Is Your Baby at Risk for Childhood Obesity? New Tool Predicts Likelihood

Save ArticleSave Article

Failed to save article

Please try again

This article is more than 7 years old.
Multiple factors influence childhood obesity and a new algorithm can predict a kid's chance of being overweight. (iStock)

What if you could predict whether you'd be overweight five years from now?

Adults face varying degrees of difficulty in avoiding a beer belly. But once a child becomes obese, he or she is likely to stay that way into adulthood, with an increased risk for hypertension, diabetes and metabolic disease.

That's why UCSF has developed an algorithm it says can predict the likelihood a baby will be obese by age five.

Pediatrician Jacob Robson, lead author of a study on the algorithm published in The Journal of Pediatrics Friday, said age five was the focus of the research because that's the age children enter school and become more susceptible to making poor food choices, and because that was the data that was available.

Robson and UCSF epidemiologist Janet Wojcicki recruited a high-risk demographic—pregnant Latina women and their children—and tracked them over five years.


Of the 166 children they followed, almost one-third were obese by age 5.

"Low-income minority groups are disproportionately impacted by obesity," says Wojcicki. "Disadvantaged groups have less access to care, and they’re often living in food deserts and may not have a great-built environment where they have opportunities for physical exercise."

But Wojcicki added that all groups in the U.S. are at risk for obesity.

Approximately 12.7 million children and adolescents are obese in the U.S., according to the Centers for Disease Control and Prevention. The CDC defines obesity in children and teens as a body mass index, or BMI, at or above the 95th percentile for their peers of the same age and sex.

But now, Wojcicki and Robson say, their predictive model can lead to early intervention.

The team's algorithm scans electronic health records and analyzes 10 data points, like a mom's weight before pregnancy, a baby's birth weight and whether she's being breastfed—all information that is routinely collected.

Based on this data, the algorithm states the percentage that a child will likely be obese by age 5.

Using the algorithm, researchers found that 94 percent of infants whose risk was ranked below the 25th percentile landed in the normal weight range by age five. In contrast, 61 percent of those whose risk level was scored above the 75th percentile were obese by age five.

Wojcicki and her team found two factors played the biggest role in predicting childhood obesity: higher-than-average birth weight and the amount a baby had gained six months after birth.

"Any super-accelerated weight gain is concerning," says Wojcicki.

Wojcicki and Robson envision a system where doctors receive alerts on their phones if the algorithm detects an at-risk patient. Then they could inform the mother before her child reaches kindergarten.

The key is to start early.

"A lot of the apps people are working on in Silicon Valley are for targeting kids that are already overweight," says Wojcicki. She says by that time it may be too late.

"Most prevention efforts are not harmful," jokes Robson. "Decreasing screen time, getting more exercise, eating healthy and targeting kids early could play big role."

The researchers' next step is to test more mothers and their children to confirm the study's findings.