Competing interest statement
Conflict of interest: the authors declare no potential conflict of interest.
Over twenty years ago, Sidney Blatt (1995) wrote a seminal paper discussing research on the deleterious nature of perfectionism and its relationship to therapy in context of three high profile suicides of individuals who seemed to exemplify perfectionistic tendencies. In his paper, Blatt discusses a conceptualization of perfectionism that incorporates two dimensions of perfectionism, self-oriented perfectionism, characterized by unrealistic expectations for oneself and intense self-criticism when one does not attain perfection, and socially prescribed perfectionism (Hewitt & Flett, 1990, 1991a, 1991b), characterized by the perception that others have unrealistic expectations that one has to meet in order to be respected or liked by others, which can lead to extreme difficulties, related to perceived failure and ultimately to depression and even suicide (e.g. Flett, Hewitt, & Heisel, 2014; Hewitt, Flett, & Weber, 1994; Hewitt, Flett, Besser, Sherry, & McGee, 2003). Blatt concluded that these forms of perfectionism both impact and are impacted by a patient’s process and outcome in therapy. Based on earlier research, he concluded that patients with high levels of perfectionism may not do well in short-term therapy (Blatt, Quinlan, Pilkonis, & Shea, 1995), but may have better outcomes in long-term therapy (Blatt & Ford, 1994; Blatt, Ford, Berman, Cook, & Meyer, 1988; Blatt, 1992). Seven years later, Blatt and Zuroff (2002) conducted an initial preliminary review of research that focused specifically on the issues of perfectionistic attitudes and therapy, summarizing the research to that date on the topic. In that chapter, they discussed their research on alliance and perfectionistic attitudes and, now, fifteen years later, we sought to further elaborate and update this review, synthesizing the extant research on the relationship between perfectionism and alliance.
Anaclitic and introjective personality
Blatt’s focus on perfectionism reflects a component of his previous theoretical work on two distinctly different personality configurations that are vulnerabilities to depression (Blatt, 1974): anaclitic and introjective (i.e. perfectionistic) personality types. According to Blatt, during normal development, a child develops a healthy balance of personality along the anaclitic line, which is concerned with relatedness and along the introjective line, which is concerned with self-definition. However, when a disruption occurs in childhood development, the child leans more heavily on either the anaclitic (relatedness) or introjective (self-definition) line and neglects the other line. This imbalance creates a predominance of either anaclitic or introjective characteristics (Blatt & Shichman, 1983). A person with anaclitic characteristics is dependent upon his relationship with another and focused on receiving gratification. He experiences a strong desire for comfort and care from others and fear of abandonment (Blatt, 1974, 2004). A person with introjective personality traits has deep-seated feelings of unworthiness and inadequacy. He is perfectionistic and critical of himself when he does not meet his unrealistically high standards and believes others will be critical of him, as well (Blatt, 1974, 2004).
According to Blatt (1974) the introjective (i.e. perfectionistic) individual is more focused on self-definition than in cultivating relationships, often resulting in underdeveloped interpersonal skills. It follows that such interpersonal difficulties may impede the process of therapy in several different ways. The introjective patient’s internal self-critic (Blatt, 1974) may keep him from fully engaging in the relationship with the therapist and meeting the therapeutic demands. He may project his self-criticism onto the therapist, expecting that he will judge him harshly, making it difficult to open up and trust the therapist. The introjective patient may also have difficulty connecting with the therapist because of his hostile dominant interpersonal style (Habke & Flynn, 2002). His perfectionistic behavior may be off-putting to the therapist and create distance between him and the therapist (Hewitt, Flett, & Mikail, 2017; Hewitt et al., in press). Additionally, the introjective patient’s drive for perfection may constrict his ability to recover from ruptures in the relationship, causing him difficulty sustaining a strong therapeutic relationship with the therapist (Zuroff et al., 2000).
National Institute of Mental Health Treatment of Depression Collaborative Research Program
Most of what we know about the connection between perfectionism and the therapeutic relationship comes from a series of analyses conducted by Blatt and colleagues on the National Institute of Mental Health (NIMH) treatment of depression collaborative research program (TDRCP). The TDRCP was a Randomized Controlled Trial that randomly placed 250 depressed patients into four conditions to examine the efficacy of four different treatment conditions (cognitive-behavioral therapy, interpersonal psychotherapy, imipramine plus clinical management, and placebo plus clinical management). The primary outcome finding of the TDRCP was that, as so often is the case in these kinds of studies, treatment condition made little difference on patient outcome (Elkin et al., 1989). Blatt and colleagues extended this work by exploring whether certain patient characteristics influenced a patient’s engagement in and success during therapy (Blatt & Zuroff, 2002). Blatt and colleagues used the conceptualization of anaclitic and introjective personality (1974) to investigate patient variables and their associations with the therapeutic relationship and outcome.
