Clinical Experiences Using Cognitive-Behavioral Therapy for Eating Disorders

Abstract

Data suggests that despite the availability of evidence-based psychological treatments for eating disorders (EDs), techniques from these therapies may be less frequently used within real-life clinical practice. The aim of this study was to provide the opportunity for clinicians to give feedback on their experiences treating EDs using cognitive-behavioral therapy (CBT) through reporting on use of CBT techniques and barriers to treatment implementation in naturalistic settings. Clinicians (N = 126) who self-identified as using CBT for EDs reported demographic information, frequency/usefulness of empirically supported treatment techniques, problems/limitations of CBT, and barriers faced while implementing CBT. The most frequently used technique reported by clinicians was psychoeducation, and the least frequently used technique was use of surveys to address mind reading. Patients’ unwillingness to follow a meal plan/nutritional guide was rated as the most impactful barrier, alongside ED severity. Of the problems/limitations of CBT, too little guidance on treating co-occurring symptoms was rated as the most impactful. This study provided a mechanism for clinicians to share their experiences using CBT for EDs in real-world settings. Overall, results regarding frequency of use and usefulness of techniques indicate a high level of endorsement. Moreover, the most frequently endorsed barriers to/limitations of CBT related to lack of guidance on treating complex ED presentations. Future research should explore ways to treat cases that go beyond the prototypical ED case and explore ways to adapt CBT to meet the needs of naturalistic treatment settings.

Eating disorders (EDs) have the second highest mortality rate of any psychiatric illness (Agras, 2001; Iwajomo et al., 2021; van Hoeken & Hoek, 2020) and can be difficult to treat, as evidenced by both high relapse rates and chronic course of illness (Fairburn et al., 2000; Keel & Brown, 2010). Further, up to an estimated 52% of patients with anorexia nervosa (AN) and an estimated 35% of patients with bulimia nervosa (BN) will relapse following eating disorder (ED) treatment (Keel et al., 2005; Keel & Brown, 2010; Khalsa et al., 2017). Given the persistent and dangerous nature of EDs, it is important to characterize barriers that may interfere with the success of existing treatments to improve treatment approaches for this population.

Several psychological treatments for EDs have demonstrated efficacy or effectiveness and thus are often considered “evidence based,” including family-based treatment (Lock & Le Grange, 2005), interpersonal psychotherapy (Murphy et al., 2012), and cognitive-behavioral therapies (CBT; Fairburn, 2008; Waller et al., 2007). Among existing treatments, CBT represents one of the most commonly used treatments for EDs and includes a range of cognitive and behavioral approaches aimed to address ED pathology (Fairburn, 2008; Waller et al., 2007). Of note, there exist several different treatment manuals that draw upon CBT principles for EDs (Fairburn, 2008; Waller et al., 2007, 2018). Enhanced cognitive-behavioral therapy (CBT-E) represents one of the more commonly used manuals within the CBT framework and has been tested across a range of diagnostic presentations, including BN (Wilson et al., 2002), AN (Dalle Grave et al., 2016), and binge-eating disorder (BED; Fairburn et al., 2015), as well as adolescents and youth with transdiagnostic EDs (Dalle Grave et al., 2020; Gowers & Green, 2009; Le Grange et al., 2020). In addition to CBT-E, there are several other published resources that outline a CBT-based approach to ED treatment, including the treatment guide developed by Waller and colleagues (2007), guided self-help manuals for BED (Striegel-Moore et al., 2010) and BN (Carter et al., 2003), and shortened versions of CBT for nonunderweight EDs (CBT-T; Waller et al., 2018). Importantly, despite some differences across manuals, CBTs for EDs possess many shared features, including a focus on tracking and understanding the functional underpinnings of cognitive and behavioral symptoms of EDs, promoting regular eating and weighing, and utilizing behavioral experiments (Fairburn, 2008; Waller et al., 2007).

Despite a robust body of work supporting the use of CBT in EDs (Fairburn et al., 2015; Linardon et al., 2017), growing data suggest that treatment providers who endorse using CBT may not deliver treatments as outlined, which may result in worse clinical outcomes. For instance, an initial study exploring adherence to ED treatments suggested that only 6% of therapists report that they adhere closely to empirically supported treatment manuals (Tobin et al., 2007). Studies focused more specifically on CBT indicate infrequent use of well-supported CBT techniques with clients with EDs (Brown & Nicholson Perry, 2018; Mulkens et al., 2018; Waller et al., 2012). For instance, Mulkens and colleagues investigated ED clinicians’ usage of specific CBT techniques and found that with the exception of regular eating, the majority of clinicians did not endorse routinely using well-supported ED techniques with clients. Data collected from patients receiving CBT for EDs further support inconsistent implementation of CBT principles (Cowdrey & Waller, 2015; Serpell et al., 2013). For instance, less than 60% of patients report therapist use of regular weighing and food monitoring records, two principal components of all existing CBT manuals (Cowdrey & Waller, 2015). Altogether, these data highlight a critical need to better understand processes that may drive infrequent use of CBT techniques in real-life practice.

Characterizing the nature and consequences of “therapist drift” from recommendations in ED treatment manuals may provide insights necessary to promote clinically driven adaptations of treatment. Additionally, understanding barriers related to therapist implementation of existing treatments can help researchers design effective efforts to increase uptake of existing protocols. Therapist drift represents one aspect of a well-characterized gap between research studies and the realities of clinical practice (Goldfried et al., 2014).

