Skip to content

Science1 publisher3 min readPublished

Five decades of research leave suicide prediction barely above a coin toss

Clinical forecasts of who will die by suicide run only slightly better than chance and have not improved in 50 years. Paul Nestadt of Johns Hopkins now works on interventions that do not require knowing who is at risk.

The Scientist · Science desk

Illustration accompanying Five decades of research leave suicide prediction barely above a coin toss

What happened

  • Research summarized by KFF Health News puts clinical predictions of who will become suicidal only slightly better than a coin toss, and that performance has not improved in 50 years.
  • Preliminary federal data show more than 28,000 people killed themselves with a gun last year, more than half of all suicide deaths in the United States.
  • Gun suicide rates have hit record highs for five consecutive years, even as overall suicide rates have dipped recently.
  • Research identifies limiting access to lethal means, including barriers on buildings and bridges and smaller pill quantities per purchase or prescription, as one of the most promising ways to save lives.
  • The Trump administration has rolled back dozens of firearms regulations, rescinded a previous surgeon general's warning that called gun violence a public health crisis, revoked community gun violence grants and cut federal staff.

Compiled by The ScientistSomething wrong?How this is made

Why it matters

  • constraint Any prevention program whose first step is naming the individual at risk inherits a forecast that half a century of research has not improved, however effective the care that follows it.
  • decision A funder deciding between targeted screening and universal means safety cannot price the swap on this record, because the ranking of means restriction comes with no measured effect size.
  • exposure The method behind more than half of US suicide deaths is also the one with constitutional protection, so the measure with the largest claimed benefit is the least available to legislators.
  • precedent New gun-owning households appeared during the pandemic faster than any clinical system could assess them. Researchers make that the case for prevention aimed at households beyond any clinician's reach.

Means restriction does not require identifying anyone in advance, and its case turns on how fast the risk moves. KFF Health News reports that people can spend less than an hour between deciding to die and acting. Sometimes five minutes. Anything imposing a delay, such as having to unlock a safe, gives them time to reconsider or gives someone else time to intervene [13].

The half-century of flat performance is specific to individuals. Group patterns are well described: gun suicides concentrate among middle-aged white men and veterans, and have risen more recently among women and among some Black and Latino men [10]. The forecast that sits near chance is the one about a named patient in a clinic. "I can't tell you which of my patients is likely to die by suicide in the next six months," said Paul Nestadt [2]. He treats suicidal patients as a psychiatrist, teaches graduate courses on the subject and has co-authored scores of research papers on how and why people kill themselves [3].

Prediction accuracy and treatment efficacy are separate quantities. A psychiatrist who cannot rank his caseload by six-month risk can still treat the patient in front of him. The KFF series examines prevention beyond medication and therapy for people at risk, and calls those two lifesaving [14].

The firearm share sets the denominator: if more than 28,000 firearm deaths were more than half of all suicide deaths [8], the national total last year was under roughly 56,000 [1]. Overall rates dipped while gun suicide rates set records five years running [9], so across the recent period where both hold, non-firearm suicide rates fell [2]. That pattern cannot test whether blocking one method pushes people toward another, because nothing was blocked. The evidence KFF cites on that question comes from places where a method was made harder to reach and people did not typically switch [12].

The reallocation case is thinner than the scientific one. "That's the one that will save the most lives," said Nestadt, medical director at the Johns Hopkins Center for Suicide Prevention, referring to limiting access to lethal means [6]. The account lists barriers on buildings and bridges, and cutting the number of pills a person can buy or be prescribed at once, among the promising measures [5]. It gives no effect size and no cost per death averted [3]. KFF also notes that means restriction may be the most difficult of these approaches to achieve [7].

The White House did not respond to specific questions about how its firearms rollbacks may affect the country's gun suicide rate, or about what it is doing to prevent those deaths [18]. Spokesperson Lauren Bis said in a statement that "President Trump is committed to Making America Healthy Again and that includes mental health" [19]. She pointed to the administration's support for developing psychedelic treatments for mental illness and an announcement of $52.5 million in grants for veteran suicide prevention [20]. Veterans are one of the two groups KFF names as having had high suicide rates for years [10].

What to watch

  • Final rather than preliminary federal counts for last year's firearm and total suicide deaths.
  • Whether any published study puts a cost per death averted on storage or barrier programs, the figure a budget reallocation would need.
  • Whether the $52.5 million in veteran grants funds means-safety work or clinical identification.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories