Is your child’s target weight a gift to the eating disorder?

Last updated on June 19th, 2020

Your child's target weight might be way off – check the method

I'm going to show you two very different approaches used by clinicians to give their patients a weight target. One — used by respected experts I quote here — is individualised to the person. The other is 'middle-size-fits-all'. I'm going to warn you against it as it could keep your child stuck with their eating disorder — whichever type of eating disorder they suffer from.

If you're in a hurry, jump to the infographic below. If you do read this whole blog, you'll be able to tell which method has been used for your son or daughter, and you'll be able to have an informed discussion about your child's needs with clinicians.

If you're a clinician and you're using a calculator or app to work out target weight, I hope that what follows will prompt you to follow the lead of some of our top experts and to change to individualised targets.

I'll explain how the middle-size-fits-all method is likely to fail around half of patients. That's because you can't beat an eating disorder if you're attached to a weight target that has been set too low. A target is one at which the brain and the mind are healthy, not just heart rate or blood pressure.

These messages from parents should ring alarm bells:

  • 'She's at 95 percent, so we're nearly there' (if I ask, 95 percent of what, they don't know, because they thought they'd be judged if they asked.)
  • 'His BMI is 20, so he's being discharged next week' ('the eating disorder is as strong as ever, but the young man has been told that if he doesn't take responsibility for his recovery the service can't help him')
  • 'My 12-year old lost 20 pounds, which according to the doctor means she's now healthy' (the weight loss puts this girl in danger)

But first, why a weight target?

Why indeed? Some experts make the case that a goal weight is useful, and some say not. My other post (here), which tells you what you need to know about weight in recovery from an eating disorder, discusses pros and cons of having a weight target.

Individualised versus One-Size-Fits-All

I'll show you the differences between the two approaches. I produced this infographic (which you're welcome to copy and share, as long as you keep it just as it is) to give you a quick idea of what it's all about. With the example I picked, the person's target weight would differ by 10 kg (22 pounds).

The top experts use an individualised approach

If your clinician is working out a weight target based on your child's personal data, then all is well. Enjoy the expert guidance.

"Target weights should be individualized"

James Lock, co-author of the FBT treatment manual, personal communication

"Yes, it is better when considering weight to take into account their weight history"


Ivan Eisler, Maudsley's Child and Adolescent Eating Disorders Service, personal communication

Cease this madness! BMI is not an accurate measure of a healthy state.


Julie O'Toole,
Chief Medical Officer of Kartini Clinic, in 'The misuse of BMI in diagnosis of pediatric eating disorders'

"Contrary to the promises of diet companies and the ideology of the war on fat, not every individual will place neatly into the 'average' weight category based on age and height."

Lauren Muhlheim in 'When your teen has an eating disorder'

"There is no one right size that fits all when it comes to healthy weight after anorexia nervosa. Thus anyone who uses an equation (such as a BMI or ideal body weight calculator) or simplifies this complicated situation is relying on tools that are inadequate."

Dr Jennifer Gaudiani, internist and expert on the medical complications of eating disorders, in 'Weight goals in anorexia nervosa treatment'

A one-size-fits-all model is still tragically common

If your child's weight target is a particular BMI or 'weight-for-height', then the clinicians are probably using a method sometimes called the 'BMI method'. As I'll show, it's a one-size-fits-all model. Or more precisely, a middle-size-fits-all.

The BMI method uses big BMI surveys from the population. The assumption is that your child will be fine in the middle: the median.

Imagine I'm about to open a shoe shop and I'm busy ordering stock. I'm looking at statistical charts and I'm thinking, frankly, why so many sizes? Too much uncertainty! I'll give all my customers the middle size.

Some clinicians — usually not specialised in eating disorders — believe that weight doesn't need to be individualised, and they would dismiss the shoe-size / weight analogy. For them, mid-range is usually all anyone needs. Some even set the goal weight lower for fear of scaring the patient — a 'small-size-fits-all' model which can lock even more patients into their eating disorder, as I discuss on this page.

If I'd been selling shoes to the unfortunate elite women in ancient China, I'd only have stocked tiny, bone-crunching shoes.

The calculator is way off

I used to think that 'weight-for-height' apps, calculators and Excel spreadsheets — all using a middle-size-fits-all (median BMI) ideal weight — would at least give a rough estimate. But you need only take a peak at a BMI chart (girls, boys) to appreciate the wide range of BMIs in a normal population. A median BMI will only fit the most 'average' patients. For everyone else the calculator is so very wrong that it doesn't even cough up a decent ballpark figure.