Specifically, both empirical and theoretical work suggests that the manner through which existing treatments are developed, tested, and manualized may limit their usefulness for practicing clinicians (Hay, 2013). For instance, empirically supported treatments are most rigorously tested using randomized controlled trials (Hay, 2013). Randomized controlled trials rely on highly controlled environments to test treatment effects on outcomes and often attempt to limit confounding variables (Goldfried et al., 2014), decreasing external validity. Within EDs, several of the most well-known randomized controlled trials for CBT-E excluded patients who were medically unstable (Fairburn et al., 2009, 2015), despite medical complications being a well-established consequence of EDs.

As another example, research trials have largely recruited participants who identified as female, White, and non-Hispanic, despite epidemiological data indicating that EDs affect people of all genders and racial backgrounds (Burnette et al., 2022; Egbert et al., 2022). Trials also report a preponderance of participants with higher socioeconomic statuses and in “low” to “normal” body mass indext (BMI) ranges (Fairburn et al., 2009; Poulsen et al., 2014), despite data supporting high rates of EDs across the weight spectrum and across all income levels (Burke et al., 2023; Duncan et al., 2017). Therefore, clinicians may diverge from existing treatments due to a perception that treatments designed using homogeneous samples may not adequately address the experiences/needs of their clients’ diverse identities and experiences (Burnette et al., 2022; Egbert et al., 2022).

While a growing body of literature has begun to explore the utility of alternative research designs, such as nonrandomized clinical trials (Le Grange et al., 2020) and naturalistic data collection (Signorini et al., 2018), there remain valid reasons for which treatments designed for use in controlled environments may not generalize to a patient population that is diverse in presenting comorbidities, medical complications, identity factors, socioeconomic statuses, and learning histories (Boisvert & Harrell, 2014; Forrest et al., 2017).

Increased bidirectional communication between researchers and clinicians presents one option to mitigate the gap between clinical research and practice and better direct future efforts to adapt treatments. Current models of clinical science typically operate using unidirectional communication, such that researchers communicate research findings and disseminate theoretical models and treatment manuals to clinicians. However, clinicians often are not afforded formal opportunities to provide feedback on these interventions to researchers to inform future research and innovation (Goldfried et al., 2014). As bidirectional communication could provide researchers with concrete feedback to help further elucidate barriers relating to treatment implementation, Goldfried and colleagues developed a “two-way bridge” initiative to provide an avenue for clinicians to describe their experiences using empirically supported treatments. To date, studies using the two-way bridge model have been completed in samples of therapists treating a range of psychological disorders, including panic disorder, obsessive-compulsive disorder (OCD), and generalized anxiety disorder (GAD; Jacobson et al., 2016; Szkodny et al., 2014; Wolf & Goldfried, 2014).

Given high relapse rates and complex presentations in EDs (Keel et al., 2005; Steinhausen, 2009), eliciting clinician perspectives regarding barriers to implementations and perceived limitations of treatments may be particularly useful in informing treatment research that maximizes external validity. As noted above, several studies have characterized the frequency of CBT technique use in ED clinician samples (Mulkens et al., 2018; Waller et al., 2012). However, this work has primarily focused on mean rates of technique endorsement, as well as therapist-related characteristics that may predict CBT technique use (e.g., clinician anxiety, clinician intolerance of uncertainty; Levita et al., 2016; Mulkens et al., 2018; Waller et al., 2018). While therapist characteristics likely exert a strong influence on reported use of CBT techniques (Waller, 2016), therapist perception of perceived usefulness of techniques, as well as therapists’ perception of limitations of CBT, have not been characterized and provide additional information to inform why technique use may be low. Understanding which aspects of current manuals are useful and identifying common clinical barriers to implementation can inform efforts in the dissemination of CBT to better adapt this treatment for nonresponders or complex presentations.

This exploratory study sought to extend prior work suggesting that CBT techniques may be infrequently used in real-life settings and provide an opportunity for clinicians to give feedback on their experiences using CBT for EDs through reporting usefulness of techniques, barriers to implementation, and perceived limitations of this approach. Our first aim was to determine which CBT techniques clinicians use most frequently and find most useful. Second, we sought to characterize rates of perceived barriers and limitations to treatment success in CBT.

Materials and Methods

PARTICIPANTS

Participants (N = 126; see Table 1) were clinicians and/or trainees who self-identified as having used or as currently using CBT for EDs, broadly defined (i.e., “Have you ever used CBT to treat EDs?”). The study sample was majority female (90.5%), non-Hispanic (91.3%), and White (88.1%). Participants most frequently endorsed a highest degree of a Ph.D. in clinical psychology (38.9%). Participants endorsed most frequently seeing patients for a total duration of 6 months to a year (23.8%), and most participants currently saw patients in private practice (54.8%). Participants were asked to rate the degree to which different theoretical orientations guided their clinical practice (totaling to 100%); on average, a behavioral orientation was rated as having the greatest influence on practice, M = 38.03, SD = 23.50.

Table 1.

