Does it get easier over time? Psychologists’ experiences of working with suicidal patients

Abstract Does working with suicidal patients become easier with time? A representative national survey of 375 psychologists (72% women, Mean age 44 years) showed that years of experience (r = −.13, p = .01) and frequency of contact with suicidal patients (r = −.15, p = .004) correlated only weakly with perceived difficulty. Thematic analysis of respondents’ descriptions of difficult suicide-related situations on an open survey-question yielded four themes: Unreachable patients, choosing between therapy and security, therapist’s boundaries and empathy with death-wishes. We conclude that improved confidence in coping with suicidality may require specific training rather than simply years of work.

Knowing how to work with suicidal patients may be one of the most important therapist skills. A review of 40 studies in Europe, Australia, and the United States showed that three out of four people who have died by suicide had contact with primary care providers during the year before their suicide and that one in five had contact with mental health services during the month before their suicide (Luoma, Martin, & Pearson, 2002). Although researchers have identified a number of factors that may increase long-term suicide risk at a group level, both in children (Gvion & Apter, 2016), adults (Rihmer & Dome, 2016) and the elderly (De Leo & Arnautovska, 2016), assessing the short-term danger of death by suicide for an individual patient is a different task. A meta-analysis of the positive predictive value of several clinical risk scales found that such instruments were unable to predict suicide with a level of accuracy that makes them useful for deciding on the right treatment (Carter et al., 2017). Another recent review concluded that suicide prediction models' accuracy of predicting a future event is near zero (Belsher et al., 2019). Signs of suicidal risk are typically present as a chronic condition rather than only shortly before the suicide (Ramchand, Franklin, Thornton, Deland, & Rouse, 2017). Moreover, the time between consideration and making a suicide attempt is often as short as 5 min to 24 hr (Deisenhammer et al., 2009;Gvion & Apter, 2011), which means that a patients' state at the time of assessment might not accurately predict his or her state at another moment in time. The fact that suicide is an infrequent rather than frequent event makes it more difficult to foresee. Most people who end up taking their lives are deemed as being of low risk, whilst people at high risk commonly do not go on to take their life (Hatcher, 2016).
Hospitalizing a suicidal patient is not always the best solution in the long run (Michel, 2016). During consultations with a suicidal patient, therapists need to balance the need for hospitalization in order to provide the patient with more supervision, against the needs for patient autonomy and continued outpatient therapy (e.g., Goldblatt, 1994). Given these conflicting concerns, and the anxiety most therapists feel making this life-or-death decision, conducting a suicide evaluation is perhaps one of the most demanding situations professional care providers face.
How does the therapist's work situation influence the perceived difficulty of working with suicidality? Previous research indicates that fragmented health care may influence the difficulty of connecting meaningfully with suicidal patients (Hagen, Hjelmeland, & Knizek, 2017). In this study, we focus on the role of length of professional experience and frequency of contact with suicidal patients. If having more clinical experience reduces the perceived difficulties of working with suicidality, this could imply that health care institutions can expect seasoned therapists to feel more confident in working with suicidal patients without specific training. If long professional experience alone is insufficient, this raises the question of why this might be so, and how psychologists can be helped to become more confident. Although there is research examining the effects of long professional experience on reactions to patients' death by suicide (e.g., Grad, Zavasnik, & Groleger, 1997;Kleespies, Niles, Mori, & Deleppo, 2016), the present study is, to the best of our knowledge, the first to examine whether longer professional experience reduces the perceived difficulty of working with suicidal patients.

Method
Procedure A web-based survey was distributed to a representative sample of Norwegian psychologists, all having at least a 6-year master's level training in clinical psychology which gives a license to work as therapists. The survey contained questions about the frequency and perceived demand for a range of potentially challenging interpersonal situations. An open item invited respondents to describe a particularly demanding situation that they had experienced in their work. No identifying personal information was collected, so participants could answer anonymously. The study was approved by the Norwegian Data Protection Services NSD (54875).

