We examined clinicians' countertransference toward BPD across client-level (client age and gender) and clinician-level (clinician demographics and professional experience) factors. Consistent with past research, we found evidence that the BPD label was associated with negative countertransference reactions (Gallop, Lancee, & Garfinkel, 1990; Markham & Trower, 2003; Rossberg et al., 2007). Specifically, despite the fact that all clinicians in the study read an identical vignette describing a client with BPD, those clinicians who accurately labeled the client as BPD exhibited lower levels of empathy toward the client and also viewed the client as more ill (i.e., less of a conduct problem) than clinicians who chose other diagnoses. The study also expanded upon prior literature by examining clinician- and client-level factors that were associated with these reactions, with results suggesting client age, client gender, and several clinician-level factors may all influence the way clinicians respond to clients with BPD.
Our findings highlight some potential concerns with regard to the use and impact of the BPD label among youth. As expected, clinicians were less likely to diagnose BPD in adolescents versus adults presenting with BPD symptoms. Lower rates of diagnosis may be because of a failure to accurately identify the symptoms as BPD or may reflect unwillingness to “prematurely” label the adolescent with a personality disorder before his or her personality has fully developed. In addition, BPD tendencies may also be conflated with normative adolescent “storm and stress” and thus not identified as indicative of early BPD. Regardless of the cause, research increasingly suggests that early identification of BPD in youth, while fraught with some costs, may provide some benefit. Several studies show that without treatment, BPD symptoms are moderately stable over development (Crick, Murray-Close, & Woods, 2005; Meekings & O'Brien, 2004; Miller et al., 2008; Rogosch & Cicchetti, 2005; Shiner, Masten, & Tellegen, 2002). Therefore, to the extent that failure to diagnose BPD may hinder access to appropriate interventions, this could negatively impact the treatment of youth with BPD. Even more troubling, among clinicians who applied the BPD diagnosis, those in the adolescent condition rated their client as less ill, less trustworthy, and more dangerous than those in the adult condition. Such findings suggest that clinicians may be more inclined to view youth with BPD as “bad” rather than “ill” as compared to adults and to exhibit more negative attitudes toward them. Future research is warranted to evaluate the direct influence of these negative attitudes on therapist–client relationships.
Consistent with prior literature (Becker & Lamb, 1994; Hartung & Widiger, 1998; Kaplan, 1983), our findings do suggest potential gender biases in the application of the BPD label. Despite receiving the same vignette except for the client's name, clinicians were more accurate in diagnosing the female client with BPD while tending to misdiagnose the male client. Thus, expanding on literature demonstrating a gender-biased overdiagnosis of BPD in women, our data suggest that there may also be a bias toward underdiagnosing BPD in men. This finding raises concerns regarding treatment delivery, particularly for men. Namely, if clinicians fail to diagnose BPD when indicated, men may not get the treatment they need.
In contrast to literature highlighting gender as a moderator of attitudes (Henry & Cohen, 1983; Klonsky, Jane, Turkheimer, & Oltmanns, 2002; Wirth & Bodenhausen, 2009), we did not find differences in clinicians' reactions to the male and female clients with BPD. Although this may suggest that clinicians do not react differently to clients with BPD on the basis of gender, another explanation for this discrepancy may be that clinicians do not feel comfortable admitting to negative attitudes toward this population and were conscious of the social undesirability of explicitly admitting to gender-based assumptions of BPD symptoms. Although the anonymous nature of an online survey helps to provide a safe environment for clinicians to report their true feelings, nonetheless, self-report assessments of attitudes rely on the willingness of respondents to answer honestly, even though doing so may be undesirable. Demand characteristics may preclude clinicians from admitting to or even being aware of stigmatizing attitudes. Indeed, examining overall mean attitude levels for the sample shows that clinicians tended to rate the client positively, with mean levels generally falling at around 2.70, indicating that they agreed with statements assessing empathy and interpersonal effectiveness and disagreed that the client would never get better. Studies using implicit measures to test attitudes might be more effective at circumventing these social desirability factors.
Interestingly, there was also an interaction of gender and age, such that adult females and adolescent males were the most likely to be diagnosed with BPD. Although this finding warrants replication, it raises some questions with regard to the confluence of gender norms and adolescent development. Following from the Klonsky et al. (2002) finding that gender-atypical behaviors may be interpreted as indicative of BPD, our findings may imply that, for men, this relationship exists only for those who present with BPD symptoms early in development. Adolescent males may be more likely than adult males to express female-typical symptoms such as affective lability, unstable interpersonal relationships, and suicidal or self-harming gestures, and thus, may be more likely to be diagnosed with BPD. Future research is needed to look at the impact of gender across age before any conclusive statements can be made.
The present study identified a number of clinician characteristics that appear to play a role in countertransference toward clients with BPD. Consistent with past research (Commons-Treloar & Lewis, 2008b; Hugo, 2001; Jorm et al., 1999), master's level therapists, psychologists, and individuals who treated more clients with BPD or who had training specific to BPD endorsed more positive reactions. Older clinicians and, to a smaller extent, psychiatrists were more negative. These factors are likely to play an important role in the therapeutic relationship such that by being cognizant of their personal and professional characteristics that may contribute to negative reactions, clinicians may be better able to manage countertransference dynamics that could adversely impact treatment.
