Psychiatric hospitalization doesn't prevent suicide, according to the data.

Psychiatric hospitalization isn’t supported by the evidence currently.

Suicide prevention. It’s important. It’s important for veterans. It’s important for youth. It’s important for older people. No one is arguing we need to let suicide run its course.

Despite endless agreement that suicide = bad, we also seem to be focused on evidence everywhere but… here. We know if someone is suicidal… we should do …something!

“What that “something” action should consist of… is not clear.

People often think of psychiatric hospitalization as the something that should be done, and so I figured it’s about time to review the evidence again.

Does psychiatric hospitalization help?

First, we should define help.

I’d argue that “less death by suicide?” is the strongest version of the argument for “is this helpful?" (Spoiler alert: no).

There is a nihilism in the arguments for our coercive system, which are something like the following: “Well, what else are we going to do?!” A lack of alternate options seems to be the argument, and it’s one I’ve heard before. The other argument is that even a little change will make a huge difference. This last reason I why people like

Wendell Potter NOW

tell me they worked for major healthcare corporations.

And as a note to you, dear readers, I did a re-review before finishing this article. This is not intended to be definitive or a scientific review… but I couldn’t find any empirical evidence for benefit that can be attributed to psychiatric hospitalization for suicide.

At best, it’s a lot of waffling about how “here is where immersive treatment can happen” which is both true for unrelated payment reasons and not a real argument. It’s one of the reasons colleagues like

Psychiatry at the Margins

point out psychiatry is having a crisis of legitimacy with the public.

Let’s begin gently, by looking at the data on a comparison between short and long-term hospitalization, dating all the way back to 1977. Treatment was longer then, on average, so that seems like a good place to start. In a galaxy that feels Long ago and far away…

Short vs. Long Term?

Is longer hospitalization better for patients?

A 3 year follow up as far back as 1977 by Mates Et. Al. Found…

Results showed the two groups to be doing equally well in most respects, but of the patients who were rehospitalized at all during the first 3 postdischarge years (30 per cent of the total sample), LT patients were rehospitalized significantly more often and for significantly longer periods.

So patients were in the hospital longer, tended to come back to the hospital more. You can make endless arguments about how people aren’t randomized to these conditions, but what you can’t say is there is a “dose dependent relationship” with length of stay and doing better. Because the opposite is true. Further:

Also, [short stay] patients felt they had received more benefit from the hospitalization.

In terms of feeling helped, it looks like the less time in the hospital the better. This was of course back in the 1970s. Things must have changed.

Does Modern Hospitalization Prevent Suicide?

A landmark largest study on the topic comes from Denmark, published in the Journal JAMA Psychiatry:

The design was a large nested case control study:

This very large study looked at all of the people in all of the country who died by completed suicide between 1981 in 1997, and compare them to controlled subjects, who didn’t die by suicide.

Participants  All 13 681 male and 7488 female suicides committed in Denmark from January 1, 1981, to December 31, 1997, and 423 128 population control subjects matched for sex, age, and calendar time of suicide.

The authors found found suicide rates go up with psychiatric hospitalizations:

Conclusions  Suicide risk peaks in periods immediately after admission and discharge. The risk is particularly high in persons with affective disorders and in persons with short hospital treatment. These findings should lead to systematic evaluation of suicide risk among inpatients before discharge and corresponding outpatient treatment, and family support should be initiated immediately after the discharge.

Take a minute with that result.

⏰

The risk goes up immediately after… admission. When people are in a when people are in a psychiatric hospital… to stay safe. But they’re actually less safe? Again, we have a problem of non-randomized individuals being subjected to this condition, but they’re not more safe. They are more likely to die by suicide, and it’s causality we can argue about. But the direction of the safety change? Less.

Of course, one could imagine that sicker people are more likely to be in hospitals. This is the kind of question that randomized controlled trials exist answer, but it’s really hard to randomize people do a hospitalization involuntarily. Usually the problem in medicine is that we keep doing some thing that we think is helpful because we’re not sure. And it’s hard to do the study. And eventually we get the evidence that this proves to helpfulness of the thing that we thought was helpful.

In this case… It seems like we have the opposite. We keep doing a thing that has demonstrable evidence for worse outcomes on the hypothesis that we are unable to understand the way in which we know it’s helpful or necessary.

What Happens After You Leave the Psychiatric Hospital?

In the JAMA Psychiatry study from Denmark, the risk for completed suicide peaks after discharge from a psychiatric hospital.