Sample Demographic and Descriptive Characteristics (N = 126)

Variable	n (%) or M (SD)	Variable	n (%)
Race		Years experience: psychotherapy
White	111 (88.1%)	0–5 years	27 (21.4%)
Black/African American	3 (2.4%)	5–10 years	40 (31.7%)
Asian	6 (4.8%)	10–20 years	39 (31.0%)
Middle Eastern/North African	4 (3.2%)	20+ years	19 (15.1%)
Prefer not to respond	2 (1.6%)	Prefer not to respond	1 (0.8%)
Ethnicity		Years experience: ED therapy
Hispanic/Latino/a	10 (7.9%)	0–5 years	46 (36.5%)
Not Hispanic or Latino/a	115 (91.3%)	5–10 years	35 (27.8%)
Prefer not to respond	1 (0.8%)	10–20 years	31 (24.6%)
Gender		20+ years	14 (11.1%)
Female	114 (90.5%)	Number of ED patients treated
Male	8 (6.3%)	< 10	10 (7.9%)
Nonbinary/nonconforming	2 (1.6%)	10–20	12 (9.5%)
Transgender male	2 (1.6%)	21–30	19 (15.1%)
Highest degree		31–40	6 (4.8%)
Ph.D.: clinical psychology	49 (38.9%)	40+	79 (62.7%)
Ph.D.: counseling psychology	3 (2.4%)	Usual duration of treatment
Ph.D.: social work	2 (1.6%)	Less than 3 months	10 (7.9%)
Psy.D.	12 (9.5%)	3–6 months	26 (20.6%)
M.D.	2 (1.6%)	6 months to a year	30 (23.8%)
LMFT	4 (3.2%)	About a year	24 (19.0%)
MSW/LCSW	15 (11.9%)	1–2 years	23 (18.3%)
MA/MS: counseling	11 (8.7%)	More than 2 years	10 (7.9%)
LPC	5 (4.0%)	Prefer not to respond	3 (2.4%)
LMHC	6 (4.8%)	Clinical setting
Ph.D.: school psychology	1 (0.8%)	Private practice	69 (54.8%)
Other	16 (12.7%)	College counseling center	2 (1.6%)
Influence of theoretical orientations (% of 100)		Community clinic	7 (5.6%)
Psychodynamic	5.98 (11.81)	Academic medical center	24 (19.0%)
Behavioral	38.03 (23.50)	Mental health clinic	1 (0.8%)
Cognitive	30.20 (17.36)	Partial hospital program	5 (4.0%)
Experiential/humanistic	5.79 (11.02)	Inpatient psychiatric unit	1 (0.8%)
Other	7.24 (16.12)	Research/academia	10 (7.9%)
Family systems	11.94 (18.97)	Intensive outpatient program	1 (0.8%)
Cognitive and behaviorala	34.12 (14.30)	Residential treatment center	1 (0.8%)

Note. M = mean; SD = standard deviation; ED = eating disorder.

^aIn a manner consistent past two-way bridge studies, we had participants rate cognitive and behavioral influences on their practices separately. However, in response to a reviewer comment highlighting how these orientations are often integrated within cognitive-behavioral therapy (CBT) treatments (including in CBT for EDs), we also summed the cognitive and behavioral ratings for each participant and report on the mean of those sums above.

Participants were recruited through e-mailing academic and clinical listservs, including the Academy for Eating Disorders, the Association for Behavioral and Cognitive Therapies, and the Society for a Science of Clinical Psychology. Participants were also recruited through postings on social media platforms (i.e., Twitter, Facebook), direct e-mailing of clinicians on websites for finding therapists (e.g., Psychology Today), and clinical internships/Ph.D. program listservs (for a full list of recruitment sources, see Appendix A).

MEASURES

Clinician Survey on Experiences Delivering CBT Survey Development.

While some research has begun to examine the gap between ED research and therapist drift from empirically supported treatments (Waller, 2009), to our knowledge, there is no current validated research measure to assess clinicians’ experiences implementing empirically supported treatments for EDs. Therefore, in a manner parallel to past two-way bridge studies (Jacobson et al., 2016; Wolf & Goldfried, 2014), A.N.G. and E.E.R. developed a list of survey items using two commonly used CBT manuals (Fairburn, 2008; Waller et al., 2007), and then held a structured group feedback session with experienced ED clinicians (see Appendix B) to edit and develop our questionnaire items. When possible, questions were modeled after previous two-way bridge study surveys (Jacobson et al., 2016; Szkodny et al., 2014; Wolf & Goldfried, 2014). After completion of the interview session, a list of questionnaire items was distributed to clinicians to provide an opportunity for one more round of feedback. The final survey included clinician demographics, frequency of use and usefulness of CBT therapeutic techniques, variables that may limit treatment success, and perceived limitations of the CBT intervention.

Frequency/Usefulness of CBT Technique Use.

Items gauging frequency of use of CBT techniques were rated twice. First, participants rated each item on a Likert-type scale regarding frequency of use, ranging from 0 (never) to 100 (always); second, participants were asked to rate how useful they find each technique on a scale from 0 (not at all) to 100 (very useful). Although prior iterations of this study used a dichotomous rating for each technique (i.e., yes or no; Szkodny et al., 2014; Wolf & Goldfried, 2014), expert clinician feedback prompted us to shift our response format to provide a more nuanced assessment of frequency and usefulness to guide future research.

Barriers to CBT Treatment Success.

