Sprinkling Bad Stats on Thin Data in Childhood Obesity

Let’s face it. The evidence base for childhood obesity treatment is thin. We don’t need bad stats to muddy the waters even further. But a recent paper in Pediatrics does just that. The authors conducted a randomized controlled trial of metformin in 160 children with obesity. They enrolled equal numbers of male, female, prepubertal, and postpubertal subjects.

They concluded that metformin works in prepubertal children, but not so well in postpubertal children. The paper has just one problem – that conclusion is false.

“An Important Error of Statistical Inference”

Biostatistics professor David Allison quickly identified this error in a letter to the journal editor:

In short, there is no evidence for a “differential response” by pubertal status, contradicting the authors’ conclusion statement.

Notably, the authors mentioned this fact in the body of their paper, saying:

No differences were found in the impact of metformin according to the pubertal stage when the interaction time × treatment × puberty was applied to the entire population (P = .41).

Yet, they stated just the opposite in conclusions and in the abstract. Unfortunately, that’s all that many readers will see.

Few Options and Thin Data on Treating Childhood Obesity

This is an important study because it adds to a thin evidence base on very few options for treating childhood obesity.

Current regimens for behavioral therapy require intensive follow-up or they won’t work. Most health plans only pay for superficial interventions that have no effect. Most drugmakers steer clear of studying obesity meds in kids. Surgery isn’t an option until children reach adolescence. Even then, tremendous obstacles stand in the way of teens who need it.

So having good information on metformin in these patients is important. Unfortunately, the conclusion that metformin doesn’t work as well after puberty is bad information. The statistics just don’t support it.

Responding to Allison, the authors conceded that their conclusions should be corrected. But the journal has yet to publish a correction.

Without clarity, misinformation can spread rapidly.

Click here for the study, here for Allison’s letter, and here for the authors’ response. And here, you can find a commentary on the study.

Sprinkled, photograph © las – initially / flickr

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


July 23, 2017