Can urban designers limit diabetes?
- Share
Level of physical activity of 275 UK school children, comparing those who walked to school against those who were driven (see Metcalf et al for explanation of units)
I’ve argued before (e.g. here and here) that designing physical environments to encourage or force higher activity levels is neither an effective nor an efficient way to tackle serious health issues like obesity and diabetes.
Here’s a surprising ripple in this debate. The EarlyBirds project in the UK has been monitoring insulin resistance in a randomly selected sample of children for twelve years, from the time they entered school at age five to age sixteen. So far the project has generated more than 60 published papers.
The surprise is the researchers find that children’s level of physical activity doesn’t have a key role in determining if they suffer from obesity or diabetes. That suggests urban design and transport policies aimed at greater physical activity aren’t as important as is commonly assumed.
What matters most they say are genetic factors and food intake. It’s crucially important to understand that “calorie reduction, rather than physical activity, appears to be the key to weight reduction.”
The project effectively involves lining up “300 children at the starting line, firing the gun, and taking snapshots every year as they move down the track to the finishing line at age 16.” Great care has been taken in the design of the study:
The snapshots are sophisticated measures that include BMI, body composition, energy expenditure, physical activity (electronic activity monitors) and metabolic health (blood sugar, cholesterol, insulin sensitivity, blood pressure, arterial stiffness etc). Uniquely, EarlyBird has been taking annual fasting blood samples from these children since they were five years old.
The researchers are forthright about the implications of the project for prevention of diabetes. There are many findings, but I’m primarily concerned with those that relate to physical activity because of the widely assumed connection to potential urban design and transport solutions.
Three findings in particular seem to go against the received wisdom. According to the researchers:
- Inactivity does not lead to obesity – it’s the other way around!
- Children who’re driven to school are just as active over the course of the day as those who aren’t (see exhibit)
- Environmental opportunities, like the provision of open space and sports centres, have no influence on children’s activity levels.
In the researchers own words:
Using time-lagged correlation to imply direction of causality, weight gain appears to lead to (precede) inactivity, rather than inactivity to weight gain. (This is) crucially important because it suggests that calorie reduction, rather than increased physical activity, may be the key to weight reduction …..
The activity cost at the age of 7y of being driven to and from school during the hours 8am-4pm is 16%, but is nil (<0.1%) over an entire 24h. As in the schools study, those who lack the opportunity for physical activity at one period of the day appear to compensate for it at another …..
The evidence we are accumulating suggests the activity of children is “programmed”, either genetically or as a result of very early experience. There is little evidence from EarlyBird studies that the physical activity of free-living children is linked to recreation or environmental opportunity.
The study suggests children’s level of activity is “programmed” prior to them starting school. Most of the excess weight carried by young boys and girls was gained prior to age five.
Rather than being concerned about issues like how much TV school-age children watch, or how much they use open space, the researchers say public health initiatives should be directed to an earlier period of childhood. By the time children get to school it’s probably too late for initiatives to be effective.
Here’s one of the papers written by the researchers and published in the British Medical Journal. It looks at the physical activity cost of driving children to school and contains the data used in the exhibit (including an explanation of the units). It’s the only directly pertinent paper I could find that’s ungated.
The project has some other interesting findings. It indicates social inequality is not associated with physical inactivity and is not a major factor in obesity; the average pre-pubescent child is no heavier than 25 years ago; children who gain excess weight are taller and more insulin resistant than their shorter peers; dietary habits are established early in life and retained throughout childhood; and parents of obese children tend to be unaware and unconcerned about their child’s condition.
The project also indicates the so-called child obesity epidemic is a teenage (i.e. post puberty) problem. The widely-held belief that the entire childhood population is at risk of obesity appears to be wrong:
The trends imply that, before puberty, only a subgroup of the population is at risk. With puberty, the skew in fatness that characterises early childhood gives way to widening of its variance, suggesting a new wave of obesity which this time involves most of the population. The changing pattern suggests that different factors at different ages are responsible for today’s childhood obesity – some operating early in life, and others much later. More than one solution may be needed to prevent obesity in childhood.