Blatt and colleagues had previously examined these two vulnerabilities in a study of therapeutic outcome in severely mental ill individuals at a long-term inpatient facility (Blatt & Ford, 1994; Blatt et al., 1988) and in a study of outpatients receiving either psychoanalysis or supportive-expressive therapy (Blatt, 1992). Both of these investigations found different treatment responses for patients with one or the other personality structure. In fifteen months of inpatient treatment, with a minimum of 4-times weekly supportive-expressive therapy, introjective (i.e. perfectionistic) patients demonstrated greater improvement than did anaclitic patients. On the outpatient level, over two years, introjective patients progressed more in psychoanalysis (average of 5 sessions/week) and anaclitic patients progressed more in psychoanalytic psychotherapy (average of 3 sessions/week; Blatt, 1992).
In the TDRCP, Blatt and colleagues were able to utilize the Dysfunctional Attitudes Scale (DAS: Weissman & Beck, 1978) to retroactively perform analyses on two factors of the DAS, perfectionism (i.e. introjective) and need for approval (i.e. anaclitic). Consistent with the terms utilized for the DAS, and for the sake of simplicity, from this point on, we will use the term: perfectionism, to denote perfectionistic beliefs and attitudes, and in place of introjective and need for approval in place of anaclitic. When Blatt and colleagues divided up the TDRCP sample by the two vulnerabilities, they found that short-term therapy was not very effective for individuals with high levels of perfectionism (Blatt, Zuroff, Quinlan, & Pilkonis, 1996). They wanted to learn more about what caused therapy to be less successful with these patients and, because of the perfectionist’s interpersonal difficulties, they looked toward the therapeutic alliance, as a possible mediator of this effect.
The therapeutic alliance, defined as the collaborative relationship between patient and therapist (Bordin, 1976), has been consistently related to therapeutic change, more that any other clinical factor examined in psychotherapy research (Horvath, Del Re, Flückiger, & Symonds, 2011). The important aspects of alliance have been identified as agreement on goals and tasks of therapy between therapist and patient (i.e. goals and tasks), as well as the co-creation of a bond between therapist and patient (i.e. bond) (Bordin, 1979).
As is the case in the current literature (Horvath et al., 2011), for the TDRCP sample, a strong therapeutic alliance was predictive of a positive outcome for patients (Krupnick et al., 1996; Blatt et al., 1996). Blatt and colleagues performed several analyses utilizing the TDRCP sample that shed light on the relationship between perfectionism and the therapeutic alliance, as well as some of the mechanisms that may impact this relationship.
The relationship between perfectionism and therapeutic alliance in the Treatment of Depression Collaborative Research Program
How does the patient’s level of pre-treatment perfectionism affect the patient’s perception of their therapist’s Rogerian traits?
In Blatt and colleagues’ (1996) initial analysis of perfectionism and the therapeutic alliance, they examined the relationship between the patient’s pre-treatment perfectionistic attitudes and the patient’s perception of his therapist’s Rogerian alliance-building traits (i.e., empathic understanding, level of regard, unconditionally of regard, and congruence). The patient’s perception of the therapist’s Rogerian attributes is most related to the bond (Bordin, 1979) component of alliance.
The authors found no relationship between the patient’s pre-treatment perfectionistic attitudes and the patient’s perception of her therapist’s Rogerian alliance-building traits (i.e., empathic understanding, level of regard, unconditionally of regard, and congruence). In a re-analysis, published a few years later, it was found that change in patient perceptions of the therapist’s alliance-building traits over the course of treatment also was not related to pre-treatment perfectionism (Zuroff et al., 2000). From these data, it seemed that high perfectionism levels were not affecting the patient’s ability to perceive her therapist in a positive manner.
However, in 2010, Blatt and colleagues again explored if pre-treatment perfectionism affected patients’ early perception of the therapist, utilizing a within-therapist analysis (Zuroff et al., 2010). This type of analysis examined the relationship between the patient’s early perception of the therapist and pre-treatment perfectionism levels for all the patients within each individual therapist’s caseload. That is, if a given therapist treats five patients in the trial, this analysis looks at the variability in perception of the therapist of the five patients within this therapist’s caseload. The authors found that when examining within individual therapists’ caseloads, there was a negative correlation between pre-treatment perfectionism and patients’ early perception of their therapist’s Rogerian alliance-building traits. This relationship was not seen in prior analyses because of variance related to therapist effects.
Based on these new findings, we can conclude that already early in therapy, patients higher on perfectionism perceive their therapist as less empathic and more judgmental than other patients would and that individuals low on perfectionism perceive their therapists as more empathic and less judgmental than other patients would. These findings are consistent with other research on perfectionism, whereby concealing facets of perfectionistic self-presentation (i.e., not displaying or disclosing imperfection; Hewitt et al., 2003) is associated with viewing the therapist as threatening, unrealistic, and judgmental (Hewitt et al., 2008). As discussed earlier, it is possible that patients with higher levels of perfectionism project their own harsh self-criticism onto the therapist. For that reason, they may believe the therapist is more judgmental and less caring than others would.
