted by police for stalking. Cho subsequently committed suicide 2 hours after killing a woman and minutes after he slaughtered 30 in a murder-suicide spree."

Sadly, no joke. Did you read my example how medical doctors can't prevent a heart attack one year in the future by admitting someone for bronchitis when someone else felt they had pneumonia? Same in psych. They can't prevent homicide one year in the future when they eval someone for suicidality and they turn out not to be acutely suicidal. How could they? Cho could have been held if he was an IMMINENT danger to self, or others, or demonstrated an inability to care for himself. That is the language of the STATE, written to protect the rights of patients from psychiatrists. He was an imminent nothing by that doctor's assessment, and 1) it's hard to know what Cho looked like at the time so we ought to be careful about second guessing and 2) in retrospect, he clearly WASN'T an imminent danger... he was a longterm risk. The STATE law does not say, "forcibly restrain this man if he MIGHT kill someone in a year." It cannot, because the psychs would have to hold nearly everyone.

Just as people can die of a heart attack walking out of a doctor's office where they were told there was no evidence of heart disease. The doctor is NOT at fault unless signs were overlooked. You haven't presented sufficient warning signs of ACUTE danger to self or others. He stalked someone; stalking is sadly common and a report of harassment is not a sufficient reason to forcibly hospitalize someone (or jail them, even). Communication issues do not deprive you of your personal freedoms either.

"What would have happened, or what SHOULD have happened? There is a difference."

Sorry Bill, but what you're saying here is that we SHOULD have hospitalized this guy for being suicidal on the off chance it might have prevented a homicide (different thing) a year later. That does NOT fly from someone who rails against the unsubstantiated bone marrow transplant for breast cancer (an unproven nice thought at the time, which turned out not to work). You cannot take away a man's rights because it could conceivably avert disaster later; our health care system cannot support the caseload that would result and the law would not permit it.

You always seemed willing to acknowledge that when a medical error occurs, the best approach is to understand why and to fix the problem, not the blame--by looking at the system that delivered the care and not the individual. Here, you're committing the name-calling equivalent of the malpractice lawsuit by calling this guy a bozo. He was one person who saw the patient once (or so it appears). He was part of a SYSTEM which failed to address Mr. Cho's behavior over an extended period of time, that includes a lack of aggressively funded counseling, school policies that prevented Tech from getting rid of this guy which are common to most schools (ref: CNN), and a legal landscape that demands a clear and present danger before people are deprived of personal freedoms. Speculation about impacts on far future events is not permitted. This psychiatrist, based on what was written, had no cause to detain Mr. Cho, and no single evaluation by a physician is a guarantee that no future problems will occur. You know this, I hope, and it's just the horrible situation that's allowing you to overlook the limitations of psych evals.

"Someone needed to untie the Gordian knot, and deal with the myriad issues. Unfortunately not every health care professional has the talent and passion of the shrink in Good Will Hunting."

And unfortunately not every health care professional has a bright, interactive, nonpsychotic patient like Matt Damon's character to see in repeated, lengthy visits funded by an award winning mathematician and who only has to overcome a negative attitude to accept the fame and love that a hollywood script laid at his feet, right? C'mon. That was a wildly unrealistic movie... Do you expect every public defender to go Johnny Cochrane on every defense? Do you expect your san kyus to pull a Steven Segal and disarm an armed gang in a jewelry store? In real life, someone with matt damon's "problems" wouldn't have received insurance approval to see a psychiatrist. He would have been referred to a few psychologyt counseling sessions perhaps, then told his benefits had elapsed. Who would be the bozo then?

[Bill Glasheen]

You are rationalizing this physican's actions, Ian. I on the other hand see the big picture pretty clearly here.

I have dealt with such people, and have seen to it that they got the help they needed. It took something more than CYA logic to do it. It took creative, persistent, and caring actions.

I'm saddened that more people don't act this way - in the name of the patient.

- Bill

[IJ]

"You are rationalizing this physican's actions, Ian."

I am saying the brief excerpt is not damning enough to justify your vitriol. What were the actions? We haven't seen the whole eval, or what was done. If, as is quite possible, Cho was brought to an ER doc who consulted psych at 2am, and that psych doctor concluded (rightly) that the patient could not be legally held, then the patient may leave. The doctor may also have denied the option of offering voluntary admission by insurance, or Cho may have declined. The doctor may have had no option but to offer meds (frowned upon for psychiatrists with no therapeutic relationship) or an outpatient referral, which he may have done, and may or may not have worked. Remember that we have Tarasoff laws because Tarasoff was killed by a psych patient under active treatment. Either way, you still havent suggested in the slightest how admitting patients like Cho for depression might make a dent in their subsequent massacre rates. That's an unjustifiable stretch.

