Who Gets Treatment? Your Ethnicity Matters

In 2010, I wrote a literature review on eating disorders in women of color in North America. I expected to find only a few articles on this subject – every lecture in my undergrad psychology classes, every piece of information targeted to the public, every discussion I had, it seemed, either omitted the existence of EDs in non-stereotypical (white, female, heterosexual, adolescent, upper/middle-class) populations altogether – or glossed over it with a footnote on “acculturation” that reductively attributed the disorder to a misguided desire to fit into the dominant culture, much as other women might aspire to look like the images of female bodies in mass media. (Acculturative stress is actually far more complex than this, and furthermore is not necessarily the sole or even the primary cause of EDs for all people of color.)

[Some people] think that I hate being Asian and want to look white, which I find incredibly insulting, as I’m actually quite happy with the physical traits associated with my ethnicity and have a lot of respect for its culture. Eating disorders are a maladaptive way for people to cope with stressful emotions and situations. It has nothing to do with wanting to look like a model on the cover of a fashion magazine. There’s already a lot of shame associated with having an eating disorder, and when you add these stereotypes into the mix, it makes it even harder for sufferers to reach out and seek help.1

Contrary to expectation, my first database search turned up pages of articles. This was obviously an area that was being explored by researchers; there was a good body of work confirming the existence of EDs in people of color, so why wasn’t this knowledge being put into practice?  Why were people of color vastly underrepresented in every treatment program I’d ever been in, inpatient or outpatient?

And it’s not just clinical practice – Gordon, Perez & Joiner’s 2002 article The impact of racial stereotypes on eating disorder recognitionestablished that in a sample of 160 undergraduate students who read a passage about a girl who exhibited ED symptoms, participants were far more likely to recognize these descriptions of disturbed eating behavior as signs of an eating disorder when the girl in the story was white, rather than black or Hispanic. (160 undergrads is obviously a limited sample, numerically and demographically, so while one can’t draw any conclusive widespread generalizations from that study, it does document the existence of this phenomenon in a quantifiable way, rather than just anecdotally.) This can become a self-perpetuating vicious cycle:

Saren - 2012 - Figure 1

That dynamic has contributed to one of the problems plaguing much prior research: EDs are assumed to affect only certain groups; this assumption is disseminated in popular and professional press, so clinicians are less likely to recognize EDs in non-stereotypical populations and those populations are less likely to conceptualize their behavior as an eating disorder, thereby creating a health disparity around treatment accessibility.

This phenomenon not only has serious consequences for individuals, but also influences research. A large number of studies which established the “canon” of ED knowledge (e.g., “fear of becoming fat” being the core psychopathology of the disorder) have drawn their samples from individuals in treatment or who have been diagnosed with an ED. And who is most likely to a) be referred to or b) seek out ED treatment? You got it. So we end up with confirmation bias: hey, looks like our sample is overwhelmingly white, adolescent and female, so it looks like EDs mainly affect that population.

The other problem this creates is regarding generalizability: we can’t be sure that results obtained from research using a particular demographic group can be applied equally to people in other demographic groups, as I mentioned above. Not every study will be perfectly applicable to the larger population – that doesn’t mean it’s not potentially still useful or valuable. But it doesn’t mean it always is, either. And since what we know about eating disorders has historically been somewhat socially constructed (because, to quote Dr. David Alexson on the DSM revisions, “The brain is so darn complicated!”) it’s important to be a critical consumer of knowledge. (That’s one reason this blog exists!)

A number of women I knew shared their stories of having an eating disorder as a person of color, and for a majority, their experiences reflected these disparities. In their own words:

Unless I come out and tell a medical professional, a friend, co-worker, professor, etc. that I have an eating disorder, they don’t suspect anything. A lot of conditions have been suspected: crack addiction (by far my favorite), diabetes, hyperthyroidism, other drug addictions. But no one ever suggests that I have an ED.

When I first started purging, I didn’t consider myself “bulimic”. Eating disorders weren’t something that black girls had. I told myself I was just ‘watching my weight’ and there was nothing wrong with it. After a few months of restricting, purging and over-exercising, I ended up in the hospital with a swollen leg. Doctors couldn’t figure out what was wrong, but didn’t address the fact that my resting pulse rate was 38, that I didn’t eat anything while I was there, and that I kept asking for laxatives.

When I began to realize I had a problem as a pre-teen and did some research, I came upon Anorexia Nervosa. I thought to myself “I can’t have this, I’m not white! I just have weird habits.” … A year later my parents would end up putting me in [inpatient treatment]. I was the only person of color but it didn’t make a difference. The doctors didn’t treat me differently; in fact they told me eating disorders are more common in minorities than we’re lead to believe.