A second related possibility is that it is not a distortion in the perception of the patient, but the therapists may actually be displaying lower levels of these Rogerian attributes, (i.e., empathic understanding, level of regard, unconditionally of regard, and congruence) when working with the patient with a high level of perfectionism. This possibility is consistent with the Hewitt et al. (2008) finding that patients with elevated perfectionism are liked less by therapists and with the idea that perfectionistic individuals can produce negative reactions in others (Habke & Flynn, 2002; Hewitt et al., in press). Both of the above possibilities can be true; the patient with high perfectionism may perceive the therapist as non-empathic because of his projection and the therapist may implicitly or interpersonally react to this perception in a negative manner and the patient, who is vigilant for such reactions, may perceive this.
How does the patient’s level of pre-treatment perfectionism affect the patient’s and therapist’s contribution to alliance across treatment?
After learning about the effects of pre-treatment perfectionism on patients’ perception of their therapists, Blatt and colleagues explored if pre-treatment perfectionism affected independent clinical ratings of patient and therapist contribution to the alliance (Zuroff et al., 2000). In other words, they wanted to examine if patients’ pre-treatment perfectionism affected the amount patients or therapist engaged in the therapeutic relationship and process. Patient and therapists contribution to alliance most closely relates to the patient’s and therapist’s goals and task (Bordin, 1979) dimension of the alliance. The authors found that early patient contribution to the alliance was not related to pre-treatment perfectionism (Zuroff et al., 2000). Early in treatment, it seems that pre-treatment perfectionistic attitudes are not a deterrent for patient input and commitment to the process (Zuroff et al., 2000).
However, pre-treatment perfectionism was negatively correlated with patient contribution to alliance later in therapy (15th session). Pre-treatment perfectionism predicted decreased patient contribution to alliance as therapy progressed, especially during the second half of therapy. In other words, high levels of perfectionism negatively affect a patient’s ability to increase his contribution and grow in the therapeutic relationship throughout therapy. From the beginning of therapy, the patient with a high level of perfectionism has more negative views of his therapist (Hewitt et al., 2008; Zuroff, Kelly, Leybman, Blatt, & Wampold, 2010), and no matter the therapist’s role in fostering this, these feelings may make it difficult for the patient with a high level of perfectionism to continue engaging in the therapeutic relationship. Perhaps, as Blatt and colleagues hypothesize, the patient needs longer in therapy to develop the capacity to trust and learn to create strong emotional bonds with another. If this hypothesis is right, this may be why long-term therapy may be more effective for patients high on perfectionism (e.g., Blatt & Ford, 1994; Blatt et al., 1988; Blatt, 1992).
It is curious why this decrease in engagement would only take place later in treatment. The authors posit this could be because the patient with a high level of perfectionism’s initial contribution to the alliance may be at an upper limit, as far as the patient can go, when it comes to building intimacy with the therapist (Zuroff et al., 2000). As seen through their lower than average ratings of their therapist’s empathy and supportiveness (Zuroff et al., 2010), patients with high perfectionism have difficulty with trust and interpersonal relationships, which may hold them back from forming deeper relationships (Hewitt et al., 2008, in press). Another possibility of why patient contribution to the alliance is only affected by pre-treatment perfectionistic attitudes later in therapy is that early in treatment, the goals and tasks of therapy may seem simpler and the patient who is high on perfectionism can engage in them without his self-criticism disrupting his willingness to engage. However, as goals and tasks become more demanding, he may experience greater difficulty with engagement because his inner self-critic is more active and distracts him from the goals and tasks at hand.
Notably, the authors found that pre-treatment perfectionism was not related to the therapist’s contribution to the alliance at any time during treatment, suggesting that therapists were able to continue engaging in a similar level of alliance building no matter the patient’s level of perfectionism. It seems therapists were not deterred from doing their part at relationship building and moving forward the therapeutic process, even with patients high on perfectionism (Zuroff et al., 2000). However, as discussed above, it is still possible that there are subtle changes in therapists’ demeanor when working with patients high on perfectionism, which were not measured by independent raters, which may have led to these patients not engaging in the alliance to the extent others patients engaged in it. Hewitt et al. (2008) found that therapists do not like patients with high levels of perfectionism as much as other patients, and it is likely that therapists’ reaction to patients with high perfectionism would manifest in some subtle way. It is possible, as discussed above, that patients high on perfectionism, who are often vigilant for signs of rejection, perceive even the subtlest signals of distance or discomfort in the therapist, and this may lead them to withdraw from the relationship (Hewitt et al., in press). More process research is required to explore this.
How does the patient’s level of pre-treatment perfectionism affect the relationship between the patient’s perspective on the therapist’s Rogerian traits and symptom reduction?
After investigating how pre-treatment perfectionism affected the patient’s rating of the therapist’s Rogerian attributes, Blatt and colleagues (1996) examined if pretreatment perfectionism affected the relationship between patients’ rating of the therapist’s Rogerian attributes and symptom reduction at termination. The authors found that patients experienced a greater decrease in symptoms when their initial perception of the therapist’s alliance-building traits was more positive. Their aim was to discover if this association would remain at different levels of pre-treatment perfectionism. They found that the association remained true for patients who had moderate pre-treatment levels of perfectionism wherein the patient’s positive view of the therapist’s alliance-building capabilities helped bolster the patient’s ability for improvement during therapy. However, the relationship between initial ratings of alliance and outcome was marginal at either high or low levels of perfectionism. Those high on perfectionism achieved poorer outcomes and those low on perfectionism achieved better outcome, no matter the patients’ early perception of the therapist’s alliance-building qualities (Blatt et al., 1996).
