Introduction
Eating disorders (ED) are characterised by extreme weight, shape and eating concerns and associated behaviours, which often have significant medical and psychological sequelae (Hudson et al., Reference Hudson, Hiripi, Pope and Kessler2007). The impact of eating disorders is associated with impairment across psychosocial domains for both individuals with the condition and their families (Swanson et al., Reference Swanson, Crow, Le Grange, Swendsen and Merikangas2011). Recent evidence suggests that the prevalence of eating disorders has become more common over the last 20 years, and the average age of onset of illness has become younger (Galmiche et al., Reference Galmiche, Déchelotte, Lambert and Tavolacci2019; Herpertz-Dahlmann and Dahmen, Reference Herpertz-Dahlmann and Dahmen2019).
Among adolescent eating disorder populations, the effectiveness of family-based treatment (FBT) has been comprehensively established across several randomised clinical trials (RCTs) among adolescents with anorexia nervosa and bulimia nervosa (Le Grange et al., Reference Le Grange, Hughes, Court, Yeo, Crosby and Sawyer2016; Lock et al., Reference Lock, Le Grange, Agras, Moye, Bryson and Jo2010; Madden et al., Reference Madden, Miskovic-Wheatley, Wallis, Kohn, Lock, Le Grange, Jo, Clarke, Rhodes, Hay and Touyz2015). However, whilst around three-quarters demonstrate some treatment response, less than half of those receiving FBT achieve full recovery, indicating the need for improved or alternative treatments (Lock and Le Grange, Reference Lock and Le Grange2019). For various reasons, some adolescents’ families are unable, unavailable, or unwilling to participate in FBT, leaving a significant gap in treatment options for this population.
Although less studied than FBT, the effectiveness of enhanced cognitive behaviour therapy (CBT-E) adapted for adolescents with eating disorders has been demonstrated in a series of cohort studies (Dalle Grave et al., Reference Dalle Grave, Calugi, Doll and Fairburn2013; Dalle Grave et al., Reference Dalle Grave, Calugi, Sartirana and Fairburn2015; Dalle Grave et al., Reference Dalle Grave, Conti and Calugi2020; Fursland and Byrne, Reference Fursland and Byrne2013). CBT-E and FBT have been compared in a non-randomised effectiveness trial of adolescents with a restrictive eating disorder (excluding avoidant and restrictive food intake disorder; Le Grange et al., Reference Le Grange, Eckhardt, Dalle Grave, Crosby, Peterson, Keery, Lesser and Martell2020). Results showed that the slope of weight gain was significantly higher for FBT compared with CBT-E, but not at follow-up. There was no difference between the treatments with regard to improvements in eating disorder psychopathology across time points, highlighting the treatments were similar on most clinical outcomes.
There are several notable differences between FBT and CBT-E (for detailed discussion, see Dalle Grave et al., Reference Dalle Grave, Eckhardt, Calugi and Le Grange2019; Dalle Grave, Reference Dalle Grave2023). In FBT the ED is conceptualised according to the medical or disease model, in which ED features (e.g. restriction, concerns about weight, shape and eating) are considered symptoms of an illness (i.e. the ED). The young person is thought to be controlled by an illness which is separate or external from the individual. According to this model, the support of parents or significant others are required to regain control over the illness by ensuring the young person follows the treatment prescribed by the FBT clinician. As such, parents are vital to this process and play the major active role in treatment and the young person has a much more passive role. A significant advantage of this model is the agnostic and external view of the ED that absolves the parent or young person of any blame as to the development of the eating disorder and motivates the resources of the entire family to help the young person recover.
CBT-E, on the other hand, adopts a transdiagnostic psychological model that views the illness as belonging to the individual. According to this model, core psychopathology based on a dysfunctional self-evaluation schema reliant on excessive control of eating weight and shape gives rise to a range of ED symptoms which interact to maintain extreme concerns about eating, weight and shape. In this model, recovery involves the young person gaining an understanding of the maintaining mechanisms at play (regardless of diagnosis) and making the decision to change using a set of individualised strategies. The young person takes an active role in working with the CBT-E clinician to understand their eating disorder and learn how to control their behaviour and mindset, with parents or carers playing a supportive role which is useful but not essential. A significant advantage of the psychological model is that it enables the young person to understand that their method of self-evaluation (i.e. based on controlling eating, weight and shape) is dysfunctional and detrimental in the long term, and learn other ways to develop a more functional self-evaluation schema that works in the long term (Dalle Grave, Reference Dalle Grave2023).