It is extremely easy to go under the radar. I lost [a significant amount of weight] in like three or four months. I was also constantly in the E.R. because I was having other issues with inflammation. They weighed me all the time, and no one mentioned anything about the fact that I lost [so many] pounds so fast … I also went eight days without eating and no one noticed except my girlfriend at the time.

In their 2003 article, Becker, Frank, Speck & Herzog examined ‘Ethnicity and differential access to care for eating disorder symptoms’ among participants in the 1996 National Eating Disorders Screening Program (NEDSP), a voluntary educational program held on college campuses in the US. They wanted to look at two questions:

(1) whether there are disparities in access to care between ethnic minority women compared with non-ethnic minority women for eating disorder symptoms and concerns, (2) whether patient-specific help-seeking factors or clinician factors have an impact on ethnic disparities in access to care for eating disorder symptoms.

While ethnicity-based disparities in treatment for eating disorders had not been studied extensively at the time the article was published, the authors based their hypotheses on past literature establishing that those in ethnic minority groups are significantly less likely to receive or seek out care for health issues (including mental health.) They provide three main reasons for this inequality:

(1) diagnostic mismatch: definitions that don’t take into account the different ways that symptoms may be manifested or expressed by diverse populations,

(2) variance in motivation to seek out help from medical professionals “for a variety of reasons, ranging from a differential in perceived seriousness of the symptoms or illness to a differential in perceived clinician receptivity,  empathy, and trustworthiness”, 

(3) clinician error or bias, which “ranges from clinician prejudice (van Ryn & Burke, 2000), to clinician unawareness of or insensitivity to cultural differences in presentations of psychopathology (Neighbors, Jackson, Campbell, & Williams, 1989), to selective information processing for clinical decision making (e.g., use of subjective judgments about base rates of illness among certain patient groups; Lopez, 1989).”

Becker et al. used two studies to examine these phenomena in relation to eating disorders, using the descriptive data gathered during the NEDSP.


  • 9069 participants from 409 college sites in the US completed self-report questionnaires.
  • Participants were sorted into 5 ethnicity groups: Caucasians (81.6%), African-Americans (6.6%), Latinos (3.4%), Asians & Pacific Islanders (3.0%), Native Americans or Alaskan Natives (.09%). 2.5% identified themselves as an ethnicity other than those listed.
  • 84% were female.
  • 9% had previously received either an ED diagnosis or treatment, while 84% had not. More Caucasians had been diagnosed or treated than African-Americans or Asians. 
  • 5787 participants met with a counselor in person to discuss their results and if necessary, be referred or recommended for further evaluation.
  • These participants were sorted by self-reported symptom severity, using a standardized algorithm for classifying into one of three groups: No referral necessary (24%), Further evaluation indicated (74.8%), or Urgent evaluation indicated (1.2%)
  • Using logistic regression, ethnicity was shown to play a significant role in predicting whether or not a referral was made. African-American and Latino participants were less likely to receive a referral, compared to Caucasians. 
  • Scores on cognitive symptoms, behavioral symptoms  and distress accounted for some of the variance in referral patterns, but even after controlling for differences in severity of symptoms, Latino and Native American participants were still less likely than Caucasians to receive a referral.
  • BMI was also found to predict some variance and was added to the regression equation, but even after controlling for BMI (in addition to symptom severity), “Latino participants were still less likely to be referred for further evaluation than were Caucasian participants, revealing significant ethnic disparities in access to care for eating disorders symptoms in this setting (Table 1)”:
Becker - 2003 - Table 1


