Reports that eating disorder (ED) rates are rapidly increasing seem nearly ubiquitous, but are rates actually increasing? Are EDs at an “epidemic” level? I came across a recently published study suggesting that this may not be the case; indeed, ED rates might actually be decreasing, at least in the Netherlands.
In the study, Smink and colleagues (2015) followed a group of general practitioners (GPs), servicing roughly 1% of the total population, asking them to record all the newly diagnosed patients with anorexia nervosa (AN) and bulimia nervosa (BN) between 1985-1989, 1995-1999, and 2005-2009. They were interested in whether incidence rates changed or remained stable over time.
Incidence refers to the number of new cases of a disease or disorder in a population over a certain time period; it is not the same as prevalence, which refers to the total number of individuals suffering from the condition at a given point in time (or over a certain period of time).
I have previously blogged about a review of prevalence and incidence rates in EDs here.
A total of 82 GPs participated in the study, seeing between 135,000-150,000 patients during each of the three study periods.
In the Netherlands, 99% of the population has health insurance and GPs play a critical role as “gatekeepers” to specialized care. Even when individuals contact specialists directly, GPs are always notified.
The same ED criteria was used throughout all the three study periods. For each patient who might have an ED, GPs reported information regarding ED symptoms, height, weight and comorbidity, among other things. The researchers then made ED diagnoses using the DSM-IV criteria based on the information provided by GPs.
- Incidence of AN stayed generally the same throughout the three study periods:
- During 1985-1989, the incidence rate (IR) was 7.4 per 100 000 persons per year (person-years)
- During 1995-1999, the IR was 7.8 per 100 000 person-years
- During 2005-2009, the IR was 6.0 per 100 000 person-years
- During 2005-2009, the average age of detection was 23.4 years (median: 19.7)
- Incidence was highest among 15-19 year olds during all three periods
- Incidence of BN decreased throughout the three study periods:
- During 1985-1989, the IR was 8.6 per 100 000 person-years
- During 1995-1999, the IR was 6.1 per 100 000 person-years
- During 2005-2009, the IR was 3.2 per 100 000 person-years
- During 2005-2009, the average age of detection was 24.8 years (median: 21.6)
- Incidence was highest among 15-19 year olds during all three periods
HOW DO THESE RATES COMPARE TO PREVIOUS STUDIES?
In a previous paper by the same group, the authors reviewed many prevalence, incidence, and mortality papers. The rates varied widely, which is unsurprising given the methodological differences among the studies. Looking at studies assessing incidence in general/primary practice (as this is what the current group did), I do think the rates are fairly consistent:
In 2005, Currin et al. reported AN incidence rates in the UK of 4.7 per 100 000 in 2000. For BN, Currin et al. reported incidence rates of 6.6 per 100 000 in 2000.
Curren et al. (2005) concluded:
The incidence rate of AN in primary care has remained extremely stable over the 12 years studied. In contrast, reported cases of BN increased during the same period. However, the peak in bulimia cases seen in 1996 was followed by a subsequent decline for the remainder of the study. This decline was almost entirely explained by the decrease in incidence rates for females aged 20-39 years … In contrast, the incidence of bulimia nervosa in women aged 10-19 years has remained relatively stable.
The authors provide some interesting hypotheses about these trends (the article is open access, so you can freely read their discussion here).
In 2013, Micali and colleagues also reported that incidence rates in the UK of BN and AN remained stable between 200 and 2009. However, they did note that rates of EDNOS increased — a diagnosis that the current authors did not assess.
In 2006, Keel et al. reported a lower point prevalence (prevalence assessed at a single time point) of BN among college students in 2002 and 1992 relative to 1982, concluding that “BN point prevalence decreased significantly in women over the period of observation.” Similarly, in 2008, Crowther et al. reported relatively stable point prevalence of (probable) BN cases between 1990 and 2004.
WHAT TO MAKE OF THESE FINDINGS?
First, Smink et al. suggest that these (and other) findings dispute the notion that BN is more prevalent than AN. Personally, I’m skeptical of this conclusion. It is not that I don’t believe the data, but I think something else must be going on.
Perhaps there is something with the way we are diagnosing EDs that is making it seem like rates of BN are decreasing when, in actuality, they are not really decreasing — they might just be shifted to EDNOS, or maybe even BED?
Or maybe this has to do with the fact that the majority of individuals diagnosed with AN go on to develop bingeing and purging symptoms, and many experience diagnostic crossover to BN (see this post and this post). However, they are first diagnosed with AN and may never be re-diagnosed with BN. (And if they were re-diagnosed, this might complicate measuring incidence rates given that the same person would be essentially counted twice.)
Second, Smink et al. also refer to Keel et al.’s explanation of what could be underlying the decreasing prevalence of BN:
They suggested that, by normalizing being overweight, a secular trend of an increasing body mass index (BMI) of the general population (Ng et al. 2014) might reduce the risk for developing BN. … In a changing weight landscape, where a fuller-figured body is the norm, there might be less pressure to aggressively counteract the effects of binge eating by means of purging.
Smink et al. also suggest that the increased focus on obesity might have led to an increase in binge eating (as a consequence of restriction), thus increasing rates of binge eating disorder (BED) (which the authors did not assess), but not BN. Again, I have a hard time buying this explanation because I am not convinced that there has been a decrease in weight-related stigma. Yes, more people are fat now than before, but is fatness any less stigmatized? It is hard for me to say; I would be very interested to hear your thoughts on this.
I have an alternative hypothesis: Increased rates of overweight and obesity in the general population might be leading clinicians to overlook EDs among those who are overweight or obese. Although serious restrictive EDs occur at any weight, I don’t think clinicians are any less stigmatizing than the general public and I suspect that many might view disordered eating behaviours as justified among those who are overweight or obese. Sim et al. (2013) reported on a few case where EDs were missed by clinicians because the patients were overweight or obese (I mentioned the case report in this post).
Third, the authors provide other explanations, such as the surge of self-help and internet resources to help individuals overcome their EDs. They also suggest that maybe the previously higher incidence rates were due to the “influx of more longstanding cases” that were previously undetected. Again, I feel these are weak explanations for such a substantial drop in incidence rates (although these factors probably contribute).
Finally, I found it very curious that the average age of detection was early-to-mid 20s. While the incidence rates were highest among the 15-19 year old group, I would still expect the average and median ages to be lower than they are. Granted, I am not sure what the average age of detection is in other countries or what has been reported in other studies. (Maybe someone can look this up?)
The study has several limitations that are important to mention.
First, only 34.2% of GPs participated in all three study periods (64.2% participated in the first two periods).
Second, EDNOS and BED diagnoses were not included. While the authors cannot be blamed for this (the study was set up in 1984, before EDNOS or BED were defined in the DSM), I do think this is an important limitation given that EDNOS is likely the largest ED group.
The authors conclude that while the incidence of AN has remained largely stable, the incidence of BN has decreased significantly over the past few decades. While this is consistent with many previous findings, I think it is too soon to suggest that the true incidence of BN has really been halved in a 15-20 year period.
Readers, what do you think? Did I miss something in my analysis? Do you think other possible factors might explain the observed trends? I’m curious to hear your thoughts!
Smink FR, van Hoeken D, Donker GA, Susser ES, Oldehinkel AJ, & Hoek HW (2015). Three decades of eating disorders in Dutch primary care: decreasing incidence of bulimia nervosa but not of anorexia nervosa. Psychological Medicine, 1-8 PMID: 26671456