Perhaps the level of physical activity has some causal role post-puberty, but at least for younger children, issues like whether they walk to school, use public transport or live at density seem to have little to do with obesity and diabetes.
The only half sensible argument I’ve heard to support conciously designing cities to promote greater physical activity is that measures like higher density are good for other reasons e.g. sustainability. That’s true, but I think claims that aren’t supported, or are contradicted, by evidence should always be avoided on principle. More pragmatically, they can undermine support for the wider objective when they’re eventually debunked.
Please login below to comment, OR simply register here :
Thank you for registering, we have just sent you a confirmation email, which includes your new password to be entered below.
“Environmental opportunities, like the provision of open space and sports centres, have no influence on children’s activity levels.”
That completely defies belief to me, unless they making comparisons being areas where you might have to drive 10-15 minutes to find a large open space/sport centre to areas where you might be able to walk 10-15 minutes to find such space. But I just can’t accept a children that live in a houses with immediate easy access to open space (including back yards) aren’t going to get more exercise on average than children that live in small apartments where the nearest decent amount of space is 15+ minutes’ drive away.
I’d also say, diabetes and obesity are only one part of the picture. Ultimately it’s overall health/fitness that matters, and I’d still suggest it’s safe to stick with our assumption that physical exercise is important, and hence ensure that kids grow up in environments where such exercise is both encouraged and easily available.
The data from a sample of 275 UK school child is interesting, but not on its own terribly useful in extrapolating to the rest of the modern world.
Wiz Aus: Here’s a quote from the EarlyBirds site -
The researchers agree with your view that exercise still matters – elsewhere they say “Children who keep active gain no less weight, but they do become metabolically healthier”.
The extrapolation issue you raise applies to all studies. What’s impressive about this one is first, it’s longitudinal, second it’s got data for 12 years, and third, the data collected is detailed e.g. it doesn’t rely on self-reporting of exercise or just on BMI. AD
Parker Alan • OAM
This is tricky on and note that the Canadians and Americans have similar similar problems. Part of the problem is the scarcity of the data not focussing both the physical on safety due to accidents and specific deseases .Instead of focusing on both and using the measure of using childrens deaths and serious per 100,000 population.
Their is some data from Canadas attempt to deal with this problem
VANCOUVER, Canada – The start of Velo-city Global at the end of this month will also the kick off the campaign to link cycling to the universal rights of a child. The United Nation’s “Convention on the Rights of a Child” has stated that children should develop their physical abilities to their fullest potential, respect the natural environment and have opportunities for recreational and leisure activity.
Cycling provides all these.
Velo-city Global is the world’s largest cycling policy conference attended by more than 1,000 global cycling experts will take place from June 26-29 in Vancouver, Canada. Velo-city Global is organized by the European Cyclists’ Federation (ECF).
Decreasing numbers Across the globe, the number of kids that walk or cycle to school has decreased from 82% to 14% within the last 30 years. In the US, only 1% (2009) of children ride their bicycle to school. Compare this with the Dutch, who see nearly 49% (2009) of children cycle to school.
“By making healthier mobility a universal right, we can boost the case for governments to invest in cycling for children. The more children cycle, the more parents buy bicycles too. Velo-city is the place where we will provide advocates, campaigners, politicians and planners with the tools and expertise to get more children cycling,” says Kevin Mayne, Director of Development at ECF.
Cycling renaissance The global conference comes at a time when North America is undergoing a huge cycling renaissance. Canada has seen a 42% increase in the number of daily bike commuters between 1996 and 2006. In the US, the share of bike trips made for the journey has risen by 50% between 2001 and 2009.
