Relationship Between Clinicians' Inclination Toward Patients at Risk for Suicide (PRS) and Self-reported Countertransference.

We investigate the relationship between clinicians' inclination toward treating patients at risk for suicide (PRS), and self-reported countertransference (CT). We consider these observed group differences to explore two competing interpretations for observed CT patterns from a primary study; whether CT patterns are more consistent with defensive attitudes or an adaptative CT montage. We used one-way ANOVA, Tuckey post-hoc, and t-test, to compare clinicians (n = 267) grouped by self-ratings of positive, neutral or non-positive inclination toward working with PRS, with regard to their level of endorsement of the Therapist Response Questionnaire (TRQ) with PRS. We hypothesized that positively inclined clinicians would demonstrate greater CT literacy skills than other clinicians, reflected in lower endorsement of negative/hindering CT and higher endorsement of positive/facilitating CT to PRS. Compared to non-positively inclined clinicians, positively inclined clinicians endorsed significantly lower levels of two potentially negative/hindering CT dimensions, factor 1: entrapped/rejecting and, factor 5: protective/overinvolvement, and higher levels of the only positive/facilitating CT dimension, factor 2: fulfilled/engaging. Neutral clinicians reported similar CT patterns to positively inclined clinicians. Hypothesis of greater CT literacy from positively inclined clinicians appears supported. Observed differences in CT endorsement by inclination group tend to support the CT montage interpretation of our original findings more than the defense mechanism interpretation proposed. Similarities in CT patterns between positively inclined and neutral clinicians suggest that positive inclination to PRS, as assessed in this study, may not be countertransferential per se.

Keywords: Clinical suicidology; countertransference; countertransference montage; psychotherapy process; positively inclined clinicians

Patients at risk for suicide (PRS) tend to elicit intense emotional responses in clinicians that can lead the therapeutic relationship into counter-therapeutic impasses (Michaud, Greenway, Corbeil, Bourquin, & Richard-Devantoy, [12]). Identifying and managing these emotional responses to patients, which we refer here as countertransference (CT, in line with the totalistic conceptualization of the term, (Kernberg, [7]), is therefore essential to treating PRS (Barzilay et al., [2]). Yet, despite the critical role of CT management in clinical suicidology, it is only in recent years that an empirical literature has emerged (Vespa, Galynker, & Chistopolskaya, [18]).

Most of the empirical research provides evidence of the emotional hardship that can accompany treating PRS for clinicians, including feelings of anxiety, depression and anger (Macleod & Hovestadt, [9]), fear of blame and of causing harm (Awenat et al., [1]), tension, fatigue, and lack of self-confidence (Michaud et al., [13]). Given the consistency of these findings, Yaseen and colleagues considered the potential of measured CT to predict suicidal behaviors in highly suicidal patients. They found a paradoxical CT combination of "distress/avoidance" and "hopefulness" to discriminate between suicide attempters and non-attempters in the short term (Yaseen, Galynker, Cohen, & Briggs, [21]; Yaseen et al., [20]). This finding was included as a subscale of a novel multimodal assessment of suicide risk (MARIS) (Hawes, Yaseen, Briggs, & Galynker, [5]), and developed into a short structured psychometric measure, the Therapist Response Questionnaire – Suicide Form (TRQ-SF) (Barzilay et al., [3])

Within the same timeframe, our team investigated the nature of CT in response to perceived suicidality in a naturalistic sample of patients treated in varied clinical settings, using also the Therapist Response Questionnaire (TRQ) (Tanzilli, Colli, Del Corno, & Lingiardi, [17]). Consistent with Yaseen et al.'s findings, a combination of similar contradictory feelings (factor 1: entrapped/rejecting, and factor 2: fulfilled/engaging) dominated our novel seven-factor structure of the TRQ with PRS (Soulié, Bell, Jenkin, Sim, & Collings, [15]). Unexpectedly though, factor 1—entrapped/rejecting, representing most variance in aggregate, was only mildly endorsed by clinicians explicitly; while factor 2—fulfilled/engaging, by far less representative in aggregate, was the most endorsed by clinicians explicitly. In our primary study, we ventured two alternative interpretations for these findings, contending that they could represent either Defense mechanisms or an adaptative CT montage (Soulié, Bell, et al., [15]).

Our follow up qualitative study attempted to shed new light on the issue of CT to PRS by interviewing 12 positively inclined clinicians, who represented the minority of clinicians (14.6%) in declaring "liking" working with PRS (n = 267) (Soulié, Bell, et al., [15]). Using a grounded theory analysis we developed a model proposing that positively inclined clinicians synchronize emotionally with PRS before regulating the shared emotion, resulting in distress reduction for PRS and deep satisfaction for clinicians themselves, thereby reinforcing their positive inclination to working with PRS (Soulié, Bell, et al., [15]). Given the general reluctance amongst many clinicians to treat PRS (Levi-Belz, Barzilay, Levy, & David, [8]), these findings invited consideration that positively inclined clinicians might be distinguished by their ability to connect emotionally with PRS. Enabled through greater CT literacy skills (i.e., CT awareness and management), this capacity for connection was likely associated with the ability to bring about positive therapeutic change (Soulié, Bell, et al., [15]). However, this possibility has not been examined empirically.

