A Qualitative Examination of Clinician Anxiety about Suicide Prevention and Its Impact on Clinical Practice

Abstract

Clinician distress about working with patients at risk for suicide is well documented in the literature, yet little work has examined its pervasiveness across clinical settings. We conducted a secondary analysis of qualitative data gathered from 26 clinicians in primary care and outpatient mental health clinics serving both adult and child clients on their perception of evidence-based practice use for suicide screening, assessment, and brief intervention. Qualitative data were coded for any mentions of clinician anxiety or emotional response, and brief quantitative measures were collected to characterize our sample. When discussing broader barriers to implementation, 85% of participants spontaneously mentioned anxiety or heightened emotional responses related to delivering suicide prevention practices to those at risk for suicide. Common themes included low self-efficacy in suicide prevention skills, distress related to escalating care, efforts to alleviate such distress, and difficulty related to tolerating the uncertainty inherent in suicide prevention work. Similarly, while standardized anxiety ratings for participants were consistent with those of non-clinical norming samples, clinicians reported mild to moderate anxiety when screening for suicide risk (M = 3.64, SD = 2.19, Range = 0–8) and engaging in safety planning (M = 4.1, SD = 2.88, Range = 1–7) on post-interview surveys. In contrast, survey responses reflected generally high self-efficacy in their ability to screen for suicide risk (M = 7.66, SD = 1.29, Range = 5.25-10) and engage in safety planning (M = 8.25, SD = 0.87, Range = 7-9.5). Findings highlight pervasiveness of clinician distress when implementing suicide prevention practices and can inform future suicide prevention implementation efforts.

Background

Suicide is a leading cause of death in the United States (Center for Disease Control, 2023). Although substantial advances have been made in the development of evidence based practices (EBPs) for suicide screening, assessment, and intervention (SSAI), these practices are often implemented inconsistently (Davis et al., 2023; Diamond et al., 2012; Gamarra et al., 2015), limiting their impact. Barriers to SSAI implementation span system- and workplace-level factors, such as time and ability to bill for these services (Diamond et al., 2012; Petrik et al., 2015), and individual factors like clinician self-efficacy, comfort, and emotional reactions (Davis et al., 2023; Diamond et al., 2012; Hendin et al., 2006; LeCloux et al., 2021; LoParo et al., 2019; Vannoy & Robins, 2011). Clinician heightened emotional response when working with individuals at risk for suicide is well-documented in the literature (e.g., Barzilay & Apter, 2022; Ellis et al., 2018; Quinnett, 2019; Yaseen et al., 2013). While clinician emotional response such as anxiety may be expected and even helpful in small doses, in excess it can interfere with a clinician’s ability to effectively engage in suicide prevention practices. The impact of heightened emotional response when working with patients at risk for suicide includes decreased therapeutic alliance (Barzilay et al., 2020; Michel & Jobes, 2011), impaired clinical judgement (Barzilay et al., 2022), reduced expression of empathy (Foster et al., 2016; Yao et al., 2024), and decreased patient willingness to disclose suicidal thoughts (Blanchard & Farber, 2020; Love & Morgan, 2021), which may contribute to worse patient outcomes (Barzilay et al., 2022; Yaseen et al., 2013). For this reason, recent work focuses on novel strategies to support clinicians in modulating their emotional reactions in response to working with patients at risk for suicide (Barzilay et al., 2022; Becker-Haimes et al., 2022, 2023; Foster et al., 2016; Yao et al., 2024).

To date, research has largely examined how emotional reactions may lead clinicians to avoid screening for or intervening on suicide risk or avoid accepting patients with a history of suicidality to their clinical practices altogether (Diamond et al., 2012; Hendin et al., 2006; Quinnett, 2019; Vannoy & Robins, 2011). Additional work is needed to understand the nuanced ways in which clinicians experience and attempt to alleviate emotional distress when implementing evidence-based practice recommendations for suicide prevention (e.g., use of standardized suicide screeners, delivery of the Safety Planning Intervention (Stanley & Brown, 2012) when indicated, escalating to higher levels of care, etc.). Such work is essential to informing efforts aimed at supporting clinicians to optimally regulate their emotions to support their suicide prevention work. This is particularly relevant in light of increased focus on system-wide suicide prevention efforts in recent years (e.g. Zero Suicide; Labouliere et al., 2018). Existing work also has typically focused on a single setting (e.g., outpatient mental health clinics, emergency rooms, primary care) or patient population (e.g., adolescent, adult). Additional work is needed to characterize the pervasiveness of clinician distress across settings and clinical populations. Further, given the range of ways that heightened emotional response have been shown to impact clinician behavior (e.g., decreased expression of empathy, impaired clinical decision making), we build on initial work to identify ways that clinicians respond to distress in their delivery of SSAIs to suicidal clients. This work is critical to informing strategies to optimally support clinicians to interact with at-risk individuals competently and confidently.