Bike sharing has also taken the North American continent by storm, with schemes popping up in Washington, Montreal, Toronto and Minneapolis. New York is set to have a 10,000 strong fleet of public bicycles in place this summer and Vancouver is also looking to do the same.
Cycle Industry Club “The growth in cycling in Canada and the US is in large part due to the hard work of bicycle advocacy groups supported by the bicycle industry,” explains Mayne. Mayne is encouraging people in the bicycle industry to join the Cycle Industry Club, which seeks toin Europe by 2020. Big names like SRAM and Trek have already pledged funds, and are seeking €1 million for bicycle advocacy by the end of the year.
As a bicycle advocate I believe that states need triple the number of cyclists.
Alan, there is simply not enough here to come to any conclusion. The statements you cite and the graph you show just seem counter to experience and common sense. Of course as a scientist I am prepared to be convinced that intuition or longheld perceptions can be debunked, but I didn’t find it here.
For example: “The activity cost at the age of 7y of being driven to and from school during the hours 8am-4pm is 16%, but is nil (<0.1%) over an entire 24h. As in the schools study, those who lack the opportunity for physical activity at one period of the day appear to compensate for it at another …..”
I get the whole genetic programming thing (I’m a geneticist) but nothing in my experience provides an explanation for that innocuous appear . I don’t find personally the compensation theory to be compelling–probably the opposite. As adults drivers always want to drive anywhere, even relatively short distances. And to be frank I would say it is completely unacceptable in a study to uncover this kind of result then not examine the cause. So on the face of it I will remain in disbelief.
On the other hand, the “explanation” may lie in what I found in the BMJ paper (actually a 2 page letter or Brief Communication). I seriously question whether that “paper” is worth anything, other than as a stimulus to do a proper study or something…
Here are a few of the critical sentences to get to the core:
So, the thing is that the kids are only walking 6 minutes and this represents 2% of total weekly activity, and clearly this is not only just trivial but does not fall within statistical detection limits as defined in their own methods! On top of that the kids are just 5 years old. I would probably need to spend hours studying all papers from this Early Birds Project, and unfortunately the only accessible one from this BMJ tidbit doesn't bode well. I'd like to see the prospective study of kids who have always walked over the whole 12 years versus the non-walkers; but I would also want/need to know why/how total activities were not significantly different.
The short distance of the walk to school may be partly explained by the different arrangement of British towns (I don't know about you but I reckon my walk took at least 4x to 5x their 6 min.) and a study of 5 year olds may have other biases; eg. the short distance may be self-selection by parents for such very young kids–maybe the distance increases as the kids get older? (Of course it is also true that how you begin often determines how you continue; ie. if young kids get driven they will probably always want to be driven.)
Of course I do believe genetics is really important, but having said that our population genetics hasn't changed in the last 30-40 years but everyone has got much fatter. Diet is clearly extremely important, and on this point the Brits are perhaps a terrible "background" to do such studies in which the activity is so minimal and the food particularly awful.
So, really honestly, I just cannot be persuaded of anything from this. Certainly your closing paragraph is not supported. And, AD a personal request, please don't file this article so you can reference it in future pieces as if it has proven something (sorry, but it is a habit of yours).
The quote you gave from the Early Birds study:
Again, this just seems nonsensical. First thing is I would want a detailed description of the available space at the schools and suburb. Since Maggie Thatcher allowed state schools to sell off “spare” land, lots went and sold off sporting fields in one of the most short-sighted ridiculous, almost criminal, actions you could imagine. (And come to think of it, the same neo-con mindset would be happy to agree with this study that says it makes no difference.) Ask any Brit immigrant, especially those from lower socio-economic levels (of course Maggie’s policy was selectively destructive of state schools in poorer areas who were most under financial stress; your public–ie. private–schools or state grammar schools in better areas did not suffer), what they think of the difference between Australian schools and their Brit equivalents. It is night and day–almost all our schools have pools, fields, basketball & tennis courts etc. And they can be used all year round.
