To the Editor In their analysis of US obesity time trends, Dr Flegal and colleagues1 found no temporal variation in men, but they found both linear and quadratic time trends in women that were independent of age, race/Hispanic origin, education, and smoking status. They stated that “the results … suggest that attempts to extrapolate from past data to possible future trends in obesity prevalence may not provide valid estimates,” and cited 9 published studies of obesity projections as examples. I believe that this conclusion could be misinterpreted and would like to elaborate on the intended purpose and rationale behind projection analyses.
Lo E. Use of Projection Analyses and Obesity Trends. JAMA. 2016;316(12):1317. doi:10.1001/jama.2016.11962