"I have dealt with such people, and have seen to it that they got the help they needed. It took something more than CYA logic to do it. It took creative, persistent, and caring actions."

Sigh... well, you know, I happen to have run into a few of these patients myself during relevant lectures and then 10 weeks of psychiatry in med school, and 6 years of clinical medicine seeing such people in the ER, the wards, 3 san diego jails, and even a supermax security prison in virginia, and my partner is 3 years into a psych residency that I hear much about. I wager I've been exposed to a bit more information about medicine, the law, and their intersection with such patients.

"I'm saddened that more people don't act this way - in the name of the patient."

Well, if you can't substantiate your implication that this doctor missed something and should have known Cho was one of say 10 (out of 260 million) to shoot up a school, you might as well disparage the intentions and concerns of any one who disagrees with your assessment.

Fact is that if my partner took a "Good Will Hunting" personal interest in all of the patients with problems of Cho's magnitude, he'd have to work 100 hours a day doing it. ER consultants called to assess risk to self others or inability to care for themselves are NOT the primary physicians caring for the patients (unless we're at a county mental health facility) and they have substantial limits on their ability to act imposed by time, the law, and other aspects of our healthcare system.

I'm not saying that we couldn't have done more for Cho, but if you think you can push expansive mental health services in an era of exponentially rising healthcare costs through legislatures, or believe that the best response to a system that delivered suboptimal care is to call one individual working in it a "bozo" after reading a few lines from his assessment, you've actually missed the "big picture." This guy is innocent until proven guilty in my mind.

[Bill Glasheen]

I'm not saying that we couldn't have done more for Cho, but if you think you can push expansive mental health services in an era of exponentially rising healthcare costs through legislatures...

Thank you, Ian. You help me make my point by making this ridiculous assertion.

If anything I am a proponent of evidence-based care. I devise the algorithms to measure that on a daily basis while also measuring economic efficiency. And you know what? Now and then physicians actually save money when they follow their own medical guidelines, or do what they know is the right thing to do. Wow!

You're a smart guy - smarter than most people I've ever met. But sometimes you argue just to argue.

I stand by my assertion. You may disagree; it is your prerogative. And that's what we do here. Through debate and occasional disagreement, we do our best to get our hands around "truth."

I KNOW you are not that cynical, and I am not THAT certain of my diagnosis. But with my perfect 20/20 hindsight, I feel pretty comfortable calling it like I see it.



- Bill

[IJ]

"If anything I am a proponent of evidence-based care."

Ok, what is your evidence that Cho could have been held (under the law, that he presented an imminent threat to self or others or could not take care of himself (eg, beg for food)) when this psychiatrist saw him? And what is the evidence that admitting a patient for suicidality averts massacres a year later?

Recently, the CDC advised testing all adults for HIV, and the tests are 99.9% reliable. Sounds great! But when the prevalence is low, the number of false positives goes up. If your prevalence is one per....

100 patients: you get no false positives and a real HIV.
1000 patients: you get 1 false positive and 1 real HIV.
10,000 patients: you get 10 false positives and 1 real HIV.
100,000 patients: you get 100 false positives and 1 real HIV.

Now you have lots of money spent and lots of scared people getting followup tests, but little detection to show for it.

Same logic applies to holding people against their will. What risk of suicide justifies holding someone? Let's say the average risk is 1%, and hospitalization cuts this by half. [I'm guessing at these numbers but we know that only 1/3 patients taking an SSRI (prozac type) benefits; 1/3 don't get better and the other 1/3 get better no matter what you do; psych interventions just aren't surefire]. At these numbers you have to detain 200 people to save a life. If the risk is half a percent and the benefit is 33%, now you need to detain 600 people. What rate of return justifies loss of liberty? What rate of return is feasible? We can't hold everyone who's depressed, for example... not enough beds.

So why dont we each do some research and come back to the table with evidence that Cho should or should not have been held? I've been talking about suicide prevention, but you've been claiming his restraint would have prevented his homicides. Perhaps you can cite some evidence that some feature of Cho's presentation predicted a homicidal outburst and that hospitalizing him would have prevented the subsequent killing? And we can see what the ROI is.

"Now and then physicians actually save money when they follow their own medical guidelines, or do what they know is the right thing to do. Wow!"

You're implying that this psychiatrist
1) didn't follow his own guidelines and or
2) did something deliberately wrong.