In this Brief Report, we describe results of a secondary qualitative analysis of interviews with clinicians in outpatient mental health and primary care settings that required or encouraged use of gold-standard SSAIs at the time of data collection. Original analyses, described elsewhere, asked about perceptions of barriers and facilitators to implementing these practices (Davis et al., 2023). Our secondary analysis reports on specific aspects of working with patients at risk for suicide that evoke emotional reactions for clinicians and describes clinician efforts to alleviate this distress in a sample of clinicians working in a range of settings and patient populations. We also report on quantitative ratings of clinician anxiety, self-efficacy, and anxiety sensitivity that were collected as part of the initial study to characterize our sample’s overall level of anxiety (i.e., trait and state) and self-efficacy related to suicide prevention practices. We report on these measures as a clinician’s own anxiety sensitivity has been found to be related to lower use of recommended practices (e.g., Becker-Haimes et al., 2022; Meyer et al., 2014; Mulkens et al., 2018).

Method

Setting

We recruited clinicians from a convenience sample of four primary care clinics that offered behavioral health services (two pediatric, two adult) and two community mental health mental health programs (one adult, one child). Each setting either required or recommended use of SSAIs including use of the Columbia Suicide Severity Rating Scale (Posner et al., 2011) and the Safety Planning Intervention (Stanley & Brown, 2012). More specific details about settings can be found in Davis and colleagues (2023).

Participants

Leadership at each clinic provided names of those involved in suicide prevention work at their clinic to the study team; study team members then reached out to those identified and invited them to participate in a one-time study visit that included a qualitative interview and brief set of questionnaires. Study participants (N = 26) included practice leaders (n = 4; physicians), primary care clinicians (n = 8; including physicians, nurse practitioners), behavioral health clinicians in primary care (n = 6; including social workers, psychologists, and case managers), and behavioral health clinicians in outpatient mental health clinics (n = 8; including social workers and psychologists). All participants reported providing direct clinical services as part of their role and are hereafter referred to as clinicians. Participants averaged 38.96 years of age (SD = 8.72) and identified as 77% female, 65% White, and 15% Hispanic/Latinx. On average, participants reported 10.84 years of clinical experience (SD = 7.25).

Measures

Qualitative Interviews

The qualitative interview asked about general barriers and facilitators to SSAI implementation. General themes related to barriers and facilitators that emerged from these interviews were the focus of the primary paper (Davis et al., 2023). Themes that emerged related to self-efficacy and anxiety are reported for the first time here. At the conclusion of the qualitative interview, clinicians also were asked to characterize state anxiety related to suicide screening, assessment, and intervention on an 11-point Likert scale (0 = none to 10 = extreme anxiety). Clinicians also rated self-efficacy about delivering SSAI (0 = absolutely no confidence to 10 = complete confidence). Oftentimes, clinicians elaborated on their Likert ratings during the interview; these responses were also transcribed and coded in analysis. The full interview guide is provided in Appendix 1.

Anxiety Sensitivity Index-3 (ASI-3;(Taylor et al., 2007; Wheaton et al., 2012))

To capture trait anxiety and benchmark participant anxiety to the general population, participants completed the 18-item psychometrically validated Anxiety Sensitivity Index-3 (ASI-3) to characterize their fear of sensations associated with anxiety and their consequences. Ratings were on a 5-point scale (0 = very little to 4 = very much). Internal consistency for the ASI-3 was acceptable (α = 0.74). Mean ASI-3 scores for this sample (M = 11.91, SD = 6.23, Range = 2–26) were comparable to those seen in non-clinical norming samples (Osman et al., 2010) and below clinical cutoffs observed in the literature (Allan et al., 2014; Osuji et al., 2022), suggesting a sample that was overall comparable in level of anxiety to the general population.

Demographic Information

Participants completed brief demographic questionnaires to capture age, gender, race, ethnicity, job title, and experience working with patients at risk for suicide.