This may not be enough to over-ride terrible eating habits but it does have an effect.
michael r james @ #3 & #4:
All fair points for debate michael, but “nonsensical” seems a bit strong given the team has published 60 papers and been going for 12 years. I suspect they might know a little more about the issue than you or I.
BTW you unintentionally put me on to EarlyBirds when you cited this article, Why our food is making us fat, from The Guardian by Jacque Peretti (great article). Peretti in turn cities the EarlyBirds project in support of his argument.
My intention in referring to earlier articles I’ve written is to show I’ve addressed the issue before, not to claim that my position is proven.
“I suspect they might know a little more about the issue than you or I.”
Well, not really sure of that. No one knows much about this stuff. I can only judge the BMJ letter and it was very unimpressive. I wouldn’t put my name to such an incomplete study, and I kind of suspect it was one of those jobs done to justify putting in a grant to get some serious money for a bigger study. That does not excuse it.
The thing is this kind of study is driven by epidemiologists and (gross generalization ahead…) I often have trouble with them, and I have worked with quite a few. Too many of them are obsessed with the numbers, and trying hard to extract meaning out of minimal or dodgy data. And often impatient with someone like me–for example my demand to either get more data, more answers to specific questions (that you cannot just conjure up with a quick regression analysis) or in fact conclude that the data are just not good enough (which was my conclusion in this case).
I’ll take another look at the Peretti piece. It is good that the EarlyBirds project is at least trying to address the issue, but a real bore if it is all mostly inaccessible. And alas, good intentions does not by itself make for good science. I wonder if more definitive analyses may come out of the giant longitudinal study by Institute of Child Health Devon on every birth in the county, funded by the Wellcome Trust. Huge numbers involved (I forget but like 30,000 maybe more), I think it is about 20 years old now but I am out of touch with it since I left Oxford (Wellcome Trust Ctr Human Genetics where some of the people were associated with the study).
Incidentally, one can invert these results and say that if true, then it tells urbanists that suburbs should be planned so that the walk to school is longer than 6 minutes!
And of course, even if 6 minutes of walking made no measurable impact on health or total exercise, it is till worthwhile if only so that young people realize they can actually get themselves from A to B using their own legs and do not always need a SUV or 4 wheels.
Note that I said measurable , because that BMJ was not capable of measuring the impact of the school walk, which they claimed (and I remain dubious but ….) was 2% of total weekly walking and the study only had 8 to 12% detection limits.
michael r james @ # 6 & #7:
I think you’ve got the wrong end of the stick there on the Metcalf paper, Michael. It’s a median walking time of 6 minutes (not a max) and a median distance of 0.7 Km (not a max). The average walking distance for that 25% of kids who live farthest from school is 1.2 km. Those distances are consistent with the four primary schools in my neck of the woods (I’ve simply measured how far apart they are).
Also, I think that 12% relates to the confidence you can have in the measures of physical activity recorded by the accelerometer – it’s accurate to within 12% (presumably + or – 6%). So the 2% might be as much as 2.25% or as little as 1.8%.
Where do you get the idea it’s not a paper? My take is the BMJ does some short papers. I agree though that it’s a bastard not being able to read the other 59 papers because they’re gated (Australian academics are often bad at this too – it seems they’re unaware there’s a major international move on, pushed by the likes of Harvard, to bring academic articles out from behind the gates).
The other thing to bear in mind is there’s much more in the EarlyBird project that’s relevant to the physical activity and urban design question than just the Metcalf paper.
Alan.
I can’t see that I referred to the 6 mins walking time as a max. The median is indeed the most significant measure to use, and I used it.
About the publication. It is what I would call a MPU: Minimum Publishable Unit. One small set of data (one age group, observed for short period) that is not only seriously incomplete but raises more questions than provides even one answer to one question (at least that I would accept).
