A 10-year follow-up study of completers versus dropouts following treatment with an integrated cognitive-behavioral group therapy for eating disorders

We compared the 10-year prognosis of patients who completed integrated group CBT with those who dropped out and examined the effect of completing group CBT on prognosis. After group CBT, eating attitudes and depressive symptoms decreased, and self-esteem increased significantly. We also compared the clinical features and 10-year prognosis of patients who completed the study and those who dropped out. The completer group had a significantly higher number of patients who had a good outcome, and the dropout group had a significantly higher number of patients who had a poor outcome. The integrated group CBT experience might have positive long-term effects.

Eating disorders are serious illnesses that can be difficult to treat [1]. Many serious problems are associated with eating disorders [2, 3], such as 1) significant mortality rates [4], 2) suicidal behavior [5, 6], 3) high medical costs [7], 4) high rates of comorbidity [8], and 5) high association with childhood trauma, such as abuse. Nonetheless, the existing evidence-based treatments are still limited in their effectiveness.

In the National Institute for Health and Clinical Excellence guidelines [9], cognitive behavioral therapy (CBT) is recommended for anorexia nervosa (AN), bulimia nervosa (BN), and binge eating disorder (BED). Fairburn [10] reported that most eating disorders have a mixed clinical presentation, in which the features of AN, BN, BED, and those of incomplete types overlap in various ways. Fairburn adapted CBT to make it transdiagnostic, i.e., appropriate all eating disorder psychopathologies, and reported that this Enhanced CBT (CBT-E) outcome did not depend on the type of eating disorder involved [11]. Accordingly, we conducted CBT in a transdiagnostic sample.

We conducted CBT in a group setting for the following two reasons. First, group therapy is time-efficient. In Japan, outpatient clinics treat many patients and it is not possible to commit large amounts of time to individual treatment. With group therapy, time problems can be solved to an extent. In addition, group therapy reduces patient payment to 54–70% of that of individual psychotherapy; therefore, group therapy is cost-effective. Many reports have indicated that group therapy for eating disorders is cost-effective and cost-beneficial [12–14]. Moreover, both CBT and interpersonal therapy have demonstrated good outcomes [15]. The second reason is that group psychodynamics work therapeutically. Group therapy gives patients the opportunity to be supported by others and to support others. Yalom et al. [16, 17] reported that interpersonal learning is an important and dynamic factor in group-based psychotherapy that allows group members to share aspects of their emotions, thoughts, and perceptions while receiving feedback from fellow group members in a safe and collaborative setting. Because recovery from eating disorders takes considerable time, many patients lose motivation during treatment. We believe that group therapy can contribute to enhancing and continuing motivation for treatment as the patients support each other.

In this study, it was difficult to apply Fairburn’s CBT-E in its entirety because in group therapy, it is difficult to carefully intervene in patient’s cognitions as is done on a one-to-one basis in individual therapy. Therefore, we partially adopted CBT-E and added a variety of approaches for conducting group therapy that motivate recovery efforts. In the group CBT, a self-assertion session seemed necessary. This is because Japanese individuals, and especially those with eating disorders, are not typically self-assertive, and it is important for patients to address this in order to have acceptable experiences within the group. A Japanese study reported that a combination program of CBT with assertive training and self-esteem enhancement was effective [18]. In addition, we considered problem-solving techniques important. Patients with eating disorders feel pressured by numerous problems, isolated, and helpless. Clarifying problems and considering solutions within the group setting helps alleviate such feelings. Pettersen [19] reported that recovery from eating disorders included improved acceptance of oneself, interpersonal relations, problem solving, and body satisfaction, which were not entirely dependent on symptom absence.

In contrast, eating disorders may be resistant to change, and dropout rates from treatment are high [20, 21]. A review of 19 studies reported that the therapeutic alliance might be an important factor in eating disorder treatment [22]. Another systematic review emphasized the importance of early-stage motivation for change as a predictor of treatment outcome [23]. A 2-year follow-up study reported that temperament is related to dropout and one of the reasons for dropout is lack of motivation [24]. Therefore, motivation for change is important to prevent dropout from treatment; we believed that integrated group CBT might help support motivation for change. We hypothesized that an integrated program incorporating CBT might have a positive impact on continued treatment and prognosis, even for short-term interventions.

The purpose of the current study was to compare the 10-year prognosis of patients who completed integrated group CBT with those who dropped out and to examine the benefits of completing the group CBT.

As reported by Fairburn [10, 11], eating disorders, i.e., AN, BN, BED, and those of incomplete type, all have a common core psychopathology. We agree with Fairburn that transdiagnostic CBT (CBT-E) is appropriate for all eating disorders. Thus, we conducted a CBT intervention in a group setting for all types of eating disorders. Fairburn’s CBT-E manual is written for use with individuals, and not for group intervention. Individual therapy has merit, in that we can intervene in response to the situation of the individual patient, and it can be used for patients who tend to avoid group therapy. However, for other patients, group therapy is cost-effective and efficacious due to the social experience and the expectation of approval within the group. Therefore, we conducted group therapy to supplement short-term outpatient treatment. We devised a therapeutic method for group therapy based on the CBT-E protocol of Fairburn.