You haven't justified these claims. I've explained what the psychiatrist has to show to hold someone, and you've not suggested in the slightest how Cho met these criteria. You then take a huge leap of faith and imply that holding someone has a violence preventing effect that lasts at least a year. If you would like to state what guidelines the doc violated, that would be helpful. If you would like to explain how an excerpt of an evaluation proves negligence, that would be helpful. If you would like to explain how you know this doctor deliberately ignored the standard of care, that would be helpful.

But all you're really doing, apparently without any experience in the practice or law of psychiatric holds or any citation of relevant literature, is concluding that a bad outcome automatically means a bad choice was made long before. This, from a self identified evidence enthusiast who has repeatedly excused the entire Iraq war disaster and brushed aside the tens of thousands of deaths it led to by saying that prewar intel is fallible??

You can tell me that I'm arguing for the sake of arguing, but that would be the pot calling the kettle black.

PS: I showed the article excerpt to 3 psychiatrist acquaintances and asked how the psychiatrist should have intervened to prevent the shooting, and they were all incredulous as to how this person would have anticipated the event or provided some life-course altering drug, therapy, or preventive incarceration to ward it off.

PPS: the fact that quality improvement methods CAN lead to cost savings does not invalidate my argument that Virginia didn't/wouldn't allow taxes to skyrocket to over expansive mental health services. Institutions make shortsighted decisions all the time, but the single psychiatrists working in those institutions are not responsible for the rules that govern psych admissions. Further, you've yet to show that expansive mental health services WOULD lead to cost reductions.

[IJ]

Here's the best source I could find about suicide prevention, a nice review published last year.

Predicting and Preventing Suicide: Do We Know Enough to Do Either?
Author: Joel Paris
Published in: Harvard Review of Psychiatry, Volume 14, Issue 5 October 2006 , pages 233 - 240

Relevant excerpts:

"Suicide completers have a profile that differs in important ways from that of suicide attempters. Completers tend to be older and male, to use more lethal methods, and to die on the first attempt.4 Moreover, many patients, particularly in younger age groups, who complete suicide tend either not to seek help or to avoid it. In a psychological autopsy study of young adult suicides, Lesage and colleagues5 found that among 75 completers between ages 18 and 35, less than a third were in treatment at the time of their deaths, that fewer than half had seen therapists during the previous year, and that a third had never even been evaluated. Hawton and colleagues6 reported similar findings: among 174 suicides in a cohort of adults under age 25, only 22% were in treatment.

Perhaps the safest conclusion is that somewhere between 3% and 7% of all attempters will eventually kill themselves, higher rates being associated with severe attempts and with repeated attempts. At the same time, attempters also have excess mortality due to a variety of other causes.32

The problem is that none of these data tell us how to assess specific attempts: we cannot know whether individual attempters are at risk, or whether patients who eventually kill themselves could have been identified at the time of a previous attempt. Moreover, attempters who complete suicide might seek help for particular episodes, but not for the ones that lead to their deaths.

Nevertheless, while scales measuring intent are capable of predicting completion with statistical significance, they identify only populations at risk, not individuals who are likely to die by suicide. For this reason, they may not be useful in making clinical decisions.

Finally, diagnosis can be used to assess the risk associated with a suicide attempt. Many mental disorders are associated with high suicide rates: schizophrenia, melancholic depression, bipolar illness, and alcoholism all have completion rates as high as 10%,37 and borderline personality disorder has a similar rate.38,39 But diagnosis, like other population-based risk factors, cannot determine whether the likelihood of suicide is high in any individual case or at any particular moment.

To prevent suicide, it must first be predicted. Clinicians have been trained to identify patients who have the risk factors associated with completed suicide. But they cannot know which patients will eventually die, or use this information to save them.

The key issue is the relative rarity of completion: when an outcome is uncommon, prediction has to be difficult. The lifetime rate of completion in the general U.S. population has been estimated at 1%,1 which stands in contrast to a 5% lifetime rate of attempts and a 15% lifetime rate of suicidal ideation.40 However, since populations of completers and attempts are distinct (albeit overlapping),4 most people who think about suicide or attempt it never die by their own hands.

Thus, the problem in predicting suicide from ideation, from attempts, or from other risk factors is the predominance of false positives. Most people who carry population-level risks never commit suicide. The same principle applies to intent: even if it predicts completion in large samples, it cannot assess the risk in an individual. There are also false negatives, since patients who do kill themselves may not always have commonly identified risks. Goldney41 concluded that "the sobering reality is that there has not been any research which has indicated that suicide can be predicted or prevented in any individual." Two large-scale studies42,43 nicely demonstrate this point. Each was designed, using standard risk factors, to predict suicide in populations of patients admitted to hospitals and followed over several years.