Patient-focused variables that limited treatment success were rated on a Likert-type scale from 0 (not at all) to 100 (is a major challenge) and were drawn both from previous two-way bridge surveys, as well as from expert feedback during survey development. Items fell into the following categories: (a) patient’s ED symptoms, (b) patient’s other problems/characteristics, (c) patient’s treatment expectations, (d) patient’s beliefs about the nature of EDs, (e) patient’s motivation, and (f) patient’s social systems.

Problems/Limitations of CBT.

Finally, clinicians were also asked to select problems/limitations associated with the CBT intervention (e.g., “select all that apply”). These items were drawn both from past two-way bridge studies in other fields (Jacobson et al., 2016; Szkodny et al., 2014; Wolf & Goldfried, 2014), as well as from expert feedback during survey development.

PROCEDURES

Participants were provided with a direct link to the survey in recruitment materials. Following the provision of informed consent, participants were asked to fill out demographic information and whether they have ever used CBT to treat patients with EDs. All participants who answered yes were then directed to the above-described clinician survey.

DATA CLEANING

To protect against bots and increase data integrity, participants were allowed to take the survey only once from a specific IP address. We also put one attention check item to ensure that participants were paying attention throughout the survey. One participant was removed for failing the attention check.

DATA ANALYSIS

Consistent with previous research in this area (Wolf & Goldfried, 2014), we first report descriptive statistics for most and least frequently endorsed clinical techniques used in CBT for EDs, alongside ratings regarding the usefulness of these techniques. We then explore mean levels of endorsement for treatment barriers in the eight domains listed above and explore the most and least frequently endorsed barriers to implementation of these treatments in each of the domains, as well as most frequently endorsed limitations of CBT.

Results

FREQUENCY OF USE OF CBT TECHNIQUES

Preliminary results (see Table 2 for a full description) indicated a wide range of mean frequencies for CBT technique use, ranging from infrequent technique use (e.g., M = 26.48, SD = 31.30) to almost always using this technique (M = 91.65, SD = 19.93). Notably, 81% of the techniques listed were used by clinicians over 50% of the time and half of the techniques were rated as being used over 75% of the time, indicating overall high levels of usage. Generally, techniques rated as more frequent were those that focused on behavioral change (e.g., regular eating) and/or represented general CBT techniques (rather than those specific to CBT for EDs). Psychoeducation was the most frequently endorsed CBT technique, followed by monitoring progress/checking in about progress throughout treatment, and promoting regular eating. The least frequently used techniques were surveys to address mind reading, followed by using imagery to rescript body image memories, and the interpersonal difficulties module of CBT-E.

Table 2.

Therapist-Reported Frequency of CBT Technique Use and Technique Usefulness

Frequency of technique use (0 = never use, 100 = always use)
Technique	M (SD)	Technique	M (SD)
Psychoeducation	91.65 (19.93)	Identifying, tracking, and challenging body image sx	74.93 (26.60)
Monitoring progress throughout treatment	88.96 (20.90)	Exposure exercises	71.98 (28.02)
Promoting regular eating	87.58 (22.35)	Motivational enhancement	70.22 (29.60)
Self-monitoring behavior and cognitions	85.83 (21.31)	Behavioral experiments	68.28 (32.35)
Assigning out-of-session homework	85.07 (22.86)	Stimulus control for ED behaviors	63.52 (28.20)
Identifying momentary antecedents and consequences of ED sx	84.18 (19.47)	Interpersonal skills training	61.52 (29.51)
Relapse prevention	83.54 (23.44)	Incorporating loved ones	55.85 (30.13)
Identifying and challenging dietary rules	83.37 (23.75)	In-session weighing	54.67 (39.96)
Developing a personalized formulation of the client’s ED	82.58 (26.17)	Perfectionism module	38.30 (34.13)
Cognitive restructuring of negative/distorted beliefs	81.70 (24.59)	Low self-esteem module	36.21 (34.93)
Increasing importance of other life domains	81.01 (23.82)	Mood intolerance module	35.77 (34.50)
Emotion regulation skills	77.95 (26.80)	Interpersonal difficulties module	34.44 (34.21)
Problem solving	76.82 (26.45)	Using imagery to rescript body image memories	27.47 (32.30)
Increasing awareness of shape and weight overevaluation	76.19 (28.05)	Surveys to address “mind reading”	26.48 (31.30)
Usefulness of techniques (0 = not at all useful, 100 = extremely useful)
Technique	M (SD)	Technique	M (SD)
Psychoeducation	91.29 (17.72)	Cognitive restructuring of negative/distorted beliefs	75.36 (28.53)
Monitoring progress throughout treatment	87.33 (23.06)	Identifying, tracking, and challenging body image sx	71.92 (28.92)
Promoting regular eating	87.31 (21.28)	Motivational enhancement	70.55 (31.40)
Identifying momentary antecedents and consequences of ED sx	84.37 (20.57)	Increasing awareness of shape and weight over evaluation	69.98 (30.39)
Developing a personalized formulation of the client’s ED	83.69 (22.07)	Stimulus control for ED behaviors	65.25 (28.33)
Identifying and challenging dietary rules	83.31 (22.63)	Interpersonal skills training	64.79 (28.64)
Self-monitoring behavior and cognitions	82.53 (23.95)	Incorporating loved ones or caregivers in treatment	63.10 (30.51)
Increasing importance of other life domains	82.50 (23.46)	In-session weighing	55.13 (37.68)
Relapse prevention	80.85 (26.31)	Perfectionism module	45.00 (36.45)
Exposure exercises	80.48 (25.12)	Mood intolerance module	43.62 (37.03)
Assigning out-of-session homework	79.64 (27.42)	Low self-esteem module	42.90 (35.98)
Emotion regulation skills	78.82 (24.84)	Interpersonal difficulties module	39.14 (34.50)
Problem solving	76.72 (26.42)	Using imagery to rescript body image memories	32.56 (33.67)
Behavioral experiments	75.52 (27.89)	Surveys to address “mind reading”	32.44 (33.34)

