An Emergency Department Intervention and Follow-Up to Reduce Suicide Risk in the VA: Acceptability and Effectiveness

Suicide is a leading cause of death in the United States (1). Veterans are a particularly high-risk group (2), although the suicide rate among veterans has slowly decreased in recent years (3). Emergency departments (EDs) often serve as the primary—and sometimes only—point of contact for suicidal individuals (4). Unfortunately, suicidal patients who are discharged from the ED often fail to connect with follow-up care or experience significant delays between ED discharge and outpatient treatment (5–7). Up to 50% of suicide attempters seen in EDs do not follow up with postdischarge treatment recommendations (6). Of those who do follow through, approximately 40% discontinue treatment within three months (7). This is particularly concerning because the period of greatest risk for additional suicidal behavior is one to six months after an initial attempt (7).

Because the ED may be the sole point of contact with the mental health system for suicidal patients, it can serve a crucial role in mitigating short-term risk by implementing brief psychosocial interventions targeting suicidal behavior (4). Studies examining such interventions are few in number but promising. Psychoeducation and follow-up contact interventions are shown to improve treatment engagement and reduce the risk of suicide attempts in the month to years after the index ED visit (8,9).

Given that both brief in-person and follow-up contact interventions for suicidal individuals show promise, we combined the in-person ED-administered Safety Planning Intervention (SPI) with structured follow-up and monitoring (SFU) by telephone to increase the potency of the intervention. Conceptually, the aim of SPI-SFU is to provide patients with a crisis management tool. For many suicidal individuals, acute suicidal crisis lasts for a very brief period, and thus professional intervention during this window is difficult or even impossible (10). SPI is a brief, evidence-based clinical intervention (11) identified as a best practice in the Suicide Prevention Resource Center–American Foundation for Suicide Prevention Best Practices Registry for Suicide Prevention. SFU consisted of highly structured follow-up phone contact that was designed specifically for this study (12). The intervention, which was identified in the VA as SAFE VET (Suicide Assessment and Follow-up Engagement: Veteran Emergency Treatment), was implemented in EDs of five Department of Veterans Affairs (VA) medical centers as a demonstration project with veterans who did not require hospitalization as a result of their ED visit. SAFE VET has been found to enhance treatment engagement (12). The purpose of the study reported here was to evaluate the acceptability and perceived effectiveness of the intervention by the veterans who received it.

Methods

Participants who presented to the ED for a suicide-related concern but who did not require psychiatric hospitalization were eligible to participate in the demonstration project, details of which have been reported elsewhere (12). To participate in this interview study, a convenience sample of SAFE VET participants was developed by contacting participants at random from the list of participants enrolled from February 2010 to May 2012 (that is, they agreed to participate in the project prior to receiving the intervention) at two project sites (Manhattan and Philadelphia). One hundred participants completed a semistructured interview between February and June 2012 to assess the acceptability and effectiveness of the intervention.

SPI helps to mitigate future risk by providing suicidal individuals with a specific prioritized list of coping skills and social supports that can be used if suicidal thoughts emerge. SPI has six key components: identifying warning signs of an impending suicidal crisis; utilizing internal coping strategies; engaging social contacts and social settings to distract from suicidal thoughts; contacting social supports, including family members and friends for support and assistance in resolving the suicidal crisis; contacting mental health professionals for support and assistance; and minimizing access to lethal means. SPI was adapted for use with veterans and is required for high-risk suicidal veterans seen in the VA. We extended SPI use to patients who expressed suicidal concerns in the ED but who were not subsequently hospitalized for psychiatric concerns.