  • 289 participants (708 were randomly selected from a subgroup (~49%) of the original NEDSP subjects that had agreed to be contacted in the future. 310 met criteria for the second study, of which 21 declined to participate.) Subjects in Study II on average scored higher on measures of impairment, distress and symptoms (see Table 1 above) than those in Study I.
  • 92.1% Caucasian, 4.3% African-American, 1.8% Asian, 1.1% Latino. Because of the small number of ethnic minority participants, they were grouped into a single category (n=22 people).
  • Ethnicity and help-seeking behaviors:
  • Subjects were asked to what extent they had volunteered information about their eating disorder symptoms to their friends, family members, and professionals before the NEDSP. Among the individuals who acknowledged eating disorder-related concerns, no significant differences were found between minority and non-minority participants on the frequency with which they voluntarily disclosed their symptoms in any of these settings.” No significant between-group differences were observed in those individuals recommended by an NEDSP counselor to seek further evaluation.
  • However, 1.5 to 2 years after the NEDSP, those with ED-related concerns in the minority group were less likely to have sought further treatment.
  • Ethnicity and clinician response: 224 of the 289 participants acknowledged concerns about ED-related symptoms. This group was asked how often they were asked about their concerns by family, friends or professionals. While the minority and non-minority groups were asked by their family members at similar rates: “fewer ethnic minority participants reported that their friends expressed concern… they were asked by their doctors about eating behaviors significantly less frequently than were non-ethnic minority participants…and also were asked less frequently even by mental health professionals,” demonstrating a trend.
  • The minority and non-minority groups were equally symptomatic, yet among those with ED-related concerns, only 31% of the minority group received a recommendation to a mental health professional (compared to 60% of the non-minority group.)

The authors found this evidence of clinician bias especially disturbing because it occurred in the highly standardized setting of the NEDSP, where they were given a specific set of instructions to adhere to for making referrals. Without this structure, would the difference between groups have been greater, or would the rates of referrals fall for both?

There were several limitations to this study. First, the authors acknowledge that the sample was drawn from a select population: Those in college who chose to participate in the NEDSP, and that self-selection of those who probably already identified their behavior to some extend as being eating disordered may have impacted the very small number of participants who were ethnic minorities. This small number of individuals meant that they had to be combined into a single group, meaning that differences between, say, African-Americans and Asians could not be examined. (This is important, as “ethnic minority” is hardly a homogeneous group!)  And “Finally, the methods for this study only demonstrate that help seeking patterns and clinician response differ between ethnic minority and non-minority individuals with eating disorder concerns. They do not address the important and complicated question of how those differences originate.” (Additionally, this study used data from over a decade ago, and I wonder if there’s been any progress made since then, given the number of articles published on eating disorders and ethnicity in the past decade – or whether clinical practice has not yet caught up.)

The authors conclude:

The disconcerting results of these studies should highlight a need for clinicians to examine carefully their biases and behaviors toward ethnic minority and non-minority patients concerning eating disorders that are (mistakenly or not) believed to be more prevalent among certain ethnic groups. This is a particularly exigent matter given the projections that minority populations will approach 50% of the U.S. population in the next 50 years (Bureau of the Census, 1992; Passel & Edmonston, 1992)… Further research is necessary to gain a better understanding of ethnicity-specific factors that may influence the perceived seriousness of eating disorders or the receptiveness of clinicians to addressing them.

Any thoughts on what these “ethnicity-specific factors” might be?
Have your personal experiences with family, friends or clinicians been similar or different to the results of this study?
What do you think needs to change in order to equalize access to treatment resources for eating disorders? 


Becker, A., Franko, D., Speck, A., & Herzog, D. (2003). Ethnicity and differential access to care for eating disorder symptoms International Journal of Eating Disorders, 33 (2), 205-212 DOI: 10.1002/eat.10129


Saren has a BA in psychology and is currently pursuing a PhD in clinical psychology. Her research involves using neuroimaging (fMRI) and other psychophysiological methods to study social, cognitive, and emotional processes in affective disorders.


  1. Saren, I’m confused at this: “9% had previously received either an ED diagnosis or treatment, while 84% had not” – why doesn’t this add up to 100%?

    I think the limitations you noted, particularly about this already being individuals who are already self-selecting. I bet the numbers are worse for the population as a whole.

    I wonder if part of it has to do with being open about the ED to family members? It is hard to participate in treatment when you are an adolescent or even young adult, without parental knowledge. Perhaps ethnic minorities don’t feel comfortable coming out about this to their families, who may view it as a “Western” disease?

    • I went back to check – from p. 208:

      “Ethnicity and History of Accessing Care for an Eating Disorder before the 1996 NEDSP: Of the subjects completing a self-report screen, 84% had never been diagnosed with or treated for an eating disorder whereas 9% had received a diagnosis and/or treatment for an eating disorder. Differences were found between minority and non-minority participants on this variable, which was used as a proxy for previous access to care for an eating disorder. Specifically, more Caucasian participants (9.7%) responded in the affirmative than did African-American (3.9%;  2¼ 20.17) or Asian participants (5.9%; 2¼ 4.16). Caucasian participants did not differ from Latino (8.2%) or Native American (12.8%) participants.”