Note. CBT = cognitive-behavioral therapy; M = mean; SD = standard deviation; sx = symptoms; ED = eating disorder.

USEFULNESS OF CBT TECHNIQUES

A wide range of usefulness ratings for differing CBT techniques were endorsed, ranging from M = 32.44 (SD = 33.44) to M = 91.29 (SD = 17.72)—however, 81% of techniques had a mean endorsement of usefulness rating over 50 and 53% received a mean rating of 75 or higher, indicating clinicians found the majority of techniques useful. In a similar manner to frequency ratings, general CBT strategies (e.g., relapse prevention) and self-monitoring strategies were rated as most useful, rather than affective or cognitive strategies. The most useful CBT technique was psychoeducation, followed by monitoring progress/checking in about progress throughout treatment, and promoting regular eating/meal planning. The CBT techniques with the lowest usefulness ratings were the interpersonal difficulties module of broad CBT-E, surveys to address mind reading, and using imagery to rescript body image memories (see Table 2 for a full description of results).

BARRIERS TO IMPLEMENTING CBT

A wide range of ratings for barriers encountered in treatment were endorsed (see Table 3), ranging from M = 24.40 (SD = 25.99) to M = 77.27 (SD = 25.55). Of the 69 items rated as barriers in using CBT-E for the treatment of EDs (rated on a Likert-type scale from 0 = not at all to 100 = is a major challenge), most barriers received mean ratings toward the middle of the scale, suggesting some impact on treatment, with very few barriers receiving extremely low or extremely high mean problem ratings. The highest-rated barriers to treatment generally were those related to ED symptoms (e.g., “unwillingness to follow a meal plan”), beliefs about the nature of EDs (e.g., “internalized diet culture”), and some treatment expectations (e.g., “only willing to give up part of the ED”). Some factors external to EDs, including co-occurring personality disorders, also received higher mean ratings.

Table 3.

Therapist-Rated Barriers to Implementing CBT

Barrier	M (SD)	Barrier	M (SD)
ED symptoms		Patient characteristics
Unwillingness to follow a meal plan	77.27 (25.55)	Refusal to complete homework	72.01 (26.96)
Severity of ED symptoms (sx)	73.82 (26.03)	Personality disorders	68.43 (25.87)
Medical instability/risk	73.42 (27.84)	Substance use	62.09 (30.43)
Chronicity	71.92 (25.42)	Low distress tolerance	60.80 (28.51)
Perseveration	58.53 (27.77)	Resistance to directedness of treatment	60.09 (29.53)
Subtypes and/or diagnosis	49.58 (28.30)	Comorbid symptoms	59.25 (26.40)
Social systems		Obsessive-compulsive disorder	57.86 (25.82)
Family reinforces weight stigma	73.18 (25.43)	Life stressors	56.81 (25.83)
Symptoms are reinforced/supported	64.11 (26.80)	Depression	55.42 (27.56)
Family does not support treatment	58.68 (33.02)	Lack of financial resources	54.47 (33.12)
Social isolation of patient	55.87 (30.39)	Posttraumatic stress disorder	54.10 (27.42)
Family history of EDs	49.20 (30.93)	Anxiety	51.93 (28.35)
Loss of family member/partner/job	39.44 (26.08)	Functional impairment	51.23 (29.97)
Patient motivation		Alexithymia (emotion identification)	50.10 (30.18)
Premature termination	56.34 (33.69)	Perfectionist/obsessive style	49.13 (28.17)
Minimal motivation at outset	53.10 (29.53)	Intellectual/cognitive ability	45.89 (31.66)
Motivation ↓: no focus on weight loss	48.20 (29.89)	Low self-esteem/self-efficacy	43.33 (29.06)
Motivation ↓: sx improvement	43.69 (28.09)	Dependency/unassertiveness	37.80 (27.82)
Motivation ↓ : learns reasons for ED	25.14 (24.85)	Physical problems	31.94 (26.29)
Treatment expectations		Beliefs about the nature of EDs
Weight gain/fluctuation unacceptable	71.01 (28.83)	Internalized diet culture	71.05 (25.49)
Only willing to give up part of ED	68.04 (26.74)	Internalized weight stigma	67.97 (25.58)
Behavior change (BxΔ) not required	62.63 (31.57)	Regular eating = catastrophic weight gain	67.08 (30.13)
Wants ↓ ED thoughts before ΔED bx	55.63 (32.25)	ED gives sense of achievement/control	63.23 (30.05)
BxΔ not possible until thoughts/urges/affect↓	54.89 (33.35)	ED is part of identity/unchangeable	57.46 (30.62)
Therapist will do all the work	53.73 (29.68)	Fears related to weight gain are realistic	56.82 (29.94)
Co-occurring sx are more important	52.23 (30.53)	ED makes their life better	55.54 (31.63)
Will be free of all ED symptoms	48.75 (29.14)	Self-monitoring = ↑ weight/shape concern	47.14 (33.23)
Pessimism due to past therapy	48.22 (27.87)	ED is due to external factors	35.79 (26.29)
Desire to focus only on past events	48.20 (31.37)	EDs are a choice	31.49 (29.67)
Treatment will be brief and easy	43.37 (27.87)	ED are caused by sociocultural stressors	30.30 (27.92)
Symptom reduction not enough	42.16 (27.99)	EDs are biologically based	24.40 (25.99)