As part of the SFU, participants received their first call within seven days of their ED visit, during which the following took place: mood check and suicide risk assessment; assessment of adherence to follow-up mental health care treatment recommendations; problem solving of barriers to follow-up treatment engagement; and review and revision, if needed, of the SPI. Additional calls were conducted weekly until the first outpatient treatment appointment took place. Within one week of this visit, participants were contacted to assess satisfaction with their appointment and address any problems that would hinder future attendance. Participants were contacted multiple times if initial efforts to reach them were unsuccessful. When imminent or elevated suicide risk was detected, procedures included contacting a project principal investigator or emergency services. The intervention was administered by highly trained project staff. Treatment fidelity was increased through role play during the training phase and weekly group supervision by a licensed doctoral-level psychologist who was familiar with SPI and SFU.

Participants were invited by mail to participate in a 45-minute interview after completion of SAFE VET participation. All participants provided informed consent and received $50 compensation. Interviews were completed by phone by master’s-level clinicians and consisted of open-ended questions about the acceptability, usefulness, and helpfulness of the intervention. Two 5-point Likert rating scale items were used to quantify satisfaction ratings of the SAFE VET components. Interviews were recorded and transcribed. Two senior investigators (BS and GKB) developed a coding strategy based on a review of the first 12 interviews, with some additional codes developed for responses that were frequent in the remaining interviews. Interviews were coded by using thematic analysis. Themes were summarized with descriptive analysis. Coders were trained on the coding strategy and demonstrated good interrater reliability (SPI questions, κ=.82, p<.001; SFU questions, κ=.87, p<.001; and SAFE VET questions, κ=.81, p<.001). Data were analyzed with SAS software, version 9.2. The Manhattan and Philadelphia VA institutional review boards approved this research study.

Results

Veterans interviewed for this study were not clinically different from those who participated in the demonstration project. The sample of 100 veterans was predominantly male (N=89, 89%) with at least some college education (N=46, 46%). The mean±SD age was 44.9±13.9, with roughly equal numbers of white (N=45, 45%) and African-American (N=44, 44%) participants. Sixty-eight veterans (68%) were living with others, and 14 (14%) were homeless. Periods of service were as follows: Vietnam, N=24, 24%; post-Vietnam, N=19, 19%; Persian Gulf War, N=20, 20%; and Operation Enduring Freedom/Operation Iraqi Freedom, N=37, 37%. Chart diagnoses were as follows: mood disorders, N=66, 66%; posttraumatic stress disorder, N=35, 35%; alcohol use disorder, N=44, 44%; and drug use disorder, N=36, 36%. Forty-nine participants (49%) were suicide attempters, with 25 (25%) reporting multiple attempts. Forty-five (45%) reported having attended one or more VA mental health appointments in the year preceding their index ED visit.

With respect to acceptability of the SPI, most participants reported being satisfied (N=97, 97%); three (3%) reported that they were neutral. All but one participant (N=99, 99%) identified at least one helpful component, which included social contacts and places for distraction (52%), social support for help (47%), professional contacts (45%), internal coping strategies (27%), warning signs (8%), and ways to restrict access to lethal means (1%). For those who found the SPI unhelpful (N=5), reasons provided are listed in Table 1.

Almost all participants recalled completing the SPI (N=98, 98%), and most could identify the current location of their plan (N=88, 88%). The most frequently recalled components of the SPI included internal coping strategies (N=73, 73%), social contacts and social places (N=85, 85%), social support (N=83, 83%), and professional contacts (N=87, 87%). Of note, very few participants (N=6, 6%) recalled that their plan included restricting access to lethal means.

Most participants (N=61, 61%) reported using the SPI to reduce suicide risk (daily use, N=10, 16%; frequent use, N=40, 66%; and occasional use, N=11, 18%). Among the 61 participants, 28 (46%) reported that they used the plan to identify professional support, 24 (39%) to identify social support, 16 (26%) to identify internal coping strategies, and 13 (21%) to identify warning signs. Those who did not use the SPI (N=18, 18%) gave the following reasons: use of strategies not on the plan, N=5, 28%; feeling too distressed, N=4, 22%; and thinking it would be unhelpful, N=2, 11%. Several participants (N=20, 20%) reported updating their safety plan, either on their own or in collaboration with a treatment provider, including updating social contacts (N=9, 45%), internal coping strategies (N=5, 25%), and warning signs (N=1, 5%). [A table presenting quotes of participants about the acceptability and utility of the intervention is included in an online supplement to this report.]

