Personality-based subtypes of anorexia nervosa

As I’m sure most of my blog readers know, the DSM has two subtypes of anorexia: restricting and binge/purge (abbreviated here as AN-R and AN-BP. Because psychologists love acronyms almost as much as they love asking you about your mother). Behaviorally, they look easy to tell apart. But that’s not always the case.

New research is also showing that these behavioral subtypes may not be the best way to distinguish between different types of anorexia nervosa. Instead, a growing number of psychologists are saying that personality differences in people with AN are actually more meaningful than the dichotomous AN-R and AN-BP. As researchers noted in a previous study (Wildes et al., 2011):

Longitudinal research has found few differences between AN-R and AN-BP in treatment utilization, time to recovery, rates of relapse, or mortality. Moreover, several studies have documented that a majority of patients with AN-R eventually develop binge/purge behaviors, suggesting that AN-R and AN-BP represent alternate phases in the course of illness rather than distinct subtypes.

This type of thinking isn’t limited to AN alone. When I posted this study to Facebook, one of my readers rightly pointed out that Aimee Liu covered this topic in her book Gaining. In fact, I also covered it briefly in Decoding Anorexia. For a more readily accessible explanation that’s well worth reading, check out Tetyana’s analysis over at Science of EDs.

A very brief condensation of the research:

Scientists looked at the various personality profiles of people with EDs and used these various characteristics to group AN and BN sufferers into three main groups:

Three Groups of ED Personality Traits:

  • Overcontrolled: The overcontrol extends far beyond food, but to virtually every area of their life. They tend to be rigid, depressed, and lack a sense of who they want to be in life.
  • Undercontrolled: Again, the undercontrol extends beyond just eating behavior. This group tends to be impulsive, emotional, and express their anger both outward (rages at others) and inward (self-injury)
  • Perfectionistic/High-functioning: Besides the obvious perfectionism, this group has a tendency towards depression, but also a larger number of healthy traits.

{Above descriptions from Westen & Harnden-Fischer, 2001}

In this initial study, the goal wasn’t just to look at differences in psychopathology and personality. The other goal was to look at how well these traits predicted outcome. Not surprisingly, the perfectionistic/high-functioning group reached remission first, followed by the overcontrolled group, and then the undercontrolled group. However, this was just one sample, and the data was recorded after the fact (the therapists were reporting on previous patients).

Researchers wanted to do a prospective study, to see how these traits would play out as people entered treatment vs. trying to recall what they were like after they were done. They also wanted to see how it would affect the course of treatment.

Last week, researchers published a study in Behaviour Research and Therapy that used ecological momentary assessment (you can read more about EMA in this post) to follow 116 women with full- or subthreshold AN being treated as outpatients and measure the relationship between personality features and AN symptoms (Lavender et al., 2013). They first administered a battery of personality and ED questionnaires and assessments. Then, over two weeks, they used the EMA protocol to ask patients how they were feeling in the moment, as well as any ED behaviors, at 6 points during the day. Over the course of the study, the researchers gathered 14,690 data points. The study participants were grouped into overcontrolled (14.7%), undercontrolled (47.4%), or low psychopathology (37.9%) groups.

The groups didn’t differ in terms of age, BMI, or full- or sub-threshold AN diagnosis. They also didn’t differ on daily measures of binge eating, vomiting, or exercise. Not surprisingly, the low psychopathology group had fewer comorbid mood and anxiety disorders, as well as lower measures of ED thoughts and behaviors. The AN-BP subtype was most frequently found in the undercontrolled group, and least likely found in the low psychopathology group. The overcontrolled group was most likely to be diagnosed with OCD at some point in their life, and had the highest perfectionism scores. The undercontrolled group, however, reported the highest levels of negative moods.

The authors conclude:

These findings further suggest that there may be utility in addressing heterogeneous subgroups in the treatment of AN, and that subtyping EDs by personality dimensions may provide a valid and clinically meaningful strategy for classification.

Translation: your personality traits had a lot more to do with how you were likely to behave than your actual subtype diagnosis. On the other hand, the personality subgroups didn’t do so great at distinguishing between patterns of ED behaviors as a DSM diagnosis. The question remains as to how these personality groups affect how AN is (or should be) treated. Is it more useful to target these personality traits during treatment, or do the actual behaviors affect treatment outcome more?