As we know from Zuroff et al. (2010; see also Hewitt et al., 2008), on average, individuals high on perfectionism will see their therapist in a more negative manner and individuals low on perfectionism will see their therapist in a more positive manner. Since individuals at both ends of the spectrum of perfectionism seem to see the therapist in a manner which is biased by their levels of perfectionism (more positive than average perception of therapist for patients with low levels of perfectionistic beliefs; more negative than average perception of therapist for patients with high levels of perfectionistic beliefs), it makes sense that their symptom reduction will not be linked to how they perceive the therapist initially.
Rather, someone who has an average level of perfectionism may see the therapist in a more realistic manner, which is not affected by her high or low level of perfectionism, and may be more in line with the therapist’s real level of Rogerian attributes. In other words, at the beginning of therapy, the person with a moderate level of perfectionism may be more accurately assessing her therapist’s alliance-building attributes, as well as her relational interactions with others. Therefore, this may be a positive prognostic sign (i.e. the ability to accurately assess relational interactions) for later symptom reduction than it is for those high or low on perfectionism.
Does level of patient’s contribution to alliance drive the relationship between level of pre-treatment perfectionism and symptom reduction?
With their new understanding of how the patient’s perception of the therapist affects outcome for patients with different levels of pre-treatment perfectionism, Blatt and colleagues next examined if the amount of change in patient contribution to alliance over treatment was a key variable in the negative predictive relationship between pre-treatment perfectionism and symptom outcome (Zuroff et al., 2000). They found that the lower than average increase in patient contribution to alliance throughout treatment by patients with high initial levels of perfectionism, explained some of the variance in the negative relationship between pre-treatment perfectionism and outcome. Patients high on perfectionism experience less symptom reduction than your average patient, in part, because of their stunted level of involvement in the therapeutic relationship as treatment progresses (Zuroff et al., 2000).
Blatt and colleagues (Shahar, Blatt, Zuroff, Krupnick, & Sotsky, 2004) explored what other factors could be contributing to the relationship between pre-treatment perfectionism and outcome? They decided to investigate if another variable related to interpersonal ability was also contributing to this relationship: that being level of social satisfaction. The negative relationship between increase in level of social satisfaction throughout treatment and pre-treatment perfectionism was demonstrated in an earlier study (Shahar, Blatt, Zuroff, & Pilkonis, 2003). In the subsequent study, the authors found that lower than average improvement in social satisfaction throughout treatment also contributed to the negative relationship between pre-treatment perfectionism and outcome (Shahar et al., 2004). Together, the lower than average improvement in social satisfaction and patient contribution to alliance, predicted by pre-treatment perfectionism, accounted for almost all of the negative predictive relationship between pre-treatment perfectionism and outcome. From this research, we learn that deficits in social functioning, the ability to connect and find social fulfillment both in and out of the therapy room, account for the poor outcome in therapy of the patient with a high level of perfectionism (for a description of these processes in psychotherapy, see Hewitt et al., in press). A likely implication of this finding is the importance of focusing treatment on improving interpersonal functioning, as well as, building social support and social skills for individuals high on perfectionism in order to improve outcomes.
How does perfectionism decrease throughout treatment?
Does the patient’s perception of the therapist’s Rogerian traits predict change in perfectionism over treatment?
After re-analyzing how pre-treatment perfectionism affects the relationship between the patient’s perception of the therapist and outcome, Blatt and colleagues performed a more complex set of analyses to explore if positive early perception of the therapist’s Rogerian traits could lead to a decrease in patients’ perfectionism level and symptoms throughout therapy (Zuroff et al., 2010). They conducted both within-therapist and between-therapist analyses (Zuroff et al., 2010). Their within-therapist analyses examined the relationship between early perception of the therapist’s alliance building traits and perfectionism and symptom levels for patients within each individual therapist’s caseload, while their between-therapists analyses examined the relationship between averages of early perceptions of the therapist’s alliance building traits and perfectionism and symptom levels for the patients of each individual therapist. Thus within-therapist analyses focus on the patient’s contribution to the effect, while between-therapist analyses focus on the contribution to the effect by the therapist.
In their between-therapist analyses, the authors found that patients of a therapist who was rated higher on Rogerian alliance-building qualities at the beginning of therapy experienced greater decreases in their perfectionistic beliefs and symptoms throughout therapy than patients of therapists rated lower on these Rogerian attributes. The authors did not find a predictive relationship between alliance and perfectionism levels throughout treatment in their within-therapist analysis (i.e., when they looked within individual therapists’ caseloads; Zuroff et al., 2010). This makes some intuitive sense as we know from looking at individual therapists’ caseloads, pre-treatment perfectionism is correlated with the patient’s initial perception of the therapist. That perception seems to be, in some way, a facet of the patient’s pre-treatment perfectionistic attitudes and, since this poor perception of the therapist is a type of symptom of the condition, you would not expect a more positive initial perception of the therapist to be related to greater change in perfectionistic attitudes throughout treatment. This is especially true, since the individuals who have the most space for revising their perfectionistic beliefs are inevitably beginning treatment with poorer perceptions of their therapists. Therefore, it seems logical that when looking within therapist caseloads, that individual patients’ perceptions of their therapist are not correlated with changes in perfectionism levels throughout treatment.