These differences in the conceptualisation of EDs, nature of involvement of parents and the child/adolescent, therapeutic strategies and procedures, and proposed mechanisms of action (Dalle Grave et al., Reference Dalle Grave, Eckhardt, Calugi and Le Grange2019) hold promise that either treatment may be effective where the other is contraindicated, not accepted, or has not achieved full recovery. However, there is limited evidence as to whether those who have not achieved recovery in either FBT or CBT-E are more or less likely to benefit from the alternate therapy. This is important, as given the remission rates of both therapies (CBT-E=20–45%, FBT=22–49%: Dalle Grave et al., Reference Dalle Grave, Calugi, Doll and Fairburn2013; Dalle Grave et al., Reference Dalle Grave, Eckhardt, Calugi and Le Grange2019; Lock and Le Grange, Reference Lock and Le Grange2019), evidence-based follow-up treatment recommendations following sub-optimal response are crucial to optimise early treatment(s) to prevent an elongated duration of illness.
One study has investigated the effectiveness of CBT for eating disorders among an adolescent cohort (52% anorexia nervosa, 31% atypical anorexia nervosa) consisting of a group for whom FBT had been undertaken but had not achieved full recovery (n=32), and a group for whom FBT had not been undertaken due to non-acceptance or engagement in the treatment (n=22; Craig et al., Reference Craig, Waine, Wilson and Waller2019). Results showed that both groups exhibited significant reductions in eating disorder psychopathology and clinical impairment, with moderate to large effect sizes. However, there was no between group analysis to compare if the effect of the intervention varied depending on previous engagement in FBT. Another limitation of this study was the intervention used a combination of protocols from the Fairburn et al. (Reference Fairburn, Cooper, Shafran and Wilson2008) and Waller et al. (Reference Waller, Cordery, Corstorphine, Hinrichsen, Lawson, Mountford and Russell2007) manuals, which limits the ability to draw conclusions, replicate or disseminate the intervention given the combined approach.
The current pilot study aimed to compare the relative effectiveness of CBT-E for adolescents between those who had completed FBT but had not achieved full recovery and those who had no prior FBT engagement. In this effectiveness design, outcome variables of eating disorder psychopathology and clinical impairment were evaluated for adolescents completing manualised CBT-E (Dalle Grave and Calugi, Reference Dalle Grave and Calugi2020) in a real-world public specialist eating disorders clinic.
Method
Participants
Participants consisted of a transdiagnostic eating disorder sample of 69 young people (female=95%, mean age 15.77, SD=1.13, range 13–17) who had been referred for treatment to a public out-patient child and youth specialist eating disorder service which sees young people up until the age of 18 (no lower age limit). Further details of the service have been described in Lim et al. (Reference Lim, White, Withington, Catania, Wilson, Knight, Rees, Middeldorp and Krishnamoorthy2023). The no previous FBT group consisted of 38% atypical anorexia, 24% bulimia nervosa, 24% anorexia nervosa, and 14% eating disorder, unspecified according to the International Classification of Diseases 10th edition. The group with previous FBT consisted of 96% anorexia nervosa and 4% bulimia nervosa. Diagnosis was taken at the initial intake assessment at the clinic (i.e. prior to starting either FBT or CBT-E) by expert clinicians, which may explain the higher than expected starting body mass index (BMI) centile of the previous FBT group (given high anorexia nervosa diagnosis rate), who may have achieved some weight regain during their FBT treatment prior to starting CBT-E. Additionally, the clinic utilises the ICD 10th edition, in which young people who would not meet the anorexia nervosa low weight/BMI criteria of Diagnostic and Statistics Manual-5, but are 15% below expected body weight (through weight loss and failure to achieve) are given a diagnosis of anorexia nervosa. BMI centile change across treatment is reported but is not included as a main outcome variable due to the variability in eating disorder diagnoses and lack of information as to weight history and the low number of low weight patients.