Beware of 'One-size-fits-all'

I'll take you through an example using both the individualised and the middle-size-fits-all approaches, so you can see how very different the results can be. I'll start with the individualised approach.

How to use a growth chart to predict an individualised weight target

Growth charts are the main tool in predicting the weight at which your child will start to be healthy. (Even then, you'll only get an approximation, so you will have to observe and adjust before declaring your kid 'weight restored'.)

Gather any growth data you have, to help estimate an individualized weight target

Below is an example of a growth chart for girls, with curves indicating percentiles.

* I explain percentiles here *

* Download CDC (american) growth charts for boys and for girls. And here are UK ones. There's also access to WHO (World Health Organization) and CDC charts on mygrowthcharts.com *

If a girl was on the 75th percentile curve for weight when young and healthy, that's roughly where she'll need to be throughout life to continue being healthy.

Example of 75th percentile girl - growth chart used towards goal weight

Dig out all the weight data you have from your child’s early years. Stature (height) data are useful too as growth may be stunted by malnourishment. Your child could have started restricting quite a bit earlier than you think, or they may have lost weight during various illnesses or just because of metabolic factors that form part of the eating disorder risk. So go back as many years as you can. Indeed, people who have anorexia in their teens have often dropped off their growth curve as toddlers.

On the example below I’ve plotted the weight of Jo, a girl who tracked on the 75th percentile for weight until age 12. The next time she was weighed, on her 14th birthday, she weighed 50kg (110lbs). Can you see how she dropped from the 75th to the 50th percentile curve? 

Target weight girl-anorexia-chart
Because this child's weight hardly increased for 2 years, she dropped from 75th to 50th percentile

I'm not showing her height data, but perhaps her height has also dropped off its percentile curve, indicating some growth stunting. Her body will need fuel to catch up.

(Note: We could also plot her growth on a BMI chart, by the way. I mention it because some therapists do that. Again, we'd expect her to track on a particular percentile curve — the 60th percentile for example. So a therapist might say the target BMI is 60th percentile, and work out the target weight from that. But BMI is a bit volatile (rather wiggly on a chart), so you'll get a clearer picture sticking to a height chart and a weight chart.)

If she has anorexia or similar, this girl might be horrified that she's gained a couple of kilos since her twelfth birthday. She'll not understand the logic of needing to stay on her weight curve. She will argue that she's fine, but her body is reacting to a state of famine, because she's lost a lot of weight and is hardly eating. Biological emergency measures are kicking in, affecting both her physical and mental state.

So we get to work on refeeding, which takes time. Let's look at where she'll need to be in, say, six months' time. No point in giving her a target that will be obsolete by the time she gets there.

From the chart, we look at the 75th percentile curve and find a target weight of 58 kg (128lbs).

individualised personalised weight target eating disorder

Health is not a number

In practice, when we plot growth charts we discover that over the years, our children have oscillated around a range of percentiles. So the target weight is, more sensibly, a target range. Also, the 'rule' that a child's healthy height or weight stays on one curve is pretty good, but there's still human variability. We've all seen kids who used to be average height and suddenly became giants. And then the genetic heritage is considered, based on the physique of family members. In short, take all predictions with a pinch of uncertainty.

Adjusting for physical and mental state

A growth chart only gives part of the story. The weight at which your child is healthy may turn out to be a few kilos above or below what the growth chart indicated. In THIS post I say more about that. Remember, also, that the chart is only showing growth patterns for a particular population — probably mostly white-Caucasian.

Sate not weight - a healthy weight isn't just a weight off a chart

How is a 'middle-size-fits-all' weight target calculated?

BMI calculator for eating disorder weight target

Let's move on to the approach which I am arguing is doing more harm than good: the middle-size-fits-all. In scientific papers it may be called the BMI (or median BMI) method, or it may refer to the use of 'normative population data'.

This approach doesn't look at your child's personal situation but at population surveys. Clinicians share with each other apps or Excel spreadsheets to make it simple.

Let's return to our 14-year old girl. We've just seen how according to her growth chart she urgently needs to increase her weight from 50 to 58 kg. But what weight target could a calculator would come up with?

You type in her sex, age and height. The answer pops up: her 'ideal' or 'target' weight is 51 kg.

100% weight for height or median BMI

More terminology for you: that's the weight that corresponds to the median BMI, or the 50th percentile BMI, or to 100 percent weight-for-height.

For this girl to get a 51 kg target is like her to walking into a shoe shop and being given a middle-sized shoe. (I explain BMI and weight-for-height (WFH) here. )

Let's be clear: any girl of that age and height would get that exact same target weight. Yet out there in the real world, girls of that age and height come in a wide range of weights:

BMI target weight method yield low weight for half the population

The only bit of personalisation that goes into a weight-for-height figure is sex, age and height.

