Suicide remains the second-leading cause of death among adolescents, but many at risk remain unidentified. One solution is universal screening in the ED, a place nearly 20% of all U.S. adolescents visit annually. Investigators developed the Computerized Adaptive Screen for Suicidal Youth (CASSY) tool, which enables teens to undergo the screening on a tablet computer, with results immediately available.
- To boost accuracy, CASSY is adaptive, gradually personalizing the questions asked based on previous responses.
- While not all patients are asked the same number of questions in CASSY, the average number is 11, making the process brief and efficient for the emergency setting.
- In a validation study, investigators found CASSY predicts an attempted suicide within three months of screening with 88% accuracy.
- There are two other instruments: the Ask Suicide-Screening Questions (ASQ) and the Columbia-Suicide Severity Rating Scale (C-SSRS). If a patient screens positive on ASQ, then he or she takes C-SSRS.
- When conducting a suicide screen on an adolescent, ask questions in private, away from parents or friends who may influence answers.
The rate of suicide among U.S. adolescents has been rising. Experts note that the suicide rate among adolescents has grown by 62% over the past two decades, making it the second-leading cause of death among teenagers in the United States.1 In 2019, 1,580 youngsters between the ages of 12 and 17 died by suicide, according to the CDC.2
Sadly, many teens at risk remain unidentified, explains Cheryl King, PhD, a professor of psychiatry at the University of Michigan. “So many teens who die by suicide [about half] have never received any mental health services,” she says. “The risk has really gone unrecognized.”1
Experts believe there is a need for an efficient and accurate method for providing universal screening for suicide risk to every teen who presents for care to an ED for any reason, a group that includes almost 20% of all adolescents in the United States.1 “Then, it would be possible to screen a broader swath of the community,” King says.
King and colleagues have been working to perfect the Computerized Adaptive Screen for Suicidal Youth (CASSY), a tool that can reliably identify which patients are at risk and perform this task through a brief and efficient approach that does not disrupt care in the ED.
Push for Accuracy
To develop the algorithms in the tool, King and colleagues studied a cohort of more than 2,000 adolescents age 12 to 17 years who presented to one of the 13 geographically diverse EDs that are part of the Pediatric Emergency Care Applied Research Network (PECARN). Patients completed questionnaires regarding suicidal ideation, past suicide attempts, depression, self-harm, substance use, and other factors associated with suicide risk.
Participants and their parents received follow-up calls at three months following the screening to determine whether there had been a suicide attempt during that period. Researchers constructed CASSY using these data.
The authors validated the tool with a second cohort of 2,754 teenagers. There, CASSY predicted a suicide attempt with 88% accuracy within three months. A total of 165 adolescents in the group attempted suicide during this period.3
While many traditional screening instruments work, King and colleagues are striving to improve accuracy with an adaptive instrument that is more personalized.
“After the first few questions, the next question that a teen is presented with depends on how [he or she] responded previously,” King says. “We have learned that there are different profiles of risk. There is no one risk factor that is necessary or sufficient to suggest someone is going to make a suicide attempt.”
To the contrary, King notes teenagers who are suicidal are heterogeneous in terms of risk factors. “We may have a teen with bipolar disorder who is emotionally up and down with strong emotions, and may even be abusing substances, who is at risk for suicide. But we may have another teen who is severely depressed, withdrawn, and is not bothering anyone,” King explains. “That teen is also at risk for suicide. By having these algorithms [for] different people with different questions, we can probably get a more accurate prediction of risk.”
Assessing suicide risk in adolescents can be tricky because the way teens think and what they fear can vary day to day.
“There is no time in the human lifespan where there is such a high range of suicidal thoughts and suicide attempts as during the teen years,” King says. “We have more deaths by suicide among adults, but we have more reported suicidal thoughts and suicide attempts among teens, which makes prediction of a suicide attempt really tough.”
Further, when teens do attempt suicide, it is difficult to know if they intended to die by suicide.
“Sometimes, we meet them in the ED when they first come in and they say they did have intent,” King says. “Then, they get hospitalized, and you talk with them the next morning, and they say they were never trying to kill themselves; they had some other motive.”
However, accurately assessing risk is important because parents do not want to hear their adolescent is at risk when he or she really is not. Such a message can be highly disruptive to a family. Further, a false-positive assessment can lead to using mental health resources that often are in short supply in busy EDs.
“In general medicine EDs ... if 35% of all teens screen positive [for suicide risk] on a tool, they simply don’t have the resources to do full mental health evaluations on that many teens every day,” King notes.
In addition to seeing either a “positive” or “negative” indicator of risk from the screen, CASSY also delivers information indicating what level of risk a patient has displayed. “You can get a level ranging from negligible to severe,” King says.
Further, the system includes different risk thresholds. “What I understand health systems like is that depending on their resources, they can choose their threshold,” King adds.
For example, if there is plenty of space, mental health resources, and personnel, leaders can set a high threshold, capturing almost all the youth at risk. But they will see more false-positives, for which they are ethically obligated to follow up.