With respect to acceptability of the SFU, most participants reported being satisfied (N=96, 96%); four (4%) reported that they were neutral. Most reported that the SFU was helpful (Table 1). The most helpful components were having someone check in regularly (75%), feeling cared for (58%), and reinforcing use of the SPI (12%). Most participants (N=69, 69%) reported finding the calls acceptable. Others identified logistic (N=28, 28%) or mood-related barriers (N=3, 3%). In terms of utility of the SFU, all participants recalled the phone contact, and almost all (N=98, 98%) identified at least one topic discussed. Most participants (N=83, 83%) recalled the mood check, whereas fewer recalled the safety assessment (N=44, 44%), SPI review (N=33, 33%), or discussion of upcoming mental health appointments (N=21, 21%).

Most participants found the SAFE VET intervention (SPI and SFU combined) useful in mitigating suicide risk (very useful, N=55, 55%; fairly useful, N=38, 38%; and neutral, N=7, 7%) and in increasing the likelihood of attending follow-up mental health appointments (very helpful, N=33, 33%; and fairly helpful, N=44, 44%). The overwhelming majority (N=99, 99%) found the SAFE VET intervention very acceptable, reporting that they would recommend it to a friend in crisis.

Discussion and Conclusions

This study assessed the acceptability and perceived effectiveness a brief two-part intervention that combined safety planning in the ED and structured follow-up by telephone after discharge for suicidal individuals who did not require hospitalization. This intervention was designed to fill a crucial gap in the current system of care. Typically, when patients are seen in the ED who in a suicide crisis, with escalating ideation or recent suicidal behavior, they are either hospitalized or discharged with a referral to outpatient care. The ED is often seen as a site where it is not feasible to implement any sort of intervention for this population. This is problematic because for the months after a suicide crisis, particularly a suicide attempt, individuals remain at increased risk (7), and the failure rate of the referral process after ED discharge is high (5,6). Therefore, suicidal individuals can have minimal or no care at the time of greatest need for help.

The intervention we evaluated is an effort to provide a basic level of care for this patient population. As an initial step, we examined whether patients found the intervention acceptable, useful, and helpful. Overall, patients found the combined intervention to be both acceptable and helpful. Generally, completing the SPI was perceived as helpful in handling their suicidal crises, and the SFU was perceived as helpful because the phone contacts conveyed a sense of being cared about. Most patients found both components helpful, whereas a few voiced some dissatisfaction, such as that the SPI was too long or that the SFU did not facilitate treatment engagement. There were also some indications that the intervention was not a good fit for some patients, who suggested that it did not target their primary issue or that it was not provided at a convenient time. Also, few participants remembered discussing or reducing their access to lethal means as a result of the intervention. Given the effectiveness of lethal means restriction as a suicide prevention intervention (13), it is important to find ways to refine SPI-SFU so that it will foster greater patient consideration and implementation of lethal means restriction. Nonetheless, everyone completed both components of the intervention, suggesting that even if some areas could be improved, any shortcomings did not result in refusal to remain in the intervention.

Although the combined intervention was found to be both acceptable and helpful, it is important to note that a convenience sample was recruited for this study. Also, results from the ongoing controlled trial are necessary to confirm that this intervention is more effective than usual care. These positive results were found in two large VA medical centers located in the Northeast; however, it is not known whether the results would be the same in other VA medical centers or in EDs outside the VA system. Finally, the mental health professionals who provided the intervention were dedicated members of a suicide prevention team, and their flexible schedules allowed them to monitor veterans after they were discharged from the ED. Further research is required to determine whether ED staff would be able to implement the intervention as successfully given other job responsibilities.