Beautiful black russian nesting dolls (matryoshka) dolls with white, green and red painting in front of dark background

    Childhood obesity figures vastly exaggerate the scale of the problem

    8 March 2018

    I argued last month that Britain’s childhood obesity statistics are worthless. They are based on an unjustifiable assumption about the scale of child obesity in 1990 and that error has plagued every subsequent measurement. A flawed methodology has led to the number of obese children being greatly exaggerated.

    I suggested that the true number is less than half that suggested by the official statistics. We are told that one in five children are obese by the time they leave primary school. I believe that the true figure is closer to one in twenty and I have the evidence to back me up.

    To recap, the normal body mass index (BMI) cut-off of 30+ does not work for children so we have to improvise. The current British system defines children as obese if they have a BMI that would have put them in the heaviest five per cent of their age group in 1990.

    Were five per cent of children obese in 1990? No. How many children were obese in 1990? We don’t know, but we do know that barely one per cent of 18 year olds were obese and we know that obesity tends to increase with age, so it seems certain that fewer than two per cent of children were obese in 1990.

    There is no reason to think that kids who were in or above the 95th percentile (ie. the heaviest five per cent) in 1990 were obese, so why should we assume that kids who weigh the same amount are obese now?

    Likewise, children at the 85th percentile in 1990 were not overweight (insofar as ‘overweight’ means anything at that age) and yet that is the assumption upon which the endlessly repeated claim that ‘a third of children are obese or overweight by the time they leave primary school’ hangs. It is patently obvious that this definition is wrong, not only because kids at the 85th percentile were not classed as obese in 1990 but also because thousands of bemused parents are sent letters from schools and hospitals informing them that their perfectly healthy children are ‘dangerously overweight’ based on this arbitrary benchmark.

    Why, then, do we base all our child obesity statistics on the assumption that anybody who exceeds the 95th percentile as recorded in 1990 is obese? It seems that the Americans started using this system and we copied them, despite the fact that childhood obesity was much more common in the USA than it is in the UK.

    The USA’s figures are more-or-less plausible. If you look their obesity stats at different ages, there is some sort of gradient. The rates rise with age, albeit with a slightly lower rate at age 12 to 19 than you might expect.

    This is in stark contrast to the graph I showed in my previous post in which British kids get remarkably fat in secondary school before half of them suddenly lose weight for no apparent reason. There is a reason, of course. Once they leave school, we start measuring obesity in a reasonable way.

    If you take Britain’s obesity figures at face value, we now have a higher rate of childhood obesity than the USA. This is almost certainly untrue. The USA has a much higher rate of adult obesity than Britain (38 per cent vs. 26 per cent). It is almost inconceivable that its rate of child obesity would be lower – you only have to visit the place to query that factoid – but this is what happens when you apply the same relative measure to countries which start from very different places.

    A few years ago I gave evidence to an Australian select committee on childhood obesity. I was amazed to hear that their rate of child obesity was just seven per cent. On paper, this is two-thirds lower than Britain’s, despite Australia having a slightly higher adult obesity rate. But once you understand that the method of counting obese children comes from the think-of-a-number school of statistics, it starts to make sense.

    As these examples illustrate, using the 95th percentile to define childhood obesity makes it impossible to make meaningful comparisons between different countries. Campaigners who are unaware of this, or who simply don’t care about the facts, breathlessly claim that London has a higher rate of child obesity than New York. It might do, although probably not, and we don’t have the data either way.

    In order to make valid international comparisons, the World Obesity Federation – previously known as the International Obesity Task Force – uses international cut-offs devised by Tim Cole and colleagues who looked at children’s body weight in six countries in the 1980s. These international cut-offs allow researchers to make comparisons that are broadly like-for-like. Furthermore, the International Obesity Task Force has always based its definition of obesity on the 98th percentile, as Cole et al. recommended in a series of studies and as clinicians use when diagnosing patients.

    The 98th percentile is a more realistic benchmark than the 95th, but it is still likely to exaggerate the scale of childhood obesity in Britain for two reasons. Firstly, the core assumption in all these estimates is that the proportion of children who were obese in 1990 was the same at all ages. If two per cent of 18 year olds were obese, it was assumed that two per cent of five year olds, ten year old and 15 year olds were also obese. This is an assumption of convenience which leads to over-counting. It is very likely that obesity rates are lower among children than among 18 year olds and that obesity rates among children rise with age.

    Secondly, Cole et al. found that obese 18 year olds were in the 98.9th percentile in 1990. This is obviously closer to the 99th percentile than the 98th. Since 18 year olds between the 98th and 98.8th percentile were not obese, there is no reason to assume children at these percentiles were obese either.

    Using the 98th percentile is therefore likely to capture all the obese cases, but it is also likely to create false positives. It is not perfect, but it has the advantage over the 95th percentile in not being utterly ridiculous.

    When this measure is used, a very different picture emerges. The World Obesity Federations’s figures suggest that the proportion of 10 to 13 year old boys in England who are obese is 4.5 per cent. Among girls of the same age, 5.2 per cent are obese. In another estimate, they find that the obesity rate among 10-11 year olds is 2.8 per cent for boys and 2.1 per cent for girls. These are vastly lower figures than the NHS’s estimates for 11 year olds of 20 per cent and 18 per cent respectively.

    In these posts, I have argued that using the 98th percentile is a much more accurate way to define child obesity than the 95th. The 98th percentile is the measure used in medical practice, it is the measure used to estimate child obesity rates internationally and it is the measure favoured by the academics who devised the cut-off system in the first place.

    There was never any justification for the British government dropping the threshold to the 95th percentile. The result of this error is that at least three out of four children who are currently classed as obese in Britain would not be described as such by a doctor. You would not describe them as obese if you saw them. The system used by our statistics agencies is not fit for purpose – unless the purpose is to vastly inflate the scale of the problem.