Predictors of Suicidal Behavioral in College Students Seeking Treatment

This blog is a summary of a CCMH peer-reviewed study that examined predictors of suicide behavior amongst student clients who received services at counseling centers nationally.

Hayes, J.A., Petrovich, J., Janis, R.A., Yang, Y., Castonguay, L.G., & Locke, B.D. (2020). Suicide among college students in psychotherapy: Individual predictors and latent classes. Journal of Counseling Psychology, 67, 104-114. doi.org/10.1037/cou0000384

Suicide is the second leading cause of death among traditionally aged college students in the U.S. 

Each year, 10% of college students seriously consider suicide and roughly 1.5% engage in suicide behavior; these rates have increased slightly every year since 2010. Students who seek help at counseling centers are three times more likely than their classmates to report high levels of suicide ideation and five times more likely to have made a previous suicide attempt. Although many studies have identified factors associated with suicide ideation among college students, fewer studies have examined actual suicide behavior in this population, and virtually no research has investigated specific risk factors among college student clients. Given the serious consequences of suicide behavior, empirical data that help psychotherapists determine which clients possess elevated risk for suicide behavior may be valuable.

The Current Study

Data from 809,875 clients were gathered by CCMH from 2012 to 2018. Of these clients, 101,570 self-reported suicide ideation on the Counseling Center Assessment of Psychological Symptoms (CCAPS-34) at some point during treatment and had information available regarding whether or not they engaged in suicide behavior, as assessed by their clinician via the Critical Incident Tracking Form or Case Closure Form. We decided to focus only on clients who reported any level of suicide ideation to help improve predictive accuracy. That is, because suicide behavior is a relatively low frequency event, we thought that limiting the sample to clients who reported suicide ideation at some point during treatment would increase the proportion of true positives in the data set (clients who reported suicide ideation and engaged in suicide behavior), at the risk of excluding false negatives (clients who did not report suicide ideation but engaged in suicide behavior). The 101,570 clients were primarily women (62.5%) who were White (67.0%) and heterosexual (75.0%). Nearly a fifth of the sample (18.6%) reported engaging in suicide behavior at some point prior to the onset of treatment.

Results

A total of 391 clients engaged in suicide behavior during treatment, 12 of whom died by suicide. Clients with prior suicide behavior, prior self-injury, and high levels of depression were approximately 3.5 times, 2 times, and 1.5 times more likely, respectively, than clients without such histories to engage in suicide behavior during treatment. None of the other 17 variables that we examined significantly predicted client suicide behavior.

We also wanted to determine if there were various types, or categories, of clients with suicide ideation and the likelihood of suicide behavior within each type. Based on clients’ response patterns to seven risk variables on the CCAPS and SDS, the following four distinct categories emerged:

1. “Prior Ideation” Group: clients in this group are characterized, relative to other classes, by a high probability of previous serious suicide ideation and non-suicidal self-injury with low probabilities of prior treatment (counseling, psychotropic medication, or hospitalization). Overall, 23% of clients fell into this category and 0.55% of clients in this group engaged in suicide behavior during treatment.
2. “Extensive Risk” Group: clients had a high probability of endorsing each of the seven risk variables (high levels of depression, prior counseling, psychotropic medication, and hospitalization, history of self-injury and serious consideration of suicide, and a prior suicide attempt). This class included 16% of all clients and 0.99% engaged in suicide behavior during treatment.
3. “Prior Treatment” Group: clients with a history of serious suicidal ideation, counseling, and psychotropic medication treatment. The group comprised 23% of clients, 0.18% of whom engaged in suicide behavior during treatment.
4. “Circumscribed Depression” Group: clients who had high Depression scores at intake with no other risk factors endorsed, which included 38% of clients, and 0.25% engaged in suicide behavior during treatment.

Implications

This study examined predictors of suicide behaviors amongst students receiving services at college counseling centers. The following takeaways were noted:

• Consistent with previous research, at-risk college student clients exhibit symptoms of depression and have previously engaged in self-injurious behavior, either in the form of a suicide attempt or non-suicidal self-injury. That being said, the overwhelming majority of clients with these characteristics will not, of course, engage in suicide behavior during therapy, and our data do not provide insight into why clients engaged in suicide behavior or the lethality of their attempts.

• The findings notably challenge the view that self-injurious behaviors are no more than a “cry for help.” It could be that repeated painful experiences may lead to increasingly lethal self-injurious behavior. In fact, there is a movement within the field of suicidology to not distinguish suicide behavior from non-suicidal self-injury, as the distinction tends to obscure the fact that some self-injurious behavior (e.g., cutting) may result in suicide, whether intentional or not. Thus, it may behoove therapists to inquire about clients’ reports of prior non-suicidal self-injury, especially since data from the current study indicate that clients with such histories are more than twice as likely as clients without such histories to attempt suicide while in therapy.

• Contrary to expectation, a number of client characteristics were found not to predict suicide behavior, including important clinical variables such as anxiety, spirituality, and alcohol use.

• In terms of the four categories we identified, the most frequent type of suicidal client (those with “circumscribed depression”) carried a relatively low risk of engaging in suicide behavior (0.25%). In contrast, students who might be characterized as having “extensive risk” factors comprise the smallest portion of suicidal clients and yet carry the highest likelihood of engaging in suicide behavior during treatment (0.99%).Perhaps not surprisingly, clients in the Extensive Risk group were more than five times as likely as clients in the Prior Treatment group to engage in suicide behavior during counseling, four times as likely those in the Circumscribed Depression group, and as almost twice as likely as clients in the Prior Ideation group. On the whole, clients in the Extensive Risk and Prior Ideation classes (i.e., those characterized by a history of non-suicidal self-injury) comprised nearly two-thirds of the students who attempted suicide.

• The group with the second highest risk of suicide behavior was composed of clients who had not sought treatment previously but had thought seriously about suicide and had engaged in non-suicidal self-injury at some point prior to intake. Because these clients are likely unfamiliar with therapy and how to benefit from it, therapists may need to be particularly mindful of early interventions (e.g., explanations of how therapy effectively addresses client problems, referrals for psychotropic medication, monitoring the alliance) so that clients experience therapeutic gains before potentially dropping out of treatment. The clients in the “Prior Treatment” class serve as a point of contrast. These students had been in counseling previously and had been on psychotropic medication of some kind in the past. Despite these indicators of potential long-standing mental health problems, these clients were at the lowest risk of engaging in suicide during treatment, likely because they knew how to take advantage of the services available to them.

• It is important to keep in mind that the data do not shed light on effective aspects of treatment for clients who are suicidal. Of the roughly 100,000 clients who indicated at some point during treatment that they had suicidal thoughts, fewer than 0.5% engaged in suicide behavior. This raises important questions regarding what therapists did, and can do, to prevent suicide with this large group of student clients with suicidal ideation.

This blog post was written by CCMH Business Team member, Jeff Hayes, Ph.D. Jeff is Professor of Education at Penn State and a faculty research consultant to CCMH who has been working with CCMH since its inception.