“However, a health system could set a different threshold, knowing that it won’t identify all the youth at risk, but it will do better than it is doing now,” King says. “The health system won’t get such a high proportion of false-positives, and [the caseload] might be more manageable.”
There is no point in screening for suicide risk if there is no mechanism in place to act on the result. Consequently, investigators have turned their attention to developing triage recommendations for patients who screen positive based on four risk levels: negligible, low-moderate, moderate-high, and severe.
“It is really clear that for the high level of risk, we are going to recommend that there is a full mental health evaluation and steps taken to ensure safety that day,” King says. “Systems may have some choice points for the low-moderate and moderate-high [risk levels].”
CASSY is geared toward patients age 12 to 17 years, and questions are designed to be answered on a tablet computer. The results are available to the provider immediately, and can be reviewed through an application programming interface or linked to an electronic medical record, if desired. Results pop up in real time.
Also, while the number of questions in CASSY can vary given its adaptive nature, the average number of questions is 11, making the screen easy to complete while patients wait for a provider.
Some pediatric EDs already maintain universal suicide screening protocols, using tools typically administered in person.
For instance, every patient who presents to the ED at Cohen Children’s Medical Center (CCMC) of Northwell Health in New York will be screened at some point during the encounter, explains Vera Feuer, MD, director of pediatric emergency psychiatry.
“We use a screening tool called the ASQ [Ask Suicide-Screening Questions],” she says.4 “If kids screen positive [on the ASQ], then they are further assessed with the Columbia-Suicide Severity Rating Scale.5 If their concerns are severe or acute, then they are seen by our psychiatry team that is in the ED.”
The ASQ is a secondary screening tool. Feuer says one ASQ item regards suicidal ideation. “There are kids who have had these thoughts in the past, but they no longer have them, so they are not necessarily acutely at risk. However, they still come up as positive,” she says.
How a screening tool is administered can make a big difference in how youngsters respond. “The recommendation is to always do [the questioning] in private, even among children as young as 10,” Feuer says. “Do not have the parent there when you are asking these questions because we know from adolescents that this really influences the way that they answer.”
In some cases, emergency staff may not be trained properly to conduct the assessments, but this is an issue that should be relatively easy to address.
Also, these computer-based questionnaires, such as CASSY, offer another way to ensure questions on suicide risk screens are answered privately.
With its busy, fast-paced environment, the ED is not an ideal setting for patients with mental health concerns. Some administrators have found a way around this environmental concern.
Feuer notes CCMC has established an ambulatory crisis center one floor above the ED.
“We try to assess kids there as much as possible. We have worked with our community, pediatricians, and schools in the area to be aware that there is another option, and that people don’t have to present to the ED to see a child psychiatrist,” she explains. “People can come to this ambulatory center and be assessed there in a very different kind of environment.”
Most kids with mental health concerns do not require the “super-restricted environment” of the ED in terms of safety. “They are cooperative, they are talking to us, and they are not out of control or unsafe. They really can be assessed elsewhere,” Feuer says.
However, in cases where patients present to the ambulatory center, but it is determined they do require the ED- or hospital-level care, they are easily redirected since the ED is in the same building.
“The set-up has been extremely successful in our health system. I have worked with other organizations to establish similar structures,” Feuer says.
Many EDs do not have the space or resources to create a separate center for behavioral health. In these departments, there may be a room or two, or a less-chaotic corner that can be dedicated to patients who present with behavioral health needs.
“Ideally, you have dedicated space, and you have dedicated, trained staff who know how to engage patients and know how to verbally de-escalate or coach kids on how to utilize coping skills in real time,” Feuer says. “This is something we try to do. That seems to really help in terms of [preventing] things like the need to medicate kids or the need to use restraints.”
Small changes or interventions can improve the ED experience for families with behavioral health needs. Feuer notes providing educational materials to guide parents can be beneficial. Such a step is not costly or staff-intensive.
“We try to do the same thing for the kids so that every visit has some sort of a therapeutic component,” Feuer says. “We now have these little cards with motivational messages and a hotline number, just small things to provide support and help [parents] cope.”
The idea is to guard against such a negative experience that patients or families are turned off from ever again seeking help.
“We don’t want that,” Feuer says. “We want them to have a different feeling when they walk away. My experience is that even small things can make a big difference in how the patient experiences the ED visit.”
- University of Michigan Medicine. Personalized screening to identify teens with high suicide risk. Feb. 2, 2021.
- National Institutes of Health. Computerized adaptive screener may help identify youth at risk for suicide. Feb. 3, 2021.
- King CA, Brent D, Grupp-Phelan J, et al. Prospective development and validation of the computerized adaptive screen for suicidal youth. JAMA Psychiatry 2021; Feb 3;e204576. doi: 10.1001/jamapsychiatry.2020.4576.
[Online ahead of print].
- National Institute of Mental Health. Ask Suicide-Screening Questions (ASQ) Toolkit.
- The Columbia Lighthouse Project. The Columbia-Suicide Severity Rating Scale.