Thursday, May 24, 2007

Impact of Community Based Diet Intervention, More than BMI

There's been a lot of fuss over the publication of the Shape Up Somerville: Eat Smart Play Hard.™, initiative results in the May issue of Obesity. (abstract)

As noted in the abstract, the objective "was to test the hypothesis that a community-based environmental change intervention could prevent weight gain in young children" as reflected in the change in BMI z-score.

In total, 1,178 children were enrolled - 385 in Somerville (the intervention group), and two control communities, one enrolling 561 children to be followed, the other 232 children to be followed. All the children were in grades 1 through 3 at the beginning of the study period, school year 2003-2004 (September to June).

The intervention was intensive, and expensive.

Overall, the researchers started with a $1.5-million dollar CDC grant. And, if that were not enough, the Tufts website notes the grant "received supplemental support from The US Potato Board, Dole Food Company, Blue Cross and Blue Shiled of MA. It is also supported through generaous donations by Whole Foods Market, WGBH, New Balance, Gaining Ground, Annie’s Homegrown, Stonyfield Farm, Organic Valley, White Wave, Earthbound Farms, Shaw’s Supermarkets, Cabot Creamery, Friendship Dairies, Newman’s Own, The Vermont Bread Company, and the Kashi Company."

The Wall Street Journal included some idea of this supplemental support in quantifiable terms...

"...the Tufts researchers helped the city win a grant from the Robert Wood Johnson Foundation for a bigger-ticket item, an extension of a bike path that will eventually go all the way to Boston. "

"...the City Council came up with its own ideas: reimbursements on gym membership for city employees and dozens of new bike racks for schools and streets."

"...Dr. Economos persuaded a Whole Foods store to donate about $35,000 in fresh produce."

"...a Department of Homeland Security grant is providing fitness equipment at fire stations and chefs to train the firefighters about nutrition and healthy meals."

"...a doctor sponsors the community fun run."

So, it's hard to add up all the additional monies and in-kind donations to know the true cost of the initative, but Somerville has secured an additional $1.5-million dollars to continue the efforts.

What exactly happened in Somerville that so much attention is focused on this small town?

Throughout the city of Somerville, the efforts to place the towns children on a diet and increase activity included:
  • School lunches, classroom snacks and fundraisers revamped

  • Teachers were trained to implement the new health curriculum, called The HEAT Club (Health Eating and Active Time)

  • School Staff Role Modeling implemented

  • After school program policies created for snacks

  • After School policies for physical activity developed

  • Increased time for recess, lunch, and physical education

  • Annual height and weight data collected with reports for parents

  • Monthly newsletter was sent to parents

  • Monthly community newsletter to community members

  • Parent education forums held for non-english speaking parents

  • Area restaurants developed "Shape Up Approved" menu items

  • Safe Routes to School maps were created

  • Crosswalks were re-painted in thermoplastic material

  • Bike racks installed at elementary schools

  • School nurses and area Pediatricians were educated about taking weight and height of children

I don't think they missed an opportunity to intervene, do you? No Child Left Behind - heck, this was an "in your face" intervention, with no one left behind.

Throughout Somerville, it was impossible to escape the Shape Up Somerville messages.

So you'd think the results would be impressive given the media's attention the last few weeks.

The researchers trumpted "A community-based environmental change intervention decreased BMI z-score in children at high risk for obesity. These results are significant given the obesigenic environmental backdrop against which the intervention occurred. This model demonstrates promise for communities throughout the country confronted with escalating childhood obesity rates."

Based on their results that found over eight months, "At baseline, 44% (n = 385), 36% (n = 561), and 43% (n = 232) of children were above the 85th percentile for BMI z-score in the intervention and the two control communities, respectively. In the intervention community, BMI z-score decreased by –0.1005 (p = 0.001, 95% confidence interval, –0.1151 to –0.0859) compared with children in the control communities after controlling for baseline covariates."

Simply put, the kids in the intervention group gained about 1-pound less than the kids in the two control towns who received no intervention.

They didn't lose weight, they just gained less, which is expected - growing children gain weight each year (or are supposed to).