Note. CBT = cognitive-behavioral therapy; M = mean; SD = standard deviation; ED = eating disorder; sx = symptom; bx = behavior; ↓ = decrease; ↑ = increase; Δ = change. Barriers rated on a Likert-type scale: 0 = not at all, 100 = is a major problem.

PROBLEMS/LIMITATIONS OF CBT

The percentage of participants endorsing problems/limitations of CBT ranged from M = 5.3% (n = 9) to M = 26.6% (n = 45), indicating that overall, clinicians infrequently endorsed problems/limitations with the treatment itself. Of the 10 items listed as problems/limitations of CBT for ED (see Table 4), too little guidance on treating co-occurring symptoms was the most frequently endorsed problem/limitation to treatment (26.6%), followed by absence of guidelines for responding to resistance/noncompliance (24.9%), and insufficient focus on affect tolerance/regulation (23.7%).

Table 4.

Therapist-Endorsed Problems and/or Limitations of CBT for EDs

Problems and limitations	n (%) endorsed
Too little guidance on treating co-occurring symptoms	45 (26.6%)
Absence of guidelines for responding to resistance/noncompliance	42 (24.9%)
Insufficient focus on affect tolerance/regulation	40 (23.7%)
Not enough time for patient to respond to treatment within time frame of manual	38 (22.5%)
The expected frequency of sessions and/or length of treatment is unrealistic for my patients	34 (20.1%)
Insufficient guidance regarding nutritional education	25 (14.8%)
Too little guidance regarding coordination with other providers	24 (14.2%)
Treatment instructions challenging to implement logistically	22 (13.0%)
Too much between session homework assigned	9 (5.3%)
Prefer not to respond	5 (3.0%)

Note. CBT = cognitive-behavioral therapy; EDs = eating disorders.

Discussion

Given a well-documented disconnect between controlled clinical research and naturalistic treatment implementation (Goldfried et al., 2014), the current study sought to characterize utilization of CBT techniques for EDs in real-life clinical practice and identify barriers and limitations of existing treatments. This study extends past literature exploring the use of CBT techniques in EDs (Brown & Nicholson Perry, 2018; Mulkens et al., 2018; Waller et al., 2012) through evaluating therapist endorsement of usefulness of these techniques and by identifying patient-related characteristics that serve as barriers to implementing empirically supported treatments.

Interpretation of our results and potential implications for clinical practice are best placed within the context of our sample characteristics. Specifically, our sample primarily comprised early career clinical psychology Ph.D.’s—further, a small, but significant number of the sample reported having seen 20 or fewer ED patients in their practice (potentially secondary to being early career). While most existing studies in this area have also reported high representation of psychologists, our study reported fewer years of experience than other samples in the literature (e.g., Waller et al., 2012). Therefore, any comparisons noted below between our findings and past work could be secondary to these differences. As we discuss further in the limitations, our work must be followed by replications in samples of clinicians with more diverse training experiences, degrees held, and experience with ED populations.

With this context in mind, our first aim was to replicate and extend findings from past work on the frequency of CBT technique use in treatment for EDs (Brown & Nicholson Perry, 2018; Mulkens et al., 2018; Waller et al., 2012) through characterizing both the frequency and usefulness of a range of CBT techniques. Overall, our findings indicate high average use ratings for CBT techniques, as well as high mean levels of technique usefulness. Specifically, around half of assessed techniques had a mean frequency/usefulness over 75, indicating generally frequent use and high perceived usefulness. However, many ratings also demonstrated high standard deviations, suggesting notable variability. Considering past findings, our ratings of CBT technique use were higher than those observed in past two-way bridge studies in other areas of psychopathology (Jacobson et al., 2016; Szkodny et al., 2014; Wolf & Goldfried, 2014).

Within past research on CBT use in EDs, frequency ratings for CBT technique use were strikingly similar to those reported by Brown and Nicholson Perry (2018), and some techniques, including regular eating, were endorsed with a similar frequency to those reported in other studies on CBT for EDs (Cowdrey & Waller, 2015). On the other hand, at first glance, other techniques, including exposure exercises and cognitive restructuring, appear to be endorsed more frequently than those reported by other research (Cowdrey & Waller, 2015; Mulkens et al., 2018; Waller et al., 2012) with mean levels of usefulness above 75 (out of 100). It could be the case that our findings reflect increasing use of empirically supported techniques in therapist samples and success of ongoing efforts to increase CBT use (e.g., Fairburn & Wilson, 2013).