For a clinician relying on a calculator, this girl, who weighs 50kg, is so close to the app's target weight of 51kg that she really doesn't need to gain weight. It's common for therapists to say that 95 percent weight-for-height is plenty — why stress out a patient by aiming for more? This girl is already at a 'generous' 98 percent (in case you're wondering, the app coughs up the 98, but you can also work it out by dividing 50 by 51).

The clinician hasn't asked for the girl's height or weight history, so all they see is a girl with an eating disorder who doesn't need weight gain 'because' her BMI is perfectly average. They don't see any need for weight gain, and most worryingly, they might not see any need to check her health. Yet this girl's health is very much at risk right now, given how much she's dropped off her weight curve. (To read more on this from an expert paediatrician, see The misuse of BMI in diagnosis of pediatric eating disorders by Julie O'Toole)

Truly, this happens a lot, especially with GPs. As I was writing this piece I got this naive email from a parent: 'My daughter (13 years) has lost at least 10 kilos in the last 3 months [that's 22 pounds], but she is still at the 25th centile, so no need to admit her or do bloods.' Luckily, the GP who gave this flawed advice did refer her to the eating disorder service, and from there the girl got energetic and competent care.

In summary: According to her growth chart the girl in our example needs to gain 8 kg (20 pounds, or more than 1 stone) to reach 58 kg in six month's time. She is currently very underweight and needs to be carefully monitored. A clinician relying on a calculator would say she is fine and doesn't need any weight gain as she is already at 98 percent weight-for-height.

It's even more wrong with height-stunting

With a restricting eating disorder like anorexia, it's common for our children's growth to be halted or slowed down. So, what if in the past year, this girl hadn't grown? Let's plug into the app her current 50 kg, but this time let's say her height is only 158cm.

Now her ‘ideal’ weight comes up not at 51 kg but 48kg. Remember that right now she weighs 50kg. Her eating disorder now has confirmation that she's fat. It tells her she must lose more weight.

With height-stunting, the meidan BMI weight target is even more off
The one-size-fits-all target could be 10kg lower than the individualized target

The app doesn't know that her growth was stunted. It doesn't know that she's been restricting, that she's lost out on weight, and her body is in 'famine' mode. It's just looking up the middle BMI for girls that height.

This girl's body is crying out for nutrition and energy for growth, yet the 'middle-size-fits-all approach would set her healthy weight 10 kg (22 pounds, or 1.5 stone) lower than indicated by her personal growth chart.

Can you see how a calculator doesn't even give a useful ballpark figure? And I didn't even choose an extreme example. With another child, with different height and weight parameters and a different history, the difference could be even bigger.

Half of our children are kept underweight by the 'middle-size' approach

The median BMI is in the middle
The 50th percentile BMI, or 100 percent weight for height

Because the median BMI, or 100 percent weight-for-height, is in the middle, it's going to be wrong for almost all our kids. That's because by definition, half the kids that sex and age have a lower BMI, and half have a higher BMI. You shouldn't need scientific studies to convince you of this — it lies in the definition of the 'median'.

Weight goal based on 100 percent weight for height or median BMI - anorexia recovery

I'm not worried about the kids whose true healthy weight is below the calculator's target. The 'extra' weight will be used for growth. It won't do any harm. As I explain in THIS post there are some good arguments for aiming high.

I'm not worried about the kids whose historical BMI has tracked close to the 50th percentile. Both methods for estimating target weight will give roughly the same result.

The kids who are in danger are those who need to track at a higher BMI. With the middle-size-fits-all method, the weight target is a gift to their eating disorder.

When a therapist puts their faith in a calculator, the tail is wagging the dog. This teen doesn't yet have her periods at 100 percent weight-for-height? They'll return later. This girl is still skipping school lunch? It's time she took responsibility. This boy is still compulsively exercising? Well, you know, some people never recover.

The question for those who use a calculator is, how can a middle number suit the wide range of human beings that come into their office? It is an extraordinary claim. It needs evidence, not opinion. In science, a theory is invalidated when there is evidence to the contrary. And we have plenty of that. I quote studies further down, and additionally we have numerous first-hand accounts of children who remained very ill at median BMI (and got well when they were eventually supported to gain more weight).

Let's be clear. For roughly half of patients, the calculator will come up with an adequate or generous weight target. That's fifty percent of patients with a chance of a successful treatment outcome. That's statistics, and it's a gamble when it comes to an individual patient. Without considering their growth history, you can't know if they're in the half who will be well served, or failed, by the calculator.

