Scientists love classifying and categorizing things they study. But it can be a double-edged sword. Classification can lead to new insights about etiology or new treatment methods. But classification can also hamper our understanding. For example, researchers like to classify and study anorexia nervosa and bulimia nervosa as if they are two wholly separate disorders, but clinicians know that most patients fluctuate between diagnoses, and as a result often fall into the eating disorder not otherwise specified (EDNOS) category.
Nonetheless, if we keep in mind that the way in which we classify things can be very artificial and may not necessarily reflect some fundamental truths about the subject matter, we can focus on extracting the insights gained from the classifications.
In the case of eating disorders, classifying patients into subtypes may be useful for developing successful treatment approaches suited for particular patient subgroups.
Previous research on this topic has identified three personality subtypes that seem to “cut across eating disorder diagnoses” (Westen & Harnden-Fischer, 2001):
Three Personality Subtypes in Eating Disorder Patients:
- “dysregulated/undercontrolled pattern: characterized by emotional dysregulation and impulsivity”
- “constricted/overcontrolled pattern: characterized by emotional inhibition, cognitively sparse representations of self and others, and interpersonal avoidance”
- “high-functioning/perfectionist pattern: characterized by psychological strengths alloyed with perfectionism and negative affectivity”
Heather Thompson-Brenner and Drew Westen wanted to build upon the initial findings and find out if this classification was valid. What makes a classification valid? In this case, a classification would be valid if it held some predictive value about the patients’ treatment responses, for example.
If the subtypes are valid and clinically relevant, they should differ in adaptive functioning, aetiological variables, patterns of comorbidity, treatment response and therapeutic interventions selected by the treating clinician.
In this study, Thomspon-Brenner and Westen focused on patients who exhibited bingeing and/or purging behaviours, so the data I’m going to talk about is coming mostly from bulimia nervosa patients. They asked a random sample of clinicians, and specifically members of the American Psychiatric Association and American Psychological Association with 5+ more years of post-residency experience, to fill out a questionnaire about their most recently terminated course of psychotherapy with a female patient who had “clinically significant symptoms of bulimia'” and they instructed them “not to choose a case based on outcome, to sample both successful and unsuccessful cases.”
In one of the sections of the questionnaire, the clinicians had to read three paragraphs describing the “prototypical” patient that fit the specific personality profile. Then, the clinicians had to rate the patient on how well they fit each of the three personality subtypes (from 1-5).
- mean duration of clinical experience: 16 years
- theoretical orientation: 37% “CBT” or “primarily CBT”; 34% as “psychodynamic” or “primarily psychodynamic” and 29% as “other”
- mean age 28.5
- 17% rated as “poor”; 46% as “middle class”; 31% as “upper middle class” and 6% as “upper class”
- 72% met criteria for bulimia nervosa; 14% for bulimia nervosa non-purging type; 6% for anorexia-nervosa binge/purge type; and 8% for EDNOS
- 38% of the patients were previously diagnosed with anorexia nervosa
- 42% had at least one admission to a psychiatric hospital
- average number of binges/week: 4.6; purges: 4.2
- half of the patients engaged in excessive exercise
- a third took laxatives
- over half of the patients fasted
The clinicians were forced to make a categorical choice about which personality subtype fits their patient. In this sample, 42% of the patients were categorized as “high-functioning / perfectionistic,” 31% as “constricted,” and 27% as “dysregulated.” What’s more, 84% of the patients strongly resembled one of the personality subtypes (considered as a rating of 4 or 5 on a 5-point scale).
After the patients were categorized into the groups based on which of the personality prototypes suited them best, the authors wanted to find out if the subgroups of differed in terms of hospitalizations, history of child sexual abuse, and Axis I and Axis II disorders?
It seems that they did.
The dysregulated group had the highest rates of hospitalizations (62%), followed by constricted (40%) and high-functioning (29%). The dysregulated group also had higher rates of “clinician-reported childhood sexual abuse” at 42% (versus 20% and 19% for the constricted and high-functioning groups, respectively).
