Novel Suicide Prediction Tool Goes Beyond Patient Self-Reports

Nancy A. Melville

April 15, 2019

CHICAGO — The majority of individuals who die by suicide fail to disclose their suicidal thoughts to healthcare providers, a fact that highlights the need for better predictive tools that go beyond patient self-report.

In new research, a novel two-part assessment tool that includes both patient and clinician data may fit the bill.

This new, briefer version of the Modular Assessment of Risk for Imminent Suicide (MARIS) tool includes the Suicide Crisis Syndrome Checklist (SCS-C), which is a patient self-report assessment, and the Therapist Response Questionnaire–Suicide Form (TRQ-SF), which measures clinician emotional responses to potentially suicidal patients.

Investigators from Mount Sinai Beth Israel in New York City administered the SCS-C to more than 400 adult patients with psychiatric disorders and self-reported suicidal ideation (SI). In addition, they administered the TRQ-SF to 59 clinicians who were treating/assessing these patients.

Results showed that a high TRQ-SF score was significantly associated with an increased risk for suicide attempts and with plans by patients, a finding that underscores the value of looking beyond patients' own reports.

"Can we rely on patients with bipolar disorder to diagnose their mania, or on those with schizophrenia to diagnose their psychosis? No. Then why do we rely on patients in suicidal crisis to diagnose their suicide risk?" asked study investigator Igor Galynker, MD, PhD.

"It's quite absurd, but that's what we've been doing for a long time, and we need new approaches in suicide risk assessment," he said.

Galynker presented the findings here at the Anxiety and Depression Association of America (ADAA) 2019.

New Diagnostic Tool

No existing diagnostic tool has been found to be of clinical value for the assessment of short-term suicide risk among high-risk individuals. Galynker and his colleagues consequently developed MARIS and assessed the tool in a proof of concept study.

In a follow-up interview, Galynker told Medscape Medical News that the new version is a simplified prediction tool that can be used for a broader range of patients.

"The instrument is highly efficient. Just one patient module and one clinician module are sufficient," he said.

In separate research, Galynker and his team reported that clinicians' emotional responses, such as hopelessness or distress, are clinically significant predictors of a patient's immediate suicide risk. On the basis of this information, they developed the TRQ-SF.

In the new study, the investigators wanted to determine whether combining the TRQ-SF with the SCS-C would improve predictive value above and beyond each instrument alone.

These two measures are among four included in the original MARIS tool. The other two measures — acceptability and explicit risk — have been shown to have lower predictive value, and so the researchers developed a shorter version, which they dubbed the Mini-MARIS.

For the study, the SCS was administered to 452 adult psychiatric patients within 72 hours of intake at Mount Sinai Beth Israel in New York City.

All participants admitted experiencing SI and were considered to be at moderate risk for suicide. Fifty-nine clinicians were administered the TRQ-SF.

An evaluation of the patients' suicidal behaviors at 4 to 8 weeks from the initial assessment was made using a variety of measures, including the Columbia Suicide Severity Rating Scale, which assesses past-month and lifetime suicidal ideation; the Beck Scale for Suicide, which examines SI over the past week; and the Modified Sad Person Scale (MSPS). On the MSPS, the investigators used total scores and scores related to individual items of suicidal risk, such as rational thinking, previous suicide attempts, and state self-harm.

Independent Predictive Value

Results of a bivariate analysis of 359 patients who reached follow-up showed that a high score on the TRQ-SF scale was significantly associated with an increased risk for patients' suicide attempts (P = .01) and suicide plans (P = .008).

Meeting DSM criteria for SCS was only associated with suicide attempts (P = .01).

Responses that met either SCS criteria or high TRQ-SF scores were associated with short-term suicide plans (P = .001) and attempts (P = .001). However, the two measures were not correlated.

"Interestingly, there was no correlation between the patients' and clinicians' assessments in predicting imminent suicidal behavior," Galynker said. He suggested that with no overlap, each has independent predictive significance.

On the TRQ-SF scale, clinician responses of affiliation and distress but not hopefulness about the patient's state were predictive of suicide attempts and plans (P < .05).

The results show that "on multivariate analyses, meeting either TRQ-SF or SCS improved the models in predicting both suicide attempt and plans when compared with traditional suicide risk factors," Galynker said.

"These are moderate-risk outpatients rather than high-risk inpatients [in the study]. This means that Mini-MARIS can be used in both groups of patients," he added.

The importance of having a reliable prediction tool is underscored by notable previous research that shows that although 50% to 70% of people who die by suicide have contact with health services a month before their death, more than 75% do not report SI.

This highlights the critical need for multi-informant and multimodal assessments, Galynker said.

Common Clinical Challenge

Commenting on the findings, session discussant Jill Harkavy-Friedman, PhD, associate professor of clinical psychology at Columbia University, New York City, and vice president for research at the American Foundation for Suicide Prevention, noted the value of efforts to improve suicide-prediction models.

"Just 30 years ago, no one was even talking about this, and now we have not just instruments but interventions," Harkavy-Friedman said.

"I think the field has come a long way, and it's because of these types of discovery of models and testing, etc," she added.

She summed up an all-too-common clinical challenge that psychiatrists face.

"If you've ever had a patient that you were worried about but they were not worried, and your heart is racing and you're hoping they make it to the next session, or the opposite scenario — a patient is talking about suicide, but you aren't that worried — that's the conundrum we have here," she said.

"Suicidal ideation is very frequent and in the moment. Sometimes it's informative, but sometimes not, and there is the question of whether we should even ask people if they are thinking of suicide," Harkavy-Friedman noted.

"I think the bottom line is no matter how you think about suicide risk, it's important to know if they're thinking about it or if they've ever thought about it before. It's not surprising that lifetime suicidal ideation is a better predictor of suicide risk than current ideation," she concluded.

The study was funded by the American Foundation for Suicide Prevention (AFSP). The study authors have reported no relevant financial relationships. Harkavy-Friedman is vice-president of the AFSP.

Anxiety and Depression Association of America (ADAA) 2019: Abstract 1172. Presented March 31, 2019.

For more Medscape Psychiatry news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.