Alternatively, it should be noted that different studies have used a range of different metrics for frequency of technique use (i.e., frequency of clinicians reporting use across different proportions of patients; frequency of samples reporting that they “always” use specific techniques), which may account for some differences in findings. Further, past studies on CBT technique use included samples of clinicians reporting a greater number of years practiced than our sample, potentially reflecting a documented increase in drift following more years of experience and higher age (Speers et al., 2022). Altogether, while high rates of use across the sample are somewhat promising, mean use and usefulness ratings ranging from 26 to 92 (and large variability in these means), suggest that there remains a subset of patients with whom many clinicians do not employ therapeutic techniques outlined in empirically supported treatments. Accordingly, further research is necessary to better characterize clinical decision making that may influence technique use, as well as continued efforts toward addressing perceived barriers and/or beliefs that may affect patterns in use.

Regarding patterns in specific technique ratings, we observed that while ED-specific CBT techniques were rated as frequently used in treatment and useful (e.g., regular eating), clinicians also frequently incorporated more general CBT techniques (e.g., cognitive restructuring; emotion regulation skills) that may not be specifically outlined in CBT for ED manuals (Fairburn, 2008; Waller et al., 2007). This finding is consistent with past research (Brown & Nicholson Perry, 2018). Several techniques rated relatively less frequently used and/or useful were modules from the broad version of CBT-E (i.e., interpersonal difficulties, low self-esteem, perfectionism); low endorsement of these techniques may represent adherence to published recommendations. Specifically, given that the broad version of CBT-E is suggested for use only in patients who have clinically significant, treatment-interfering levels of these challenges (Fairburn, 2008), infrequent use may not indicate therapist “lack of adherence” to the treatment per se, and instead represent a discerning selection of modules based on patient profile. However, this interpretation is speculative and must be tested directly in future research.

Two of the techniques reported as least frequently used and less useful were surveys to address mind reading and using imagery to rescript body image memories. Consistent with our findings, past research has also suggested low rates of survey use in therapists who treat EDs (Cowdrey & Waller, 2015; Mulkens et al., 2018, Waller et al., 2012), despite some empirical support for this technique (Waller et al., 2007). There are several potential reasons for low endorsement of use and usefulness of these strategies. First, surveys and rescripting of body image memories are from Waller and colleagues’ (2007) CBT manual but are not included in Fairburn’s (2008) CBT-E manual. Although we did not assess the specific manuals clinicians used, it may be the case that clinicians surveyed in our study were primarily using CBT-E with patients, but we cannot test this possibility directly. Alternatively, both surveys to address mind reading and using imagery to rescript body image memories requires clinicians to address body image-related concerns, which previous research has identified to be the area of treatment that clinicians reported being the most nervous to address (Turner et al., 2014). Therefore, it could be the case that clinicians have less comfort with techniques used to target body image. Indeed, techniques targeting shape and weight concern generally received mean use and usefulness ratings below techniques targeting eating behaviors, such as regular eating, challenging dietary rules, and self-monitoring.

In our sample, in-session weighing received frequency and usefulness ratings around 50, with high variance across the sample, despite the inclusion of this technique in all CBT manuals for EDs (Fairburn, 2008; Waller et al., 2007). Although our mean level of patient weighing was higher than that reported in several past studies (Mulkens et al., 2018; Waller et al., 2012), high levels of variability in mean ratings may reflect a bimodal distribution in a manner to past work (Mulkens et al., 2018; Waller et al., 2012), with therapists reporting either extremely consistent or inconsistent use of weighing. A range of past empirical and theoretical work has discussed potential barriers to weighing patients (Waller, 2016; Waller & Mountford, 2018); future work might consider expanding on these findings to explore how clinicians make decisions to weigh (or not weigh) their patients, as well as how consistently these policies are applied in practice. Additionally, consistent with past findings (Brown & Nicholson Perry, 2018; Mulkens et al., 2018; Waller et al., 2012), our sample reported relatively frequent use of motivational enhancement techniques, despite limited and inconsistent data supporting this approach (Waller, 2012).

Our findings also highlight clinician perception of patient-focused barriers that interfere with disseminating CBT for EDs. In our sample, the highest-rated perceived barriers to implementing CBT (i.e., those that interfere most frequently) were related to ED symptom severity (e.g., unwillingness to follow a meal plan, symptom severity, and medical instability), beliefs about EDs (e.g., internalized diet culture), and family beliefs about weight and eating (i.e., family reinforcing weight stigma). Factors related to symptoms outside of the ED (e.g., OCD) and/or more structural in nature (e.g., lack of financial resources) received ratings around the midpoint of the response scale, suggesting some impact, yet relatively less than the impact of ED-related factors.

To our knowledge, few studies have explored clinician-reported barriers to implementation of CBT for EDs; our findings thus provide a complement to past work (Waller et al., 2016) demonstrating a range of clinician factors that can influence technique use. Results also suggest that future work is necessary to better understand these self-reported barriers. For instance, it would be important to further characterize how barriers interfere with response to CBT and clinicians’ views on what resources may be most useful in addressing these barriers (e.g., increased availability of training resources focused on reported barriers, addendums to manuals that highlight creative ways to address barriers while remaining adherent, webinars wherein clinicians can discuss creative solutions with researchers who can implement empirical tests of commonly used solutions).