Note that the main analysis was run excluding the those with a BMI centile less than 25 with no difference to the interpretation of the results. Similarly, the main analysis was run with and without controlling for diagnosis as a covariate, with no difference to the interpretation of the results. Splitting the groups into low vs higher weight groups as per previous studies (e.g. Le Grange et al., Reference Le Grange, Eckhardt, Dalle Grave, Crosby, Peterson, Keery, Lesser and Martell2020) was not considered as there were insufficient number, with samples sizes of seven and one young person having a BMI centile of less than 25 in the no previous FBT and previous FBT group, respectively.
Procedure
Protocol at the clinic is to offer FBT as front-line treatment, which is described to families by an accredited FBT clinician(s) who receive ongoing supervision by international experts in delivering FBT. For those for whom FBT was not accepted, contraindicated or unsuitable based on initial clinical assessment, CBT-E was offered as their initial treatment at the service (n=42). Frequent reasons for non-engagement with FBT included: parents/carers not being available to engage in treatment; disagreement with proposed FBT treatment; and/or young people being highly individuated from their parents.
The remaining 27 individuals consisted of those who had previously engaged in FBT at the clinic, which had been discontinued without full recovery. FBT is delivered by accredited FBT clinicians (masters or doctoral level social workers, psychologists, nurse practitioners or psychiatrists) who receive ongoing monthly supervision. The decision to discontinue FBT and offer CBT-E was based on clinical judgement by the existing FBT clinician, which was discussed at case review by a multi-disciplinary team with experience in eating disorder treatment. Reasons for FBT discontinuing was typically stalled progress which could be indicated by: limited success empowering parents to support regular eating with their young person; lack of compliance/availability of the family; limited reduction in eating disorder behaviour; psychopathology, distress and conflict within the family environment impacting upon engagement or progress; and/or evidence of highly individuated young people which presented a barrier to FBT processes. Exclusion criteria for starting CBT-E included: young person not agreeing to start CBT-E; medical or psychiatric instability that contraindicated out-patient treatment; and anything that would prevent regular attendance for the period of treatment.
Participants engaging in CBT-E received the manualised intervention (details outlined in the treatment manual: Dalle Grave and Calugi, Reference Dalle Grave and Calugi2020). The intervention is transdiagnostic, designed to address the core psychopathology of eating disorders over three main steps of treatment, which takes between 20 and 40 sessions depending on the young person’s need for weight regain and presence or absence of external maintaining mechanisms of their eating disorder psychopathology. Interventions were delivered by clinical psychologists or masters/doctoral level psychology students. Clinicians all had training in cognitive and behavioural interventions, general training in treating young people with eating disorders, and had completed CREDO online CBT-E training (Khera et al., Reference Khera, Viljoen and Murphy2021). Clinicians and students had regular individual supervision with a senior clinician (last author) experienced in CBT-E, as well as monthly group supervision with the developer of the adolescent manual for CBT-E. A single therapist delivered the intervention where possible with a substitute available if the primary clinician was unavailable or leaving the clinic (e.g. students ending placement). A consort diagram of the participant flow is shown in Fig. 1.
Figure 1. Participant flow during treatment.
Measures
Body mass index centiles
BMI centiles were calculated using individual’s weight, height, and the Centre for Disease Control and Prevention growth charts (www.cdc.gov/growthcharts). Measurements were taken in clinic.
Eating Disorders Examination Questionnaire (EDE-Q)
The EDE-Q adapted for youth was used (Goldschmidt et al., Reference Goldschmidt, Doyle and Wilfley2007). The EDE-Q is a 36-item measures that yields four subscales: Restraint (e.g. ‘Have you been trying to cut down on food to control your weight or shape?’); Eating Concern (e.g. ‘Have you been scared over losing control over weight?’); Shape Concern (e.g. ‘How unhappy have you felt about your shape?’); and Weight Concern (e.g. ‘Has your weight affected how you think about (judge) yourself as a person?’). Items are scored on a 7-point Likert scale (0–6) with higher scores indicating greater severity of eating disorder symptoms. Mean scores are used for each subscale. A global subscale score is derived by summing the mean scores from the four subscales and dividing by four. Prior studies have reported acceptable psychometric properties and excellent Cronbach’s alphas (0.93–0.96; Forsén Mantilla et al., Reference Forsén Mantilla, Birgegård and Clinton2017).