Participants
A link to the survey was sent by email to 1994 psychologists representative of the Norwegian population of psychologists with regard to gender, age, and type of work. The 512 psychologists who responded to the invitation (25.7%) were fairly representative of the original sample and the Norwegian population of psychologists with regard to age (M ¼ 43.75 years, SD ¼ 12.2), gender (71.5% women) and place of work (70.2% worked in secondary mental health services and 57.3% worked with adults at least part-time). About threequarters of these responders (75.4%, n ¼ 386) completed all questions relevant to this article. Their mean number of weekly consultations was 14.11 (SD ¼ 7.36) and the mean of their years of experience was 14.5 years (SD ¼ 11 years, range 0-63 years).

Survey
The survey was piloted by asking a small group of psychologists to answer the items and comment on them. The survey featured questions about demographic variables (gender; age), work situation (e.g., number of patient consultations per week and place of work) and years of professional experience as a psychologist. Participants were then asked to rate a series of 15 potentially challenging interpersonal therapeutic situations on a 6-point Likert scale ranging from "not at all difficult" (0) to "extremely difficult" (5). The descriptions of difficult situations were mainly based on situations from the Facilitating Interpersonal Skills Scale (FIS; Anderson, Crowley, Himawan, Holmberg, & Uhlin, 2016), a few were based on suggestions from clinical psychologists who piloted the questionnaire. Respondents were also asked to rate the frequency with which these situations arose in their own practice. An open item asked respondents to describe a particularly difficult interpersonal situation they had encountered in their work.

Statistical analyses
Respondents that indicated that they never had contact with suicidal patients were removed and variables were tested for skew, kurtosis, and outliers before correlations were examined. None of the continuous variables had a skew or kurtosis greater than j1j and there were also no clear outliers.

Thematic analysis
Responses to the open-ended survey question were subjected to thematic analysis (Braun & Clarke, 2006). First, two authors familiarized themselves with the qualitative data and one identified the qualitative segments with content related to suicidality. Then, one generated initial codes to describe the situations that featured difficulties associated with patient suicidality. Then, two authors organized these codes into thematic categories. Then, all reviewed and discussed these themes and reached a consensus.

Results
Approximately half of the 386 respondents (45%, n ¼ 173) saw suicidal patients at times, often or very often, approximately half (52.3%, n ¼ 202) reported seeing suicidal patients rarely or very rarely, and 11 (2.8%) reported that they never saw suicidal patients. We removed these 11 from further analyses. Among the 375 respondents who saw suicidal patients, approximately half (48%) rated seeing suicidal patients as very difficult (4) or extremely difficult (5) on the six-point scale (M ¼ 3.37, SD ¼ 1.24). Only 2% rated seeing suicidal patients as not at all difficult (0). Table 1 shows correlations between background variables (gender, age and whether the respondent had reported that he or she worked in secondary mental health care), years of professional experience, and frequency and perceived difficulty of working with suicidal patients. Of the background variables, only gender was significantly related to perceived difficulty of working with suicidal patients: women (M ¼ 3.5, SD ¼ 1.1) found working with suicidal patients more difficult than men (M ¼ 2.99, SD ¼ 1.4) did (t ¼ 3.8 (373), p < .001). In addition, women (M ¼ 2.48, SD ¼ 1.1) reported that they saw suicidal patients more often than men (M ¼ 2.08, SD ¼ 1.0) (t ¼ 3.2 (209.5), p ¼ .002). Gender was also related to age and years of professional experience, with women in this sample tending to be younger (r ¼ .21, p < .001) and to have fewer years of professional experience (r ¼ .19, p < .001).
Working in secondary mental health care was positively correlated with the frequency of working with suicidal patients in this sample (r ¼ .35, p < 001, n ¼ 375), but was not associated with the perceived difficulty of this work (r ¼ À.05, p ¼ .36, n ¼ 375).
There was a significant association between how often respondents saw suicidal patients and how difficult they rated these situations. Those who saw suicidal patients more often found situations involving suicidality somewhat less challenging (r ¼ À.15, p ¼ .004, n ¼ 375). According to Cohen's (1988) criteria, this is a weak association. There was also a weak but significant association between years of professional work and the perceived difficulty of working with suicidal patients. Respondents with longer professional experience found suicidality less difficult to deal with (r ¼ À.13, p ¼ .01, n ¼ 375).
To investigate the possibility that experienced psychologists tended to be asked to see the more challenging patients, we examined whether years of professional experience were associated with seeing suicidal patients more often. No such association was found. In fact, there was a non-significant tendency for more experienced psychologists to encounter suicidal patients less often than their less experience colleagues (r ¼ À.09, p ¼ .08). We found the same pattern for age: older psychologists saw suicidal patients somewhat less often than younger psychologists did (r ¼ À.11, p ¼.03).
As gender was related to several of the variables we wanted to examine (years of professional experience, frequency of working with suicidal patients and perceived difficulty of working with suicidal patients) we ran a simultaneous multiple regression with gender, frequency and years of working with suicidal patient as independent variables and perceived difficulty of working with suicidal patients as the dependent variable. Both frequency (b ¼ À.19, p < .001) and years of professional experience (b ¼ À.11, p ¼ .03) contributed to variance in perceived difficulty of working with suicidal patients when the effect of gender was controlled for (Table 2). To summarize, more experienced psychologists and those who saw suicidal patients more Table 1. Correlations between background variables (gender, age, and working in secondary mental health care), years of professional experience, and frequency and perceived difficulty of working with suicidal patients (n ¼ 375). On the open survey item, respondents were asked to choose one or more difficult situations they had experienced in their work as a psychologist and to describe this or these situations. A wide range of difficult situations was described by the 253 psychologists who responded to this question, and 40 of these responded with descriptions of difficult situations that involved suicidality. The thematic analysis of these 40 responses yielded four main thematic categories, which we called: "Unreachable patients (19 responses), choosing between therapy and security (six responses), therapist's boundaries (five responses) and empathy with death-wishes (six responses)." The remaining responses were too nonspecific or short to be classified (e.g., "I've experienced several situations where patients have communicated suicidal intent, and those situations are always challenging"). We describe the four thematic categories below.

