HSR&D Home » Research » IIR 11-298 – HSR&D Study
Increasing Treatment Seeking Among Suicidal Veterans Calling the Crisis Line
Tracy A. Stecker, PhD
White River Junction VA Medical Center, White River Junction, VT
White River Junction, VT
Funding Period: May 2014 - April 2015
Reducing suicide is a national priority and an urgent concern within the Department of Veterans Affairs. The passage of the Joshua Omvig Veterans Suicide Prevention Act highlights the importance that stakeholders including Veterans groups, VA Central Office, and Congress place on developing and implementing a comprehensive program to reduce suicide among Veterans. By addressing suicide risk among Veterans, the current proposal is responsive to VA HSR&D "Priority E-Mental and Behavioral Health". In particular, the goal of the proposed study is to improve initiation of behavioral health (i.e., mental health, substance use) treatment services among at-risk Veterans. This goal recognizes that most individuals who die by suicide are not in behavioral health treatment despite evidence that the large majority have one or more behavioral health disorders (i.e., depression, alcoholism) that are known to promote risk. The goal to facilitate behavioral treatment is also consistent with recommendations provided in the U.S. National Strategy to Prevent Suicide, the Institute of Medicine report on reducing suicide and the recent Department of Defense Suicide Prevention Task Force report.
Veteran populations are at increased risk for suicide. Moreover, rates of suicide among active duty service members have increased dramatically since 2005, causing great concern that increased risk will carry over to individuals who leave the military. In response, in recent years the VA has implemented a number of strategies and services to address suicide risk, among them the creation of the Veterans Crisis Line in 2007. Although by calling the Crisis Line thousands of at-risk Veterans are coming forward, nonetheless many express resistance to behavioral health treatment services. These Veterans are likely to remain at risk unless strategies can be developed to link them to care in a timely manner. By identifying and treating at-risk Veterans as soon as possible, the negative impact on Veterans and their families can be minimized.
Building on theory and previous empirical findings, we proposed a randomized controlled trial to pilot test the efficacy and feasibility of a brief, individualized cognitive-behavioral (CB) intervention to promote entry into behavioral health treatments for at-risk Veterans who call the VA's National Crisis Line and are resistant to initiating behavioral health treatment. The goal of this research was to evaluate the efficacy of the intervention for increasing engagement to behavioral health treatment services.
AIM 1: PILOT TEST THE INTERVENTION ON ATTITUDES TOWARD TREATMENT. The objective of this aim of the study is to conduct a pilot trial of the intervention on attitudes toward behavioral health treatment among at-risk Veterans who call the Crisis Line and are resistant to behavioral health treatment. The intervention was tested on 19 at-risk Veterans to determine whether the intervention modified beliefs and the intention to initiate behavioral health treatment.
AIM 2: PILOT TEST THE INTERVENTION ON INITIATION OF AND ADHERENCE TO TREATMENT. The objective of this aim is to test the intervention on the initiation of and adherence to behavioral health treatment. A secondary objective of this aim was to test the efficacy of the intervention on the degree of suicidal ideation (SI) as measured by the C-SSRS.
This pilot trial was conducted at the Canandaigua VA. Participants (n=19) were recruited by Crisis Line responders at the end of calls from among callers to the Veterans Crisis Line. Eligible participants were those who indicated suicidality during the call but who were not in treatment and refused a transfer to their local Suicide Prevention Coordinator. The potential participant was transferred to speak with a research assistant who is also an experienced clinician. Potential participants were given the option to be mailed a recruitment letter instead of being directly transferred to a researcher. If interested and eligible, they went through the consent process and were administered a brief pre-intervention assessment.
Participants determined to be eligible and agreeing to participate went through the informed consent process with the research assistant. Eligible participants who consented were assigned a random binomial number representing intervention or control condition. The member of the research team determined to assigned group was blind to participant data. All interested and eligible participants were asked to schedule a time with the research assistant to complete an initial assessment by phone. Completing this assessment took approximately 75-90 minutes by phone. During the scheduled appointment, intervention participants were scheduled to receive the brief, individualized CB intervention.
