Are EDs really on the rise? No, 10% of teen girls DON’T have an ED

It’s a question I get a lot: how many people out there have eating disorders? I can take some reasonable guesses, but there’s not much in the way of good, population-wide prevalence studies.

But new estimates of ED incidence (the number of new cases diagnosed in a particular population in a specific time period) and prevalence (the total number of current cases of disease in a population at any point in time) were just published from the UK. Already, the study is being horribly, awfully, make-me-see-red misinterpreted in the news media. {For the record, I was going to blog on this study even before my blood pressure skyrocketed.}

Luckily for you, the research was published Open Access (meaning you don’t have to pay a ginormous fee to subscribe to the journal or read an individual article) so you can read the research for yourself. If only some journalists out there would take the time to do the same. Although the incidence of AN and BN in females remained stable during this time period, EDNOS diagnoses increased significantly in females. Males showed similar patterns of increase in EDNOS, but not AN and BN (Micali et al., 2013).

What the researchers found

The scientists counted the number of ED diagnoses entered in the General Practice Research Database between January 1, 2000 and December 31, 2009. The GPRD “contains anonymised records representing about 5% of the UK population.” Although it’s a fairly small fraction of the population, the GPRD was designed to be representative of the overall population in terms of geography, gender, and age. The researchers identified all subjects between ages 10-49 who had data in the GPRD because they believed it would capture the great majority of new ED diagnoses in the population. The patients received diagnoses of AN, BN, or EDNOS according to ICD-10 criteria.

ed incidence by age

The researchers identified 9120 first-time ED diagnoses in the GPRD during the study period. A small subset of these patients received diagnoses of both AN and BN. All but 21 were classified by the researchers as having either AN or BN. The remaining 21 were classified as EDNOS based on symptoms and BMI. Of the 9120 cases, 2134 (23.5%) were AN, 3433 (37.8%) were BN, and 3505 (38.6%) were EDNOS.

Over the decade, rates of AN and BN in females remained constant. However, overall rates of EDs increased, driven entirely by an increase in EDNOS diagnoses, from 17.7 cases per 100,000 population in 2000 to 28.4 per 100,000 in 2009. During the 2000s, EDNOS became the most common ED diagnoses in females age 10-49.

Incidence rates for EDs remained significantly lower in males, but showed a similar pattern: an overall increase in ED diagnoses driven by an increase in EDNOS diagnosis. Males and females did show slightly different patterns at the age of diagnosis, however. The highest incidence of all diagnoses in females was between 15-19: 164.5 per 100,000 people (or 0.164% of females between 15 and 19 were diagnosed with an eating disorder during this time period).  Remember this percentage- I will come back to it shortly. In males, AN diagnoses peaked between 15-19, whereas BN peaked between ages 20 and 29. EDNOS, just to make the data more confusing/interesting (circle one), peaked in males between 10-14.

ED incidence for females

ED incidence for males

This type of study is really important because we honestly don’t know that much about EDs in the community. The problem with this study (that the authors acknowledge) is that it is looking at formally diagnosed eating disorders rather than actual ED incidence. Many people with EDs don’t come to medical attention for their disorder for any number of reasons, not the least of which are stigma, an unawareness of the illness, an unwillingness to disclose the specifics, and other larger social barriers that interfere with a person’s ability to seek care. Then there’s the issue of GPs and PCPs actually recognizing and diagnosing an ED. How many of us have been fobbed off with “It’s just a phase,” or “She’ll eat when she’s ready,” or “He’s not that thin- it’s no big deal,” or “People like you (males, minorities, lower socioeconomic status) don’t get EDs.” I’ve told physicians that I thought I had an ED and they told me that it was just a thyroid problem and (at the ripe old age of 20), I was too old to have AN.

So yes, these numbers are likely significantly underestimating the true prevalence of EDs in the community.

As well, the reasons for the increase could come from two main sources: we’re doing a better job of diagnosing EDs (that is, people with EDs are actually getting the diagnosis they need, but there’s the same general portion of the population that has an ED) and/or EDs are actually increasing. It’s impossible to say for sure. Let me repeat this: the study never says that EDs are on the rise. Instead, it says that the diagnosis of EDs is increasing. There is a VERY LARGE difference between these two things.