As I say, there is immense pressure on researchers to publish–and frequently, what with grant bodies annual reports, institutional reports and endless official reviews by innumerable bodies with good intentions and charged with proving no one is wasting tax-payers money (an American disease). This is especially awful in Australia and leads to loads of MPU papers in unread and uncited zero-impact local journals.
Another new trend in journal articles is the way in which they allow you to download. Even though this article was in principle available as PDF (and it loaded as a PDF in the web window) it only downloaded as rubbish. It would only save directly from the good displayed copy as a .html (now I hate this format because as far as I can tell it is not a permanent archive copy, instead everytime you click the file on your hard drive, it actually goes and (tries) recovers the dozen little files comprising a html! Super-dumb and hyper-irritating. So I haven’t actually got a copy of that paper–I gave up after wasting my time and have no intention of accumulating/filing dead .html files!
/rant
As I said earlier I agree the EarlyBird project does sound interesting. But you know, if the body responsible does not want to make even a summary of its work available free and easy, then it starts to lose my interest. Certainly I cannot and will not just believe broad statements as gospel. Because none of those ostensibly good intentions guarantees anything as to output or quality (oops, guess my rant wasn’t quite over).
michael r james:
“I don’t know about you but I reckon my walk took at least 4x to 5x their 6 min” AD
I have written on the scientific publishing thing before and will probably write another more detailed piece for Crikey one day. It is a very tricky problem and I wonder about the Wellcome Trust’s attempt to “encourage” its researchers to publish in open-access journals. I use the quotes because what the WT request politely is more or less law. It is also not a bit hypocritical because they want to support the creme-de-la-creme and the major definition of that is publishing in the very top journals (and people can hope but PLOS is not there yet; in fact it is too broad and wide to probably ever get there).
Any scientist still working hard on grants and career etc is still going to want to publish in Nature, Science etc. And get this, you even often pay for the privilege! It is not so different to the Gina Rinehart issue: most of the top journals are for-profit and for anyone to propose public-minded alternatives will need to cough up the dosh, which is perhaps where WT comes in (as the world’s largest private biomedical research funder after Bill Gates if his foundation is counted in the same category).
Will WT understand when you publish in PLOS-Genetics (IF 10) instead of Nature-Genetics (IF 36) not to mention citation counts and h-index score? As I am sure you understand a recent single Nature-stable article is like a dose of steroids or viagra in your cv.
I have often thought that the only solution is for WT to actually buy MacMillan but even if they could afford it, it would be considered a misuse of research monies. Perhaps Bill Gates should buy it because it is after all a very profitable company; part of the griping is that its profit margin of >45% is way above almost any other publishing company in the world, and it is getting this “super-profit” from publicly funded research!
Well, yes, it suggests that my school walk was way out on the shoulder of the normal curve?
BTW, I even wonder about that statistic: 600m in 6 min is 6 kmph which is pretty damned fast for a 5 year old, don’t you think? My walk to work was just short of 3km and it took me about 30 min (best time ever 26 min) which is ≈6 kmph. I am not a slow walker–and whenever with other people I always have to slow down (equally I can walk a lot faster than 6 kmph but not all day; I can pretty much keep up the 6 kmph all day).
Frankly there are several things I can identify and others I suspect, in that “paper” that don’t smell right. Yep, very unscientific (except for my earlier explicit comments).
Chiming in to this discussion rather late, but it’s in response to Alan citing this in another blog post.
Like Michael James above, I’m sceptical about drawing such sweeping conclusions from one study of five-year-olds. FWIW I dug up the original paper and some of the related sources to find out what was actually done: it was based on attaching accelerometers to 275 five-year-olds for a week and recording the counts.
