Diagnosed eating disorders are more common in some schools than others: schools with greater proportions of female students, and schools with higher numbers of children with university-educated parents. These were the headline results of our study, published last week in the International Journal of Epidemiology (open access).
Eating disorders are serious illnesses (someone with bulimia nervosa is around twice as likely to die young as someone without it; someone with anorexia nervosa about six times more likely), so this might make us worry about the effect of all-girls private or selective state schools on the mental health of young women – but should it?
In clinic, we do seem to see more girls with eating disorders from some schools than others – but no one had studied this before. We wanted to find out whether it really was the case. We also wanted to try to unpick whether any differences were because the schools themselves are different from each other, or because different types of individuals attend different schools.
Eating disorders are more common in the children of highly educated parents, so we would expect to find more eating disorders in a school with more children of university-educated parents. If the differences between schools are still there after we have taken account of each individual’s parental level of education (and other individual risk factors), then that suggests that the school environment itself might make a difference over and above this. The UK doesn’t collect the sort of data you need to answer these kinds of questions, so we worked with the Karolinska Institutet in Sweden, where everyone has a personal ID number and health and social care data is routinely collected.
We used data on 55000 girls from Stockholm County who left “Gymnasium” (school for 15-18 year olds) between 2002 and 2010, including information about who had been seen in an eating disorder clinic or diagnosed with an eating disorder, which school they attended, and individual characteristics like family history of an eating disorder and parental level of education.
We found that there were differences in rates of eating disorders between schools, even after taking into account the individual characteristics of the students within them, suggesting there may be something important about the school environment itself, not just different types of students attending different schools. A girl at a school with 50% girls and where 50% of the parents have university education has a 2.1% chance of developing an eating disorder, whereas at a school with 75% girls where 75% of the parents have university education, a similar girl with the same individual risk factors would have a 3.3% chance.
What we have found is an association – we don’t know if these types of schools actually cause eating disorders. Also, Sweden has strict gender equality laws, meaning that all-female schools do not exist there, so we don’t know how these findings would relate to all-girls schools in the UK. In the article, we speculate on possible reasons why eating disorders might be more common in some schools than others. One reason could be that eating disorders might spread within a school: if some students have eating disorders, that might make it more likely that others develop eating disorders too.
There is some evidence from other studies that extreme weight loss behaviours, body image concerns, and binge eating, all of which can be components of an eating disorder, spread within friendship groups. Another possible explanation is that very aspirational schools (with lots of highly educated parents) encourage their girls to try their best at all times. This might lead to some of them developing perfectionistic traits – which have been found to be linked with developing eating disorders.
Although large data sets like this one are really useful, they do have limitations. They rely on everyone with a health problem seeking help for it – people who don’t go to a clinic don’t get counted in the clinic numbers. Numbers receiving treatment are very low in the United States, for example, where only up to 28% of people with an eating disorder get help for it. We hope they are higher in Sweden, where there is a comprehensive, government-provided health service; but stigma, denial, and a lack of public awareness about eating disorders may still prevent people seeking the help they need. This means that a third possible explanation of our findings is that schools with high numbers of girls and high numbers of highly educated parents may seem to have more students with eating disorders because they are better at encouraging their students to seek treatment.
This leaves us in that common but frustrating position of concluding that we need more research before we know for certain where and how we might be able to intervene in schools to prevent eating disorders. Do we need to increase input into schools with lots of students with eating disorders? Or help those with fewer cases get better at identifying students in need of help? Either way, I think Child and Adolescent Mental Health Services (CAMHS) need to be working closely with school pastoral care teams to make sure that students get the help they need, as a recent initiative in Oxford Health has been doing. Nationally, this also means we need to address the chronic underfunding and cuts to CAMHS, and the fact that some areas have very long waiting times to get help, which leaves some schools trying to pick up the pieces, but that is another story.
Of course, eating disorders are complex illnesses, and what leads any one individual to become ill will probably involve genetic predisposition, personal characteristics and the environment and stressors they encounter. School environments are just one factor that may play a part. The more we can understand about each possible risk factor, the closer we can get to designing better interventions to try to prevent and treat them. And when you are working with illnesses this common, and this potentially serious, this is vital.
This article first appeared on ‘The Guardian’ on 29 April 2016.