Luckily, many of the same researchers have looked at that question using 154 AN inpatients and day hospital patients (Wildes et al., 2011). They gathered personality data at admission to intensive treatment. The participants were primarily female Caucasians, with an average age of 25.6 years and an illness duration of roughly 8 years.

Again, the researchers were able to separate out the inpatients into the same three personality clusters: overcontrolled (36.4%), undercontrolled (42.9%), and low psychopathology (20.8%). The undercontrolled group were more likely to have a history of substance abuse, previous hospitalizations, and higher ED psychopathology. As in the previous study, there were no differences in age, BMI, or illness duration between the three groups, which indicates that these personality traits don’t help predict AN severity or chronicity.

These personality clusters did affect treatment outcome, however. The undercontrolled group did worse in three ways: poor outcome at discharge, discharge against medical advice, and readmission to intensive treatment during the three month follow-up. And the odds of this happening were extremely significant: the undercontrolled group had a 3.56 greater odds of poor outcome at discharge than the overcontrolled group, and a 11.23 greater odds than the low psychopathology groups. The overcontrolled and low psychopathology group didn’t differ on AMA discharge and readmission to treatment.

 AN personality outcome

When the researchers looked at the patients based on AN subtype, the AN-BP group was more likely to have worse outcome at discharge than the AN-R group, but there were no other differences in outcome.

So what does this mean for you?

For one, these personality groups are based on clusters of personality traits. While some people fit perfectly into one group, others may mostly have traits of one group but still have traits of another. The analysis the researchers used to place people in these groups was sort of a best fit, not a perfect fit. A problem that has yet to be solved is whether diagnostic crossover significantly alters your personality traits. It’s the question of whether the personality traits are driving the ED behaviors, or whether the ED behaviors in which you are engaging can also impact your personality traits. Whereas many personality traits are lifelong, many are much more malleable, and it’s not impossible to think that there can be variance over time, especially as your ED behaviors change.

As well, it’s important to remember that response to intensive treatment doesn’t always correlate with ultimate AN remission. Write Wildes et al:

One possible explanation is that factors that predict initial treatment response may differ from those associated with longer-term outcomes. For example, inhibited and constrained personality traits may enable overcontrolled AN patients to tolerate the structured environment of an intensive treatment setting, facilitating better initial outcomes and a decreased risk of premature discharge relative to undercontrolled patients. A shy and cautious interpersonal style also may decrease the likelihood that overcontrolled patients will seek additional intensive treatment in the short-term. In the longer-term, however, overcontrolled personality traits may make it more difficult for individuals to relinquish eating disorder symptoms, leading to a chronic course of illness.

Which was my experience to a T. I fit mostly into the overcontrolled group, though with a heaping dose of perfectionism added in. I did really well (in general) with intensive treatment BUT those changes never lasted. I was so stuck and entrenched in my illness that I had a really hard time making meaningful changes as an outpatient and keeping the changes I could make while inpatient.

But personality isn’t totally destiny. Finding a career I wanted was really helpful (although it didn’t prevent a really messy, nasty relapse), as was getting good treatment for co-occurring disorders. Learning how to work with my personality, and having a treatment team that was willing and able to do so, was also key. I’m still fairly overcontrolled, but I’m trying not to let it run my life, either.


Lavender, J.N., et al. (2013). Personality-based subtypes of anorexia nervosa: Examining validity and utility using baseline clinical variables and ecological momentary assessment. Behav Res Ther. doi: 10.1016/j.brat.2013.05.007

Westen, D., & Harnden-Fischer, J. (2001). Personality Profiles in Eating Disorders: Rethinking the Distinction Between Axis I and Axis II. American Journal of Psychiatry. DOI: 10.1176/appi.ajp.158.4.547

Wildes, J.E., et al. (2011). The Clinical Utility of Personality Subtypes in Patients with Anorexia Nervosa. J Consult Clin PsycholDOI: 10.1037/a0024597

Twitter Digg Delicious Stumbleupon Technorati Facebook Email

9 Responses to “Personality-based subtypes of anorexia nervosa”

  1. So, Carrie, to extrapolate (or perhaps the researchers extrapolated), was there a sense that the undercontrolled group might also have a dx of BPD?