The ‘school journey’ refers to a two-hour period during which the journey took place. But yes, for those who walked to school the interquartile range of journey time was 5 to 10 minutes and that of distance was 400 to 1200 metres. It’s a bit rich to expect such a short trip is going to markedly affect the physical activity of a five-year-old: particularly when the average activity level recorded in the study was consistently in the range of 700 to 800 counts per minute over all waking hours during the week. According to another paper describing this particular accelerometer (Puyal 2002), this level is roughly the threshold between ‘sedentary’ and ‘light’ activity, yet that’s the average – it would appear these are the classic kids who can’t sit still!
The study also tries to account for the differential amount of ‘moderate to high intensity’ activity – the sort that is supposed to really count for good health – by excluding all records with less than 1000 counts per minute. But again, according to Puyal 2002, the threshold for ‘moderate’ activity with this type of accelerometer is 3200 counts per minute, not 1000. So the results aren’t really accounting for the effect of walking specifically on moderate-level activity.
What I think we’d need to see to be convinced is a similar study (a) carried out on 10-year-olds rather than 5-year-olds, (b) distinguishing higher activity levels with a higher threshold. and (c) considering typical Australian school journeys. When my own child goes to school it’s either a 1.4km or a 3km walk (depending whether we get the tram part of the way) – so the minimum is longer than 75% of the British trips.
Strewth #12:
Good points. I could only access one of the papers, but there are 59 others so perhaps there’s a more satisfying explanation for the various claims made here than the paper I referred to.
The walking distances cited aren’t that much shorter than my kids primary school’s catchment (Alphington PS). I checked out other schools in my area and your 3 km is at the far end. Could be further in newer suburbs though.
A local study along the lines you suggest would be good. I find it quite plausible though that the energy expended in walking to and from school (say 1 km each way) by pre adolescents wouldn’t, in the context of their overall day’s activities, be significant enough to be an important obesity factor .
Full disclosure: the primary school my child attends isn’t the closest one to home. But if it was, it would still be an estimated 1km journey, and there are others in the area for whom it would be quite a bit longer. I’m in a 100 year old suburb about 7km from Melbourne GPO – not unlike Alphington really.
As for the benefit of the daily walk, I’d say that by the time a child is 8-10 years old and highly attracted to the idea of spending all day on the couch watching TV or playing games online, they’re surely going to benefit from going for a walk to school or to the local park even if it’s only 1km away.
Strewth #14:
It’s the relative benefit that matters. If a 5-10 yr old is getting adequate daily exercise already, then they’re not going to get much, if any, additional benefit from walking/cycling to school. Certainly not enough to offset the risk to their health from traffic.
Over the nine year span my two attended APS (and I did most drop-offs and pick-ups), I reckon I could count the number of “obese” kids at the entire school (enrolment about 300) on two hands. And that’s despite most parents, and the canteen, giving them ready access to lots of cheap sugar.
Alan which suggests to me a) your schools weren’t representative or b) judging obesity by sight isn’t very reliable. And sure, simplistic measures like BMI have their own issues, but I’d still go by the judgement of a medical professional rather than a lay observer.
I’d also say while it may be true that kids that don’t walk/cycle to school tend to compensate elsewhere (even though that doesn’t fit with my own personal observations), regular scheduled activity that’s part of your daily routine is more likely to lead to life-long behaviours that are beneficial to your health and fitness. Plenty of kids do all sorts of after-hours activities during their school years but give them up once their parents stop insisting on it or ferrying them around everywhere.
Dylan Nicholson #16:
Where are these primary schools with a high % of obsese children? And what’s considered “high”?
And where are these judgements by medical professionals that relate to primary schools? I reckon a lot of the stats that get cited in the media conflate pre adolescents with adolescents.
And if there actually is reliable evidence of worrying levels of obesity in some primary schools, I’ll bet it’s down to social class. Which means that walking/cycling won’t effectively address the problem and isn’t the answer – what the children are eating, and the lifelong example they’re getting from their parents, will be far and away the key problem.
Please login below to comment, OR simply register here :
Thank you for registering, we have just sent you a confirmation email, which includes your new password to be entered below.