On 11 August 2026, Germany’s Federal Statistical Office published its weekly figure. The subject was eating disorders, and the numbers are exact: 12,400 inpatient treatment cases in 2024, 2.2 per cent more than in 2023 and 8.0 per cent more than in 2004. Women accounted for 92.9 per cent of those treated. Girls aged 10 to 17 made up 48.1 per cent of all cases, against 24.9 per cent in 2004. In absolute terms, from just over 2,800 to just under 6,000.
This is an official statistic. It appears annually, it is broken down by age and sex, and the underlying database goes back to 2000. For our subject, that is unusual. On most of the questions this magazine asks, Germany offers no official number at all, only estimates, surveys and references to US research.
Which makes it worth looking closely at what the figure measures, and what it does not.
What is actually being counted
Germany’s hospital diagnosis statistics count treatment cases, not people. Anyone admitted twice in the same year under the same primary diagnosis appears twice. The statistical office says so explicitly in its methodological note. Reporting the figure as “6,000 girls” reports it wrongly.
The count also captures only those who reach a hospital bed. Inpatient case numbers are not a measure of how common an illness is. They are a measure of how common it is, multiplied by referral pathways, admission thresholds and available capacity. A rising number can reflect more illness, a more attentive diagnostic practice, expanded child and adolescent psychiatry capacity, or thin outpatient provision that pushes people towards inpatient care. The statistic alone cannot tell you which share belongs to which factor.
And one claim that circulated this month is not supported by these data: that the current thinness backlash in fashion and social media produced these hospital figures. The rise runs across twenty years. You cannot draw a causal line from a two-year trend to a time series that starts in 2004. We are not drawing it.
We checked the underlying data ourselves in Genesis, table 23131-0002, reporting year 2024, primary diagnosis ICD-10 F50. The result: 11,500 female and 878 male treatment cases, 12,378 together, rounded to the 12,400 of the press release. Among girls, 2,602 cases fall in the 10-to-under-15 band and 3,358 in the 15-to-under-18 band, 5,960 together, the “just under 6,000”. For boys in the same span the figures are 183 and 155, 338 together. The press release checks out.
The code list decides what exists
Eating disorders sit in diagnosis group F50 of the ICD-10-GM, the German edition of the diagnostic code. That group has eight entries: anorexia nervosa in three variants, atypical anorexia nervosa, bulimia nervosa, atypical bulimia nervosa, overeating associated with other psychological disturbances, vomiting associated with other psychological disturbances, other eating disorders, and eating disorder unspecified.
There is no entry of its own for binge eating disorder.
This is neither an oversight nor a matter of dispute. It is how the register reads at the Federal Institute for Drugs and Medical Devices, which issues the ICD-10-GM. Binge eating disorder is listed as an independent diagnosis only in ICD-11, as 6B82. German hospitals still bill under ICD-10-GM; the institute publishes the 2026 edition.
In practice: anyone treated as an inpatient for binge eating disorder does not appear in the statistics as such, but in one of the residual categories, or not in F50 at all. The question “how many people in Germany are treated in hospital for binge eating disorder” has no official answer, because the code list does not know the diagnosis.
A second entry in the same register is just as telling. The code for overeating associated with other psychological disturbances carries an explicit exclusion: overweight, to be coded under E66. The fat body does have a place in the diagnostic code. Just not in the chapter on mental disorders, but in the one on metabolic disease.
What the numbers then show
The most common diagnosis in 2024 was anorexia nervosa with 9,000 treatment cases, ahead of bulimia nervosa with 1,300. Together that is 10,300 of 12,400, a little over 83 per cent. The remaining 2,100 or so are spread across the atypical forms and the residual categories.
Put differently: Germany’s eating disorder statistic is overwhelmingly a statistic about two diagnoses, both of which are defined through weight loss and weight control. That does not make it wrong. It makes it something other than what it is read as, namely a picture of disordered eating in Germany.
That matters for fat people. If the only robust figure on eating disorders captures almost exclusively diagnoses that presuppose a thin or actively weight-reducing body, then disordered eating among fat people is systematically absent from the official record. It is missing from the press release, from the headlines built on it, and from the debates that lean on those headlines.
The same pattern, two attributes
We described this mechanism yesterday, in a different place. Germany’s largest discrimination survey never recorded body weight as an attribute, because the General Equal Treatment Act does not name it. So the data are missing. So the argument for adding it to the law is missing. A closed circle.
With eating disorders it runs in parallel, through the diagnostic code instead of the statute. What is in the code gets counted. What gets counted goes into the press release. What is in the press release becomes the public picture. And what is not in the code does not exist for that chain.
So in both cases Germany does not have a data gap in the sense of an oversight. It has a definition gap, from which a data gap follows. The difference matters, because it says where you would have to start: not with better surveying, but with the category that comes before it.
What we are not turning this into
No reproach to the Federal Statistical Office. It counts correctly what the code prescribes, and it publishes its methodology openly, including the note on multiple counting. No reproach to the child and adolescent psychiatry services that treat these cases. And explicitly no claim about how common any given eating disorder actually is in Germany. That is precisely what this source cannot tell you, and that is the point.
What remains is an observation worth keeping to hand the next time this figure travels through the news: it is real, it is well made, and it describes a section whose edges were drawn decades ago in a classification register.
If this affects you: In Germany, the Federal Institute of Public Health (BIÖG, formerly BZgA) runs a free eating disorder helpline on 0221 89 20 31, Monday to Thursday 10:00 to 22:00 and Friday to Sunday 10:00 to 18:00. Only standard call charges apply. Email counselling is available at essstoerung@bioeg.de. Outside Germany, look for the national helpline in your own country.
Sources
- Federal Statistical Office, Figure of the Week No. 33, 11 August 2026: https://www.destatis.de/DE/Presse/Pressemitteilungen/Zahl-der-Woche/2026/PD26_33_p002.html
- Genesis-Online, table 23131-0002, hospital patient diagnoses, reporting year 2024, ICD-10 F50: https://genesis.destatis.de/datenbank/online/statistic/23131/table/23131-0002
- Federal Institute for Drugs and Medical Devices, ICD-10-GM version 2024, block F50-F59: https://klassifikationen.bfarm.de/icd-10-gm/kode-suche/htmlgm2024/block-f50-f59.htm
- Martina de Zwaan, ICD-11: Veränderungen im Bereich der Essstörungen, Verhaltenstherapie 34 (2024), issue 3-4, pp. 139-146, DOI 10.1159/000542466: https://karger.com/ver/article/34/3-4/139/918978/ICD-11-Veranderungen-im-Bereich-der-Essstorungen
- BIÖG, eating disorder helpline: https://www.bioeg.de/service/infotelefone/essstoerungen/




