The Suicide Risk Assessment Makes Me Want to Kill Myself

"There is but one truly serious philosophical problem, and that is the suicide risk assessment."

Albert Camus, The Myth of Sisyphus

Did you read the title and think I needed a psychiatric admission? Ha! Well, you aren’t wrong. Important though: I'm not expressing suicidal ideation with my title, I'm expressing homicidal ideation. If you're a psychiatric practitioner and you're still paying attention to the content of the words, and not the form, you don't deserve your license. Sadly, the more likely situation is you're on track for a director position.

First I should clarify what I mean by Suicide Risk Assessment (SRA from here on out). I'm not referring to your earnest assessment of the risk of the patient in front of you. I'm referring to an institution-mandated risk-stratification tool based on pre-set questions that must be documented and acted upon based on the determined risk level. I focus on the SRA here just because of its increasing popularity, but I hope the points generalize to any similar mandated clinical assessment. And I should also note the volume of my anger is secondary to my fear that the SRA is cementing that hospital administration has free rein to meddle in how clinicians practice with paperwork that increases clinician liability.

I view the SRA as a bureaucratic tool that benefits suits/corporate, masquerading as patient benefit, bought in by physicians due to their fetishism for positivism and all-too-ready submission to authority. Too fluffy? I'll give it another go. The SRA is a "cover my ass" for the hospital, pushing liability to the physicians, who naively believe they are doing something for patients while they waste precious time better spent elsewhere. The admin's goal isn't to prevent suicides, their goal is to prevent suicides from having a financial impact. So they have 2 things they're trying to decrease 1) litigations for suicide AND 2) bad press from suicide. That these goals overlap somewhat with preventing suicide is pure coincidence, but they know it's a damn good selling point to doctors. This is called good marketing. My personal bias is that I don't get that upset by the suits/lawyers... they didn't go to school to learn about patient care. I get upset by the obedient physicians who don't care enough to even question it.

Danger to self or others has become the un-official official credo of psychiatry, and this mode of thinking is downloaded early in training and now rests deep in the collective psychiatric unconscious, at the cost of serious societal harm. Medical students spend years being clueless wanna-bes... they come out of medical school starving for 2 things: answers and authoritative respect. Simple models of the world and simple goals provide this. Tell me my goal is to stop the person in front of me from killing themselves/others, and I'll know exactly what to do. Tell me my goal is to maximize the mental health of the person in front of me, and I'm even more useless than Seroquel (zing!).

The problem is compounded by the fact that the average student of psychiatry doesn't experience outpatient care until after 4 years of medical school and 2 years of residency. That means the entire foundation of their knowledge, the blueprint on how to practice, is built on the emergency/inpatient mindset. When a resident gets to the outpatient world, they are so used to thinking through the lens of is this patient a danger to self or others.. of what are the legal ramifications of my decisions... that they continue to act as if the crux of their job is to decrease suicide/murder. And mandatory SRAs on all patients re-enforces this shitty model. I imagine a psychiatrist's utopian world is one where the whole population sits in a dark room, passionless, lobotomized, tube-fed. Low-risk worms that show up on time every month to their 5 minute tele-med-checks silently billed high-complexity+45 min therapy add-on.

In an ideal world, good attendings teach residents how to practice in such a way that satisfies the administrative/legal/hospital overlords, while making decisions and interacting with the patient based on their own Ubermenschian value system. But as suits/lawyers creep further and further into how providers practice, free-thinkers are incentivized to get the fuck out of the hospital (and consequently out of teaching).

But let me get back to the main point here, the problem with SRAs.

1) it doesn't add information that isn't already completely obvious

Understanding the clinical uselessness of the SRA requires 2 tools lacking in the average provider's armamentarium/arsenal (me pandering to the blowhard psychiatrists who all use the same stupid words): statistics, and common sense.

It is incredibly easy to misconstrue certain statistics that sound to be transmitting clinically meaningful information. When you're trying to detect a low-probability event, really-important-sounding-statistics can become impressively useless when you break them down to intuitive statistics that stupid-human-brains can process.