Clinical Impairment Assessment (CIA)
The CIA is a 16-item self-report measure of psychosocial impairment due to eating disorder features (Bohn et al., Reference Bohn, Doll, Cooper, O’Connor, Palmer and Fairburn2008). Items are scored on a 4-point Likert scale (0–3), and are summed to form a total score, with higher scores indicating higher level of impairment. Prior studies have reported acceptable psychometric properties and excellent Cronbach’s alphas (0.94; Reas et al., Reference Reas, Rø, Kapstad and Lask2010).
Analytic plan
Data were analysed using SPSS version 28. A two (time: pre-treatment vs post-treatment) × two (group: previous FBT vs no previous FBT) mixed model analyses of variance was used to test for differences on the main outcome variables of EDE-Q and CIA, as well as for BMI centile. Within-group effect sizes for completer and intention-to-treat (ITT) were calculated for each group (previous FBT, no previous FBT) using Cohen’s d, where the effect is derived from the mean at pre-treatment, minus the mean at post-treatment, divided by the pooled standard deviation. Treatment completion was defined by completing a full course of CBT-E of 20–40 sessions as indicated per the manual (Dalle Grave and Calugi, Reference Dalle Grave and Calugi2020). Independent sample t-tests were used to assess pre-treatment differences between groups for continuous measures, and chi-square was used for categorical data. Due to insufficient sample size, a low weight group analysis was not performed. The results of this cohort are described in the text. Reliable and clinically significant change for completers were examined as per recommendations by Jacobson and Truax (Reference Jacobson and Truax1991).
Three unique outliers were detected for the pre-treatment EDE-Q, CIA and BMI centile that were more than 1.5 box lengths from the edge of the box in a boxplot. Inspection of the values revealed them to be low scores (i.e. non-clinical) scores on the pre-measures and one low initial BMI centile (4th). As this is consistent with this population, and removal of the outliers did not affect the interpretation of the results, the outliers were retained. There were no outliers as assessed by examination of studentised residuals for values greater than ±3, normal distribution was confirmed by Q-Q plot. Levene’s test showed no violation of the assumptions of equality of variance. There was homogeneity of covariances as assessed by Box’s test of equality of covariance matrices (p=.53).
Results
Pre-treatment group differences
Pre-treatment measures for the sample are shown in Table 1. There were no significant differences for: age, t 67=–.06, p=.96, d=–.01; BMI centile, t 67=.11, p=.31, d=.46; CIA, t 65=1.95, p=.06, d=.49, or treatment completion χ2(1)=1.51, p=.22. Significant group differences were found for the EDE-Q, t 67=2.39, p=.02, d=.59, with the previous FBT group scoring lower than then no previous FBT group, Mdiff=.72, SE=.30, 95% CI [.12:1.32].
Table 1. Intention to treat means, standard deviations and effect sizes for pre- and post-intervention measures
BMI, body mass index; EDE-Q, Eating Disorders Examination Questionnaire; CIA, clinical impairment assessment. Pre and post mean scores are displayed, with standard deviation in parentheses.
Intent-to-treat analysis
Repeated measures ANOVA showed no significant time × treatment group interaction for the EDE-Q: F 1,67=1.32, p=.26, partial eta squared=.02, or the CIA F 1,65=1.97, p=.17, partial eta squared=.03. The time x treatment group interaction was significant for BMI centile F 1,67==7.81, p < .01, partial eta squared=.1.
Follow up analysis of BMI centile showed that there was no significant difference between groups at the start of treatment F 1,67=.01, p=.92, partial eta squared <.01, or at end of treatment, F 1,67=2.27, p=0.14, partial eta squared=.03. Simple main effects of time showed that the group with no previous FBT treatment significantly increased in BMI centile across treatment, F 1,41=22.47, p<.01, partial eta squared=.35, whereas the group with previous FBT treatment did not change in BMI centile, F 1,26=.22, p=.64 partial eta squared <.01.
