A Comparison of Treatment Outcomes: AN, BN, and EDNOS

My previous posts described treatment outcomes and correlates of my patients with various eating disorder diagnoses, including Anorexia Nervosa (AN), Bulimia Nervosa (BN), and Eating Disorder Not Otherwise Specified (EDNOS). Some findings were similar across diagnostic groups:

• The vast majority of patients with any diagnosis who completed treatment achieved full remission.

• The attrition rate, overall, was high.

Family-Based Treatment (FBT) was associated with treatment retention, meaning that patients who participated in FBT were less likely to discontinue treatment prematurely than those receiving individual therapy.

• Patients who paid a reduced rate for services were much less likely to complete treatment and much less likely to achieve remission than those who paid full rate.

• Hospitalization during treatment with me was not related to treatment completion or treatment outcome, regardless of diagnosis.

• History of intensive eating disorder treatment (e.g., residential treatment, day treatment) was related to treatment non-completion and poorer outcome for patients across diagnoses.

Quite surprisingly, the differences among the diagnostic groups outnumbered the similarities:

• Patients with AN who completed treatment attended more sessions, over a longer period of time, than patients with BN or EDNOS. Completing treatment for AN involved an average of 28 sessions over 17 months, whereas completing treatment for BN or EDNOS involved an average of 15 sessions over 10 months.

• Type of treatment had a significant impact on outcome for patients with AN but not for patients with BN or EDNOS. All AN patients who achieved remission did so through FBT. However, patients with BN and EDNOS achieved remission through various means, including individual therapy with no family involvement, individual therapy with family involvement, and FBT.

• Younger age at intake and shorter duration of illness predicted treatment completion and remission for patients with AN. However, neither age nor duration of illness affected treatment outcome for patients with BN or EDNOS.

• All “treatment non-completers” with AN (those who quit prematurely, moved away, or were referred to other treatment settings) discontinued treatment within the first six months. In contrast, a number of “treatment non-completers” with BN or EDNOS remained in treatment for more than a year.

• Among patients with AN, those who completed treatment attended significantly more sessions than those who did not (28 sessions vs. 8 sessions) and remained in treatment for a much longer time than those who did not (17 months vs. 3 months). In contrast, for patients with BN and EDNOS, treatment completers and non-completers both remained in treatment for an average of 10 months, and treatment non-completers actually averaged more sessions than treatment completers (20 sessions vs. 15 sessions). Thus, it appears that for BN and ENDOS, treatment completers did not remain in treatment longer, but rather reached remission more quickly.

• Caucasians with AN were slightly more likely than Hispanics with AN to complete treatment and achieve remission. In contrast, Hispanics with BN or EDNOS were significantly more likely than Caucasians with these diagnoses to complete treatment and achieve remission.

• The presence of a comborbid disorder was not related to treatment completion or treatment outcome for patients with AN. However, the presence of a comorbid disorder was a significant predictor of treatment dropout and poorer outcome for patients with BN and EDNOS.

• History of hospitalization for an eating disorder or related psychiatric issue was related to poorer outcome for patients with BN and EDNOS, but not for patients with AN.

• Taking psychotropic medication during treatment was related to treatment completion and positive outcome for patients with AN but not for patients with BN or EDNOS.

Please bear in mind that these findings are specific to my practice and my patients, and are not intended to be generalized to other treatment settings or other patients.

5 Replies to “A Comparison of Treatment Outcomes: AN, BN, and EDNOS”

  1. > The vast majority of patients with any diagnosis who completed treatment achieved full remission.

    Dr Ravin, I’m certain this question has come up previously . . . I’m interested in knowing what data you have if any regarding how these patients fared in the months after leaving your treatment for the last time . . . as I understand your definition of remission a patient could leave your office in full remission, but a relapse 1 or 6 months later would not be included in the assessment.

    We have a daughter who has struggled with an eating disorder since high school, and she is now 27. She is now in outpatient having undergone residential and PHP treatment that was very effective in helping turn her life around (to the point of an individual who questioned residential treatment, these programs were absolutely vital in keeping her from engaging in binging and purging . . . ). One of the most significant issues we’ve encountered in building her outpatient team – and I think this is a pervasive problem – is finding qualified in-network providers, and we eventually gave up and assembled an entirely out-of-network team.

    Your Website/blog stands out for the information it provides on your credentials, experience and focus re: ED. In our experience it is very common for therapists who list themselves as being able to treating ED to provide no readily information on their practice or no readily information on this aspect of their practice . . . this is a significant red flag in my view.

    In any event, I’m happy to have found your site . . . and you.

    1. James,

      Building a competent outpatient team is indeed quite difficult, and families often need to spend more money, time, and effort than they planned to in order to assemble the best team.

      In response to the question you asked about relapse after completing treatment with me – I am currently in the process of collecting that follow-up data. It will be quite a while before I have finished collecting it, analyzing it, and writing it up. This is the first time I have collected follow-up data and I am very interested in the results. I plan to continue collecting follow-up data on former patients every summer.

      My goal in ED treatment is to provide the family with the knowledge, tools, and confidence they need to keep their loved one well, spot any early signs of struggle, and intervene effectively. In essence, it is my job to make my job obsolete. For this reason, I genuinely hope the rate of relapse is very low.

  2. I believe the most important data is what happens after treatment as well as to identify what the person and family did to help foster remission.

    For my D she’s had the most growth post-treatment because she has a strong wellness coaching plan that helps her with long term positive behavioral change. That helps to build her resilience and to learn more effective coping for comorbid social anxiety.

    Unless the follow up survey asks those specific questions, how would you determine what has helped them remain in remission?

    Although staying within their accurate target weight range is critical, it is only the first step.

  3. Wendy,

    I completely agree that full weight restoration is necessary, though not sufficient, for recovery. I also agree that data about what happens post-treatment that fosters continued remission or triggers relapse is very important. However, that question cannot really be answered through a survey, even if those questions are asked. So many variables are involved, and they all interact with one another.

    I think the question of what helps people maintain remission vs. relapse could best be answered through a true experimental design with a large sample which isolates specific variables.

    With survey data, the closest we could get to answering your question would be to determine which variables are associated with continued remission post-treatment. Then we may be able to predict, with some accuracy, who will remain in remission based on those variables, although we couldn’t assume causality.

  4. Dr. Ravin,

    Would you please elaborate on remission needing more after full weight restoration? What have you found to be helpful for individuals with no significant co-morbids, except anxiety ( which it seems many with AN suffer from)?

    I often hear that DBT is useful, but would like your opinion on what has been effective for your patients.

Comments are closed.