From the above discussion it can be seen that there are a number of prerequisites before a meaningful cost of obesity study can be performed. There needs to be a good study of the prevalence of obesity (defined using the WHO's suggested BMI cut points). In addition, a standard group of obesity-associated disorders should be included and it is optimal if there are known local RRs of these disorders from previous studies. Such estimates of RR will account for details of ethnic variation that may otherwise not be fully detailed in an analysis. If there are no local RRs, then it would be appropriate to use published values from other countries, trying as best as possible to match ethnic groupings and perhaps using a conservative risk factor as well as a high value. This approach assumes that the physiological and pathologic consequences of obesity are consistent across communities, which is not unreasonable. There also needs to be a standard approach to cost centers used for direct costs. In this way, it will become easier to compare costs across countries and perhaps a little easier to perform the studies themselves if the type of data utilized is standardized and collected.
To estimate the proportion of disease that could be prevented by eliminating obesity, we calculate the population-attributable risk percent (PAR%) or the population attributable fraction. This is the maximum proportion of disease attributable to the specific exposure (such as obesity). PAR% is based on the incidence of disease in the exposed (i.e., obese group) as compared with the nonexposed, taking relative risks from analyses that control for confounders (e.g., age, smoking, dietary intake, etc.). PAR% is calculated using
where P is the prevalence of exposure in the population and RR the relative risk for disease.
A key issue in estimates of the proportion of disease due to obesity is the reference group. Is the disease a
Was this article helpful?