The above findings point to the importance of having a therapist who is high on congruence, empathic understanding, and unconditional regard in assisting a patient to decrease her levels of perfectionism (Zuroff et al., 2010). The authors posit that this type of therapist may have the larger arsenal of alliance-building skills necessary to create a successful therapeutic experience, resulting in a greater decrease in perfectionistic beliefs. Within this accepting relationship, the patient’s socially prescribed perfectionistic beliefs will be challenged head on, and she will have the support necessary to challenge her self-oriented perfectionistic beliefs. Based on the above treatments that focus specifically on the relational underpinnings and interpersonal behaviors related to perfectionism may be particularly effective (Hewitt et al., 2015).
Does the patient’s contribution to alliance at the beginning of treatment predict change in perfectionism throughout treatment?
After finding that pre-treatment perfectionism was not associated with early patient contribution to alliance (Zuroff et al., 2000), Blatt and colleagues examined if early patient contribution to alliance was predictive of perfectionism levels throughout treatment. That is, would a patient who contributed more at the beginning of therapy have greater success at attenuating his perfectionistic beliefs? This was in fact the case, level of patient contribution to the alliance at the beginning of therapy predicted change in perfectionism throughout treatment (Hawley, Ho, Zuroff, & Blatt, 2006). This may be because the patient who is actively engaged in a positive therapeutic relationship with his therapist is successfully learning to value trust and his connection with others. In this supportive relationship, he can begin to experience the compassion of the therapist, challenging his socially prescribed perfectionistic beliefs. When aligned with the therapist, he can achieve the safety necessary to challenge his self-oriented perfectionistic beliefs, as well.
What we know from this data is that two treatment conditions will result in a decrease in levels of perfectionism, A) if a patient is in treatment with a therapist who is seen, on average, as high on congruence, empathic understanding, and unconditional regard by his patients (Zuroff et al., 2010) and B) if a patient contributes more to the alliance at the beginning of therapy (Hawley et al., 2006). We do not know what causes some patients high on perfectionism to contribute more to the alliance than others early in treatment. It could be personality characteristics that were not measured in these studies. Alternatively, these results on the early perception of the therapist’s Rogerian attributes and on patient contribution to alliance are not independent of each other. It’s possible that patients who contribute more early in therapy are often the patients of therapists who are higher on these Rogerian variables. More process research is necessary to examine if there is a connection here.
Do perfectionism levels throughout therapy drive the predictive relationship between early level of patient’s contribution to alliance and outcome?
Blatt and colleagues also explored if and how perfectionistic beliefs throughout therapy, as opposed to pre-treatment perfectionistic beliefs, contribute to the relationship between early patient contribution to the alliance and decrease in depressive symptoms (Hawley et al., 2006). As discussed earlier, the authors found that early patient contribution to alliance predicted change in perfectionism levels throughout treatment (i.e. higher patient contribution to alliance early in treatment predicted greater decrease in perfectionism and vice versa). The authors also found that decreases in depressive symptoms followed decreases in perfectionistic attitudes throughout treatment (Hawley et al., 2006, 2015; Rice, Sauer, Richardson, Roberts, & Garrison, 2015). These results demonstrate an indirect association between early level of patient contribution to alliance and decrease in depressive symptoms with perfectionism levels between both of them. In other words, a patient’s early contribution to alliance predicts a decrease in his perfectionistic beliefs and as the perfectionistic beliefs decrease, depressive symptoms will decrease. It seems that decreasing levels of perfectionistic beliefs are a precursor for symptom change. This study demonstrates another reason that therapists should be especially conscientious of trying to involve their patients in the alliance-building process from the beginning of treatment. However, there is some research outside the TDRCP that we will explore that calls into question if the patient’s early contribution to therapy is enough to decrease symptoms or if it is only enough to decrease perfectionistic beliefs. The patient may also need to have a positive relationship with his therapist to achieve decreases in symptoms.
The relationship between perfectionism and alliance beyond the Treatment of Depression Collaborative Research Program
While most of the research examining perfectionistic attitudes and alliance is derived from Blatt and colleagues’ work with the TDRCP sample, there are two other works that contribute to our understanding of this issue. The first of these studies was by Rector, Zuroff, and Segal (1999) who explored the relationship between the therapeutic alliance, the DAS, and depressive symptom reduction in their sample of 47 outpatients with either a depressive disorder, anxiety disorder, or both diagnoses, receiving 20 weeks of cognitive therapy in a community-based mental health center. When measuring alliance, they utilized a patient self-report that separately examined the goals and tasks dimension and the bond dimension (Bordin, 1979) of alliance. Their intention was not to examine perfectionism per se, as much as dysfunctional beliefs, which included perfectionistic beliefs. However, their study is still informative on perfectionism and alliance, as they examined the relationship between the DAS perfectionism scale (along with the need for approval scale) and alliance.