The main effect of time showed a significant effect of time for the EDE-Q: F 1,67=37.52, p<.001, partial eta squared=.36; and CIA: F 1,65=36.42, p<.001, partial eta squared=.36, showing a significant reduction in scores from pre- to post-treatment. There was no significant main effect of group for the EDE-Q: F 1,67=2.57, p=.12, partial eta squared=.04, or the CIA: F 1,65=.88, p=.35, partial eta squared=.01. Dependent sample t-tests were run for each group to develop an effect size for the effect of time on EDE-Q, CIA and BMI centile, which are shown in Table 1. This procedure was repeated for the cohort as a whole, with the Cohen’s d of .78, .80 and .29 for the EDE-Q, CIA, and BMI centile respectively.
Completer analysis
Repeated measures ANOVA showed no significant time × treatment group interaction for the EDE-Q: F 1,40=1.01, p=.43, partial eta squared=.02, CIA; F 1,40=.99, p=.33, partial eta squared=.03, and a borderline interaction for BMI centile; F 1,40=4.28, p=.05, partial eta squared=.10.
The main effect of time showed a significant effect of time for the EDE-Q: F 1,40=91.64, p< .001, partial eta squared=.58; CIA: F 1,40=51.49, p<.01, partial eta squared=.57 and BMI: F 1,40=8.53, p<.01, partial eta squared=.18, showing a significant reduction in scores and increased BMI centile from pre- to post-treatment. There was no significant main effect of group for the EDE-Q: F 1,40=3.85, p=.06, partial eta squared=.09, CIA: F 1,40=0.40, p=.53, partial eta squared=.01 or BMI centile: F 1,40=.03, p=.86, partial eta squared >.01. Dependent sample t-tests were run for each group to develop an effect size for the effect of time on EDE-Q, CIA and BMI centile, which are shown in Table 2.
Table 2. Means, standard deviations and effect sizes for pre- and post-intervention measures among treatment completers
BMI, body mass index; EDE-Q, Eating Disorders Examination Questionnaire; CIA, clinical impairment assessment. Pre and post mean scores are displayed, with standard deviation in parentheses.
Reliable and clinically significant change for treatment completers
Whilst effect sizes as described above provide an index of the average change across treatment intervention, reliable change (RC) provides an index of whether individual changed sufficiently such that differences between scores are unlikely to be due to measurement unreliability. Clinically significant change (CSC) suggests that RC has been achieved, and that this change has taken the person from a score typical of a problematic, dysfunctional population, to a score more reflective of the general population (Jacobson and Truax, Reference Jacobson and Truax1991). The reliable change index for the EDE-Q and CIA were calculated using the Cronbach’s alphas of normative data published using adolescent samples (Bohn et al., Reference Bohn, Doll, Cooper, O’Connor, Palmer and Fairburn2008; Wade et al., Reference Wade, Byrne and Bryant-Waugh2008), and the means and standard deviations of the pre-test data from the current sample. Using this method, RC was calculated to be indicated by changes of 0.93 for the EDE-Q and 4.67 for the CIA. The proportions of those who completed treatment and exhibited reliable deterioration, no change changed, or reliable improvement is shown in Table 3.
Table 3. Number of completers achieving reliable change and clinically significant change
EDE-Q-Global, Eating Disorders Examination Questionnaire, Global scale; CIA, Clinical Impairment Assessment; CSC, clinically significant change.
Using the method recommended by Jacobson and Truax (Reference Jacobson and Truax1991), a CSC cut-off score of 3 for the EDE-Q was calculated using the weighed mid-point between the means from the sample data from this study, and normative non-patient data (Mond et al., Reference Mond, Hall, Bentley, Harrison, Gratwick-Sarll and Lewis2014). For the CIA, the cut-off point of 16 was used as recommended by the authors (Bohn et al., Reference Bohn, Doll, Cooper, O’Connor, Palmer and Fairburn2008). CSC was defined as those who had scored above the clinical cut-off point before the treatment, achieved reliable change based on RC calculations, and finished the treatment with scores below the clinical cut-off point (Jacobson and Truax, Reference Jacobson and Truax1991). The proportions of participants exceeding the cut-off points pre-treatment and achieving RC and/or CSC are shown in Table 3. Data are shown for those completing CBT-E and outcome measures.