Let me provide a 100% real example. I got a call from my primary care doctor:

"Hey sorry to tell you, but you have syphilis, let's schedule you for a penicillin shot",
"Um can I have another test to confirm?",
"The test is 99% accurate",
"Accurate? I still want to the confirmatory test",
"It's cheaper to just have the penicillin shot",
"Yeah I'm fine with paying for another test out-of-pocket"

Do not mistake the steadfast confidence as my ability to instantaneously consider my pre-test probability/false-positive rate of having syphillis in rural bumble-fuck as a medical student who was not participating in the local lady-of-the-night scene. I just didn't want to tell my girlfriend. The confirmatory test was negative.

So when you tell people that boys are 3.7 times more likely to actually complete suicide, you're not saying anything that has any sort of actionable content. If 1.7% of deaths are suicides (and men are 3.7x more likely), then by my calculations that means .7% of women and 2.7% of men commit suicide. So the statistic that you can actually process (absolute risk).... is a 2% difference. Do you see how useless that is for the person standing in front of you?

(Side note: Women are ~4 times more likely to attempt suicide. And men are ~3 times more likely to complete suicide. Those are 2 different statistics, feel free to combine them. My theory as to what's going on? I plead the fifth.)

Summing up a bunch of these clinically unimportant risk-factors typically leads to a more robust clinically meaningless-risk factor. We're in the GIGOgian age of psychiatry, loads of garbage in, very shiny billable garbage out. The Fig-Scrubbers are ecstatic about it, but it's all faux-knowledge. I should clarify, I'm not saying the designers of these risk tools aren't aware of these limitations, I'm saying that when you force providers to fill out a pre-made tool, they'll begin to over-value the quantifiable (but unimportant) over the qualitative (and clinically meaningful).

Another important facet to consider is a reverse consideration of ecological validity. The emergency-room/outpatient-office is not a representative sample of the environment (and vice versa). If suicide risk factors are drawn from completed suicides in the environment, then they won't map well to decision-making in these settings. Example: One of the highest risk factors for completed suicide is a previous mental illness diagnosis. But a person in the emergency room who has no mental health diagnosis saying they're suicidal is more likely to acutely complete suicide than someone in the emergency room with a mental health diagnosis saying they're suicidal. This isn't a paradox of logic. Suicide risk factors are drawn from the population of people who actually kill themselves, which extremely infrequently is anti-social malingerers and borderline crises. This population only sort-of overlaps with the population that ends up in the emergency room saying they're going to kill themselves, which is heavily loaded with anti-social malingerers and borderline crises.

Now I know what an outsider might think: it sounds like a good thing to consider the risk factors, even if they're not perfect. A good analogy is obesity. Imagine in the emergency room they develop risk score based on answers to questions that factor in a patient's age, gender, ethnicity, parents, etc, etc and perfectly calculates the risk that the person in front of you is obese. Why don't we do this? Because a provider can just look at the patient and ask themselves "are they fat?" and it's equally effective. This brings up the next important point...

2) it takes up time that could be used elsewhere

The hospital follows Darwin's first law of thermodynamics: Any time spent on something is time actively NOT spent on something else (resulting in more morbidity/mortality in the area not spent time on). The hospital staffs the place so that it just barely functions. This is a key concept that you don't grasp until you're the on-call resident carrying the weight. Let me review a fun anecdote that everyone will remember: when we were using Hydroxylchlorquine to treat COVID. Might as well give people hydroxychloroquine right? Even if it doesn't do anything it's basically harmless, right?

NO. NO. NO. I repeat: Any time spent on something is time NOT spent on something else. Hydroxychloroquine ever so slightly widens the QTc (aka increases risk of cardiac arrest) causing all the staff to be running doing EKGs on all patients.

Steven was an intellectually disabled 47-year-old man. Every single day I would walk into his room and find his BiPAP machine off his face, his oxygen at 80%. I'm not lying we had this exact interaction every single day for a week.

"Steven! You need to wear your oxygen machine!"
"I do?! I didn't know that"
"You're killing me Steven"
"Am I going to die?"
"No just keep that mask on, you're getting better every day"
"Can I have an apple juice?"

Repeat this Mon, Tues, Wed, R, and Friday.

Until Saturday at 8 am, when a code was called to Steven's room. His heart had stopped. I was the first to do a compression. The first push on his chest, I crunched his ribs and I felt them break. His ribs crumbled like pieces of chalk. I'm not saying this poetically: every push I could feel his heart, the elasticity of it. His heart was no longer the life-source-pump but just a thick useless balloon, dead of its vital electricity. To do compressions, you have to ignore that the frail bones and the squishy balloon were once alive, you have to ignore your instincts and push against the deadness. You have to dissociate the previous humanhood of the person in front of you, and rely entirely on basic procedure. This is the mindset where crisis medicine is best practiced, but also the mindset where outpatient psychiatry is worst practiced.

Steven died not just because of COVID. Not just because he wouldn't keep his mask on. Not just because hydroxychlorquine increases risk of cardiac arrest. Sure these all played a part in the swiss-cheese-model, but the big point I'm making here is that the vital piece of swiss was the time the doctors/nurses/aids lost with their patients because of the additional workload that hydroxychloroquine caused. If an EKG takes 15 minutes, and a nurse has 6 patients (extremely conservative estimates), that adds 90 minutes to their workload EVERY DAY. Have you seen an inpatient nurse work before? They do not have a spare 90 minutes. So those 90 minutes must be taken from some other essential task. And that task would have kept Steven alive.

Yes, Steven died. After leaving Steven's room, I went right back to work. I didn't have a single thought about him again for a very long time.

It wasn't until 1 year later, in my childhood basement, amongst a collection of toys, I found a chalk-board and chalk-box. I removed a piece of thin chalk from the box. When I went to mindlessly sign my name on the chalkboard, the chalk instantly snapped into 2 from my hand applying too much pressure against the board. Steven.

But we must move on (even if I haven't...).

3) when the risk-score is connected to a required action, it leaves a false paper trail chock-full of liability

The SRA is a way to push full liability to the doctors for a completed suicide. Let me say that in a more fun way: the SRA is hospital administration having doctors fill out the paperwork of the evidence showing they're responsible for the error. But it gets significantly whackier than this if the hospital implements a required action based on the stratification of risk. For example the suicide risk determines the precautions, f.e.: Low risk = no precautions, Middle risk = Mild precautions, High risk = Major precautions. Now this might not sound intuitive at first glance, but a person's risk level doesn't equate to the precautions required. I meet a kind, meek 45-year-old man who lost his job, has no prospects, doesn't think he'll be able be to support his wife and kid, is a caretaker to his dying parent etc, etc. Despite a fear of heights, earlier in the day he walked to the top of his rooftop, thought of his kid growing up without a dad, walked down and came to the hospital. He feels hopeless. When initially asked if he can see his situation changing, he sits quietly and stares blankly. The man feels trapped, with the weight of his life bearing down on him, and I can feel his weight in the interview. He tells me has been thinking of ending his life for months. This man is insanely high acute risk, and absolutely requires admission. But I'm going to check "low risk" on his risk assessment when I admit him. Why? Because he doesn't need the precautions required if I check high risk, it's a waste of resources. The 45-year-old quiet family man isn't going to slam his head on the radiator to the soundtrack of screaming patients/nurses in a Kill-Bill-esque blood-bath. So if I honestly fill out the risk assessment then I waste hospital resources and make this man's experience even more miserable than it needs to be. And if I send him up as high risk, I'm guaranteed to get a call from the nurse: "We're short-staffed, why the fuck are you having us stalk poor Toby McGuire".

4) It gives the illusion of understanding/doing something

What psychiatrists are being trained to do (if you don't call an institutional policy of mandatory suicide risk assessment with pre-set questions as "training", then I don't know what else would qualify as "training") is ask a series of rote "risk factor" questions as a response to suicidality. This trains providers to respond to "I'm suicidal" with a series of monotone questions. Which trains providers to resent patients who endorse suicidality. Which trains patients to not bring up suicide to their providers. And the worst part is that this "Risk Assessment" often will take the place of genuine risk mitigation. The provider will feel satisfaction of having done something, without actually having done something meaningful. The silent killer is learners offshoring the development of clinical judgments/interventions to a risk score.

If you're in a higher-up position thinking to yourself while you read this: "there's a lot more behind the scenes that you don't get". At some point in a career of climbing-the-hospital-structure, your thought-process insidiously becomes less of you playing- the-game and more of the game-playing-you. Hospital psychiatry is embedded in a corporate structure whose values directly oppose the purported values of (good) psychiatry. The longer you breathe in the toxic air, the harder it is for you to remember what fresh air feels like. I'm not whole-sale knocking the thing, it can't exist if it can't compete, but I am worried that the individuals who are attracted to climbing this increasingly corporate structure are the ones training the physicians of tomorrow.