A second, more recent, study by Whelton, Paulson, and Marusiak (2007) also examined the relationship between pre-treatment perfectionistic attitudes and the therapeutic alliance in a sample of 169 outpatients from a community mental health clinic. When measuring the alliance, they looked at the patients’ ratings of one composite measure, which included the components of alliance described above, i.e., bond and goals and tasks (Bordin, 1979).
How does pre-treatment perfectionism predict early therapeutic alliance, as assessed by the patient?
Rector and colleagues (1999) wanted to see how pretreatment dysfunctional beliefs, including perfectionistic beliefs, affected alliance early in treatment. They found that pre-treatment perfectionism was predictive of the patient’s perception of her bond with the therapist, but not of the patient’s perception of her involvement in the goals and tasks of therapy early in treatment. This seems similar to Blatt and colleagues’ findings that within an individual therapist’s caseload, pre-assessment perfectionism was related to the patient’s perception of the therapist’s Rogerian attributes (Zuroff et al., 2010), but early patient contribution to alliance was not associated with pre-treatment perfectionism (Zuroff et al., 2000). From both the TDRCP findings and Rector and colleagues’ work, it seems that patients with high levels of perfectionism see their therapist and the therapeutic relationship as less supportive than other patients see them (Zuroff et al., 2010; Rector et al., 1999), but are still able to contribute to alliance and engage in the goals and tasks of therapy early on (Zuroff et al., 2000; Rector et al., 1999). Perhaps, as discussed earlier, early in treatment, the patient with a high level of perfectionism already reaches her upper limit in ability to sustain closeness with the therapist (Zuroff et al., 2000), especially because she does not see the therapist as supportive as others would (Zuroff et al., 2010). Additionally, it’s possible that early in treatment, the patient’s self-criticism is less active because the tasks are not as demanding and this lets the patient more easily engage in the therapeutic process.
Whelton, Paulson, and Marusiak (2007) also explored how the patient’s rating of the therapeutic alliance was correlated with pre-treatment perfectionism. They found that higher pre-treatment perfectionism scores were correlated with lower than average ratings of the therapeutic alliance by the patient at each session examined (session 3, 6, 9, and 12). This finding is difficult to interpret, as the authors examined the global levels of alliance, therefore, we do not know which dimensions of alliance are affected by pre-treatment perfectionism throughout therapy.
What is the relationship between early therapeutic alliance, as assessed by the patient, and decrease in perfectionism throughout treatment?
Rector and colleagues (1999) next explored if early alliance predicts a decrease in perfectionism throughout treatment. They found the patient’s rating of his involvement in the goals and tasks of therapy at the third session was associated with a decrease in perfectionistic beliefs at termination (the greater involvement in the goals and tasks of therapy, the greater the decrease in perfectionism), but the patient’s rating of the therapeutic bond at the third session was not correlated with a decrease in perfectionistic beliefs throughout treatment.
These results are very similar to the results of the TDRCP that examined the relationship between early alliance and change in symptoms of perfectionism throughout treatment. It was found that patient contribution to the alliance at the third session was linked to a decrease in perfectionism (Hawley et al., 2016), while the individual patient’s early perception of the therapist’s alliance-building qualities was not when controlling for therapist effects (Zuroff et al., 2010). Perhaps Rector et al. (1999) would have found a relationship between early perception of therapeutic bond and decrease in perfectionism if they separated within and between-therapist effects?
Can level of therapeutic bond affect the association between change in perfectionism and change in depressive symptoms?
Interestingly, Rector and colleagues (1999) found that if there was not a high level of therapeutic bond present, the relationship between perfectionism levels and change in depressive symptoms was greatly weakened. Namely, even if the patient is very involved in the therapeutic process early on and his perfectionism levels are decreasing, without a strong therapeutic bond, change in perfectionism did not translate into greater changes in depressive symptoms. It is of interest that we can see a drop in perfectionism that is not related to a decrease in symptoms, as we would expect (e.g., Hawley et al., 2006). Change in perfectionistic beliefs may not be enough to achieve symptom change, perhaps other related beliefs need to change, as well.
Several authors posit that a strong therapeutic bond is necessary for symptom change because it helps the patient improve his interpersonal functioning (Hewitt et al., 2017; Rector et al., 1999), an important correlate to symptom change. As we know from Shahar et al. (2004), deficits in interpersonal functioning drive the relationship between perfectionism and poor outcome. In a comprehensive model of perfectionism, Hewitt and colleagues (2017) suggest that perfectionism is comprised of deeply ingrained traits represented by both interpersonal and intrapersonal perfectionistic behaviors. Changes in perfectionistic attitudes or cognitions may not reflect changes in deeper components such as perfectionistic traits or interpersonal styles, which are necessary to improve his social functioning.