Pokorny and colleagues42 followed 4800 patients admitted to the in-patient ward of a Veterans Hospital. A logistic regression included the following predictors: attempted suicide, suicidal ideation, a diagnosis of affective disorder or schizophrenia, depressed feelings, recent history of violence, low social interest, urge to do harmful things, fear of losing control, remorseful feelings, impatience, and feelings of failure. But the model failed to identify any cases in which suicide occurred.

Goldstein and colleagues43 also attempted to develop a statistical model that would predict suicide in a group of 1906 patients with mood disorders admitted to a tertiary care psychiatric hospital. The risk factors included the number of prior suicide attempts, suicidal ideation on admission, bipolar affective disorder (manic or mixed type), gender, outcome at discharge, and unipolar depressive disorder in individuals with a family history of mania. Again, the model failed to identify any of the patients who committed suicide.

In summary, given our present knowledge, even among high-risk samples of patients admitted to hospital for mental illness, it is not possible to predict suicide with any degree of accuracy. This conclusion has been supported by a recent meta-analysis,44 as well as by a systematic review.45

[since gas, medicine, and gun restriction work best] ... the most convincing evidence that suicide can be prevented comes from interventions not carried out by mental health professionals. This approach is consistent with a population-based strategy.73"

This whole article is about suicidality, not homicidality, but that is the situation the psychiatrist faced when he evaluated Cho. My larger point is that when one cannot predict suicide in those presenting with suicidality with any success, the prediction of homicide in individuals presenting with suicidality is next to impossible. Such is the evidence.

[Bill Glasheen]

You are missing the point here, Ian. You are thinking in one dimension.

This is something I have specialized in. I have built models to predict rare events like who is likely to be admitted to the hospital or who is likely to be in the top 1% of costs next year. Want to know the first things that the inferential statistics tell you? (This would be by regression, trees, neural nets, or clustering techniques) No ONE disease will make you a zebra. It isn't a single, severe condition. It's the combination of things which create the high probability of bad things happening.

A good clinician for example knows that if you have 3 body systems that fail, you are very likely to die. The literature shows it. A good internist knows it. A good clinician knows that a depressed diabetic is trouble. Etc., etc.

It's the big picture. Go back and see the combination of things that were happening. The autism (or perhaps pre-schizophrenic condition). The stalking. The depression. The suicidal ideations. The antisocial behaviors reported by virtually everyone who knew him. It's the witch's brew which creates trouble in medicine. Screw just suicide; look for the patient in deep trouble.

Look in the literature all you want. I have years of experience with proprietary statistical models built from tens of millions of patient's medical claims history.

As you well know, the field of mental health is an inexact science in a relative stone age. But as a chief medical officer once told me, "Sometimes even a blind squirrel can find a nut."

One additional comment... You sound a lot like an MD rationalizing a bad outcome. Unfortunately our litigeous society (and trial lawyers) have gotten MDs away from open communication about process and CQI.

Here we have a KNOWN bad outcome. This is easy. Rather than rationalize why the MD did everything right, go back and look for what the "system" missed. Sometimes odd outcomes are due to mathematical chaos, and can't be predicted. But sometimes the evidence is all there - in hindsight - so plain that it would bite you in the arse if you got close enough.

Stay away from de Nile and start looking for clues. Certainly the FBI is beginning to profile such individuals by gathering a number of such cases. The patterns are starting to come forth.

How can the system get it right next time?

- Bill

[IJ]

"You are missing the point here, Ian. You are thinking in one dimension."

I am responding to your unfounded accusation that this doctor was a bozo. Don't want responses to unfounded claims? Don't make them.

"This is something I have specialized in... It's the big picture. Go back and see the combination of things that were happening."

As you're aware Bill, this is also something I specialize in. However, you're convinced that because you have worked in models that predict which people are at risk for hospitalization, that 1) you're an expert in violence prediction 2) apparently that there were indications that this guy was an imminent threat and should have been held by that psychiatrist. That confidence and after the fact judgement are unfounded.

If you would go back to the article I excerpted, you might notice that the author reviews all of the relevant literature from medline and finds that we can only comment on which populations (not individuals) are going to have these rare events. Specifically, after studying combinations (yes, combinations!) of risk factors, the best scientific approach in the area was unable to predict any of the subsequent suicides. I'll re-cite it:

"Pokorny and colleagues42 followed 4800 patients admitted to the in-patient ward of a Veterans Hospital. A logistic regression included the following predictors: attempted suicide, suicidal ideation, a diagnosis of affective disorder or schizophrenia, depressed feelings, recent history of violence, low social interest, urge to do harmful things, fear of losing control, remorseful feelings, impatience, and feelings of failure. But the model failed to identify any cases in which suicide occurred.
Goldstein and colleagues43 also attempted to develop a statistical model that would predict suicide in a group of 1906 patients with mood disorders admitted to a tertiary care psychiatric hospital. The risk factors included the number of prior suicide attempts, suicidal ideation on admission, bipolar affective disorder (manic or mixed type), gender, outcome at discharge, and unipolar depressive disorder in individuals with a family history of mania. Again, the model failed to identify any of the patients who committed suicide."

