Opinion | Is ‘Some Good’ Good Enough in Suicide Risk Assessment?

Riddle: You see a boat filled with people. It has not sunk, but when you look again you don’t see a single person on the boat. (Hint: Think of the famous Sherlock Holmes quote, “There is nothing more deceptive than an obvious fact.” See the answer to the riddle at the end of the op-ed.)

Last month, the U.S. Preventive Services Task Force (USPSTF) restated what it had summarized for several years: “The current evidence is insufficient to assess the balance of benefits and harms of screening suicide risk in children and adolescents.”

Again, again, and again, the USPSTF evidence reviews on screening for suicide risk include:

• More randomized trials are needed on the benefits and harms of screening for suicide risk among children and adolescents in primary care settings compared with no screening or usual care
• More information is needed on the performance characteristics of screening tests for suicide risk
• Treatment studies are needed in populations with screen-detected suicide risk, in all age groups
• Evidence on screening and treatment is lacking in populations defined by sex, race and ethnicity, sexual orientation, and gender identity, such as American Indian/Alaska Native youth (who are at increased risk for suicide)

Any discussion of “best practices” or “a gold standard,” including the glossy and well publicized Columbia-Suicide Severity Rating Scale (C-SSRS), Computerized Adaptive Screen for Suicidal Youth (CASSY), and countless other ideation-centric assessments, is currently meaningless.

Can ideation continue to serve as a surrogate or proxy for near-term suicidal events? Clearly, where roughly half of patients “deny” ideation, other factors hold significance. Although the latest USPSTF recommendation may be considered a significant step backward by some, especially those with vested financial and fragile reputational interests, this finding represents an opportunity to “be better than good.” By exploring new questions, factors, and methods, the standard of assessment care, practices, and ethics may be elevated. What then should be the standard of care? It is not prediction. It is not mathematical certainty. It is reasonable, probabilistic anticipation of risk based upon a systematic review of conventional and unconventional factors. It may be considered a corollary to the legal standards of evidence, for example, preponderance of the evidence, clear and convincing, and beyond a reasonable doubt.

Rather than thoughtlessly following the limitations of with often reckless and wasteful search, selection, availability, confirmatory, and representative biases, I urge the readers to utilize accelerated and probabilistic likelihood ratio (LR) protocols in research and clinical practice. I suggested this model in 2006 in the American Journal of Emergency Medicine, “Adolescent Violence Screening in the ED.” This screening version provided a tiered and rational system for reasonable foreseeability of suicidal patients with successive stage and final score LRs.

LRs have been used in:

• The deception of epidural hematoma in anemia
• The deception of myocardial pain
• The deception of flank fullness in overweight patients

Similar to the HEART emergency department protocol (history, ECG, age, risk factors, initial troponin), a suicide assessment cascade of this new type evaluates the impact of ideation, non-ideation, and dysexecutive states on attempt rates with confirmatory bedside neurological tests. The deception of ideation, for example in adjustment disorder with dysexecutive features, can be answered within reasonable medical certainty, i.e., an LR >10, in specific suicidal youth cohorts (P value lower than 0.05).

And, most critically, this fast-track method will likely interrupt the awfulness of unrecognized, misdiagnosed cases based on valueless ideation-centric assessments, fraught with false-positive and false-negative outcomes. How many times do I need to emphasize this point?

There is the stopgap USPSTF suggestion that screening for “depression and anxiety” may improve the assessment landscape. However, imprecise definitions, the consequent construction or minimization of findings, and increasing criticism of sensitivity/specificity statistics will likely perpetuate the same old problematic, proverbial wheel with tragic outcomes. We are desperately in need of new tools and instruments.

And what has been the awful cost of this valueless proposition? In that we are at a historic 50-year high in suicide deaths and attempts, and a mystifying rise of youth suicides without traditional “red flags,” the consequences to a civil society when either danger is normalized (false negative) or violence constructed (false positive) are immense. Think for a moment of the enormity of the tens of thousands of unnecessary lives lost, damage to personal reputation, school dysfunction, troubled family and peer relationships, extreme ED boarding times, and medico-legal consequences to providers and hospitals.

Add to this tragic excess the unfathomable waste of privately and publicly funded dollars, including Public Health Service monies, the caretaker of “well allotted” U.S. taxpayer funds. For example, the National Institutes of Mental Health’s CASSY study, another ideation-centric assessment referred to earlier, cost the public over $10.5 million.

Let me ask again, “Is some good good enough?”

“No.” Unequivocally “NO!” What other medical specialty would tolerate such recurring, indeed escalating personal disaster? Let me state clearly that the study of suicidology is fraught with multi-layered complexity. It is hard. It involves observation, memory, expression, and reasoning. It is richly the scientific method. I take pride in continuing to learn for imaginative and scholarly sake, but most importantly for the public welfare.

So, Dr. Copelan, “What is your value proposition?”

1. Suicide risk assessment is a challenge in that suicidal thoughts and behavior represent a complicated family of presentations. Therefore, a validated, dimensional assessment requires evidence informed definitions across medical, neurologic, and psychiatric observations, suspension of biases, precision data-driven protocols, and new application of really old (Bayes’ theorem) statistical methods (LR).
2. A comparator ED study assessing the C-SSRS and CASSY with a pioneering, LR probabilistic based, brief physician guided assessment will evaluate the benefits and harms of screening for suicide risk among children and adolescents. A description of the RCT sample, measures, data collection, data analysis method for goodness of fit, study’s strengths and weaknesses, and number of subjects, the power of the study, and main outcome and follow-up measures will be indicated.
3. In this analysis, the simple likelihood-ratio based statistic will be used and derived successively for testing that the nested model (C-SSRS & CASSY identical ideation predictor variable) does not quantify strong support, and that the alternative model yields strong to very strong evidence in favor of the hypothesis.
4. Inducting theory was developed from case and completed feasibility studies across diverse groups. The specific, rather than universal inclusion criteria will document the process of selection thinking exercised. They include suspected, new onset, or worsening suicidal or violent thoughts; abrupt change in thinking or behavior as a result of psychological, medical, or medication event; and medically cleared resuscitated suicide attempters.
5. The protected lives, reduced legal burden, lessened ED boarding load, and significant cost production efficiency outcomes, as a result of this confident, pioneering, probabilistic approach, are consistent with the sentinel new event alerts of the Joint Commission and CMS for this serious, potentially fatal, and currently inadequately evaluated and treated public health emergency.
6. You know what else you guys might like? Let’s do CME, grand, or ED rounds sometime soon.

Answer to the riddle: All the people were married. Suicidal ideation is similarly deceptive.

If you or someone you know is considering suicide, call the National Suicide Prevention Lifeline at 1-800-273-8255.

Russell Copelan, MD (Ret.), lives in Pensacola, Florida. He graduated from Stanford University and UCLA Medical School. He trained in neurosurgery and completed residency and fellowship in emergency department psychiatry. He is a reviewer for Academic Psychiatry and founder of eMed Logic, Inc., a non-profit originator and distributor of violence assessments.