I've said a lot, so let's tie it all together. When a risk score (that doesn't accurately predict anything) is used to make actionable decisions (impacting use of limited resources), the whole hospital staff just learns to play the game and translates the "risk score" (which took valuable time away from the patient) into its required action, all the while leaving behind a paper trail of lies/liability.

There needs to be a counter-movement to the Figs-wearing, stethoscope-Instagramming, Destigmatize-mental-health meme-ing, squeaky-wheel-gets-the-grease-ing culture that psychiatry has become. What then do I think is the job/skills of a psychiatric provider with regard to suicide? A good psychiatrist is a sponge. He is capable of feeling the emotions the patient is transmitting to him through the magical intersubjective field, recognizing how the patient's emotions mixes with his own, and translating it into the necessary words/actions for that moment. So he/she knows themselves sufficiently to pick up what certain emotions in themselves mean. This is a skill that needs to be developed. Every time you meet a person or a patient, pay close attention to the sensations they elicit, your heart-rate, to the theme of your thoughts in contrast to the theme of their thoughts. In the emergency room... Are they discussing suicide and you're bored? That's vital information (consider narcissism). Are they discussing suicide and you're annoyed? (consider borderline). Are they discussing suicide and spitting at the nurse because the orange juice is warm? (it's antisocial). Are they discussing suicide and you feel helpless? (admit them). Of course, it's incredibly important to pay attention to your own emotional state based on your own personal biases and your own recent emotional events. For example: I have to factor out a patient eliciting the emotion of anger/criticism in me, because I'm always angry/critical. It's a variable that reveals information about me, not the person I'm talking to. But you should already know this stuff, you fucking idiot.

So when a patient confides in me: "I'm having thoughts of killing myself", there are multiple things I know I need to figure out, and while I figure them out I need to (implicitly and explicitly) communicate a few things. What do I need to implicitly communicate to them: gratitude that they told me, that they did the right thing, that I will only choose what to do based on what is best for them. I don't care about my legal risk, I care about them. (The truth is that of course I care about legal ramifications, but I also know that if I do my job and do it correctly... even if I make a wrong decision... I'll have made the best one at that time with the information I had available. And for that, I allow myself to ignore any thoughts of legality). The meta-communication that occurs whenever mechanical questions are asked is "Sorry, but now this is a legal issue and I have to do what my manager told me". People who are MDD suicidal or bipolar suicidal will experience this as a distancing of their personhood (because it is), people who are borderline suicidal will experience this as a direct assault to their personhood (they said to you, in their language, "I need love" and you replied "Have you thought about ways of killing yourself?"). What needs to be said explicitly by the physician depends on too many factors to ever even consider breaking down into a flow-chart. But when I don't know what to do or I can feel myself getting too wrapped up in something, I just remind myself: be human. It's an infinitely more effective credo than find out if this patient is a danger to self/others. And yes, do a fucking risk assessment for Christ's sake, just not in the spirit of corporate. If you ask "do you have a gun" with your head down and pen ready to write on the clipboard, the patient will experience you as another suit. If you take in the person's pain, and then solemnly ask... "do... you have a gun?", they'll experience you as caring.

All humans want mostly the same thing: to be loved, to be heard. We don't need someone to understand, but we need to know someone out there is willing to try. When someone feels so untethered to the network that suicide becomes an option, what they need is to get re-tethered. You need to stay humble as a provider- a few well-selected words can't glue a person to the human web. But maybe you can offer a glimmer of hope... or maybe a glimmer of caring that could lead a person to feel that a glimmer of hope is not conceptually impossible. Maybe that means setting them up with a therapist, or telling them about a group therapy. Maybe it means being interested in their new job for 15 minutes every month. And yes, sometimes it means putting them inpatient where they can take a moment to breath. But eventually their discharge day will come, and I am confident they'll have better outcomes if they know they're being discharged to a psychiatrist where the goal their working on is improved functioning and not prevention of danger to self or others.

What's the big point of all this? Keep the suits and the lawyers out of psychiatry. If it's too late for that (and it probably is) then at the very least keep the suits and the lawyers out of how you practice psychiatry.