Low weight group
Seven (mean age=16.29, SD=0.76) and one (16 years) young people started treatment with a BMI centile below 25 in the no previous FBT group and previous FBT group, respectively. In the no previous FBT group, four completed treatment. Mean scores for completers changed from 4.04 (SD=1.05) to 3.14 (SD=2.18) and 33 (SD=2.71) to 23.35 (SD=14.34) from pre- to post-treatment on the EDE-Q global and CIA, respectively. BMI centile increased from 9.86 (SD=6.09) to 21.28 (SD=13.15) for those who completed treatment.
The one participant (age 16) with previous FBT treatment did not fully complete CBT-E. The pre and post scores changed from 3.53 to 0.84 and 24 to 13 on the EDE-Q global and CIA, respectively, which met criteria for clinically significant change, with weight changing from the 2nd centile to the 25th by discontinuation of treatment.
Discussion
This was the first pilot study to compare treatment outcomes for CBT-E between those who had previously completed FBT treatment and those with no previous FBT treatment. Findings from this study reinforce recent evidence highlighting the effectiveness of CBT-E for adolescents (Dalle Grave et al., Reference Dalle Grave, Calugi, Doll and Fairburn2013; Dalle Grave et al., Reference Dalle Grave, Calugi, Sartirana and Fairburn2015; Dalle Grave et al., Reference Dalle Grave, Eckhardt, Calugi and Le Grange2019), and extend these findings in an Australian community public mental health setting. A novel finding from this study was that the effect of CBT-E on transdiagnostic eating disorder psychopathology and psychosocial impact does not vary according to whether individuals have previously tried FBT. The overall effect size of CBT-E on eating disorder psychopathology (EDE-Q Cohen’s d=.78) was comparable to similar studies of CBT-E among adolescents (Cohen’s d=.82–1.03; Craig et al., Reference Craig, Waine, Wilson and Waller2019; Dalle Grave et al., Reference Dalle Grave, Calugi, Sartirana and Fairburn2015). The study found no differences in attrition rates between the groups, showing that previous discontinued FBT does not appear to be an indicator of future CBT-E discontinuation. Overall, the attrition rate found in this study (60.9%) was similar to those in recent CBT-E studies with adolescents (61–63%; Craig et al., Reference Craig, Waine, Wilson and Waller2019; Le Grange et al., Reference Le Grange, Eckhardt, Dalle Grave, Crosby, Peterson, Keery, Lesser and Martell2020).
The current results showed that the effect of CBT-E on weight gain during treatment varied according to previous FBT treatment history, although these results must be interpreted with caution due to the variability in eating disorder diagnoses and unknown data on the weight change during FBT. Those who had previously engaged with FBT showed no weight gain throughout treatment, whereas those with no previous FBT treatment gained weight. An explanation for this finding is that initial engagement in FBT, whilst not achieving full recovery, may have resulted in some degree of weight regain. As there is a focus on early weight gain across the initial stages of FBT (Le Grange et al., Reference Le Grange, Accurso, Lock, Agras and Bryson2014), it could be that this was achieved through FBT prior to discontinuing, therefore this group may have been less underweight or weight suppressed at start of CBT-E compared with the group that had no prior FBT.
These findings are important and have implications for clinical theory and practice, future research and public health policy. The findings that CBT-E can elicit a treatment response after discontinuing FBT supports the notion that treatment suitability may vary according to individual differences in young people and their families. Encouragingly, it shows that having a sub-optimal response to FBT does not appear to reduce the likelihood of positive response to CBT-E. Alternatively or in addition to the above point, it may be that FBT and CBT-E have different roles to play at different times across young people’s treatment journey, and that for some people sequencing treatments at the optimal time for each is a way to facilitate recovery (e.g. FBT early in course of illness, followed by CBT-E at a later stage).
Implications for clinical practice and health policy would include offering CBT-E to young people for whom FBT is contraindicated or have not responded to FBT at individual clinician and structural (i.e. model of care) levels.
Future research could focus on identifying the differences in mechanism of action or acceptability to further understand why CBT-E was successfully completed where FBT was discontinued. Such investigations could also benefit from qualitative studies, as well as collecting data on reasons for discontinuing FBT/reasons for referral to CBT-E, which were not collected in this study. Future studies could collect data on weight history to draw conclusions about the need for weight regain and differences in the treatments to achieve the same. Additionally, controlled trials investigating moderators of outcome between the treatments could inform treatment allocation. Investigation the alternate temporal sequencing (i.e. FBT after discontinued CBT-E), would be equally as important as the findings in the current study.