Analysis Plan

We quantified Likert scale ratings related to state anxiety and self-efficacy to characterize our sample. During our primary analysis, we observed frequent spontaneous mentions of clinician anxiety and other heightened emotional responses that clinicians described would occur for them when delivering SSAIs. To further explore this theme, we conducted a secondary round of qualitative coding to capture all specific mentions of clinician heightened emotional responses when engaged in suicide prevention care and any descriptions of how they responded to this distress. Consistent with an integrated approach (Bradley et al., 2007), we used both deductive methods to identify content related to heightened emotional reaction and ground-up inductive code development to develop subthemes. We first had three coders (JJ, AF, CB) review three transcripts to refine a definition for reliable coding of what constituted clinician mention of heightened clinician emotional response or response to distress that was recognizable across coders. Each subsequent transcript was double coded by two trained members of our research team (AF, CB). Transcripts were coded independently in sets of three, and both coders and the first author met to review all codes, discuss any areas of disagreement, and reach a consensus. Once all mentions of emotional response were coded, we employed an inductive approach with two members of the research team (JJ, MB) to independently summarize data into subthemes. Broad categories of subthemes generated based on the data were: (1) low self-efficacy in suicide prevention skills, (2) distress related to escalating care, (3) discomfort tolerating uncertainty in suicide prevention work, and (4) clinician strategies to alleviate distress related to suicide prevention work.

Results

Clinician Self-Reported Anxiety

Self-reported state anxiety about SSAI use, assessed via the 11-point Likert scale, indicated a wide range in subjective anxiety, with average scores indicating mild to moderate anxiety for suicide screening (M = 3.64, SD = 2.19, Range = 0–8) and engaging in safety planning (M = 4.1, SD = 2.88, Range = 1–7). On average, participants reported high self-efficacy in their ability to screen for suicide risk (M = 7.66, SD = 1.29, Range = 5.25-10) and engage in safety planning (M = 8.25, SD = 0.87, Range = 7-9.5). Mean ASI-3 scores for this sample (M = 11.91, SD = 6.23, Range = 2–26) were comparable to those seen in non-clinical norming samples (Osman et al., 2010) and below clinical cutoffs observed in the literature (Allan et al., 2014; Osuji et al., 2022).

Qualitative Themes

During qualitative interviews, 85% (n = 22) of participants spontaneously mentioned anxiety or heightened emotional responses related to working with patients at risk for suicide and using SSAI before being prompted to share their ratings of distress on the Likert scales. Descriptions of each identified subtheme along with illustrative quotes are described below.

Low Self-Efficacy in Suicide Prevention Skills

In contrast to self-reported quantitative scores indicating overall high self-efficacy in screening and engaging in safety planning, participants qualitatively shared about experiences of low self-efficacy to engage in SSAI use. Consistent with theoretical models that hypothesize that anxiety is driven by low self-efficacy (Becker-Haimes et al., 2022), participants shared lack of self-efficacy in their own work (e.g., if their assessment was “thorough and accurate”), decision-making during screening (e.g., they did “not know what to take seriously”), or decision-making during assessing (“what to do with that data”), and tied this to experiences of anxiety. Some clinicians without specialized mental health credentials (e.g., medical doctors, nurse practitioners) described feeling ill equipped to engage in SSAIs because “I’m not as good at [safety planning].” Other participants used more subtle language that illustrated relief when patients denied suicide risk or felt like they were “lucky” when they did not have to work with patients at high-risk for suicide and were able to avoid SSAI use. As one participant shared:

My anxiety is really raised, to be honest with you, when they start endorsing things. I’m already a little nervous going into it because I know it’s like- “Okay, there’s suicidality present. Let’s just see to what degree it’s there”. When they start endorsing things, I’m like- “Oh, sh*t” I’m very anxious. Then, it’s like- “Oh my god, someone has to help me. I can’t do this by myself.”

Distress Related to Escalating Care

Participants identified personal experiences of distress related to navigating processes related to escalating care (e.g., involuntary commitment, transportation to higher level of care). As one clinician shared, “[…] getting that family to the right avenue of care is always a little anxiety-provoking. Not necessarily in my skills to ask the question, but to make sure that we get them to the safest place possible.” This was noted as a particular source of anxiety within the context of limited time during the session and managing multiple patient appointments. As exemplified here:

It’s really nerve-wracking—it’s scary to do that. I just think the more exposure on an ongoing basis—the more we do this, the easier it will be. Hopefully we don’t have to [involuntarily hospitalize] that many kids. Again, because it doesn’t happen that often— because we’re not doing it, it just makes it—I think the avoidance is pretty high and people are pretty anxious about it.

