Junkfood Science: Does the evidence really show that school obesity policies and weigh-ins don’t increase taunts against fat kids?

October 09, 2008

Does the evidence really show that school obesity policies and weigh-ins don’t increase taunts against fat kids?

Should we be concerned that today’s anti-obesity programs could be having unintended consequences for school children? Since Arkansas passed Act 1220, launching the country’s most comprehensive school-based anti-obesity policies, parents, educational professionals, and child psychology and eating disorder specialists have voiced concerns that the program could be focusing negative attention on children’s weights and putting young people at risk for dangerous weight loss practices and eating disorders, and heightening stigma against fat children.

According to a new study published in the current issue of Archives and Pediatric and Adolescent Medicine, the Arkansas program hasn’t led to any increase in weight-based taunting among children or teens.

How many looked closely at this paper and recognized that such a conclusion could not be credibly made from the data?

As we know, studies can be artfully designed to arrive at whatever conclusion might be desired. Yet, few busy pediatricians probably had time to inspect this study to find out how it managed to arrive at conclusions different from what others have shown. Let’s put on our detective hats and take a look.

What are the concerns over the teasing of fat kids?

Since the classic studies done in the 1960s, the special stigma and shame fat children face from other kids has been recognized as being more severe than even that faced by children with handicaps, amputated limbs or facial deformities. Bullying of children has been shown in the medical literature as having significant physical, academic and emotional health consequences.

Fat prejudice, like all prejudices against discriminated groups, is a learned behavior that’s been shown in children as young as 3 years of age. When those classic studies of children were recently replicated forty years later, researchers found that prejudice against fat children has only worsened. Nearly all fat teenage girls also describe being tormented, stigmatized and called names by their peers because of their size, resulting in lasting painful memories and internalization of those negative labels.

Bullying doesn’t just come from other children. Fat children can encounter it from adults and in media. Even children’s books portray fat characters so negatively, they have been shown to be fueling bullying of fat kids. Adults are even using stigma of fat children through storytelling to attempt to motivate them to eat ‘healthy’ and exercise to lose weight. Most young people may not even recognize this as bullying, they only know they feel worse about themselves and their bodies.

The authors of this new study noted that concerns have been raised over weight-based teasing increasing as a result of the Arkansas policy. “Weight-based teasing is of particular concern because it has been associated with negative psychosocial behaviors, including disordered eating, and anxiety, as well as depression, suicidal ideation, and suicide attempts,” they wrote. A prospective study had found that being teased “about their weight was a significant predictor of disordered eating behaviors 5 years later such as fasting, taking diet pills, self-inducing vomiting, and using laxatives or diuretics.” So, for this study, their primary objective was to see if rates of weight-based teasing had increased among young people since the implementation of the Arkansas Act 1220 in 2004.

Brief overview of methodology

This study* was led by Rebecca A. Krukowski, Ph.D., at the Fay W. Boozman College of Public Health, University of Arkansas for Medical Sciences in Little Rock. As we’ve learned, this academic department has been tasked through the Arkansas legislation, with official oversight of the progress of the implementation of Act 1220, to conduct evaluations that impact its implementation, and prepare an annual report summarizing the results. It is funded by the Robert Wood Johnson Foundation and has received funding for the next five years.

As such, these authors had conducted baseline telephone surveys of parents of children in the Arkansas public school system during the summer of 2004, which they repeated in 2005 and 2006. “A stratified multi-stage sampling procedure was used to ensure that the sample included parents whose children attended public schools across all geographic regions, school types (elementary, middle, and high school), and school sizes.”

Parents were called and completed the telephone interview for their children under age 14. Concerning teasing, they were asked: “Do others tease, joke, or make fun of your child because of his/her weight?” (yes or no).

Older teens were interviewed over the phone themselves, with parental consent. They were asked: “Do others tease, joke, or make fun of you because of your weight?” (yes or no)

The abstract reported that the prevalence of weight-based teasing among children and teens didn’t change significantly from baseline during the two years following the start of this program, with 14% of parents reporting their children had experienced weight-based teasing.

After a closer look, you may not be so ready to buy that conclusion.

The fine print

As with any survey or clinical study comparing two groups, how the participants are selected can easily bias the results. In this study, we see several indications of selection bias.

In selecting participants for the baseline survey in 2004 and the follow-up survey 2 years after the program began, the authors controlled for school size and geographic location. However, they did not control for the gender, age, weight status or minority status of the children! These differences add up to slant the results and minimize the weight-based teasing seen in the 2-year survey. For example, compared to the first survey, the 2-year follow-up survey had:

1. Fewer obese and overweight children. There were 15% fewer children with BMIs ≥95th percentile on growth curves [percentage fell from 14.6% to 12.4%]; and they included 40% fewer children with BMIs ≥85th percentile [percentage fell from 17.4% to 10.6%]. Yet, as we’ll see, weight-based teasing is dramatically more prevalent among fat kids.

2. Fewer girls. Girls made up a 4-5% lower percentage of the kids. Yet, as we’ll see, girls experience more weight-based teasing.

3. Fewer teens. 17% fewer teens, who self-reported the teasing they experienced, were included in the second survey. This weighted the second survey to include more young children whose teasing was reported by their parents.

4. Fewer minority children. There were 21% fewer minorities included in the second survey. As we’ll see, the white children experienced less teasing.