It’s hard to argue that hospitals aren’t helpful, but at the same time, you’re more likely to die when you go in. And you’re more likely to die when you come out. Compared to matched controls people who didn’t have that exposure at all.

If you were more likely to die from a heart problem when you were admitted to a hospital to attempt to fix it, well, we would say that hospitals are not safe places for people with heart problems.

What About Veterans?

How about in another population, like, for example, veterans? I have a vested interest here, as both a physician, and a member of the Veteran Mission Possible team working to any end veteran suicide. I would want to help vets, and if hospitals were helpful, I would agree that we need more of them.

It’s not safer or more effective in veteran populations, as published in 2013 in the American Journal of Psychiatry:

A total of 153 suicides occurred among the 68,947 service members. The overall suicide rate in the cohort was 71.6 per 100,000 person-years, compared with the rate of 14.2 per 100,000 person-years in the general active-duty U.S. military population. Personnel released from a psychiatric hospitalization were therefore five times more likely to die from suicide. The risk of dying from suicide within the first 30 days after a psychiatric hospitalization was 8.2 times higher than the risk at more than one year after hospitalization.

Conclusions

Active-duty U.S. service members who are released from a psychiatric hospitalization are a group at high risk of suicide. Aggressive safety planning and targeted interventions during and after hospitalization are recommended.

What I find particularly unsettling about that finding is that the conclusion doesn’t actually follow the results. Because although it acknowledges the high risk, it doesn’t really touch on the fact that psychiatric hospitalization don’t seem to have demonstrable benefit. We can do the correlation versus causation dance some more, but the correlation is with worse outcomes.

The problem is that those large studies, including the study in Denmark, date back a number of years. Taking a look at some more modern studies, we have a meta analysis published in 2016 specifically related to suicide risk, again in the journal JAMA Psychiatry:

The cohort included 770 643 adults in the mental disorder cohort, 1 090 551 adults in the nonmental disorder cohort, and 370 deaths from suicide from January 1, 2001, to December 31, 2007. Data were analyzed from March 5, 2015, to June 6, 2016.

This is a lot of people. It’s not better:

Results  In the overall population of 1 861 194 adults (27% men; 73% women; mean [SD] age, 35.4 [13.1] years), suicide rates for the cohorts with depressive disorder (235.1 per 100 000 person-years), bipolar disorder (216.0 per 100 000 person-years), schizophrenia (168.3 per 100 000 person-years), substance use disorder (116.5 per 100 000 person-years), and other mental disorders (160.4 per 100 000 person-years) were substantially higher than corresponding rates for the cohort with nonmental disorders (11.6 per 100 000 person-years) or the US general population (14.2 per 100 000 person-years).

And again, these are only looking at the 90 days after discharge suicide rates, comparing the general population to people who had been discharged from the psychiatric hospital. A total of 370 people died by suicide, and there was 90 day intervals with those populations above.

That is a terrifyingly high completed rate of suicide after psychiatric hospitalization.

To put that in perspective, the rate of civilian death in Ukraine after the Russian invasion per 100,000 is 16.7 in the past year. It is 14 times more lethal to be hospitalized for depression, and then discharged than to be a civilian in a country that Russia is actively shooting missiles at and murdering civilians in the street.¹

It’s not just about the US health care system either, Brazilian hospital have higher completed suicide rates after discharge too:

This sample comprised 1,228,784 adult patients admitted to psychiatric hospitals. Of these, 3201 died by suicide within 365 days of discharge. The risk of suicide was positively associated with male sex, age between 18 and 29 years, living in the South region, and living in rural or intermediate municipalities. The highest risk of suicide was among patients with depressive disorders (aHR, 3.87; 95%CI, 3.41–4.38) follow by opioid-related disorders (aHR, 2.71; 95%CI, 2.00–3.67), particularly among female patients.

Suicide rates remain terrifyingly, higher not just in the 90 days after discharge, but a year after discharge. We still haven’t seen any compelling evidence that psychiatric hospitalization decreasing suicide risk. Let’s keep looking. How about people who look like they’re doing better when they’re in the hospital? It turns out this is a red herring, and those “I’m doing great now” people are at the highest risk!

[Authors] examined a sample of 2,970 adult’s ages 18–87 admitted to an extended length of stay (LOS) inpatient psychiatric hospital. We used group-based trajectory modeling via the SAS macro PROC TRAJ to quantitatively determine four suicide ideation groups: nonresponders (ie, high suicide ideation throughout treatment), responders (ie, steady improvement in suicide ideation across treatment), resolvers (ie, rapid improvement in suicide ideation across treatment), and no-suicide ideation (ie, never significant suicide ideation in treatment).