Of the problems/limitations of CBT, the highest endorsed problems and limitations were only endorsed by ~25% of therapists, indicating that clinicians report infrequent limitations of CBT. Notably, previous two-way bridge studies had higher endorsement of problems/limitations, ranging from 38% to 56% (McAleavey et al., 2014; Szkodny et al., 2014; Wolf & Goldfried, 2014). In our sample, the most frequently endorsed issue with CBT was “too little guidance on treating co-occurring symptoms.” Given that up to 65% of individuals with an ED will have either an Axis II diagnosis/and or a co-occurring anxiety disorder (Khosravi, 2020; Swinbourne et al., 2012) and meta-analytic data suggesting relationships between comorbidity and outcome (Vall & Wade, 2015), it is important that clinical research characterize (a) how clinicians may be adapting treatment to address co-occurring concerns in ED populations, and (b) whether these approaches are effective.

In addition to treatment of co-occurring symptoms, more than 20% of the sample endorsed barriers related to the length of treatment (i.e., needing more time in treatment than described in the manual) and insufficient focus on how to deal with affect regulation and/or noncompliance. Again, while the infrequency of these ratings is cause for cautious optimism, these areas also offer important information for researchers in CBT for EDs regarding areas that may be important for future investigation and incorporation into manuals and/or treatment guides.

STRENGTHS AND LIMITATIONS

Despite several strengths of our investigation, our study has limitations that are important to note. First, consistent with other prior two-way bridge studies, therapists’ self-report of technique use may be unreliable, and adherence should be assessed by using other modalities (e.g., behavioral coding). Future research might consider exploring the relationship between self-reported use of techniques and behavioral measures of adherence to inform how closely self-reported adherence relates to other modalities of assessment. Second, most respondents to our survey had a Ph.D. in clinical psychology or were in a Ph.D. training program. While we attempted to recruit a wide range of mental health professionals through using public websites for finding therapists (e.g., Psychology Today), there exist other databases for ED-related care that were not included (e.g., International Association of Eating Disorder Professionals). Inclusion of a wider range of databases for recruitment may have resulted in greater representation of master’s-level providers. It is critical that future research test whether patterns of CBT technique use vary when sampling across providers with different degrees and training experiences.

Third, this survey was administered during the COVID-19 pandemic (i.e., data collection began in summer of 2021); therefore, reports from respondents may reflect alterations in delivery of CBT in response to pandemic restrictions (e.g., in-session weighing may not have been possible during virtual sessions; Murphy et al., 2020). Fourth, since the authors did not collect data on where participants were recruited from, or their specific CBT-related training, the authors are unable to explore any relationship between recruitment source or participant training that may have influenced the results. As noted above, the study primarily consisted of early career therapists who may have had more recent training and/or increased training in newer published CBT manuals (e.g., Waller et al., 2018). In support of this possibility and consistent with our results, there is some evidence to suggest that younger and less experienced therapists are more likely to adhere closely to empirically supported treatment protocols (Waller et al., 2012).

CONCLUSIONS

In conclusion, this study builds upon previous literature by exploring both frequency and useful ness of empirically supported treatment techniques and identifying the most prevalent barriers to implementing ED treatments in real-world settings (Jacobson et al., 2016; Szkodny et al., 2014; Waller et al., 2012). Our findings highlight that while therapists endorsed overall high usage of most CBT techniques, several techniques (e.g., surveys to address mind reading) are still not frequently used. Furthermore, therapists endorsed minimal limitations of the treatment itself but did endorse a range of patient-related characteristics that serve as barriers to treatment implementation. Future research should explore novel adaptations to existing treatments that better account for the complex barriers therapists face in delivering treatment in real-world settings.

Appendix A. Clinician Recruitment Sites Internship/Training Programs

• Memorial University

• Michigan State University

• Montclair State University

• Ohio University

• PGSP–Stanford

• Rutgers

• Ryerson University

• SDSU/UCSD joint program

• Simon Fraser University

• St. Louis University

• SUNY–Albany

• SUNY–Buffalo

• Temple University

• University of Calgary

• University of Hawaii

• University of Iowa

• University of Kansas

• University of Kentucky

• University of Louisville

• University of Miami in Ohio

• University of Missouri–Kansas City

• University of Montana

• University of North Carolina–Chapel Hill

• University of Oregon

• University of Pittsburgh

• University of South Florida

• University of Tulsa/Laureate Institute for Brain Research

• University of Waterloo

• University of Western Ontario

• University of Wyoming

• Virginia Commonwealth University

• Virginia Consortium

• Washington University in St. Louis

• Yeshiva University

Professional Organizations/Websites

• Association for Cognitive and Behavioral Therapies

• Society for a Science of Clinical Psychology

• Academy for Eating Disorders

• Psychology Today

• Alma

Appendix B. Eating Disorder Consultant Clinicians and Affiliations

• Sasha Gorrell, Ph.D.—Assistant Professor at University of California, San Francisco (UCSF)

• Marita Cooper, Ph.D.—Research Postdoctoral Fellow at the Children’s Hospital of Philadelphia

• Lisa Anderson, Ph.D.—Assistant Professor at University of Minnesota

• Nicholas Farrell, Ph.D.—Clinical Psychologist at Rogers Behavioral Health

• Jamal Essayli, Ph.D.—Assistant Professor at Penn State University

• Katherine Schaumberg, Ph.D.—Assistant Professor in the Department of Psychiatry at University of Wisconsin—Madison