Follow-up assessments: Follow-ups were performed at one and three months following the baseline. Participants were assessed on suicidality, symptoms of PTSD and depression, beliefs about treatment, and treatment utilization. These sessions took approximately 30 minutes to complete.
The PASS was administered to all participants in order to assess beliefs about behavioral health treatment and the extent to which they intend to initiate treatment. The PASS has solid psychometric properties which were previously described.
The Colombia Suicide Severity Rating Scale:
The Columbia Suicide Severity Rating Scale (C-SSRS) is mandated for use in medication trials by the Federal Drug Administration and has been uniquely recommended as a standalone measure of change in suicidal thoughts and behavior over time. Data also support reliability and validity of phone administration. We used the C-SSRS suicidal ideation "intensity" item as a secondary outcome measure. The item rates intensity of suicidal ideation at baseline and during each follow-up on a five-point scale. The intensity ratings are further assigned to the "most common" ideation and the "most severe" ideation during follow-up. We used the "most severe" rating of ideation (rather than the "most common" rating) for analysis because even short bouts of suicidal ideation may be intense and life-threatening, for example short-lived but severe ideation may occur following a stressful life event or during alcohol intoxication.
We assessed symptoms of the most common disorders (depression, alcohol, drug, PTSD) among Veterans who die by suicide. 1) Depressive symptoms: The Physicians Health Questionnaire or PHQ-9 assessed depressive symptoms over a two-week period. 2) Alcohol use severity: The widely used Addiction Severity Index or ASI assessed symptoms and behaviors over a 30-day period. To assess alcohol use severity, we used a reliable and valid "alcohol composite score" based on frequency of alcohol use, frequency of intoxication and perceived alcohol-related problems. 3) Drug use severity: We used a reliable and valid ASI "drug composite score" based on frequency of use of 11 types of drugs and perceived overall drug-related severity over a 30-day period. 4) PTSD symptoms: The PTSD Checklist (PCL-M) assessed the diagnostic symptoms of PTSD over the previous month. Respondents were asked to rate how much they were bothered by each item on a five-point scale.
The follow-up assessment interview has been modeled from the Treatment Services Review, TSR (McLellan, 1992), and participants were asked about their use of substance use and mental health treatment services. Participants were asked whether they have initiated treatment (in terms of both scheduling the appointment as well as attending the appointment), and the number of treatment sessions attended. Information regarding the treatment session was assessed (i.e., whether the appointment was at the VA or a non/VA setting, whether the appointment was with a psychiatrist, psychologist, or other, etc.). In order to explore treatment outcomes, we assessed the type of treatment received (i.e., CBT, medications, etc.). The C-SSRS follow-up and symptoms of the most common disorders among VHA patients who die by suicide (i.e., depression, alcohol, drug, PTSD) were also assessed at each follow-up. We considered using the VA electronic medical record to assess for service use. However, since a large number of participants used non-VA treatment facilities for treatment either due to preference or because they do not have access to the VA system (e.g., due to discharge from the military for substance abuse), our most efficient method to conduct this assessment was self-report.
All participants completed all measures. Participants were randomly assigned to either the standardized usual care or intervention. Randomly assigned usual care participants received standardized care as usual per Crisis Line protocol.