Attrition bias took the form of compliance rates. A total of 1,551 parents were interviewed at the baseline survey — 76.1% of the parents contacted had agreed to participate. By the 2-year follow-up survey, nearly half of all parents refused to participate [55.2% cooperated]. We are left only to guess how this changed the results in the 2-year survey. But like most intervention studies, those who drop out of participating are usually those with the least favorable outcomes. While it might seem natural that if someone in an official capacity called asking you to provide them with more data, you might be less tempted to cooperate if you were having a bad experience with the program, there is some indication that this is the case. The highest rates of weight-based teasing of the fattest children had been reported by parents in the interim 1-year survey, when cooperation rates were higher [62.8%].

Reporting biases surfaced in the results they failed to report, those they did and how they were interpreted. The percentages of children under age 14 whose parents had reported weight-based teasing was given in detail and broken down by weight status. But there was no such breakdown for weight-based teasing as reported by the teens themselves. Nor was any trend data on weight-based teasing among the teens provided - for teens, the three surveys were lumped together with only an odds ratio reported, making it impossible to know if teasing had increased or decreased over the years among them.

Girls <14 years. According to the parental reports, 27.8% of the girls with BMIs ≥95th percentile on growth curves [termed ‘obese’ by the study authors] and 21.5% of girls with BMIs ≥85th percentile on growth curves [termed ‘overweight’ by the study authors] had experienced weight-based teasing. This compared to 9.3% of nonfat girls.

Boys <14 years. According to the parental reports, 28.9% of ‘obese’ boys and 14.6% ‘overweight’ boys had experienced weight-based teasing, compared to 8.7% of nonfat boys.

Overall, 28.5% of ‘obese’ children had reportedly been bullied about their weight, according to their parents, compared to 9% of nonfat children — more than a 3-fold difference. When the authors adjusted for race/ethnicity, gender, age and school size, the ‘obese’ children were 4.63 times as likely to experience weight-based teasing compared to nonfat children.

Minority status. During the years of the Arkansas program, weight-based teasing among the minorities compared to the white children reversed places. At the 2-year survey, weight-based teasing was 10% higher among minority children compared to whites. In contrast, it had been 37% higher among whites before the program began. Once again, minority children appear to be less helped by this program.

Teens. The authors reported: “While ‘obese’ children were greater than 4 times more likely to be teased than their nonoverweight peers (OR, 4.26; 95% CI, 3.16-5.76), ‘obese’ adolescents were almost 9 times as likely as their non-overweight peers to experience weight-based teasing (8.82; 95% 4.55-17.08)weight-based teasing is more prevalent among adolescents compared with children."

The authors led readers to assume this finding indicated that weight-based teasing suddenly increases at age 14, but that may not be the case at all. There’s a much more logical explanation for this finding — one that also may explain why the rates of weight-based teasing among the younger children were so much lower than has been reported by other studies. In fact, it lends support to concerns that weight-based teasing among the younger children may be considerably higher than was reported.

Did you catch why?

Young people themselves are far more likely to know if they are being teased by their peers because of their size than their parents. The self-reported teen figures are likely to be more complete than those from parents for the younger kids. Other studies have shown that most victims of even severe school bullying never tell their parents or an adult that they are being bullied. Researchers have learned that sadly as few as less than a quarter of parents whose children have suffered bullying are aware of it.

“[M]any times kids are reluctant or embarrassed to talk about being teased or bullied,” said Wendy Craig, Ph.D., a psychology professor and researcher on bullying at Queens University in Kingston, Ontario. Parents may not know what to look for or how to broach the topic. Fat children all too often suffer in silence and are left feeling they have little support of adults. When they’re surrounded by negative information about ‘obesity’, even from adults figures of authority at school, young people may especially not feel free to speak out about how they feel. Their stories, often not shared until years later, reveal painful and traumatic experiences, including weigh-ins.

I was dying a thousand deaths — and I weighed 165 at 10 years old, probably almost 100lbs more than all the other kids. I remember everyone laughing at me, and to this day cannot believe adults could inflict this kind of trauma on a child.

The Fay W. Boozman College of Public Health authors of this study made no attempt to corroborate the validity of the parental reports by interviewing the children.

In the final analysis, the conclusions of this study as reported in the medical journal, in the University’s press release and in media, overreach the data. “[W]e found no downside to student BMI assessments,” said co-author Delia Smith West, Ph.D.

“And if the assessments and the law’s other provisions motivate kids to eat healthier foods and get more exercise,” she said [which, as we’ve seen, they didn’t],” there is the upside of a major public health benefit.” According to the press release, professor West added that school-based obesity programs “can even reduce the frequency of teasing.” But this study provided no evidence to support that, either.

Why did this study work so hard to downplay the evidence indicating that this government initiative appears to be having unintended consequences, heightening body and health concerns among the children, increasing stigma of the fat children, and increasing eating disordered and high-risk weight loss efforts among girls; had failed to improve the children’s diet or activity levels; and had failed to demonstrate any measurable effects on the children’s weight status or health?

Associated Press provided some insights when it described legislative complaints waged against this program last year that had threatened it, based on parental reports of problems and concerns it was hurting children.

Lead author, professor Krukowski, was reported as saying she hoped the Legislature would consider the survey results during its coming session next year. “The weight-based teasing rates didn't increase, they didn't change at all,” she said. “This definitely does give some evidence that doesn't need to be a concern.” The governor’s spokesperson said that it was encouraging to hear that the program isn’t “adding any additional psychological impact on the kids.” Was this study most focused on investigating to see if this program was having adverse effects on the children or on lobbying politicians for continuation of the legislature’s support of this program?

© 2008 Sandy Szwarc

* Study disclosures:

Financial Disclosure: None reported

Funding/Support: This study was supported by grants 051737, 60284, and 30930 from the Robert Wood Johnson Foundation (Dr Raczynski).

Bookmark and Share