    • Carrie Arnold June 29, 2013 at 5:31 pm

      It’s possible or even probable, but I don’t know if they measured it directly. You can read the full text of the 2001 paper that would have the best sense of relationship between BPD and ED personality traits.

    • Carrie Arnold June 29, 2013 at 5:33 pm

      Okay, I read the 2001 study again, and the undercontrolled group is more likely to have BPD although BPD diagnoses aren’t exclusive to that group.

  2. Personality disorder, according to the current version of the Diagnostic and Statistical Manual ( DSM-5 ), refers to a class of maladaptive personality traits, that is, enduring patterns of behavior, cognitions and inner experience that are exhibited across many contexts and deviate markedly from those accepted by the individual’s culture. These patterns are inflexible and are associated with significant distress or disability.

  3. Residential electrical services Tulsa happen to
    be very essential element of the society, as the electricians develop a key
    side of your home that features most of your daily use items and
    gadgets. However, if someone has freshly painted over
    or patched it, that screwdriver is one heck of a lie detector.
    In many cases this is perfectly fine, even encouraged.

  4. I’ve just found your blog- excellent book btw.

    I have another suggestion as to the poor outcomes for some of those leaving treatment.

    In my experience, there is no treatment for anorexia with bulimia (or perhaps as described on the above as an “uncontrolled” group). There is plenty around (well, in comparison!) to ANR, discussions on how and what to eat and plans to eat more and so on. But when it comes to bulimia with anorexia, no advise is given other than “just stop” or “cut down” but with no real intensity on any treatment. It would seem the best way to recover that I am aware of is to just simply restrict, which can then lead to an ANR relapse and then it’s only a matter of time…

    I say this as I have seen and myself experienced this on too many occasions to count. It is sad and depressing that some of the best and most experienced professionals I know of are clueless as to what to suggest- if indeed they address or acknowledge it at all- it seems many hope it will just ‘go away’, as if it is not a problem in it’s own right but instead a side effect of being an emaciated weight. The only time I felt people feel they needed to pay attention and realise just how suffocating the struggle is was when I or those I knew were losing too much weight as a direct result of purging and the message was to “just stop or lose treatment” or to “just keep a bit down”, as if weight was the only measure.

    It is owing to the person who wants so badly to never do this again that the will power is that strong that they do eventually recover, and stay recovered for a duration of time, but it has never been the case in all my years of recovery treatment that the issue itself is formally recognised and worked on psychologically and emotionally. Thus unlike ANR in my experience (which I have also relapsed suffered with and had treatment for (without the AN-BP being present)) the way around bulimia is will power rather than feeling understood and helped or supported in the same way.

    I hope I have experienced something quite rare and that everyone else bar myself and all the many people I have known (which is many…) but it does go to show that on recovery outcome graphs like those above, there are poor outcomes for those with bulimic traits.

    It is not the problem of the person who is suffering, nor their desire to change, it is the lack of treatment for this particular disorder which I really think needs so much support not least of all because of the shame which becomes a vice and causes more isolation then I can even begin to detail here.

    • I had AN-R leading to what was probably AN-BN (and according to this interesting blog I was probably somewhere between overcontrolled and perfectionistic, veering towards perfectionistic as I definitely had a sense of what I wanted from life, but was emotionally inhibited).

      The answer to binge-purge cycles (or binge-restrict cycles in my case) is still to eat more. If you have an AN diagnosis, you are probably overcompensating with your restrictive behaviour in order to ‘make up’ for binges. One of the first things I learned was to stop restricting. This reduces the urge to binge (which in AN with binges behaviour is probably due to HUNGER!) If cutting restrictions doesn’t work, it might be that you’re eating emotionally (my mother told me she did this as a bulimic), in which case you might try something like DBT to learn emotional regulation skills (mindfulness is excellent).


  1. News You Can Use – June 29-July 6 2013 « Eating Disorder Pro - July 3, 2013

    […] Personality-based Subtypes of Anorexia Nervosa – The DSM has two subtypes of anorexia: restricting (AN-R) and binge/purge (AN-BP). Behaviorally, they look easy to tell apart. But that’s not always the case. New research is also showing that these behavioral subtypes may not be the best way to distinguish between different types of anorexia nervosa. LEARN MORE. […]