Another possible explanation for this finding is that without a continuing, strong therapeutic relationship, perfectionistic beliefs can be altered on a surface level, but those beliefs are not internalized and, for that reason, symptom change does not follow. Relatedly, Paul Gilbert, the founder of compassion-focused therapy, a treatment for individuals with high levels of self-criticism, writes, that a therapist can help these patients develop self-compassion by creating an environment of acceptance and compassion, which can be internalized by the patient and provide her with the safety necessary to develop self-compassion (2009). Without realizing the unconditional acceptance of the therapist, the patient may not be able to restructure her perfectionistic beliefs at a more core level.
The necessity of this unconditional acceptance could explain why patients of therapists who were perceived as displaying higher levels of alliance-building qualities at the beginning of therapy experienced greater decreases in both their perfectionistic beliefs and symptoms throughout therapy than patients of therapists who were rated as lower in these Rogerian attributes. Perhaps, these patients were able to realize their therapist’s support and compassion throughout therapy and as they internalized these attitudes, their perfectionistic attitudes abated.
What might explain the relationship between pre-treatment perfectionism and alliance?
Whelton and colleagues (2007) explored potential mediators of the relationship between perfectionism and alliance. They examined if affective display could mediate this relationship. The authors found that a high level of hostility and a low level of positive affect mediated the relationship between perfectionism and the therapeutic alliance at specific sessions, suggesting that negative affect may account for a great deal of the reason that individuals high on perfectionism have difficulty forming therapeutic alliances (Whelton et al., 2007). This is consistent with the Perfectionism Social Disconnection Model described by Hewitt et al. (in press) who suggest that perfectionistic individuals can present with hostile and distancing behaviors in therapy that, if not dealt with, will interfere with therapeutic alliance and outcome. As we discussed earlier, patients with high perfectionism elicit negative reactions in their therapists (Hewitt et al., 2008). The negative emotions displayed by the individual with high perfectionism may cause the therapist to feel discomfort or annoyance, which the patient may perceive, causing the patient to partially disengage from the therapeutic relationship and therapeutic process. As stated earlier, the authors of this study examined a composite therapeutic alliance score, so it is difficult to determine, which aspects of alliance (i.e., bond or goals and tasks (Bordin, 1979) are related to patient affect.
A limitation to making conclusions from this research is that the majority of information about perfectionism and alliance is based on one study (The NIMH TDRCP) with many sets of analyses of this one sample. It is essential that we conduct more research on perfectionism and the therapeutic alliance utilizing different populations and measures. Other potential limitations regarding our synthesis of data from this research is the variability in the specific scales and types of measurement used. That is, three different alliance measures were utilized to examine different aspects of alliance across these studies. Two measures were patient rated [The Barrett-Lennard Relationship Inventory (Barrett-Lennard, 1962) & Working Alliance Inventory (Horvath & Greenberg, 1989)], while the other was evaluated by independent raters [The Vanderbilt Therapeutic Alliance Scale (Hartley & Strupp, 1983)]. Because of their self-critical biases, the response style of patients with a high level of perfectionism may be more critical than others and this may mean that their ratings of alliance may be more negative than an independent rater’s would be. Additionally, there were no measures used that assessed the therapist’s perspective of alliance directly.
Two different measures of perfectionism were used in these studies [Dysfunctional Attitudes Scale (DAS) & Depressive Experience Questionnaire (DEQ; Blatt, D’Afflitti, & Quinlan, 1976)]. Both the DAS and DEQ are self-report measures, meaning that we do not have any observer or therapist data, only the patients’ perspective on their perfectionistic attitudes. Patients high on perfectionism may perceive their progress in both decreasing their perfectionistic beliefs and their symptoms, as less drastic, than an outside rater would. One new independent clinical rating system that has a scales consistent with perfectionism would be the Anaclitic and Introjective Depression Assessment (AIDA; Rost, Fonagy, & Luyten, unpublished material), which has been recently supported in a series of construct validity analyses (Miller & Hilsenroth, 2016).
An additional limitation is that the two measures of perfectionism used in these studies (The DAS and DEQ) were not developed to measure perfectionism, but rather dysfunctional attitudes. The wider field of perfectionism research has not embraced either the DAS or DEQ, as measures of perfectionism and, instead has utilized the multidimensional measures of Frost and colleagues, (Frost, Marten, Lahart, & Rosenblate, 1990) and Hewitt and Flett (1989). Perfectionism is a broad construct and extends beyond attitudes into trait, interpersonal and psychodynamic domains. Future work with other measures and broader conceptualizations of perfectionism would provide more information on the nature of perfectionism in treatment.
Another important limitation is the generalization of these studies, as they examine different types of patient populations (depressed population, mixed depression-anxiety population, and general outpatient populations). The results also may not generalize to other broad patient populations with more severe or less severe levels of mental illness, or other highly specified diagnoses.
In order to further enhance our knowledge on the relationship between alliance and perfectionism, it is imperative to conduct more research utilizing a wide range of populations and measures, which capture different aspects of perfectionism and alliance. We hope that more studies will examine the relationship between perfectionism and therapeutic alliance from independent rater and therapist perspectives. It would also be useful to have more studies on perfectionism and alliance, which separate out therapist and patient effects.