Brief CB intervention:
The CB intervention is a brief, manualized, tailored one-on-one single session lasting 45-60 minutes and administered by phone. The intervention targets a change in the beliefs that influence whether or not someone enters mental health or substance use treatment. During the session, participants were asked to dispute beliefs about behavioral health treatment elicited by the PASS assessment instrument in a three-step process. During the three-step process, participants were asked about their stated beliefs, whether the belief is 100% true, whether there were any alternative beliefs, and whether anything could change the accuracy of the belief (how realistic is this belief). For example, if a participant indicated that they believed it would be too difficult to talk about some of their experiences at war with a mental health professional, they were asked to estimate how accurate this belief is (0-100% of the truth). They were asked about other thoughts that come to mind in terms of talking about their experiences, what this would mean about them, and about the worst thing that could happen if they did talk about the experiences. After the participant discussed each elicited thought in more depth, they were asked to dispute these thoughts. For example, a thought such as "If I start talking about what I experienced, everyone might think that I am crazy" might have been altered to "If I start talking about what I experienced, I might feel out of control momentarily but talking about it might make me feel better" or "Holding this inside has been really hard so maybe it would be better for me if I let it out in a safe environment". A maximum of three beliefs was discussed in this manner during the session. The participant and interventionist mutually agreed upon beliefs discussed during the session.
All participants assigned to the treatment condition were administered the same structured intervention; however, content discussed within the session differed for each participant based on their responses and thoughts surrounding the beliefs discussed within the session. While participants may frequently endorse similar beliefs (e.g., Some experiences would be difficult to talk about), their ideas surrounding that belief discussed in the session differed. Therefore, while the intervention is structured, it is not scripted.
Feasibility of implementing the intervention through the Crisis Line.
As a feasibility trial, the trial was successfully able to implement the intervention at the VA's National Crisis Line including systematically training approximately 100 Crisis Line responders in how to make a referral to the study, ongoing coordination with these individuals and their supervisors, and receiving one or more referrals from numerous Crisis Line responders who were trained, with these accomplishments occurring during a period of dynamic change on the Crisis Line including turnover of top leadership and major changes in procedures and personnel.
Characteristics of the sample
Nineteen participants were recruited for this trial, 9 in the intervention condition and 10 in the control condition. Participants were primarily male (78%, 91% respectively) and had a mean age of 39.8 and 44.6, respectively. Baseline scores on measures of depression and PTSD indicate that the sample at baseline reported moderate to severe depression (20.3 and 16.0, respectively on the PHQ-9) and moderate symptoms of PTSD (57.0 and 53.0, respectively on the PCL-M).
Specific Aim 1:
All data was processed and managed using SPSS Version 16.0.
Statistical Power Considerations. This pilot study was not powered to detect statistically significant differences. Basic statistical tests are described below and were conducted, but statistically significant findings could not be found because the sample size for this pilot was not sufficient to ensure stable standard errors.
Hypothesis 1: Participants receiving the CB intervention will have significant increases in positive beliefs as assessed at the one month follow-up than participants in standardized usual care.
Data Analysis for Hypothesis 1:
To assess the effect of the intervention on the change in beliefs about behavioral health treatment, a post- minus pre- change score was computed for beliefs about treatment. This variable ranged from 1 to 7. Change scores were computed from baseline to 1 month to 3 months in both conditions.
Results for Hypothesis 1:
We assessed four items regarding beliefs about treatment. The first item asked the respondent "By attending treatment, I will have fewer bothersome symptoms", scaled on a scale of 1 (unlikely) to 7 (likely). The intervention participants went from having a mean score of 4.0 at baseline to a 4.4 at the 1 month follow up. Control condition participants went from a mean score of 4.5 to a 5.3 at the 1 month. For the item "Going to treatment means that I can't handle my own problems" on a scale of 1 (unlikely) to 7 (likely), intervention participants went from a 2.7 to a 3.6 and control participants went from a 3.7 to a 3.3. For the belief "Some of my experiences will be too difficult to talk about in treatment" on a scale of 1 (unlikely) to 7 (likely), intervention participants went from a 3.7 to a 3.9 and control participants went from a 3.4 to a 2.3. A fourth belief asked respondents "Treatment is worthless/valuable" on a scale from 1 ("worthless") to 7 ("valuable"). Mean scores from intervention participants at baseline were 6.0 and moved to 6.4. Mean scores from control participants were 6.1 at baseline and moved to 6.0 at 1 month.