A strength of the study was its real-world implementation in a community public mental health setting, which demonstrates its external validity. There are several limitations to the current study. The sample is relatively small and contained missing data within both groups which could be improved in a larger sample with more rigorous data collection procedures. No data were collected on those who were offered CBT-E but declined, which limits evaluation of the acceptability of the treatment in general as well as differences between groups in acceptability/feasibility. A lack of control group was another important limitation, which limits the ability to draw conclusions about the effectiveness of CBT-E compared with no treatment or alternative treatments. Conclusions with regards to weight regain between the treatments must be very cautiously considered. There were no data on weight history collected, either pre-morbid or pre-FBT which represents a limitation. It is not known, for example, if those from the previous FBT group had achieved a healthy weight range prior to starting FBT, trajectories of weight prior to starting either treatment, or where individuals target healthy weight is relative to current. Future studies could collect more data at an individual level to establish the influence that treatments have on helping young people achieve their individual healthy weight range. Incorporating qualitative data could provide deeper insights into participants’ experiences with FBT and CBT-E, as well as reasons for discontinuation or non-engagement with FBT. Relatedly, there were no data collected prior to FBT on ED psychopathology, so it is not known if young people were improving, declining or unchanged on these measures prior to commencing CBT-E. Lack of follow-up data prevents the ability to determine if the treatment effects found were maintained. Whilst adherence measures are commonly used among clinicians at this clinic, they were not collected for this study and represent another limitation that could be addressed in future studies.
Conclusions
The current study supports recent investigations showing the effectiveness of CBT-E among adolescents with eating disorders. It is the first study to show that the effectiveness of CBT-E on ED psychopathology does not vary depending on if the young person has or has not previously tried FBT. Implications from these findings give hope for young people and families that have not achieved full recovery through FBT and suggest that considering CBT-E in these situations is warranted. Future research could replicate these results in a larger sample, and seek to further understand the mechanisms of action between the therapies as well as individual differences that moderate treatment outcome to allow for evidence-based treatment allocation.
Key practice points
-
(1) Previous lack of completion/achieving recovery in FBT does not appear to reduce the likelihood of completing CBT-E.
-
(2) Adolescents with ED benefit from CBT-E regardless of FBT history.
-
(3) Considering offering CBT-E is warranted for young people and families who have not achieved full recovery through FBT, although more research is required to understand optimal timing.
Data availability statement
The data that support the findings of this study are available on request from the corresponding author [DW]. The data are not publicly available due to restrictions from ethical review.
Acknowledgements
The authors would like to acknowledge and thank the young people and families that have participated in the research. Additionally, the authors would like to thank Rachel Reavley, Renee Calligeros, Ally Barrington, Beth O’Gorman, Nikki Jermyn, Kirsten Young, Michelle Jordan, Constanze Shultz, Samantha Morrison and Anna Gollan for their help in delivering CBT-E.
Author contributions
Daniel Wilson: Conceptualization (lead), Data curation (lead), Formal analysis (lead), Funding acquisition (lead), Investigation (equal), Methodology (equal), Resources (equal), Visualization (lead), Writing - original draft (lead), Writing - review & editing (equal); Riccardo Dalle Grave: Supervision (lead), Writing - review & editing (supporting); Tania Withington: Writing - original draft (supporting), Writing - review & editing (supporting); Melanie Dalton: Conceptualization (equal), Investigation (equal), Methodology (equal), Project administration (equal), Resources (lead), Supervision (lead), Writing - original draft (supporting), Writing - review & editing (supporting).
Financial support
Dr Wilson is supported by a Queensland Health Clinical Research Fellowship.
Competing interests
R. Dalle Grave receives royalties from Guilford Press, Routledge, Springer and Positive Press.
Ethical standards
The project was reviewed by the Children’s Health Queensland Hospital and Health Service Human Research Ethics Committee and was granted exemption of ethics review (EX/2023/QCHQ/99469). The research conformed to the Declaration of Helsinki.
Open access
Comments
No Comments have been published for this article.