by Fahad Razak, MD, Aditi Krishna, S.V. Subramanian, PhD
It is well known that in high-income countries such as the United States, average body weight, typically measured as body mass index (BMI), has steadily increased over the past few decades. It is implicitly assumed that these average BMI increases are constant and unchanging across all weight classes (i.e. normal weight, overweight, obesity). Very few studies have systematically looked at whether this is true. Does BMI increase equally across weight classes? Has the range in BMIs actually increased as average BMI rises? We used data from the largest dataset available to examine yearly changes in weight gain in the United States – the Behavioral Risk Factor Surveillance System, a survey of more than 3 million people.
Over a 20 year period, we showed that for all individuals, while mean BMI increased by 8 percent, the standard deviation (capturing the span or range in BMI) increased by nearly 30 percent. This tells us that over 20 years, individuals have responded very differently to the increasingly obesogenic environment in the US. For instance, in 1993, the 5th and the 95th percentile on the BMI distribution was 20.1 kg/m2 and 34.5 kg/m2, respectively; while in 2012 it was 20.3 kg/m2 and 39.6 kg/m2, respectively. Over time, the average means less. It oversimplifies an increasingly varied population.
Most importantly, we observed an increasing BMI range among individuals in all subgroups: sex, race, and education level. For example, the BMI range increased by 36 percent among non-Hispanic whites and 33 percent among non-Hispanic blacks. Span increased by 28 to 38 percent within groups of different educational levels. Future research needs to focus equally (and even perhaps more) on understanding the causes of increasing BMI ranges and not just average changes. It might affect how we choose to intervene to address the rapid change in population weight over the last 20 years.
About the Lead Author
Fahad Razak is a staff physician at St. Michael’s Hospital, Department of Medicine, University of Toronto, a research scientist in the Keenan Research Centre, Li Ka Shing Knowledge Institute, and a research fellow at the Harvard Center for Population and Development Studies. He has a multidisciplinary background including biomedical engineering, epidemiology and public health, and a medical degree with specialization in general internal medicine. His research interests include the causes and consequences of the changing shape of chronic disease risk factor distributions at the population level, with special focus on body weight.