Who’s Failing Whom? Treatment Retention for Eating Disorders

I find the idea of treatment retention for eating disorders to be quite interesting. Mostly, I find it intriguing to dissect the way that authors write about treatment retention – that is, how they tend to look at factors within people that make them more or less able to complete treatment, rather than things about the treatment that serve or don’t serve people’s needs. I’ve been reading a lot more about adolescent eating disorder treatment these days, given that I’m doing a practicum at an adolescent treatment centre that does things a bit differently. Resultantly, I’ve become more interested than ever in how we can better meet people’s different treatment needs and provide a more comprehensive treatment continuum.

The stark reality of treatment is that people don’t always finish it. That statement sounds fairly banal, but it’s a loaded one. Too often, I see this framed as people failing treatment. Linked to the idea of treatment resistance (which I’ve taken issue with before in one of my more-commented-upon posts), treatment dropout has a framing issue as far as I’m concerned. This framing issue is more than just an issue of semantics; it has real, material effects on people’s lives.

When you drop out of treatment, you might feel like a failure – you might be explicitly labelled so in the clinical imagination. In a neoliberal society in which non completion is framed as an issue of will or motivation, there’s little room left to consider other factors that make the treatment environment unhelpful.

I recently read an article by Hoste, Zaitsoff, Hewell & le Grange (2007) that asks “what can dropouts teach us about retention in eating disorder treatment studies?” More so than reporting on the results of the study, I am going to comment on:

  • What other factors we might consider
  • Why it is important to look beyond the individual; and,
  • What looking beyond the individual might mean for reconsidering treatment contexts

I’ll also comment on the links between treatment and trials; in this case, the article is specifically looking at the implications of exploring dropout for the ability to conduct reliable research about eating disorders, which gives us a nice opportunity to also consider what research gets counted as evidence in eating disorder treatment.

The Study

Dropout is a particularly common problem noted in eating disorder treatment efficacy studies. Hoste et al. note that rates of dropout can be as high at 63% for studies about anorexia nervosa (AN) and around 30% for bulimia nervosa (BN). They note that this has an impact on the rigour of study results; I would of course add that this is concerning in terms of program completion and the question of whether we are serving those with eating disorders well. Some have even argued that until we better understand why people drop out of treatment, randomized controlled trials (RCTs) are not terribly informative (Halmi et al., 2005).

The numbers above are for adult treatment; adolescent treatment study dropout rates are around 10% for both AN and BN, according to the authors. However, this doesn’t necessarily mean the treatment is superior for adolescents; there are some significant differences between adult and adolescent treatment in terms of voluntariness and ability to drop out against medical advice.

In this paper, Hoste et al. describe differences between adolescents who completed and those who dropped out of treatment for BN in the US. They also explain the strategies they used to retain participants in their treatment study, which led to a low dropout rate of 11%. In the study, 80 participants completed either family-based treatment or individual supportive therapy for 6 months.

First, let’s look at how the authors define and describe treatment dropout. They write:

“Nine participants (11%) did not complete treatment, defined as failing to complete at least 50% (10 sessions) of the prescribed treatment dose.” (p. 668)

This way of describing treatment non-completion is common in the eating disorders (and other treatment) literature. Sure, you might argue that I’m picking hairs by pulling out this quote, but I think it speaks to the way we frame treatment dropout. The authors do note that some who dropped out of treatment did so because they were unsatisfied with treatment, but do not expand upon why that might be or suggest changes to the treatment itself to incorporate patient feedback. Given the individualizing of the frame, I think it’s important for us to look at what the authors looked for and at and what they didn’t.

They looked at participants’:

  • Family status
  • Duration of illness
  • Ethnicity
  • Age
  • Diagnosis
  • Scores on the Eating Disorder Examination
  • Scores on the Rosenberg Self-Esteem Scale
  • Scores on the Beck Depression Inventory

They found that 2 of the 9 participants who dropped out were from intact families, and that those who dropped out had significantly longer illness duration; there were no other significant differences between those who dropped out and those who remained on the variables they measured.

To attempt to minimize dropout, they:

  • Tracked attendance on a sheet, and if participants did not schedule follow up appointments they contacted parents
  • Made reminder calls to parents the day before appointments
  • Scheduled appointments in a coordinated way (i.e. a physician and psychiatrist appointment on the same day)
  • Encouraged rescheduling of missed appointments (primarily through contacting parents)

What’s Missing?

