In almost every field of human endeavor, funding is restricted. As a result, rather than a flood of new research, there is a phenomenon called the meta-study. Remember this Yale University discovery? Sugar-sweetened beverages have a strange effect. On a day when a person drinks a lot of soda pop, he or she also tends to eat more calories worth of food. That conclusion was derived from the information contained in 88 previous studies.
A meta-study is an exercise in meta-analysis, which has to do with the statistical methods of combining evidence from different sources. (When it comes to qualitative studies, apparently the practitioners tend to favor other terms, like “systemic review,” “research synthesis,” or “evidence synthesis.”) In meta-analysis or systemic review, differences between conditions or populations, or other particulars have to be taken into account. Factors have to be weighted. Allowances need to be made for the designs and intentions on which the various component studies were based.
For instance, MedPage contributing editor Shalmali Pal describes the parameters of a recent University of Melbourne project headed by Elizabeth Waters, Ph.D. From a mass of information on nearly 30,000 kids, obtained from 55 different studies, the criteria were carved out:
For the current analysis, data was collected from controlled trials, with a minimum duration of 12 weeks that took place during or after 2005. Studies did not have to be randomized… All participants were younger than 18 at the time the studies began… the studies had to report at least one of multiple primary outcomes including height and weight as well as BMI.
For these investigations to be meaningful, a lot of human intervention is needed, a lot of tweaking. Despite best intentions and strenuous efforts toward objectivity, the meta-study is by its very nature a product of art as much as science. Still, meaningful conclusions can be attained.
Or not. Both of these news stories refer to the same meta-study from the University of Melbourne, Australia. One is titled “Obesity Programs Have Slim Effect on Kids’ Weight,” and the other, “Child Obesity Prevention Interventions Can Be Effective.” One has a negative spin and the other is positive.
From the Pal story, with the negatively slanted title, we learn that Dr. Waters and the team set out to update a 2005 review which had found that while many interventions aimed at reducing weight gain in children didn’t exactly succeed, at least they were able to get the kids to eat more healthful foods and be more physically active. Pal quotes the report from the Dr. Waters group:
This review demonstrates wide variation in the effectiveness of individual level and behavioral interventions for childhood obesity prevention. On balance, it appears that a variety of interventions can impact on either behaviors or adiposity and shift child outcomes in the desired direction.
In other words, some things work for some people, sometimes. These researchers also noted that pediatric obesity interventions, when they do work, are most effective with kids between six and 12 years of age. There has been a lot of worry over whether these interventions cause negative body image, unhealthy attitudes about weight, or hazardous dieting habits, but, according to the Dr. Waters group, those concerns are unwarranted.
The more positively titled of the two articles, credited to HealthDay, describes the Dr. Waters group’s work in a slightly different way:
Studies that used a controlled design and evaluated interventions, policies, or programs in place for twelve weeks or more were included. Data on intervention implementation, cost, equity, and outcomes were extracted. Meta-analyses using available body mass index (BMI) or standardized BMI (zBMI) scores were performed, with subgroup analysis by age group.
It mentions the same two positive effects: the effectiveness among six to 12-year-olds and the apparent absence of psychological damage from interventions.
In meta-analysis, scholars comb over and re-parse the results of previous research, perhaps noticing relationships that were previously missed, or curious anomalies that seem to be geography-based, or whatever. Sometimes they even find grounds to refute and debunk the old studies. Nothing is wrong with any of this.
But in a field like childhood obesity, where promising new leads seem to turn up fairly often, in many cases the needed information can’t be dredged from old reports. Brain scans, for example, illustrate the similarities between what happens in the brains affected by cocaine addiction and food addiction. It takes a lot of expensive equipment and highly trained people to gather this kind of information.
Even if all laboratories were funded as lavishly as their scientists would prefer, there is still a place for the meta-study. Meta-analysis is not just a poor relation, a second-rate substitute for real research, that we have to make do with in lean years.
Some things readily lend themselves to the meta-study approach. For instance, to implement the food addiction paradigm, research needs at least two prongs, both of which can probably be pursued with meta-analysis. One question is, “What causes addiction?” This would imply, eventually, figuring out how to prevent it. Another is, “How is an established addiction stopped?” While more definitive studies are being performed, it can’t hurt to meta-analyze the daylights out of existing studies.
Some ideas have proven to be effective for a meaningful number of people. They include residential treatment and the traditional 12-step program. Institutions should probably find ways to facilitate the availability of these treatments. As for research, it doesn’t need to seek out subjects who have lost the most weight, but subjects who have attained a healthy size and stayed that way for 10 years or 20 years. Who are those people, and what are their secrets?
Your responses and feedback are welcome!
Source: “Obesity Programs Have Slim Effect on Kids’ Weight,” MedPage Today, 12/09/11
Source: “Child Obesity Prevention Interventions Can Be Effective,” USA TODAY’s Your Life, 12/11/11
Image by incurable_ hippie, used under its Creative Commons license.