The Trouble With Accurate Eating Disorder Diagnoses

Almost anyone who works in a specialized field (such as pediatric eating disorders) spends a great deal of time thinking about diagnostic criteria for diseases and conditions.  This is true throughout medicine and psychiatry, of course:  is it type 1 or type 2 diabetes?  Is it autism or pervasive developmental disorder?  Is it bulimia nervosa or binge purge anorexia nervosa? Is it eating disorder not otherwise specified (EDNOS, a term no longer included in DSM-5) or anorexia nervosa?

As in other fields that deal with very young children, the patterns seen in adults can be present, absent, partly present, mostly absent… You get the idea.  Children with rheumatoid arthritis, for example, may have all or most of the symptoms that an adult with the same disease would have, or they might present completely differently (e.g. Still’s disease) and their symptoms change over time. Children are not miniature adults.

The problems of diagnostic accuracy have been well described, not least by Dr. Thomas Insel, director of the National Institutes of Mental Health (NIMH), who writes about the “gap between modern neuroscience and contemporary psychiatry.” This gap means that in the field of brain disorders we are still living in the era of the 1700’s, when the great biologist Linnaeus developed a system of taxonomy to classify all living organisms from their external characteristics. If the flower looked one way, the plant belonged in the pea family (Fabaceae), if somewhat differently, in the cabbage family (Brassicaceae), etc.  Since long before Linnaeus, physicians have classified diseases depending on the constellation of symptoms they could see with their eyes:  if the patient was driven to exercise, restricted their food intake severely, and believed themselves to be fat, despite all evidence to the contrary, they had anorexia nervosa, if they were not underweight, binged on large amounts of food and then, panicked, got rid of it by vomiting, they had bulimia nervosa, etc.  The human urge to classify, to organize, to arrange in patterns is strong and it has served us moderately well.  But classification based on external descriptors or symptoms is about to be blown apart by genetics, molecular biology and imaging technology.  Already the system of plant and animal classification developed by Linnaeus—amazingly still in use—has been altered by genetic information about the origins of living things.

Why does taxonomy matter?  Does it matter?  I would argue that it is in fact less important than the principles of treatment, especially in the world of pediatric eating disorders.  For example, whether a child has anorexia nervosa binge-purge subtype, or restricting varieties of ARFID is decidedly less important than recognizing that they need more food!  It does seems likely that prognosis could be different within and between diagnostic categories—for example, early onset anorexia nervosa may have a better or worse prognosis than ARFID.

Over the years I have become convinced that there are children, and especially teens, who have imitative forms of the condition we call anorexia nervosa (a brain disorder I hope we will one day see in neuroimaging).  These would be children/teens who refuse to eat for psychosocial reasons.  But which are which?  You can’t tell by looking, that is for sure.  And since the treatment is the same, it is of largely academic interest to try and tease this out.  Only time and longitudinal experience with their illness will tell, and trivializing the suffering of a child or that child’s parents by saying they have an “imitative” form, could do great harm.

So until we have that biological marker, that lab test, that brain imaging study, that gene assay, we are stuck with the simple observation of patterns.  Receiving treatment that works is more important than understanding why it works.

In the case of childhood eating and growth disorders we know what works: food and love.  So for now we’ll have to stick with that.