You write as if you're the only one who's aware that multiple risk factors identify the highest risk patients, as this escaped the notice of the clincians and researchers who have made suicide and violence prediction their life's work and daily practice. It's strikingly naiive. They know. They have tried to predict these rare events. It can't be done at present: no effective predictionn rule was available to that doctor, a fact you continue to ignore. The rarity of the events and the commonality of the risk factors with the inherent variation of individuals and their circumstances make this very difficult especially farther out. Those with actual experience in the field have found it frustratingly difficult, and yet without a day (correct?) of psych experience, you have the confidence to imply that you can identify the next shooter beforehand. Why haven't you worked in the stock market, weather, or if a car of yours has ever broken in your possession, why did you let it happen?

"one additional comment... You sound a lot like an MD rationalizing a bad outcome. Unfortunately our litigeous society (and trial lawyers) have gotten MDs away from open communication about process and CQI. "

Oye, Oye, Oye... not remotely applicable. I happen to be doing a fellowship in quality improvement for hospitals. I just got back from a patient safety conference and from a QI training program. I am working on a DVT prevention project, a communications improvement project, and a anticoagulant safety project at the moment. Systems improvement is the focus of my non-wards responsiilities and I speak about systems problems to any who will listen and those who must--my trainees. YOU didn't write about systems (except to call for research, which must mean you understand on some level the limited tools this doc had) and you didn't encourage systems discussion. You called an individual doctor working in a system a BOZO, and if you hadn't noticed, I responded by emphasizing the system restraints that person was working in, namely:

--brief single encounter
--insurance blockades
--inability to control med compliance
--LEGAL limitations on action, huge concerns which you have as yet failed to respond to, instead insisting that this bozo should have entertained options NOT available to him to avert distant and unforeseeable events.

Yes, I am countering your accusation of incompetence, because it was unfounded. If and when you make some kind of comment on how the system could do better, in light of the legal restraints and limited evidence for intervention I have provided references for, then we can have a production conversation. However, you've focused on blaming an individual in a deliberately provocative way. I would rather fix the problem not the blame, and we COULD go back and do a root cause analysis on this case to find the holes. But we have to get past your unfounded critiques of single people based on limited information to get there--it's not denial that's in the way, not mine at least.

You could start by telling us what you would have done upon seeing Cho for depression after accusations of harassment that would have recognized impoending disaster and averted it. Keep in mind that your intervention would either have to be blindingly precise OR able to be applied to hundreds if not thousands of people accused of harassment who have depressed mood, and that you have to have an iminent danger to hold someone.

[Bill Glasheen]

I am responding to your unfounded accusation that this doctor was a bozo. Don't want responses to unfounded claims? Don't make them.

I'm not backing away from the "accusation" so don't worry. You'll have plenty of bandwidth to have some fun with me.

I'm taking a position the way one does in a debate. I'm having fun making an interesting case. At the end of the day, I can drop it, and go home to sleep. I will have learned something in the process.

"Pokorny and colleagues42 followed 4800 patients admitted to the in-patient ward of a Veterans Hospital. A logistic regression included the following predictors: attempted suicide, suicidal ideation, a diagnosis of affective disorder or schizophrenia, depressed feelings, recent history of violence, low social interest, urge to do harmful things, fear of losing control, remorseful feelings, impatience, and feelings of failure. But the model failed to identify any cases in which suicide occurred.

He used a logistic regression, and his model failed to get a result.

So????

A logistic regression assumes linearity. Life isn't linear. Did he next try decision trees? Neural nets? Did he do exhaustive scatter plots, looking for nonlinear relationships which would suggest first transforming some of the variables?

Unfortunately most don't know how to use those tools. Give most "data modelers" an inverted "U" relationship, and the vast majority will completely miss it. They won't do scatter plots. They'll try a regression, get no result, and assume no relationship. And they would be wrong because they have just enough knowledge to make themselves dangerous in the field of data mining.

That's what I do for a living, Ian. I can see through bad science pretty quickly in this domain. Been there, done that.

Keep looking!