For instance, it is notable that psychologists and master's level therapists, although more empathic, viewed their clients' symptoms as more of a conduct problem and were less trusting of them than psychiatrists. To understand this finding better, we conducted post hoc analyses to examine whether differences in training and experience across clinician discipline might help explain this pattern. With the given literature suggesting a link between personal contact and positive attitudes toward mental illness (Commons-Treloar & Lewis, 2008a 2008b), our first thought was that psychiatrists may have less direct contact with this population. A one-way ANOVA of the number of clients treated by clinician discipline revealed that psychiatrists reported treating significantly more clients with BPD than psychologists or master's level therapists, F(2, 515) = 12.67, p < .0001, over the course of their careers, indicating that the amount of contact alone may not fully explain countertransference. We next examined the differences in diagnostic accuracy across discipline, hypothesizing that differences in the degree to which clinical disciplines emphasize diagnostic classification may impact attitudes. A one-way ANOVA of diagnosis by discipline indicated that, while psychiatrists and psychologists did not differ in their accuracy, master's level therapists were significantly less likely to diagnose BPD than the other two groups, F(2, 553) = 3.14, p < .05. These analyses seem to support our original hypothesis that differences in attitudes may reflect differences in the theoretical perspectives emphasized in each discipline. Namely, in utilizing a medical model, psychiatrists may focus more on diagnostic criteria to conceptualize clients, whereas master's level therapists, who often lean more toward a person-centered approach, put more weight on individual symptom presentations in making their clinical impressions. In doing so, psychiatrists may overemphasize the client's acting-out behaviors, which may result in a more distrustful attitude toward them. Of course, these findings are purely exploratory, and further research is needed to understand these differences better.
However, the results of the present study also make salient the complex interplay between these many clinician characteristics and the need for further research to parse apart the specific aspects of each professional characteristic that are most closely associated with negative countertransference. As an example, our findings highlight the importance of distinguishing between clinician experience and age. Whereas the former reflects active hands-on practice and training, the latter may point to individuals who have not had specialized training as recently as or as often as younger clinicians. Clinicians would be wise to take these subtleties into account when reflecting on the personal factors that might be playing into their reactions to clients.
Likewise, although our study highlights the importance of hands-on practice early in training, the question of what “hands-on practice” entails has yet to be answered. Perhaps the difference in attitudes between psychiatrists and master's level therapists could be attributed to differences in the length or intensity of the relationship with the client. Master's level therapists can work closely with a client for years conducting case management and intensive psychotherapy multiple times a week, whereas psychiatrists are limited to a 30-min session often less than once a month. Depending on which characteristics prove most influential, the type, structure, and intensity of the training model that is best will vary.
Our findings suggest that direct exposure to this population early in clinical training and continuing throughout one's career is beneficial to enhancing feelings of competence, which, in turn, fosters positive countertransference. Although this recommendation may be counterintuitive given the common view that novice therapists should not receive difficult patients before they have accumulated the experience to handle them, our results suggest that early experience when supervision is more consistent and intensive may be the most effective means of building competence with difficult populations.
The present study has a number of important strengths and weaknesses. The emphasis on both client and clinician factors as well as specific domains of countertransference toward BPD clients is unique. This approach allowed us to understand the discrete factors that play into clinicians' reactions to these difficult clients, which, in turn, provided a more in-depth understanding of factors that both strengthen and challenge the therapeutic relationship. In constructing the vignette, we took care to include no additional information beyond the DSM symptoms of BPD to avoid unduly biasing the clinicians' attitudes. At the same time, the online self-report format of the survey limited the conclusions that could be drawn from our results. Using a convenience sample, we were not able to ensure equal cell sizes, making it difficult to examine moderating effects of clinician characteristics. Likewise, our sample size precluded us from being able to examine more intricate pathways between different context variables. In understanding the null results obtained for some of our main hypotheses (e.g., gender differences), it is also likely that a vignette design is simply not a strong enough manipulation to accurately assess clinicians' implicit biases toward BPD. Instead, the association between gender and negative attitudes toward BPD may be more accurately assessed implicitly (i.e., using a computerized implicit association task) in which the purpose of the study is less transparent (Peris, Teachman, & Nosek, 2008; Stier & Hinshaw, 2007; Teachman, Wilson, & Komarovskaya, 2006). Clients with BPD are a particularly difficult population to serve because of their high sensitivity to rejection, need for acceptance, and associated self-harming gestures. However, the population is only one of many trying clientele with which clinicians interact. We focus on this disorder because it is commonly encountered in clinical settings. However, these findings are likely to apply to other disorders, and further research is needed to confirm generalizeability to other clinical populations. Finally, data were cross-sectional and causality cannot be inferred.
Nonetheless, the findings of the present study are noteworthy. The diagnostic classification system exists as a framework to guide treatment efforts, but this framework is only as effective as the diagnostician using it. Our results suggest that there may be both age and gender biases in the way that clinicians apply the BPD label. We also found systematic differences in attitudes based on different aspects of clinician experience and training. Together, these findings suggest that more education in the form of hands-on practice is necessary to improve clinicians' familiarity and comfort with the BPD diagnosis.