The authors briefly comment on the idea that there may be more to the picture than the individual patient when considering dropout. They note:

“Although speculative, it is possible that certain aspects of the therapy contributed to the current study’s low dropout.” (p. 670)

By “aspects of the therapy,” they suggest that family involvement and dynamics might lead to treatment completion or dropout, noting again that patients with intact families seemed less likely to drop out. They suggest that “intact families [may] have more resources to facilitate treatment” but it is unclear whether by resources they mean emotional, physical, or financial resources. I might infer that they mean emotional resources, given that they follow up by suggesting that involving families or other supporters in treatment may lead to improved treatment outcomes.

Other than this commentary, however, there is no expanded exploration of factors beyond the person that might impact treatment completion or dropout. We might think about these additional factors that I would argue are worth considering in terms of:

  1. Factors related to the person’s circumstances: by this I mean things that are related to the person’s social location, like their gender, ethnicity, sexual orientation, ability, socioeconomic status etc. that may limit their ability to access or continue in treatment. Think, for instance, about the challenge of accessing or continuing in a Western therapeutic environment if one is from a culture that is skeptical of seeking external help for mental illness or that frames mental illness differently. Think also about geographic and financial barriers to accepting or continuing treatment – the challenge of needing to drive over an hour for treatment when one has other life responsibilities, for intstance
  2. Factors related to the type of treatment: by this I mean
    1. How the treatment resonates with the person; this includes alliance with the treating professional, amount of surveillance and rigidity in the treatment, which can lead to secrecy or feelings of discomfort in the treatment environment
    2. The degree to which those in the treatment team listen or do not listen and respond to the patient as a person with specific needs, rather than “just another person with an eating disorder” and/or who comes from a particular social location
    3. The extent to which the treatment prepares the person for their lives beyond the treatment context and makes it clear to patients how they can transition from the treatment environment to the rest of their lives (which might help make the treatment seem worthwhile to the person)

Interestingly, though this study was published in 2007, we continue to see few studies of this ilk; ones that acknowledge that there may be good reasons for which people drop out of treatment that are not just “patient non-compliance.” What I would love to see are more studies that consider how and why people drop out of and/or cycle in and out of treatment. Moreover, I would love to see these studies frame this not as “how can we change people to make them fit our treatment models,” but rather “how can we change treatment systems to more fully meet people’s [multiple, intersecting, and individual] needs”? Of course, that’s probably too much to ask…

On another level, the fact that dropout from treatment and treatment studies is common also impacts what we see as compelling evidence for eating disorder treatment. In the popular discourse around eating disorders, I often see “evidence based” being tossed around as if that is an unproblematic term and as if we actually have strong research that differentiates the evidence for various kinds of treatment. As Halmi et al.’s suggestion that we need to forstall treatment efficacy studies until we better understand why and when people drop out would suggest, we are a LONG way from really being able to say that one type of treatment is better, across the board, than all others.

I’ve said it approximately one million and five times, but there is NO one size fits all treatment for eating disorders – people are complex, their personal circumstances are complex, and we need to listen to people’s experience as another valid form of evidence. I am so acutely aware, especially right now, of how easily the support of one kind of treatment can morph into the vehement denial of others’ lived experiences. I wonder, often, how we can both push ourselves to develop effective modes of treatment while remaining open to the possibility that all of our assumptions about “what works” might be completely wrong for different people. I suppose what we really need is an unlimited imagination, an open ear, and a lot of funds that allow for a diverse and responsive treatment continuum. Totally reasonable request, right? In the meantime, I’d be satisfied if we recognized difference and stepped up to the challenge it brings, rather than trying to fix it all the time.


Hoste RR, Zaitsoff S, Hewell K, & le Grange D (2007). What can dropouts teach us about retention in eating disorder treatment studies? The International journal of eating disorders, 40 (7), 668-71 PMID: 17607714


Andrea is a PhD candidate focusing on individual, familial, and health care definitions and experiences of eating disorder recovery. She has an MSc in Family Relations and Human Development and a BA in Sociology. In her Masters research, she used qualitative and arts-based approaches (digital storytelling) to explore the experiences of young women in recovery from eating disorders. Andrea has recovered from EDNOS. She can be reached at andrea[at]scienceofeds[dot]org.