You write as if you're the only one who's aware that multiple risk factors identify the highest risk patients, as this escaped the notice of the clincians and researchers who have made suicide and violence prediction their life's work and daily practice. It's strikingly naiive. They know. They have tried to predict these rare events. It can't be done at present:

Maybe BECAUSE THEY DON'T KNOW WHAT THEY ARE DOING!!!!!!!

I make a decent living because I have acquired such skills.

FWIW, social scientists aren't mathematicians. It's a rare few who understand how to use these tools. And you shouldn't be starting the analyses without first doing a lot of work with the data.

It's part hard science, part art, and part understanding the underlying principles which drive things.

Those with actual experience in the field have found it frustratingly difficult, and yet without a day (correct?) of psych experience, you have the confidence to imply that you can identify the next shooter beforehand.

WRONG!!!!!!!!!!!!

I said they shouldn't have let this person go. And hindsight proves me right.

Nobody could have predicted precisely what he did. But it doesn't take a genius to see that he needed help. Why did all the lay people around him see it, and yet this physician let him go with language suggesting there was little wrong with him?

Bozo!

THE BOY NEEDED HELP!!!!! A whole community was screaming for it, and they let him down right when they had him in their grasp. An opportunity to help someone was missed.

The fact that he went and shot 32 afterwords is besides the point. Horse shoes and hand grenades. Some should be held, some should be let go. This boy's past and the pleas from all around suggested there was something "interesting" going on that deserved further work. Inpatient or outpatient, THE LAW gave them the means to do what needed to be done. Act. Follow up. See it through to the end.

Why haven't you worked in the stock market, weather, or if a car of yours has ever broken in your possession, why did you let it happen?

I'm doing fine in the market (dad helps ) I live my life around the weather all the time via my own research of myriad sources, and all my cars work fine even though they have high mileage.

And I know the difference between deterministic predictability and deterministic unpredictability.

And "it" does happen. But it happens more to some than to others.

I happen to be doing a fellowship in quality improvement for hospitals.

Excellent!

YOU didn't write about systems

My bad. However in all these evidence-based medicine and litigation discussions, I write about them all the time.

this bozo should have entertained options NOT available to him to avert distant and unforeseeable events.

Wrong. THIS was the opportunity to do something. (S)he could have done a little bit of research and put some pieces together. There was no need to think "Am I going to stop a mass murderer?" All the person had to do was do his/her job.

The ball was dropped.

Yes, I am countering your accusation of incompetence, because it was unfounded.

And I disagree.

And yes, the SYSTEM failed. But often in a system, there is someone who takes charge. There is someone who gets paid the bigger bucks to sign the papers that make things happen as they should.

People should at least feel accountable for their actions.

You could start by telling us what you would have done upon seeing Cho for depression after accusations of harassment that would have recognized impoending disaster and averted it.

No need to get all cute and predict a school shooting. All you need to do is see someone spiraling out of control, and help them. Period.

When you are doing your job, "the system" is invisible.

- Bill

[IJ]

"I'm taking a position the way one does in a debate. I'm having fun making an interesting case. At the end of the day, I can drop it, and go home to sleep. I will have learned something in the process."

Ah, this is starting to make sense. You don't actually believe that this doc was a bozo, but wanted to provoke an interesting discussion. My faith in your reasoning was justified!

"Keep looking!"

Warmer! You seemed to think my point is "there is no way to predict homicide and there is no point in trying." However, my point is actually, and your response to my article confirms, "there WAS NO reliable prediction tool available to this physician at the time he saw Cho." Hence, your accusation that he was a bozo remains preposterous. He couldn't have, and he had nothing to hold Cho on. It was legally impossible for him to do what you have asked him to do, a fact you have continually ignored for unclear reasons.

"WRONG!!!!!!!!!!!! I said they shouldn't have let this person go. And hindsight proves me right."

No, Bill, it does not. IF you had met Cho with the shrink and said "hold this guy!" and he didn't, then hindsight would have proved you right (in a sense.) But hindsight didn't "prove you right." Hindsight and your opinion are the same, not an opinion and confirmation.

Further, an event occuring doesn't make your decision right even if you had said it before the event. Decisions are made with the data available and need to be judged in that light. Anecdote:

I had a patient with severe liver disease. They are vulnerable to bleeding, infection, and confusion; mine had come in with infection. We gave him IV drugs, identified the bug, and switched him to oral, per standard of care and expert guidelines based on randomized trials. The consulting liver attending wanted to keep the patient for the whole course on IV (because he had just been involved in a meritless malpractice case in which a patient sued after his infection got worse after going home on orals). I would not; the random experience of one person does not outweigh the literature and the national experience. The patient went home, his infection got better.... and 10 days later, he bled to death in an unrelated event.