And Following Along in the Results Section, This Deeply Unsettling Finding:

Resolvers [the people who acted like they were responding robustly to treatment in the hospital] were the most likely to die by suicide post-discharge relative to all other suicide ideation groups. Resolvers also demonstrated significant improvement in all clinical outcomes from admission to discharge.

To Make My Point Very Clear:

Psychiatric hospitalization, as it’s practiced now, is a failed experiment.

It doesn’t modify the risk downward, and at best, we can blame the risk going upwards, often dramatically, on something else. Experimental and unproven! C’mon UnitedHealth Group!

Psychiatric hospitalization is expensive. It is often traumatizing. I hear a lot of arguments about how sometimes we have to do it. The point, however, of science, it’s supposed to be that it lets us challenge orthodoxy when the data does not support what we previously understood to be “true.”

Why are we continuing to use psychiatric hospitalization for suicidal patients, if the outcomes are all worse?

It might have a lot to do with a treatment for the nerves of the psychiatrists on call, and for hospital lawyers who are worried about liability? It doesn’t seem to be about better outcomes for patients at all. Because the data does not support any better outcomes.

I’d be thrilled to be proven wrong. I think it’s possible to create psychiatric hospitalization that would work more like medical hospitalizations, with procedural care, rapid acting interventions, and complex treatments that require the kind of nursing care that hospitals provide. But that’s just not what happens in those places. Only 1 out of 10 hospitals in the US even has access to electroconvulsive therapy.

But we can’t just keep doing things that harm patients and pretending we’re doing the right thing.

I may have some real experts read this column, and if you have expertise, please share it with me in the comments. I would love to be wrong. I would love there to be data that supported robust better outcomes from psychiatric hospitalization. But I just haven’t seen it.

But Let’s Hear Some Common Sense…Why Would Psychiatric Hospitalization Possibly Be Helpful in Its Current Form?

Imagine a cardiac unit. And in that cardiac unit all they were able to do was prescribe medications that take months to work. All they can do is prescribe statins. And so somebody has a heart attack, and they’re lucky enough to not die, and they're admitted to an inpatient setting, and they’re given a statin. And then they spend five or six days fretting about whether the statin is working or not. And they ask the person how their heart is feeling? Or, whether if fees like your veins are a little bit less cholesterol-ish?And after four days, or maybe seven days, or maybe two weeks, then we feel comfortable discharging you. Because now you’re on a statin. The difference between us and psychiatric hospitalization, of course, is that you don’t run the risk of violence or coercion on a cardiac unit as much, and statins have long-term benefits that are well-established for mortality.

Staff at current inpatient stays in hospitals are focused on the adjustment of medications. These medications take weeks more to have a plausible effect biologically, which is much longer than the time a person is at the hospital. We are doing nothing in that timeframe to treat the underlying condition. And by and large we eschew psychotherapy on inpatient units, which might actually be helpful. It is a theater of the absurd. And there is only evidence of worsening risk. We shouldn’t be that surprised when doing nothing does worse than nothing.

The Future Could Actually Be Bright

We now have a whole array of treatments that have either gotten FDA approval, or are very close. These treatments work rapidly, and they work in ways that are dramatically better than traditional psychiatric treatment. General medical hospitals have ICUs, extracorporeal membrane oxygenation, ventilators surgical ORs, and more. In short, they are places where medical miracles can occur that can save lives.

With fMRI guided accelerated intermittent theta burst stimulation² (SAINT and HOPE tms), depression can be in remission in five days. That takes brain scanners, equipment, technicians, staffing, and is therefor well suited to inpatient settings. Rotational Field TMS is on the horizon. MDMA and psilocybin take hours of therapist time in controlled settings, which is actually pretty well suited to inpatient stays. Electroconvulsive therapy is a remarkably effective treatment that requires anesthesiologists and psychiatrist to work together, which could easily be arranged in a hospital. Inpatient psychiatric hospitalization could be the place where things are done that can’t be done otherwise. That would produce dramatically better outcomes. The hospital could be something other than a place for patients to become even more more at risk. In the future, psychiatric hospitalization can be much better.

—O. Scott Muir, M.D.

1. This is using UN estimates of civilian deaths in Ukraine and their prewar population from 2021 of around 41 million.

2. I will take this moment to continue my ongoing campaign to remind the Nobel committee that Dr. Nolan Williams should be the recipient of the 2024 Nobel prize in medicine for his groundbreaking work on Stanford accelerated intelligent neuromodulation treatment.