Overall, there were no statistically significant differences in changes in beliefs regarding treatment during the course of the trial, but low statistical power (given the very small sample) requires cautious interpretation of the nonsignfiicant result.
Hypothesis 2: Participants receiving the CB intervention will have significant increases in the intention to initiate behavioral health treatment at the one month follow-up than participants in standardized usual care.
Data Analysis for Hypothesis 2:
To assess the effect of the intervention on the change in beliefs about behavioral health treatment, a post- minus pre- change score was computed for the belief targeting the intention to seek treatment. This variable ranged from 1 to 7, with a 1 indicating strongly that the participant did not intend to seek treatment and a 7 indicating a strong intention to seek treatment. Change scores were computed from baseline to 1 month to 3 months in both conditions.
Results for Hypothesis 2:
The intervention group mean score on this item went from 6.1 at baseline to 6.8 at 1 month, and all of the intervention participants sought treatment between baseline and 1 month.
The control condition did not show movement in this item. At baseline the mean score on this item was 6.5, and at month 1, the score moved to 6.4. It should be noted that control condition participants seemed to report an intention to seek treatment at baseline.
Overall, there were no statistically significant changes in beliefs about treatment seeking or obtaining a treatment appointment, although with a larger sample such differences might be detected given the more positive trends in the data in the intervention versus control group.
Specific Aim 2:
Hypothesis 1: Participants receiving the CB intervention will be more likely to initiate behavioral health treatment than participants in standardized usual care during the three month follow-up period.
Data Analysis for Hypothesis 1:
This hypothesis will be assessed by count. The dependent variable will be specified as whether or not behavioral health treatment was initiated in the prior 30 days, the primary outcome of the study.
Results for Hypothesis 1:
All of the intervention participants (100%, 4/4) sought treatment by the 1 month follow up assessment versus (57%, 4/7) of the control condition. Participants receiving the intervention were more likely to seek treatment by the one month follow up than participants in the control condition.
Hypothesis 2: Participants receiving the CB intervention will receive more behavioral health treatment than participants in standardized usual care.
Data Analysis for Hypothesis 2:
This hypothesis will be assessed by counts.
Results for Hypothesis 2:
Overall, there were no statistically significant differences in changes in treatment seeking during the course of the trial, but low statistical power (given the very small sample) requires cautious interpretation of the nonsignfiicant result. Indeed, all of the intervention group participants sought treatment.
Secondary Analysis: Determine if participants receiving the CB intervention report lower levels of suicidal ideation (SI) post-intervention than participants receiving standardized usual care.
Results for Secondary Analysis:
Both conditions reported a mean score on the suicide severity scale of the CSSR-S at baseline of 3.8. This would indicate that participants were actively suicidal with some intent to act but without a specific plan. At the one month follow-up, intervention participants moved down to a 2.0 and control participants moved down to a 2.2, indicating non-specific but active suicidal thinking. Overall, there were no statistically significant differences in changes in suicidal ideation during the course of the trial, but low statistical power (given the very small sample) requires cautious interpretation of the nonsignfiicant result.
The main finding from this pilot trial is that we were able to successfully work through logistics involved in implementing a procedure to have callers to the VA's National Crisis Line who report suicidal ideation but refuse a treatment referral receive an evidence-based intervention session designed to improve treatment utilization. Individuals who received the intervention were 100% likely to seek treatment after the intervention session (versus 57% of the control condition), and reported a higher intention to seek treatment. There were no differences found; however in the number of treatment sessions received by those in the control versus intervention conditions, but this could be due to a low n in the trial.
External Links for this Project
NIH ReporterGrant Number: I01HX000805-01A2
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
PUBLICATIONS:None at this time.
DRA: Mental, Cognitive and Behavioral Disorders
DRE: Treatment - Efficacy/Effectiveness Clinical Trial
MeSH Terms: none