Tolerating Uncertainty about Suicide Risk

Another common theme among participants was the personal emotional reactions or anxiety associated with the uncertainty of what happens once a client at risk for suicide leaves the clinic. For example, one participant shared that despite their high self-efficacy in their ability to safety plan, they worried about what happened after the client leaves and if they would follow their safety plan: “But the anxiety comes in when I’m like, ‘Are they actually going to follow this [the safety plan]?’”.

In addition to worry about following the safety plan, participants endorsed worrying about their ability to connect with the family (in the case of child or youth clients) and to remember to follow up. One participant described the process of managing follow up with patients and the emotional weight of it: “And, maybe it’s only ten kids who are suicidal at once in the practice, maybe it feels like more because it’s such a huge emotional burden.” Others described distress and physiological sensations associated with anxiety (e.g., experiencing stomachaches) following allowing a patient identified at risk of suicide to leave the building. As shared by one participant:

I don’t know, you have to just give it up about your liability– you can’t control it. You can’t control other people and I think that’s something that’s threaded throughout our practice in family medicine, but particularly is one of the lenses here when we’re talking about suicidality that’s hard is we want to control other people. And, we want them to follow our recommendations on their anti-hypertensives, we want them to follow our recommendations on connecting with [behavioral health services]. Why haven’t they called? Why has the case been closed? This is frustrating to me as a provider. Ugh, I can’t trust them, they need to get [involuntarily committed].

Alleviating Distress Related to Suicide Prevention Work

Participants reported efforts to alleviate distress associated with suicide prevention work. Most commonly, they reported seeking consultation when anxious, emotionally distressed, or doubting their competency in the context of providing suicide prevention care. Respondents also noted urges to avoid thinking about or discussing suicidality to alleviate distress. For example, one participant described an inclination to use shut-down language (“that makes me want to say, ‘you’re not really serious, you’re not really going to hurt yourself, are you?’”) in response to suicidal disclosures. Avoidance inclinations may be strongest when consultation is not available, as illustrated by this participant: “I think if you’re here at 7:00pm and you don’t know that anybody else is around, people are going to be less likely to [ask about suicide risk] because they’re not really sure what to do after that.” Although participants did not report acting on avoidance urges themselves, they did note observing colleagues’ avoidance behaviors.

Discussion

This Brief Report adds to the growing body of literature documenting clinician experiences of heightened emotional responses when working with patients at risk for suicide (e.g., Barzilay & Apter, 2022; Ellis et al., 2018; Larkin et al., 2024; Quinnett, 2019; Yaseen et al., 2013). We replicate prior research highlighting the importance of understanding and attending to clinician distress when engaged in suicide prevention work. We extend current research by demonstrating the pervasiveness of this anxiety across settings and in a broad age range of populations served, even in a sample that had relatively low trait anxiety and report high self-efficacy in delivering SSAIs. On average, participants reported mild to moderate state anxiety scores, with a proportion of clinicians reporting more significant anxiety on quantitative measures. Though participants reported overall high self-efficacy to engage in suicide prevention practices and did not generally endorse avoiding asking about suicidality themselves (although it was alluded to by several participants), they commonly described internal experiences of discomfort and more subtle forms of avoidance and shared in qualitative interviews about lack of self-efficacy in their own suicide prevention work. These seemingly contradictory findings suggest that self-report of relatively low levels of anxiety and high self-efficacy on quantitative measures may not adequately capture the full experience of clinicians and associated emotional reactions associated with engaging in this work. Our qualitative work highlights the frequency of spontaneous mentions of anxiety when clinicians are asked to reflect on SSAI implementation, common reasons for anxiety among clinicians in these settings, and how clinicians respond to this distress. This is important as suicidal patients are able to perceive clinician emotional responses, which can lead to decreased therapeutic alliance (Barzilay et al., 2020; Michel & Jobes, 2011), and reduce client willingness to disclose suicidality again in the future (Blanchard & Farber, 2020; Love & Morgan, 2021). This, in turn, limits opportunities for clinicians to intervene and optimally support patients.

Taken together, our results underscore that clinicians across a range of contexts and working with a range of clinical populations experience emotional distress when working with patients at risk for suicide. Clinicians in this study shared reasons for this distress related to lack of confidence in their ability to effectively intervene, concerns about how to navigate escalation pathways, and the difficulty of the uncertainty inherent to this work. Importantly, this heightened emotional distress was seen among a group of clinicians who overall report high self-efficacy to engage in suicide prevention practices. Our results suggest that efforts to train and support clinicians will continue to benefit from directly acknowledging the expected and understandable emotional reactions inherent to suicide prevention work and helping clinicians develop skills to manage this distress. As one clinician in another context shared with us: “Although we have the information, we know what we know, we are still just people, we are human beings, and we are just showing up in the world as people.” These efforts may be particularly warranted given the increase in clinician burnout seen in recent years (Linzer et al., 2022). However, how to best support clinicians to do this remains an empirical question.