Within the full-text we find a curious table designed to highlight the effect of this intervention, Estimated intervention effect over 8 months on a child at the 75th percentile BMI z-score, which seems contrary to the objective of reducing the incidence of overweight and obesity, which is defined as a BMI that places a child in the 85th percentile or higher.

Nonetheless, the researchers included it to show how the intervention would effect children at the 75th percentile - children, it should be noted, already within the definition of a "healthy weight" for age.

It seems to have escaped many that with or without intervention such a child would remain within "normal, healthy" weight, as evidenced in the table data included.

Case in point - highlighted in the table - a boy, at baseline 8-years old has a BMI of 17 and is in the 75th percentile. Eight months later, without intervention (it's hypothesized) he might grow about an inch-and-a-half and gain 5.07 pounds; with intervention, just 4.25-pounds (a difference of 0.82-pounds, or about 12-ounces).

Without intervention the child's BMI is now 17.4; with intervention it is now 17.17 - in both instances the BMI increased, right?

But in both instances the child in this scenario fell on the charts, going from the 75th percentile to the 73.9 percentile without any intervention and the 71.1 with intervention, and at a cost of $1.5-million to keep this hypothetical child well within "normal" when he would have remained there anyway?

I wondered, what happened to the kids who were overweight or obese to start - the 170 of 385 children in the intervention.

As I combed through the data, another curiosity - no data was provided to show the intervention was effective for this particular subset of children within the intervention group.

Why was that data not provided?

I would think that would be the icing on the cake for the researchers to highlight the benefit of such an intense intervention, wouldn't you?

So, I got to thinking, what changes, based on the data provided, might happen for a child within this group?

The researchers neatly summed up the hypothetical child within normal, so how about we create the same table for a child who was at the 95th percentile at baseline to see what might happen.

Johnny is 8-years old, he stands 4' 2.25" and weighs 72-pounds - BMI 20.1, 95th percentile.

In eight months, like the scenario presented by researchers, he now stands 4' 3.75" (growing 1.5") and without intervention is likely to weigh 77-pounds (gained five pounds); with intervention he'd weigh about 76-pounds (gained four pounds).

Without intervention he's fallen to the 93.8 percentile; with it he's fallen to the 93rd percentile.

Yes - with or without intervention this hypothetical child went from "overweight" to "at risk for overweight" - which begins to explain the absence of data on the subject of where the children were on the charts at the end of the study period.

I'm not trying to say we shouldn't try to help children engage in more activity or eat a healthier diet, but quite frankly, these results are not as impressive as we're being led to believe.

We have no data to look to in our understanding of how this type of aggressive community-based intervention impacted children already underweight or at the bottom of "normal, healthy" weight; we have no clue as to how this type of intervention impacted those children who were well within "normal, healthy" weight; and without the hard data we don't know that any of this helped those children at risk for or already overweight!

I'd like to see the data and also see someone investigate impacts in other areas of life, like reading and math scores. Sandy Szwarc, over at Junkfood Science, pointed out that "While the school year was absorbed in diet and exercise, (after the Shape Up program was completed in 2005) the average reading test scores among Somerville kids are 15.4% below state average, and their math test scores are a whopping 26% below those of kids in the rest of the state."

I did some poking around and found the Somerville MCAS testing stats online; in the 2003/04 school year, third grade reading scores indicated that 44% needed improvement or were failing; in the 2004/05 school year - these same students, now 4th graders, tested in ELA (english language arts) showed 63% now needing improvement or failing, and 73% needed improvement or were failing math (not tested in 3rd grade).

Just some food for thought, and a hat tip to Sandy for the educational impacts that may be part of the intiatives overall impact not recorded in the results that are focused on BMI z-scores.


  1. On the one hand, getting the kids to eat more fruits and veggies and less junk is a good thing. On the other it seems like a huge waste of resources for something the parents should be doing anyway.

  2. So what this tells me is that out of about a dozen different things they could have used to proclaim their success, there were none. Except something that looks like a fabulous statistical convolution. Their lack of forthcomingness 'implies' that they actually have nothing that they want to be forthcoming about. Go figure.

  3. Anonymous5:25 PM

    It's worth mentioning that Somerville isn't quite a "small town"- it has the highest population density in Massachusetts.