Reporting error in weight and height among the elderly: Implications and recommendations for estimating healthcare costs

A large literature has examined the healthcare consequences of obesity. A major barrier to careful study of these consequences is reliance on self-reported measures of weight and height. Previous research has developed algorithms to adjust for such error among working age adults. In this study we consider elderly adults, a group likely to differ in reporting error patterns from working age adults due to involuntary weight loss and changes in cognition, muscle mass, and bone density. We first provide evidence on the degree and type of reporting error in this population. Second, we consider how well standard approaches to adjusting for such error preform in an elderly population in terms of estimating obesity prevalence and regression coefficients. These findings have direct implications for evaluating anti-obesity programs among the elderly and estimating the obesity-related healthcosts to the Medicare program.