PErfect analogy. Knowing that he was going to bleed, I would have kept him, but, that information wasn't available when the decision was made. I can't keep everyone who MIGHT have a problem, even though, based on his multiple conditions and severe liver disease, I knew he was going to live only 4-8 months on average. AND we don't know that keeping him would have saved him AND the issue at hand and the final event were different but related, like in the Cho case. We do know that keeping all cirrhotics in the hospital will lead to huge waste, hospital complications, and is practically impossible because we have so many here. It is STILL the right thing to send people home on orals when they are ready; he was. Hindsight doesn't actually make me wrong about this guy, or you right about Cho.

"Why did all the lay people around him see it, and yet this physician let him go with language suggesting there was little wrong with him?"

Yes Bill, that excerpt says there was little wrong with him, but has already been explained,

1) the pathology might ave been detailed elsewhere
2) what he did say was plausible and specific and quite likely to be accurate
3) "language" doesn't make Chos better... actions do. Referral may have been his only real option--and you're not concerned with whether he exercised it or not, you're ready to condemn him for an excerpt and a failure to do what he legally could not and you've presented no evidence would have helped: hold Cho.

"Inpatient or outpatient, THE LAW gave them the means to do what needed to be done. Act. Follow up. See it through to the end."

Interesting. And what specific law or laws are you referring to here? Please be specific. While you're at it, let us know how a consultant with a single visit to see a patient about an acute issue can become that patient's champion / advocate / case manager / etc and see the case through to the end? That is a monster task, one that psychiatrists would have to undertake with many of their patients, and one that SOME PCPs can manage. But not consultants. You don't become someone's PCP or champion because you do a psych eval on them once. It's impossible. Now, if you wanted to complain that our healthcare system was fragmented and failed to coordinate to help Cho, you MIGHT have a point (he might have been given every possible referral and chance but failed to followup, something we cannot force). THEN we'd be talking about system problems, but instead you've decided that a single doctor working in this system is a bozo for not magically overcoming the system limitations. That is the logic of the malpractice attorneys, not the quality improvement specialists.

"Wrong. THIS was the opportunity to do something. (S)he could have done a little bit of research and put some pieces together. There was no need to think "Am I going to stop a mass murderer?" All the person had to do was do his/her job."

This was "an" opportunity, true... hmmm. You do not seem to know what the psychiatrists job is, or can be... when my partner did psych call at hillcrest med center in san diego, every night, there would be constant pages about patients in the ED and in the hospital plus calls from outpatients. There was no time for sleep, perhaps an hour or two fragmented. You might do 6 capacity evaluations and get 12 total consults and run from one to the next. Collateral information can be unavailable. Outpatient records may be unavailable. The idea of doing a literature search is understandable, but to someone that's actually been on call managing up to 20 ICU patients, its laughable. There is no time to lit search all of our treatment decisions, and often no clear answer when the lit search is complete (you can see my response to a letter about steroids in sepsis that came out in this weeks Journal of Hospital Medicine for an example). This is where computerized decision support should come into play, but it must be based on knowledge which was and is sorely lacking in this case, and is poorly enacted throughout the nation. These are difficult systems problems. If you worked in this system, you would have been just as much the bozo. That's why this doctor wasn't a bozo.

"And yes, the SYSTEM failed. But often in a system, there is someone who takes charge. There is someone who gets paid the bigger bucks to sign the papers that make things happen as they should. People should at least feel accountable for their actions."

Well, this progress incorporating systems level problems is encouraging. (I still maintain that airplanes will fall from the sky at some rate, no matter what system is used to maintain them, but that is another discussion). However, you're calling the ER consultant (the marine) the bozo for not making some dramatic intervention in Cho's life (for shooting at an undercover delta force in a civilian vehicle) when they followed standard practice for their task and had no adequate violence prediction data on hand (when a friend-or-foe system had not been developed or purchased) and when they were working in a system controlled by med execs and the variable ER circumstances (when generals and iraqi terrorists interacted to create an unpredictable system). I'm sure that psychiatrist DOES feel bad (like the english teacher does) and it's perhaps your own emotional reaction to the tragedy that allows you to omit the possibility that they DO regret the events. I just don't see how you expect marines to design a better battlefield while they're being shot at.

"No need to get all cute and predict a school shooting. All you need to do is see someone spiraling out of control, and help them. Period."

Well, you must be able to either help everyone with a variety of psychosocial problems (oh, were that our jails, schools and mental health nets were so competent and well funded!) or able to identify the future disasters with precision (again, do let us know how!).