We need more research on what aspects of the multifaceted construct of perfectionism, beyond attitudes and beliefs, negatively affect alliance formation. As we learn which trait, interpersonal, and psychodynamic domains affect the therapeutic relationship, we can identify specific therapeutic interventions and/or therapist traits that could improve the therapist’s bond with patients high on perfectionism and increase the patient’s engagement in therapy.
We also need more research on how specifically the therapeutic alliance decreases perfectionistic beliefs in relation to symptom reduction and what therapists can do to bring about these changes. How can therapists help patients with high perfectionism improve their social skills and interpersonal abilities? It will be helpful to develop a greater understanding of what specific techniques and interventions the therapist who is perceived to have high Rogerian attributes is actually using to help patients decrease their perfectionistic attitudes and their symptoms.
Additionally, we do not know if there are ways in which therapist behavior is affected by the patient with a high level of perfectionism. Are there subtle differences in the way the therapist treats patients high on perfectionism? If so, this knowledge will allow therapists to be more vigilant of the subtle messages they may be communicating to patients high on perfectionism, which may deplete their trust in the therapist and discourage their involvement in therapy, and address potential issues.
Finally, there is currently no research on therapeutic alliance and perfectionism in long-term therapy. Blatt and colleagues have demonstrated better outcome for patients with high levels of perfectionism in long-term therapy (Blatt et al., 1988; Blatt, 1992; Blatt & Ford, 1994). It would be helpful to understand specifically the role of therapeutic alliance in creating better outcomes in these long-term treatments.
From the above studies we’ve examined, based on the TDRCP and otherwise, it seems that the therapeutic alliance is affected by a patient’s pre-treatment level of perfectionistic attitudes and that those attitudes affect the subsequent development of the therapeutic alliance. Patients with high levels of perfectionism will see their therapist and their relationship with their therapist in a more negative manner from the outset of therapy (Zuroff et al., 2010; Rector et al., 2000; Hewitt et al., 2008). How an individual patient perceives his therapist seems only important in leading to better outcomes for a patient with mid-range perfectionism (Blatt et al., 1996) who is not prone to see the therapist in a more extreme positive or negative manner from the beginning of treatment. Additionally, the ability of a patient with a high level of perfectionism to engage in therapy does not appear to be negatively affected, at first (Zuroff et al., 2000; Rector et al., 1999), but only later in therapy (Zuroff et al., 2000). Perhaps, early on, patients higher in perfectionism have gotten as close to their therapists as they can because of their difficulties with interpersonal relationships (Habke & Flynn, 2002; Zuroff et al., 2000).
What is effective in helping a patient decrease her perfectionistic attitudes? If a patient is fortunate to be in treatment with a therapist who is rated as higher on Rogerian constructs, the patient will likely experience a greater decrease in perfectionism and symptoms (Zuroff et al., 2010). This indicates that there are specific interventions or stances a therapist can take, which will help change their perfectionistic attitudes. Additionally, patients who engage more in treatment early on are more likely to experience a decrease in perfectionistic beliefs (Rector et al., 1999; Hawley et al., 2006). Therapists will likely see their patients’ perfectionistic attitudes decrease if they can figure out how to involve them more throughout the therapeutic process. That said, even though lower perfectionism levels are generally linked to decreases in symptoms (e.g., Hawley et al., 2006), that decrease in perfectionistic attitudes, related to early high levels of patient involvement in the goals and tasks of therapy, may not be accompanied by the expected decrease in symptoms if a strong therapeutic bond is missing (Rector et al., 1999).
How does a therapist’s reaction to the patient with high perfectionism affect the alliance? It seems that therapists contributed just as much to the alliance when working with patients who were high on perfectionism, as when working with other patients (Zuroff et al., 2000). However, Hewitt et al. (2008) found that therapists often have negative reactions to patients with high perfectionism, and it is possible that there are subtle ways in which the therapist engages differently with patients high on perfectionism because of his negative feelings toward that patient, which were not measured by the independent raters (Hewitt et al., in press). The patient’s perception of these subtle changes in therapist demeanor may result in the patient not increasing her engagement in the alliance to the same extent as other patients during treatment.
Furthermore, these studies inform us about mediators between pre-treatment perfectionism, alliance, and outcome. The relationship between pre-treatment perfectionism and alliance can be partially explained by the higher level of hostility and lower level of positive affect endorsed by the patient (Whelton et al., 2007; Hewitt et al., 2008). While, the relationship between pre-treatment perfectionism and outcome is almost entirely explained by level of patient contribution to alliance and satisfaction with one’s social network (Shahar et al., 2004). These results highlight the importance of social functioning in improving both the therapeutic alliance and outcome of patients with a high level of perfectionism. The above findings lead us to believe that it is imperative for clinicians to forge a strong therapeutic alliance with their patients with high perfectionism to assist and support them in improving their interpersonal functioning and social skills in order to help them find relief from symptoms and change their perfectionistic beliefs.