We conducted the present ancillary study to investigate differences in CT endorsement between clinicians grouped by inclination (positive, neutral, or non-positive) toward treating PRS. Drawing from the aforementioned qualitative model (Soulié, Bell, et al., [15]), we hypothesized that positively inclined clinicians would be more CT literate than other clinicians, that is more aware of and able to manage their CT responses, which we anticipated to translate into lower endorsement of potentially negative/hindering CT responses, and higher endorsement of positive/facilitating CT responses.

We conducted the analysis on the primary study dataset (Soulié, Bell, et al., [15]). In summary, 267 fully licensed psychiatrists, psychologists and psychotherapists were recruited through professional associations and five publicly funded mental health services (District Health Board) in New Zealand (NZ). Participants rated the TRQ (Tanzilli et al., [17]) online, with reference to the most recent suicidal patient they had met at least three times in the previous six months. For the purpose of the study, we defined a suicidal patient as "a person who shows or has shown suicidal behaviours (including suicidal ideation); or who has attempted suicide before AND who seems [to the clinician] to be at risk of suicide". This resulted in the previously reported seven-factor structure of the TRQ with PRS (see Table 1).

TABLE 1. Description of the seven factors of the TRQ with PRS with reference to (Soulié, Bell, et al., [15]).

1 Note. TVE = total variance explained; α = internal consistency measure Cronbach's Alpha.

Additional information included the assessment of positive inclination toward PRS through rating of the following like-statement "overall you would say that you liked working with PRS" on a five-point Likert scale. We calculated factor mean scores by the three types of inclination: positively inclined clinicians (n = 39), corresponding to the "true" range of answer group (i.e., those who rated the like-statement either true or very true), non-positively inclined clinicians (n = 119), corresponding to the "not true" range of answer group (i.e., who rated the like-statement either not true or not true at all), and neutral clinicians (n = 107), corresponding to the "neither true nor false" answer group (i.e., who rated the statement somewhat true).

We used one-way ANOVA and Tuckey post-hoc test to examine differences in factor mean scores between these three groups with regard to each of the seven factors of the TRQ with PRS. Subsequently, to get a more distinct picture of the data, we removed neutral clinicians from the analysis and conducted t-test and Wilcoxon test (non-parametric) to compare positively and non-positively inclined clinicians.

One-way ANOVA indicated a significant difference in mean scores between the three inclination groups only in relation to factor 2—fulfilled/engaging (F(2, 262) = 7.277, p = 0.001) (see Table 2). Non-positively inclined clinicians scored significantly higher (M = 3.06, SD = 0.54), than positively inclined (M = 2.77, SD = 0.54, p = 0.009) and neutral clinicians (M = 2.83, SD = 0.51, p = 0.004). This means that clinicians who rated the like-statement as not true, endorsed less fulfilled/engaging CT responses to PRS than clinicians who rated it either true or neither true nor false. Moreover, the t-test showed no significant difference in mean scores between positively inclined and neutral clinicians on any dimension other than factor 4—informal/boundary crossing (t(144) = 2.160, p = 0.03), with positively inclined clinicians endorsing significantly higher scores of factor 4 (M = 4.43, SD = 0.48) than neutral clinicians (M = 4.24, SD = 0.46). This means that as a group, positively inclined clinicians endorsed significantly lower levels of informal/boundary crossing CT than non-positively inclined or neutral clinicians.

TABLE 2. Tukey post-hoc test for factor 2—fulfilled/engaging between inclination groups.

After removal of neutral clinicians, a t-test showed significant differences in factor mean scores between positively and non-positively inclined clinicians with regard to three CT dimensions (see Table 3): factor 1—entrapped/engaging (t(156) = 2.022, p = 0.045), factor 5—protective/overinvolvement (t(156) = 2.071, p = 0.040), and factor 2—fulfilled/engaging (t(156) = −2.895, p = 0.004). Positively inclined clinicians endorsed significantly lower levels of entrapped/rejecting CT (M = 3.98, SD = 0.64) and protective/overinvolvement CT (M = 3.64, SD = 0.75) than non-positively inclined clinicians (M = 3.73, SD = 0.67; M = 3.40, SD = 0.60 respectively), and significantly higher levels of fulfilled/engaging CT (M = 2.77, SD = 0.54) than non-positively inclined clinicians (M = 3.06, SD = 0.54). Non-parametric analyses (Wilcoxon) replicated the t-test results. This means that positively inclined clinicians endorsed significantly lower levels of two potentially negative/hindering CT responses (factors 1 and 5), and higher levels of positive/facilitating CT responses (factor 2), with factors 1 and 2 being most representative in aggregate, i.e., cumulating together 33.8% of the Total Variance Explained (TVE).

TABLE 3. Comparison of factor mean scores between positively inclined versus non-positively inclined clinicians.