What about comorbidity with Axis I and Axis II disorders?
Clearly, there are differences. And with a few exceptions, the data mostly fit the predictions that Thompson-Brenner and Westen made before collecting the data. So, that’s great. But, why do we care? What’s the point of categorizing patients into these personality subtypes? Can this grouping tell us something useful about the patients?
After all, remember that
A valid psychiatric classification should ideally predict treatment response (Robins & Guze, 1970).
The short answer is, yes, it does seem that grouping patients based on these personality subtypes can tell us something about treatment response.
Thomspon-Brenner and Westen found that patients in the “dysregulated” and “constricted” groups spent longer time in treatment and had more negative responses to treatment.
Moreover, patients in the “dysregulated” group attained recovery after 92 weeks of treatment, compared to 73 weeks for patients in the “constricted” group and 51 weeks for patients in the “high-functioning” group. And, only 43% of patients in the “dysregulated” group recovered, compared to 50% in the “constricted” group and 62% in the “high-functioning” group.
I won’t bore you with the details, but essentially, Thomspon-Brenner and Westen found that these personality categorizations had a lot of predictive value in terms of treatment length and outcome that wasn’t there if we were to just look at eating symptoms or personality disorders for example.
Although the three types of patients were similar in their eating disorder symptoms (90% sharing the bulimia nervosa diagnosis), they also showed distinct patterns of comorbidity.
Patients rated as “dysregulated” had the highest rate of comorbid Axis I diagnoses and cluster B personality disorders. Patients rated as “constricted” were intermediate between the “dysregulated” and the “high-functioning” groups on Axis I comorbidity and were particularly likely to have cluster C (avoidant, dependent and obsessive–compulsive) personality disorders.
Personality patterns also predicted differences in treatment length and outcome. Dysregulation and constriction were both negatively associated with outcome. Patients in the constricted category attained recovery on average 5 months later than the high-functioning patients, and dysregulated patients attained recovery approximately 5 months later still.
Of course, there are limitations that should be noted: the data are retrospective and collected from clinicians (and it is hard to know if a rating “4” is the same for every person). It will be important to repeat these findings in a prospective study (as opposed to retrospective) and use multiple clinicians (or “informants”) per patients, if possible.
But anyway, why are these findings important? There are two main reasons.
1. These findings might explain the inconsistencies that we often see in research studies, since any given sample of bulimia nervosa patients (for example) is going to have a mix of different personality subtypes. This might explain contradictory results with regard to serotonin activity, for example, or associations with various personality disorders (say, if one sample has a higher percentage of “high-functioning” patients and another has a higher percentage of “dysregulated” patients, which may occur depending on how the sample was collected).
2. Perhaps more importantly, understanding the different personality subtypes may help clinicians treat patients better. How? By understanding the unique “patterns of thinking, feeling, and regulating impulses and emotions,” as well as knowing that a particular patient may require longer time in treatment or may have lower rates of recovery under conventional treatment modalities, the clinician might be able to change their treatment approach to focus on targeting “both personality and eating disorder symptoms.” This is in contrast to using the same CBT approach with different patients who may have very different personality characteristics (and thus, different “reasons” or perhaps “drivers” for engaging in symptoms.)
In any case, this is very interesting stuff, and I’m interested to see where it goes. And by the way, I think I would definitely fit into the “high-functioning group”: I’m quite an anxious person and I am somewhat perfectionistic (though I’m working on it, as you can probably tell: I don’t over-edit these posts, and while that means there might be a lot of annoying mistakes, it also means I’m not spending hours just editing spelling and grammar, which is a good thing in my books.)
What about you? And do you think having your therapist focus integrating both your personality characteristics and eating disorder symptoms in their treatment methodology would be (or would’ve been) helpful in your recovery?
Thompson-Brenner, H., & Westen, D. (2005). Personality subtypes in eating disorders: validation of a classification in a naturalistic sample. The British Journal of Psychiatry, 186, 516-524 DOI: 10.1192/bjp.186.6.516