After intervention, eating attitudes, depressive mood, and emotional response to stress reduced, and self-esteem increased. Through awareness of the change, the participants seemed to have recovered self-esteem and were motivated to continue treatment. In our previous report [33], depression, anger, and confusion were more marked in women with eating disorders than in healthy women. Several reports have indicated that depressive symptoms predict treatment outcome [34, 35], and can indicate the risk for suicide. It is important to reduce depressive symptoms and to prevent suicide. The CBT intervention used in this study improved depressive symptoms.

The dropout rate from eating disorder treatment is said to be high because of the psychopathology involved in these disorders, such as denial of the severity of the disorder by the patient and treatment resistance [21]. Bandini et al. reported that the dropout rate of outpatients from CBT was 33% for AN-BP, 27% for AN-R, 25% for BN, and 21% for EDNOS [36]. Several factors, such as low patient engagement, purging episodes, restrictive eating, use of several weight control practices, and psychiatric co-morbidity affect dropout rates [37]. However, others have reported that purging does not influence the dropout rate [38], and instead that low depression scores [39], divorce of parents [39, 40], a tendency to repress anger [41], and self-efficacy [42] influence dropout. In this study, the dropout rate was 33.3% for AN-BP, 27.8% for AN-R, 25% for BN-P, and 22.2% for BN-NP. We had a small sample size and so significant differences between each type of eating disorder were not found, although the presence of purging was frequent in those who did not complete CBT. We found that in the dropout group, compared to the completer group, there was higher anxiety and anger, not depression. Thus, it is important to tackle resistance to, and anxiety towards, treatment, and to conduct the introduction to treatment carefully.

Regarding the 10-year outcomes, the completer group had a significantly higher number of patients with a good prognosis, and the dropout group had a significantly higher number of patients with a poor prognosis. We could track only 10 of 17 patients who dropped out from therapy. The 7 patients whose prognosis could not be followed all dropped out during the outpatient consultation phase. This, it is likely that the prognosis of these 7 cases was poor. Furthermore, the 5-year prognosis was similar to the 10-year prognosis; therefore, we speculate that it is possible to estimate the long-term prognosis by prognosis after 5 years. The therapy was a short-term intervention; thus, this therapy alone did not influence the long-term outcome. However, since there were fewer dropouts from outpatient clinics among patients who completed the group therapy, it is possible that the group therapy experience had a positive effect on subsequent treatment continuity and prognosis. It seems that the patients continued practicing the skills and methods learned from group therapy, and reconfirmed these skills in the outpatient consultations.

People with eating disorders experience social impairments [43]. Difficulties in interpersonal relationships have been reported as potential factors in the development and maintenance of body image disturbances and altered eating behaviors [44]. Group therapy gives patients an opportunity to form peer relationships, and helps them to develop communication and socialization skills. Furthermore, in the group therapy sessions, the patients develop self-awareness by listening to other members of the group and develop interpersonal relationships. If the patients successfully work together in the group setting, they derive benefits, such as the prevention of feelings of isolation and the mutual recognition of increased self-esteem. People with eating disorders generally have low self-esteem [1, 45–47], and improving self-esteem is one of the main targets of treatments for eating disorders [48, 49]. Interventions that improve self-esteem help in recovery from eating disorders. Considering this research, many patients who completed the group CBT continued outpatient treatment and showed a good prognosis. They showed improvement in depressive symptoms and recovery of self-esteem. This effect may be temporary, but the experience of group therapy may have provided hope and had a positive effect on continuation of treatment. Many Japanese patients experience difficulties when speaking to others with the same illness and talking about themselves. Therefore, it is valuable to provide a place to gather as a group, and opportunities for group discussion. In addition, self-assertion sessions are advisable because Japanese patients are often passive and willing to adapt to others without stating their own opinions. Furthermore, Japanese individuals often have a strong sense of guilt if they suffer from eating disorders. It seemed that the motivation to continue treatment was enhanced by the mutual support in the group, and by completing group therapy, which led to a sense of accomplishment and confidence.

A limitation of this study is its small sample size. It is difficult to confidently state that the non-completers had poor prognosis, because they were few in number. Further studies are required to replicate these findings with a control group, preferably using an RCT.

Another limitation of our study is that we could not follow up 26% of the subjects for the full 10 years. In retrospect, a more robust method of follow-up could have been determined at the start of study. In addition, social adaptation was assessed using the GAF; however, the DSM-5 proposes that one should use the WHO Disability Assessment Schedule (WHODAS). WHODAS is a multifaceted assessment of living functions from 6 areas, and if it had been used, we might have appreciated better the patients’ living conditions.

It appeared that the participants of this study had a considerable number of developmental disorders including autism spectrum disorder. However, for the diagnosis of developmental disorders, detailed childhood information is necessary, and since there are limitations to retrospective investigation, it was not reported in this study.

Finally, it is thought that not only short-term interventions, but also additional CBT sessions, may be more effective. In future, we would like to test this hypothesis.

Depressive symptoms, emotion-oriented coping, and self-esteem were improved after the integrated cognitive-behavioral therapy in a group setting. Many patients who completed group therapy had a good 10-year prognosis.