We examined differences in factor mean scores between clinicians grouped by positive, neutral or non-positive inclination toward treating PRS, in relation to ratings on the TRQ with PRS (Soulié, Bell, et al., [15]). Compared to non-positively inclined clinicians, positively inclined clinicians endorsed lower levels of factor 1—entrapped/rejecting and factor 5—protective/overinvolvement, two potentially negative/hindering CT dimensions, and higher levels of factor 2—fulfilled/engaging, the only unequivocally positive/facilitating CT dimension. Neutral clinicians reported similar CT patterns to positively inclined clinicians.

These quantitative findings appear consistent with our qualitatively derived hypothesis that positively inclined clinicians have greater CT management skills, corresponding to the lower endorsement of negative/hindering CT responses, associated with greater positive involvement, as reflected in their higher endorsement of positive/facilitating CT responses. Referring back to the primary study (Soulié, Bell, et al., [15]), these findings tend to contradict the defense mechanism interpretation while weighing in favor of the CT montage explanation.

After Maltsberger and Buie's psychoanalytic formulations (Maltsberger & Buie, [11]), the defense mechanism interpretation proposed that the CT patterns observed in the original study could reflect defensive attitudes aiming at reducing the anxiety arising from experiencing "countertransference hate" (CT hate) toward PRS. The low endorsement of factor 1—entrapped/rejecting CT could be suggestive of repression (i.e., unconscious forgetting or blocking) of CT hate, while the high endorsement of factor 2—fulfilled/engaging CT could reflect reaction formation (i.e., the turning of CT hate into its opposite) (Soulié, Bell, et al., [15]). However, to be consistent with this view, non-positively inclined clinicians would be expected to be more defensive than positively inclined clinicians, resulting in greater repression of negative feelings corresponding to a lower endorsement of factor 1 and 5, and greater use of reaction formation translating into higher endorsement of factor 2. By showing the opposite patterns, the present findings tend therefore to contradict the defense mechanism interpretation.

In contrast, the findings appear consistent with the alternative CT montage interpretation. Relying on similarities observed between factor 1: entrapped/rejecting and the suicidal state, this interpretation proposed that the CT patterns observed in the primary study could reflect an adaptative CT montage, where clinicians empathize with the suicidal state at an implicit level, corresponding to the high prevalence of factor 1—entrapped/rejecting CT in aggregate, but manage this emotional resonance, resulting in relatively low explicit endorsement of this factor. At the same time they may be caring actively for their patients, leading to the high explicit endorsement of factor 2—fulfilled/engaging, although less representative in aggregate (Soulié, Bell, et al., [15]). In this perspective, given their hypothesized enhanced CT literacy skills (i.e., CT awareness and management), we would expect positively inclined clinicians to be more efficient at implementing such CT montage, which is consistent with the present findings. Compared to their non-positively inclined peers, positively inclined clinicians endorsed lower levels of negative/hindering CT responses, consistent with greater CT management skills on their part, and higher levels of positive/facilitating CT responses, reflecting their greater positive emotional involvement toward treating PRS.

Similarities in CT endorsement levels between neutral and positively inclined clinicians suggest that positive inclination to PRS, as assessed in this study, may not be countertransferential per se. This finding may be attributable to the limitations that accompany using a single-item assessment of positive inclination. In relation to this, when interviewed, positively inclined clinicians indicated that the terms "satisfying" and "meaningful" fitted their experience of treating PRS better than the term "like" (Soulié, Levack, et al., [16]). These may be more appropriate constructs to consider in future research.

Non-positively inclined clinicians endorsed significantly higher levels of two potentially negative/hindering CT responses, and lower levels of the positive/facilitating CT dimension, which suggests difficulties in CT management. It is possible that patient's diagnosis of personality disorder (PD) rather than suicidality could have influenced these results, given that 47.6% of the sample presented with PD according to their clinicians (Soulié, Bell, et al., [15]), and PD can have significant CT impacts (Michaud et al., [13]). However, because we based patients' suicidality and PD assessment entirely on clinicians' judgment, PD diagnosis itself could have been influenced by clinicians' negative CT rather than the other way around (Chartonas, Kyratsous, Dracass, Lee, & Bhui, [4]). In any case, given the critical importance of the therapeutic relationship, empathy, and emotional connectedness in clinical suicidology (Høifødt & Talseth, [6]; Maltsberger, [10]; Pompili, [14]; Waern, Kaiser, & Renberg, [19]), this finding questions whether self-assessed personal inclination could be used as a self-selective tool to promote effectiveness in the assessment and the treatment of suicidal behaviors.

Findings revealing differences in CT endorsement by inclination groups suggested that positive inclination is not countertransferential per se, in that positive inclination does not indicate an absence of CT responses to PRS, including negative/hindering CT. However, positive inclination to PRS appears to be associated with greater CT management skills, and greater positive emotional involvement, consistent with the notion of a CT montage as valid theoretical premises for advancing our understanding of CT occurrence and management in relation to treating PRS.

Acknowledgment to Dr Dalice Sim for providing statistical expertise