"When you are doing your job, "the system" is invisible.""

We sure don't notice it. But its there. And it controls a whole lot more than we know. We do recognize how unpleasant it is to be labeled a bozo for the effects of doing our best within a system when it happens to us, however. You (and everyone) might find "Set Phasers on Stun" a fascinating read... it is just a series of "human error" tales from medicine, business, war, aviation and other situations that reveal just how much we are at the mercy of our systems, but how often the ensuing investigation finds a scapegoat to blame.

When will you tell us exactly what "help" you would have provided to Cho and everyone like him?

[Bill Glasheen]

what specific law or laws are you referring to here? Please be specific.

I'm referring to the court order which provided the opportunity to assess.

A Montgomery County magistrate issued a temporary detention order Dec. 13, 2005. Cho was "mentally ill and in need of hospitalization and present an imminent danger to self or others as a result of mental illness, or is so seriously mentally ill as to be substantially unable to care for self and is incapable of volunteering or unwilling to volunteer for treatment," the order states.

One would hope in doing such an assessment that all available resources would be checked.

Do you think his medical history at VA Tech was absent any information about his troubles? From all I have heard from myriad sources, Cho wasn't a stranger either to the police or to the medical system.

No, we don't have all the information here.

Yes, I have a big problem with this - a reason given to release him.

His insight and judgment are normal

You have stated you don't have a problem with this statement. We disagree.

When will you tell us exactly what "help" you would have provided to Cho and everyone like him?

I'm not ready to save the world, Ian. Yet...

There is however this hot-looking blonde cheerleader...



< Sigh... >

Back to earth, Bill.

I act one event at a time. And I have acted in cases where people in Cho's exact predicament were held long enough to get help. I didn't directly help said people. But I did see to it that the process got started.

Two for two. Not statistically significant. But we save the world one problem at a time.

you've decided that a single doctor working in this system is a bozo for not magically overcoming the system limitations. That is the logic of the malpractice attorneys, not the quality improvement specialists.

Priceless!

- Bill

[IJ]

"I'm referring to the court order which provided the opportunity to assess."

And because of that order he WAS asessed... the next question is, did the law provide the opportunity to hold him (the next next question is would have holding him helped, but that's for you to demonstrate some other time). Would Dr. Bill please tell us which box he would have checked?

[] imminent danger to self
[] imminent danger to others
[] unable to care for self.

If you don't check a box, he is released. If you do check a box, a hearing is held to confirm your judgement. Since you are fond of the hindsight makes right argument, you would have lost this hearing; Cho did not harm anyone for over a year, and managed his food and shelter just fine (heck, he went to college!). Dr. Bill, your patient would have been released, period, unless there is some side of the story yet to be presented.

Let me also throw out the point that many assume that "if it might help, do it." This is in fact not true, because treatments have sometimes difficult to perceive risks. Some patients decompensate when hospitalized; some patients resent the experience so much they decide not to voluntarily seek treatment next time. Cho didn't seek treatment, but a psychiatrist without Nicholas Cage's ability to see what comes "Next!" who treats 100 patients as you are suggesting might do more harm than good.

"One would hope in doing such an assessment that all available resources would be checked. Do you think his medical history at VA Tech was absent any information about his troubles? From all I have heard from myriad sources, Cho wasn't a stranger either to the police or to the medical system. "

One would hope so! However, that does not mean that collateral information IS available. There might be no way to obtain information on complaints filed to the school, or no way to contact acquaintances of the patient (unles he gives their names). His medical records would remain confidential until he signed a waiver. Sometimes medical records arent available for days; holds can't be placed during the wait. Such is the reality of psych hold evals. Doctors don't have access to police records. But you've identified some great systems problems to pursue without demonizing individuals!

"You have stated you don't have a problem with this statement [that Cho's insight and judgement were normal]. We disagree."

Well, I personally wasn't there when Cho was evaluated, and neither were you, so I'm completely unclear as to how you could overrule the assessment of someone who was. Unless you thought they were referring to a trend... they assessments are of abilities NOW, when they were done. Just like medical capacity decisions, the fact that a person might have been completely nutsoid over and over has no bearing on measures of their ability to perform simple reasoning tasks at the time of the assessment. You do NOT have the information required to disagree rationally with the doc's assessment, just as you cannot say based on information presented in this thread whether Cho had used drugs that day.

"And I have acted in cases where people in Cho's exact predicament were held long enough to get help. I didn't directly help said people. But I did see to it that the process got started."

I'm curious to hear the details, of course. Especially since Cho's situation was somewhat unique.