      My guess is that the remaining 7% is the participants who either skipped, declined to answer or missed that question – pretty common on survey data, I had that happen on my harm reduction survey where the n for some questions was 161, 160, 159 because some people didn’t answer all, either intentionally or accidentally.

      Voluntary screening programs as a rule tend to capture participants who already have some level of concern or interest in the issue being screened for, so I really liked that the authors DIDN’T make a big deal of how omg 76% of participants warranted further evaluation. I wish I had a copy of the criteria they used for discriminating between those groups (no/some/urgent evaluation indicated), I’d be interested in knowing as well if false positives were an issue… but that’s tangential.

      For the women who shared their stories with me for my lit review, the difficulty of being open with family members definitely arose as a common theme. I can’t speak to that myself, as a white person whose family is fairly well-versed in psych issues (though I was extremely loathe to disclose to my family, albeit for other reasons) but perhaps some readers can?

  2. There are some other studies that may be interesting for you to add here. I don’t have the exact citations (sorry!), but I believe that Ruth Striegel-Moore was one of the authors of some studies on this topic. I think there was one specifically about seeking help for binge eating that found that Mexican-American women were more likely to be prescribed *diet pills* than referred to anywhere that might address behavioral/psychological/physiological issues.

    One of the additional findings of that study was that something like half of the people who believed they -needed- treatment actually sought treatment (sorry- don’t have exact numbers handy!). Then, a large percentage of those who actually -did- seek treatment said that their first experience was so invalidating/traumatic/frustrating/etc that they would be reluctant to ever attempt to get treatment again!

    • Thanks, I think I’ve read the ones you’re referring to and used them in the lit review. I’m a big fan of Striegel-Moore; she does some really interesting work and I’d definitely like to write about some of her articles in future posts.

    • Ineffective/”bad” treatment is worse than no treatment in some ways, in my opinion. That’s one of my problems with the current models of ED treatment which have extremely poor rates of long-term efficacy (residential, day programs, etc.): going through treatment and then relapsing, as is common, sets up a belief that if one has “failed” treatment, one is beyond hope – particularly in people who have had their EDs for a long time and have a history of being involved in the mental health system via the “revolving door” of treatment, there’s a certain learned helplessness that develops, where treatment is really not doing anything for them yet they’re dependent on it, or at least perceive themselves to need it.

      (this is mainly based on personal observation of friends & people I was in residential with years ago, though I do have a number of citations for chronicity and self-efficacy beliefs and motivation, etc. if anyone is interested. I’m still pulling together my thoughts on this area – it’s something that I feel strongly about; I so appreciate the work that others do w/r/t insurance coverage for EDs because it is a critical issue, but the state of ED treatment is just such an ineffective mess I have a hard time getting worked up about it in some ways. I have people that I was in treatment with five years ago on my Facebook, and a lot of them are still in the relapse/re-admit pattern they were in back in 2007. And these are YOUNG women, some of them not even in college yet. If something doesn’t work after several attempts, why not try something else? (What’s that definition of “insanity”?) Well, because we don’t yet have something else. (Yes, there’s FBT but that has a number of specific requirements that make it incompatible with all individuals with EDs, as it requires not only particular family dynamics and structure but also factors like disposable income, ability to take time off from work, social support by spouse/family/friends/workplace, etc. CBT works, but only for less severely symptomatic bulimics. And so on.)

    • An article from earlier this year by Stephanie Covington Armstrong that mentions both her experience as a Black woman with an eating disorder, and a few studies on the prevalence of EDs in ethnic minority populations. (I’ve read the Goeree, Ham & Iorio one she mentions and it’s quite good.)

  3. This is so interesting and I think, a very important post – stereotypes are alive and well in our communities and people DO get excluded from accessing help or even a diagnosis because they don’t fit the stereotype.

    I would love to read about what the incidence of eating disorders in third world countries is, especially places that are experiencing famine, grave poverty etc.

    I would imagine that there would be more people in these places with eating disorders than any of us imagined – because despite the old cliches of how thankful we would be for the food if we were in a third world country etc, usually eating disorders go much deeper than surface issues. I myself grew up in poverty conditions and yet, I have an eating disorder, despite the stereotype of them being something middle class/wealthier people get.

    • “I would love to read about what the incidence of eating disorders in third world countries is, especially places that are experiencing famine, grave poverty etc.”

      Incredibly difficult to study. One of the major problems is that the way their disorders present themselves might be completely different, or very different, and thus, using our criteria (ICD-10, DSM-IV) would yield very low rates, even though disordered eating might actually be much more prevalent.

Comments are closed.