Unreachable patients
When asked to describe a particularly difficult situation several psychologists described a situation where they were physically, emotionally or mentally unable to reach and help a suicidal patient, sometimes because the patient telephoned from a distance or did not turn up for sessions. A female psychologist with 33 years of experience described the following situation: The patient called me from home and said he had done it now. I asked what he meant, and it turned out that he had taken tablets with the intention to die by suicide. Speech started to become slurred. Managed to ask where he was, confirmed that he was at home (which was in a small municipality nine miles away). The patient hung up. Had to think quickly, called the local doctor who fortunately drove off, got him hospitalized and pumped. Saved his life. Made me angry afterwards, had to process my own reactions in order to continue therapy.
Respondents also described not reaching the patient mentally or emotionally. These descriptions included situations where the psychologist was not able to induce any kind of hope in the patient, situations where the psychologist did not manage to make the patient understand that it was not possible to help him or her without cooperation and situations where patients refused to talk to or listen to the therapist. A male therapist with 2 years of professional experience described the following situation: Girl (17 years) with the suspicion of borderline personality disorder speaks about suicidal thoughts but refuses to talk about the concrete plan she has (method) because then I, or her parents, will stop her. She is referred to the emergency unit, and there she tells them that everything is fine.
Therapists described situations involving an incongruence between the therapist's expectations of the patient and the patient's willingness or ability to engage. For example: The patient wants help for suicidality and self-harm, but has difficulty engaging in the process. Doesn't "own" the problems or to reflect upon her own thoughts and feelings. Devalues and complains about the therapy-"this isn't helping me"-but does not contribute to the therapy and does not take responsibility for choosing what to talk about. Strong feeling of responsibility and fear in the therapist, and very little of the same in the patient.