'My kid is back'

Here's one story from a huge pool of similar stories that are everywhere on parent forums. These stories start with a child who stayed ill for months or years while supposedly 'weight-recovered' using a 'middle-size-fits-all' figure. At long last the child recovers when a new therapist, or the parents insist on a higher weight. There are plenty of similar stories with recovered adults too.

This is the story of a 14-year old boy who was 'stuck' with anorexia at an unsuitable weight:

My son was told that he was fully grown, and that as his BMI was 19, he was now weight-recovered and could maintain his current weight. But he was still terribly ill. He couldn't even drink a glass of orange without major anxiety. Eventually we sacked the therapist and resumed high calorie feeding. It was twice as hard as he had been told a BMI of 19 was PERFECT for him.

With weight gain, he grew another 5 inches. We got him to a BMI of 24 and kept him there. It's made all the difference.  My kid is back. For the last two years he's been doing just great, fully functional, and in good recovery.

Where did this BMI of 19 come from? Looking at boys' BMI charts, I see that for a 14-year old it's the median percentile. This boy went from a disastrous 'middle-size-fits-all' to a weight that truly suited him.

Target weight
An irrelevant image to relax you after all these numbers

What do the top experts say?

As you'll have gathered from the experts I have already quoted on this page, top professionals assess each individual for their needs. If they set a target weight, it's individualised and it's open to revision.

Here are a few papers that explain the problem with a one-size-fits-all approach (it might be referred to as 'the BMI approach') and that strongly recommend that targets are individualised:

"Such methods [using median BMI] are unlikely to gauge the extent of weight recovery needed for individuals whose baseline weight was higher or lower than average, which may not only lead to partial recovery for these individuals."

Lebow et al (2017) 'Is there clinical consensus in defining weight restoration for adolescents with anorexia nervosa?'

"Treatment goal weight for a particular patient cannot simply be read off the charts on the basis of normative population data. Put another way, treatment goal weight is not necessarily the same as the weight associated with median BMI."

Golden et al (2015) "Update on the Medical Management of Eating Disorders in Adolescents"

"Professionals also use growth curves to establish target weights for recovery. Research shows that these curves, when available, are more accurate than using population averages to determine an individual's treatment goal weight."

Lauren Muhlheim in 'When your teen has an eating disorder'

One study Lauren Muhlheim cites is this:

"In most cases, [a growth curve] allows a clinician to easily estimate a patient’s healthy body weight and provides a number that is specific to the individual patient based upon their previous growth parameters, rather than on the population average for age (i.e. the 50th percentile).

Harrison et al 2013 on the utility of growth curve data in adolescents with eating disorders

Why does anyone use a middle-size-fits-all target weight?


There are plenty of lovely, caring therapists who use calculators to work out your child's target weight. And BMI charts are everywhere in the health service — I'm guessing they are so ordinary that people might not pause to wonder quite how inaccurate they are for individual use.

Clinicians might be hoping to get a ballpark figure because your child doesn't have any historical data to plot onto a growth chart. Hopefully I've done a good job of highlighting how this figure could be wildly wrong.

On the other hand, I talk to parents who were never asked for their child's growth history. I hope this post contributes to change.

There is also a tricky shift from research to clinical practice. Researchers need to define 'recovery' as systematically as possible in order to assess treatments. Most have used median BMI as their benchmark. For instance in Lock and Le Grange's landmark study of FBT, for treatment to count as successful, teens had to do well on a behaviours and beliefs questionnaire and they had to reach 95 percent weight-for-height. Let's be clear: that 95 percent is a choice made for the sake of consistency in a moderately big trial. The authors are the first to say that in clinical work, healthy weight is individual.

"Clinically we use individualized weight calculations; this is not possible in research studies as the same outcome needs to be assessed in all participants."

James Lock, co-author of the FBT treatment manual, personal communication, March 2020

Is there any reputable therapist who endorses the 'middle-size-fits-all' approach?

In the world of children and adolescents, I have come across only one reputable researcher and clinician who has put pen to paper to advocate a 'one-size-fits-all' approach. This is Christopher Fairburn, who has researched and written on CBT for eating disorders. His manual advocates a fixed low BMI figure for all adults, a fixed low BMI for older adolescents. Only for younger adolescents does he take age and sex into account by using a 100 percent weight-for-height target. He writes that in his experience, these targets work for most patients. However the success of CBT is not very high (though it is one of the better approaches available to adults — and adults have traditionally been given low weight targets). How much higher would recovery rates be if targets were not only higher, but individualised?

In summary: 95% of what?

I hope this article will help you enquire what a therapist means when they say your child is x-percent weight-recovered. Is that percentage relative to a carefully assessed personal target, or relative to a middle-size-fits-all number from a calculator?