Ongoing approaches currently being tested aim to increase clinician tolerance of uncertainty related to suicide prevention work using principles from exposure therapy and learning theory (Becker-Haimes et al., 2023), and incorporate opportunities to practice effective empathic communication using virtual human interactions (Foster et al., 2016; Yao et al., 2024). We might also consider targeted strategies to prepare individuals for emotionally intense clinical work, such as building training opportunities to mirror the intensity of clinical interactions, designing simplified tools to support decision-making in emotionally charged states, and providing in the moment coaching and support (Beck et al., 2024). Our results are consistent with the growing body of literature suggesting that developing and testing implementation strategies that acknowledge and address the emotionally intense nature of this work is an important avenue for future research to equip the mental health workforce to optimally deploy gold-standard SSAI across settings. This study also points to specific sources of clinician distress that can be targeted to optimize SSAI delivery. In addition to tolerance of uncertainty through exposure, SSAI training could target self-efficacy through repeated practice of each SSAI component across degrees of risk severity. Training focused on self-monitoring of distress and associated behaviors, such as shut down language that discourages client disclosure (e.g., “You’re not really going to hurt yourself, are you?”) may also bolster SSAI delivery.

Study findings should be interpreted within the context of the study selection criteria and study limitations. Participants were nominated by leadership at their organizations based on their engagement with suicide prevention practices. Therefore, they may have been more familiar with suicide prevention work than clinicians on average, which was further supported by high self-efficacy ratings. Overall, trait anxiety sensitivity scores for participants were also in the low range and comparable to those reported in non-clinical norming samples. Study results may therefore not be representative of experiences across clinicians more broadly, particularly those who may be more likely to experience anxious distress. That said, anxiety was frequently and spontaneously shared during qualitative interviews. Anxiety specific to suicide prevention work may be higher in a sample with less familiarity and systems-level support than clinicians in this sample. Our overall sample size was also relatively small, with a heterogenous group of clinicians in terms of profession and training. This limited our ability to examine whether qualitative themes or quantitative ratings differed by clinician characteristics. Future studies should build on our growing understanding of clinician anxiety and its impact on delivery of suicide prevention practices by sampling clinicians from a range of settings, in different roles, and at different levels of familiarity with delivery of SSAI to inform the development of tailored support strategies. Studies might also examine whether experiences differ by types of provider, setting, and training to inform implementation strategies. Finally, we did not collect data about fidelity to SSAIs, which limits the ability to draw conclusions about how anxiety, self-efficacy, and adherence and competency with SSAI delivery interact. Future studies should also gather data on fidelity of SSAI use to determine its relationship to clinician anxiety and self-efficacy. Future studies should also build on existing work (e.g., Barzilay et al., 2022; Yaseen et al., 2013) to continue to examine the impact of clinician heightened emotional response on patient outcomes. Our results also suggest that in addition to clinician self-report tools like the Therapist Response Questionnaire-Suicide Form (TRQ-SF; Barzilay et al., 2018), future work should continue to develop methods of behavioral observation to examine the subtle ways that clinician anxiety might manifest in the clinical encounter that clinicians may not be aware of themselves and are therefore unlikely to be captured by self-report. This may include building on existing schemes like the Empathic Communication Coding System (Kleinsmith et al., 2015) to include a broader range of observable clinician behaviors such as shut down language, avoidance of asking questions, or leaving the room to consult with colleagues. Finally, while a strength of our findings is the fact that issues related to emotional distress were shared without being specifically queried, we did not ask targeted questions about how emotional experiences directly influenced SSAI implementation. We therefore may have missed reasons for anxiety or manifestations of anxiety in the current sample. We also did not ascertain clinician anxiety related to other aspects of their clinical work, which limited our ability to draw comparisons between anxiety about working with patients at risk for suicide as compared to other populations.

Overall, results highlight the importance of continued work to understand how clinician emotional response interacts with SSAI implementation. Our results demonstrate the importance of learning about the experiences of clinicians currently engaging in this work through mixed methods to best develop and tailor implementation strategies to support them to optimally deliver SSAIs.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s10597-024-01364-6.