Choosing between therapy and security
The second main theme involved situations where therapists faced a choice between a course of action that might be more therapeutic in the long run and opting for a less therapeutically motivated course of action that was designed to keep the patient alive, at least in the short run. Therapists described wanting to make decisions that might lead to progress in the therapy, yet being afraid that the patient might die by suicide if not treated with "special" but not necessarily therapeutic approaches such as providing extra sessions when not indicated, or providing forms of treatment that were not evidence-based. A male therapist with 35 years of experience described a situation where the patient said she would discontinue her life if the therapist did not recommend her for a treatment that the therapist seriously doubted would be helpful.
A patient told me that she would take her life if I would not recommend her for a specific treatment, a treatment I was very unsure about and doubted. I had explained to her that I could not make any recommendations regarding further treatment until the examinations had been completed so that I could be confident as to what might be helpful for her in the longer term. After the session ended and the patient was on her way out of the office, she told me that we would not meet for our next appointment, because by that time she would have discontinued her life since I would not do as she wished.
Other examples involved whether to hospitalize a patient who did not wish to be hospitalized and whether to not hospitalizing a patient who did wish to be hospitalized. In the latter instance, the therapist, a female psychologist with 10 years of professional experience, had to make a choice between "managing" the suicide risk by hospitalizing the patient and a path that might be more helpful in the long run, but which the patient rejected. She explained: I find particularly demanding those times when the patient wishes to be hospitalized and threatens to take his life, whilst I disagree, since I believe that [in that case] hospitalization is contraindicated and will make things worse. ( … ) to persist in my decision [not to hospitalize the patient] is very demanding, even though I know, professionally, that it will be best for the patient.
A male therapist with 16 years of experience also described situations where he felt hospitalization would not be the best choice, but where he felt pressured by patients who said they would take their lives if not hospitalized and relatives who wanted a patient hospitalized. The therapist admitted that he sometimes gave in to the pressure: So those times when I give in and hospitalize, I know I have been pressured into it and that I've given the patient a treatment that I don't believe will be helpful for longer than perhaps just that one day, not to mention that I have also given somebody else [the hospital personnel] the job of setting limits.

Therapist's boundaries
Some therapists described situations in which they felt scared, angry or intruded upon, as when a patient called the therapist at night during patient crises (without this being part of the therapeutic plan) or acted in other ways that scared the therapist. In these situations, the therapists' views of their own and the patient's roles did not agree. For example, a therapist described "a patient who overstep boundaries related to [my] private life and calls during suicidal crises and self-harm episodes." Empathizing with the wish to die A few therapists described the difficulty of understanding and respecting the patient's wish to die. A female therapist with 11 years of experience described how personally demanding this could be. She wrote: Empathically trying to understand that a person would prefer death to life is a great strain on the therapist, who usually feels exactly the opposite. It may take a long time for a therapist to accept that a choice [of death rather than life] has to be respected, even if this choice goes against everything reckoned as normal for a human being. Therapists who work with suicidal people should have colleagues that can help them stay in such terrain over time.
A female therapist with 40 years of experience described empathizing with the patient's suicidal wish because of the patient's very difficult life situation, whilst at the same time finding it impossible to compromise on the goal of preventing suicide: When empathy makes you realize that, considering the patient's life situation, suicide is logical. At the same time, you are supposed to think about your responsibility as a "life saver," and the documentation requirements from the authorities.
To summarize, a common feature of the descriptions of difficult situations was a lack of agreement between the therapists' and patients' goals for their interaction. Commonly therapists can rely on collusion between the therapists and patients in finding means and goals for therapy. In the situations described this common goal seemed to be lacking.