Final thoughts

There is no number on the scales that can say a child is healthy. Let's keep up the teamwork between parents and clinicians to address our children's needs. And where there is uncertainty, let's remember: 'First do no harm'.

Where to next?

* Weight-restoration: why and how much weight gain? *

* What do BMI and Weight-For-Height mean? *

A short fun video that makes the point that one size does not fit all

* More on weight and feeding in Chapter 6 of my book *

* 'How child's growth charts could help prevent an eating disorder: identify a problem early': a short piece by Lauren Muhlheim *

* More on target weight on the excellent, parent-led FEAST site: 'Setting target weights in eating disorder treatment'

Thank you to James Lock, Daniel Le Grange, Ivan Eisler, Esther Blessitt, Rebecka Peebles, and a whole lot of well-informed parents, for helping me write this post.

6 Replies to “Is your child’s target weight a gift to the eating disorder?”

  1. Eva, excellent article and you're spot on. I found this article originally via Ref 100 on page 64 of your book (2020 edition).

    While I have zero argument with your position on the median being an inappropriate target, the rest of your article seems to focus only on one side of the error – children who should be ABOVE the median.

    You hardly discuss the other half of the problem i.e. the children who have been at, perhaps, the 10th (BMI) percentile all their lives but who are now being forced up to the 50th percentile.

    This is particularly a problem for children whose families are naturally small – for example, children of Indian or Singaporean or Japanese origin.

    Yes, the median is the halfway point. But if it's wrong for (approx) 50% of children, then it's wrong for the other 50% as well 😉

    It would be great if you added some content on that. Many thanks.

    1. Mark, you are spot on. Thank you for raising this. I'll see where I can tweak the article (I am cautious, as some of my visitors suffer from an ED and they may be quick to jump on any reason to avoid gaining weight).
      For visitors reading this I'll explain some more:
      I've not made a big deal about the 50% of kids for whom the median BMI might be an over-estimate. This is because:

      a) it won't do them any harm to have a buffer or to overshoot. It may indeed be helpful. (I write about this in https://anorexiafamily.com/weight-restoration-eating-disorder ).

      b) as there is a tendency towards 'fat-bias' or 'fat-phobia', my guess is that many of these children will not be brought to a median BMI anyway — everyone may be quite quick to point out that the family norm is for a slight build.

      Yes the sampling behind these BMI charts may well have been a mostly white-Caucasian population. I have only glanced at some of the papers explaining how some of these charts (in the UK, or CDC, or World Health Organisation) were compiled. I think the information is out there, for someone who has the patience.

      What you've highlighted is one more reason to be very cautious with predicting recovery weight, and even though it might feel 'scientific' to plot an individual chart, it's never the only criterion for weight recovery.

      Thanks for writing!
      Eva

  2. Thank you for this really important article Eva. As a clinician, I strongly believe that target weights for the purposes of eating disorder care and as a marker of “recovery” is a very dangerous tool.

    I love your explanation of this for our care of children and adolescents with eating disorders.

    Too often, the 50th percentile is just not the best benchmark for good physical or psychological recovery; and for a young person, and their family, the destructive nature of achieving this and then having the goalpost change can have a terrible impact on engagement in treatment, trust and recovery.

    A few kg/pounds sometimes makes a huge difference in overcoming the eating disorder; and we don’t know why! Being at a higher percentile is for many, physically optimal for the body to return to functioning well.

    We aren’t all built the same! Achieving the 50th percentile for weight is as absurd as insisting that the 50th percentile for height will be good enough and a measure of recovery!

    Thank you for your work,

    Dr Sarah Wells, Clinical Psychologist

    1. Dear Dr Sarah Wells, thank you so much for taking the time to write this. The more clinicians speak up, the more it will become unacceptable to use arbitrary benchmarks.
      Eva

  3. Thank you, Eva, for writing this blog. My 11 year-old daughter’s (non-ED trained) physician told me when my daughter reached the 50th percentile that she was weight recovered and any existing angst/anxiety was who she was and that we were basically done. I had read otherwise and kept pushing as I knew she was still struggling immensely in a way she had not pre-ED. My daughter did not reach a good state / weight until a full 25 lbs later and a return to her historic growth curve at around the 85th percentile for weight for her age. That was just under a year ago and my daughter is now a healthy and happy 12 year-old rocking her curves and not looking back.

    1. Thanks so much for taking the time to comment, Anna, it's been most useful. Parents and therapist need to hear about something like 85th percentile being right for some of our kids, and the arbitrariness of that 50th percentile mark.

Leave a Reply