The study hits the media

I like to rip on the Daily Mail (or the Daily Fail). I don’t feel bad about it because they really, honestly deserve it. Take their headline from this study:

One in ten teenage girls has an eating disorder and boys as young as ten are also at risk

WHAT?!? Did they read the same study as me? Remember that percentage I asked you to remember? The 164 per 100,000 people? This means that the study found that 0.164% of teen girls are diagnosed with an ED each year.  The lead author was quoted as saying ‘The absolute numbers affected by eating disorders are large. It’s 5 to 10 per cent among adolescent girls, and closer to 10 per cent.’ Which may be true (if it is, my guess is that the numbers are probably on the lower end), but that’s not at all what the study found.  Even so, studies in the US have found that the lifetime prevalence of all ED diagnoses is around 4-5% or so- far less than the 10% statistic that’s being bandied about.

The rest of the article wasn’t all that bad, outside of the usual blaming of pop culture for the increase in EDs. Certainly it’s possible, and certainly it’s a contributor, but the fact is that we don’t know this.

The problem is that the headline is what got picked up by other news outlets.

Study: One in 10 Teen Girls Has an Eating Disorder (The Fix)

Anorexia and bulimia rise 15% in under a decade (The Mirror) (AN and BN were the only diagnoses to remain constant, actually)

Granted, not all news stories appeared to horribly botch the coverage, but plenty did, too. Just another friendly reminder not to believe everything you read.

Twitter Digg Delicious Stumbleupon Technorati Facebook Email

9 Responses to “Are EDs really on the rise? No, 10% of teen girls DON’T have an ED”

  1. And the usual pick up in the Guardian comments page suggests banning television… I’m not entirely sure how “the media” or any of the social commentators out there can balance the rise in ED-NOS with a media drive to show skinny models. Surely if that was the driving factor, you would have expected to see the rise in AN rather than in ED-NOS.
    (Carrie – I think there’s a typo in the last section, and a . snuck into the 164/100,000 figure…)

    • Carrie Arnold May 22, 2013 at 9:26 am

      Actually, I would expect the increase to be in BN/EDNOS. BN appears to be a culture-bound syndrome: that is, the genes are only expressed in certain environments. For one, you have to have readily available food that someone can eat. Whether it’s more sensitive to sociocultural pressures is unclear. But the data on AN incidence/prevalence has shown that this diagnosis has been fairly steady across many locations over time.

  2. 164.5/100000 is actually 0.1645% or 1.645 per thousand. In the study they round it to 0.2%. They note the incidence of depression in the same age group as 11/1000 or about 1%. These are still low numbers, but of course this is the new diagnosis incidence and not the prevalence in the population which will be much higher due to the chronicity of ED’s. No idea where 10% comes from.
    I am not sure what to make of the figures to be honest. It does seem a very low incidence compared with what you see, but then again maybe I am a bit biased. Maybe it just says that GP’s are really lousy at making the diagnosis. I was surprised that only 1% of adolescent girls were being diagnosed with depression. They must be a happy bunch in the UK.

    • Carrie Arnold May 22, 2013 at 9:24 am

      Typos fixed. Thanks!

      You’re right- these incidence rates seem very low, both for depression and for EDs. The chronicity factor accounts for some of the lowness, but not all. I think many MDs aren’t that good at diagnosing mental health problems.

  3. Just eyeballing the graphs I’m not convinced about the linear association. What’s the 2002 peak in EDNOS due to for example?

  4. I hate it when the paper itself isn’t too bad, but the press release or the media commentary seems to be talking about a totally different study. Andrew Wakefield being a case in point…

  5. …jesus….


  1. Some statistics: how common are the various types of eating disorder? | Anorexia and other eating disorders: How to help your child eat well and be wellAnorexia and other eating disorders: How to help your child eat well and be well - March 21, 2014

    […] Carrie Arnold’s blog and most importantly, the comments and discussions it raised:  and also […]