Discussion
The quantitative data showed that years of professional experience had a significant but weak relationship with the perceived difficulty of working with suicidal patients, such that more experienced therapists tended to feel slightly more confident in such situations. Frequency of contact with suicidal patients was weakly and negatively associated with the perceived difficulty of situations involving suicidality, indicating that such contact may slightly reduce such difficulty, as has also been found in other studies (Notø, Olsen, Normann, & Løberg, 2005). Long professional experience may provide therapists with exposure to situations involving potential suicidality, thereby enabling them to learn how to manage such situations and develop some confidence in their skills in working with suicidality. Longer professional experience may also expose therapists to suicidal patients who do not go on to take their own lives, thus increasing their confidence that they will usually be able to help patients who are contemplating suicide. In addition, the longer experience might help therapists to acknowledge and accept without undue self-reproach that it is not always within their power to prevent patients from taking their own lives.
The observed association between years of professional experience and perceived difficulty of working with suicidal patients was weak, however, accounting for only 1.6% of the variance in perceived difficulty. The weakness of this association is in agreement with studies that show that, although more experienced therapists may have lower rates of early termination, in general therapists do not become more skilled in their work as they gain experience (e.g., Goldberg et al., 2016). A search of the literature suggested that this may also apply to work with suicidal patients. We found no clear evidence that more experienced psychologists or psychiatrists were more proficient in working with suicidal patients than their less experienced colleagues. A study which rated the appropriateness of psychologists' responses to a vignette describing a potentially suicidal person showed, although advanced students and psychologists were more helpful in their responses than beginning graduate students, psychologists were no better than advanced psychology students in responding appropriately in spite of their more extensive formal training and experience (Richards & Range, 2001). This indicates that general training in psychology and professional experience might not be enough. Other possible reasons are that the respondents' professional skills had increased during the years but that they still perceived working with suicidal patients as difficult, or that the more experienced respondents had encountered situations that helped them understand the challenges of such work, disenchanting any illusions and clinical naivety.
We were surprised to find that in our sample, the female gender was positively associated with the frequency and perceived difficulty of seeing suicidal patients. The female gender was also negatively associated with age and clinical experience. One might speculate that this reflects a demographic change over time in the population of psychologists in Norway which used to be dominated by male students (87% were male in 1970) but is now dominated by female students (84% were female in 2016). This means that our older respondents were slightly more likely to be male. Other potential explanations are a female response bias or actual gender differences. Gender differences in the impact of patient death by suicide on the therapist were found in a study by Grad et al. (1997), as women reported more guilt, shame, and professional self-doubt after a patient's suicide than men. Further studies are needed before we can conclude that there are gender differences in therapists' perceptions of the difficulty of working with suicidal patients.
The descriptions of difficult situations involving suicidality in the open part of this survey showed that a lack of established common means and goals of therapy was characteristic of many of these situations (e.g., when the patient could not be reached, when the patient did not talk, when there was disagreement on whether hospitalization was indicated, and when continued life did not seem a better choice than suicide). The therapists found themselves in situations where they were not seen as being helpful, sometimes actually seen by the patient as working in opposition to the patients' wishes. Our therapists also described how emotionally taxing it may be to empathize with patients' suicidal wishes. Some of these difficulties have also been noted in prior research on therapists' perceived challenges (Hagen et al., 2017).
With all patients, including those that are suicidal, it is recommended that therapist should aim to develop a strong working alliance (Jobes, 2016;Michel, 2016). The working alliance is associated with a reduction in subsequent suicidal thoughts and, to a lesser degree, attempts (Dunster-Page, Haddock, Wainwright, & Berry, 2017) However, as described in the present study, some patients may be unable to share their needs and goals with the therapist, sometimes therapists and patients struggle to agree on treatment goals, and some patients may experience desperation when the therapist's suggestions for treatment-approaches do not agree with their own. Also, a secure bond might not be available, as some suicidal patients may strive to end relationships with others, including the therapist, while therapists attempt to build a relationship with the patient (Aherne, Coughlan, & Surgenor, 2018).
Based on the findings of this exploratory study, we suggest that merely working as a therapist for several years might help somewhat but is probably not enough to reliably increase therapists' confidence in working with suicidal patients. Based on the qualitative data, we suggest that this might be partly because working with suicidal patients poses a challenge to the way therapists commonly are trained and work. Psychologists are mostly trained to work in cooperation with patients who have as an overarching goal to increase their own wellbeing. Therapists are to a lesser degree trained to work with adult patients whose behavior and choices conflict with the therapist's goals and where there are legal and other reasons why the patient cannot be allowed to direct what the therapist does. In those cases, establishing a working alliance may be difficult, if not impossible. Knowing what to think, do or say as a therapist when there is a weak or non-existent working alliance with a suicidal patient might require more than just years of clinical experience.
Due to the anxiety-provoking aspects of this work and the feelings of helplessness when one is unable to "reach" a patient as described by respondents in this study, there is probably a need not only for training that imparts knowledge and changes attitudes but for training that offers opportunities to practice skills in real-world scenarios. Although therapist anxiety in such situations is expected and therapist trait anxiety may even help therapists to respond wisely (Brown & Range, 2005), practical training might enable therapists to cultivate ways to use and soothe their emotional arousal (anxiety, anger or other feelings) and think clearly, be empathic and act in the patient's best interest, even if the patient is rejecting and uncooperative (Cramer et al., 2013;Kene, Yee, & Gimmestad, 2018). Deliberate practice has been incorporated into a training program designed to increase psychiatrists' ability to assess the risk of suicide in psychiatric patients (Silverman & Berman, 2014). As well as teaching practitioners how to respond to difficult situations on a cognitive level, the deliberate practice may also represent a form of repeated emotional exposure (in the context of a caring training relationship) that might reduce the arousal experienced during an encounter with a suicidal patient to a functional level. Hatcher (2016, p. 409) suggested that: "there should be a space where clinicians can practice assessments in different circumstances such as interviewing the patient who insists on leaving or who does not talk" and likened practicing how to proceed in difficult situations such as these to flight simulation training to enable pilots to cope with difficult circumstances, such as landing in bad weather. Future research may help us to understand how the therapist's anxiety or anger interacts with the patient's distress when the patient is suicidal (and in other therapeutic situations), and how psychotherapists can best be trained to tolerate the stress of evaluating and working with difficult-to-reach suicidal patients and feel a sense of mastery in such cases.
An intervention called "Collaborative Assessment and Management of Suicidality" (CAMS; Jobes, 2016) offers a specific solution to the challenge that suicidality poses to the working alliance. The philosophy behind the approach is to avoid discussion with the patient as to whether suicide is an option for the patient since it always is. It is maintained that since clinicians have limited control over the patient, the option of dying by suicide is always there, even after efforts to avoid suicide through hospitalization. The clinician's goal is to cooperate with the patient in identifying and treating the problems that compelled the patient to consider suicide, whilst also discussing with the patient how the risk of suicide can be lessened, for example by removing easy access to lethal means and taking specific steps such as calling a hotline or emergency department should suicidal urges arise (Jobes, 2016).
Further studies are needed to explore in detail which specific challenges that working with suicidal patients may represent, how therapists and suicidal patients may manage to resolve different types of challenges (e.g., Østlie, St€ anicke, & Haavind, 2018), and which training programs are effective in increasing therapist confidence as well as skills in this area.

Limitations of the study
There are several methodological limitations to this study. The majority of the psychologists who were invited to take part in the study failed to respond, so generalizing to other Norwegian psychologists (and even more psychologists from other cultures) must be done with caution. Although years of clinical work predicted somewhat less perceived difficulty of working with suicidal patients, the direction of the relationship is not clear since the study was correlational. Some therapists may choose work that involves less frequent contact with such patients because they experience this work as difficult. Others may feel so uncomfortable with the emotional challenges of working with suicidality that they oversee obvious warning signals, and report that they see suicidal patients less often.
The association between the frequency of contact with suicidal patients and the lesser perceived difficulty of working with them might also be due to unmeasured variables such as conditions in the working environment. In some work environments, there may be more resources available, for example, easy access to training and consultation, better hospitalization routines and greater access to alternative interventions such as home visits by trained nurses.
Other limitations are that the survey format did not allow for an in-depth exploration of informants' subjective experiences of working with suicidal patients and that this study does not provide information on therapists' actual skills or the effectiveness of therapists' work. Also, all responses described situations involving adult patients.

Conclusion
The findings from a mixed-method survey that attracted 375 quantitative and 40 qualitative responses indicated that working with suicidal patients is not much easier for experienced psychologists, as compared with less experienced colleagues. The qualitative responses suggested that working with suicidal patients may require ways of working that differ from how psychologists are commonly trained to work. Psychologists are trained to respect the patient's need for autonomy but to guide toward "therapeutic" solutions. However, in the current descriptions, patients, perhaps out of desperation, did not consistently show the therapy-congruent behaviors which therapists are trained to expect, such as showing up, speaking, trusting and respecting the boundaries of the therapist's role.
We agree with those clinicians and researchers that have argued that working collaboratively with suicidal patients is essential (e.g., Jobes, 2016) Most psychologists are trained to respect and try to increase the patients' autonomy and establish cooperation rather than fighting with or overruling the patient. However, psychologists may at times need to cope with crises where the patient is disillusioned, desperate or angry and there is no possibility of developing a workable alliance. In such situations, psychologists may need to find unusual solutions, sometimes accepting that no ideal, familiar, collaborative and "therapeutic" solution exists at that moment. In other words: they may need to disengage from their usual manners of working and find new ways to cope.
Working with suicidal patients remains challenging even for experienced psychologists. Improved confidence in coping with suicidality may require specific training, rather than simply years of work. Further research could explore the use of deliberate practice using concrete situations such as those described in the present study. In addition, to practice ways of establishing a working alliance with desperate patients, psychologists might need to practice how to cope when not being able to form a working alliance, while still passionately wanting to help.

Disclosure statement
The authors declare that they have no conflict of interest.