Overly broad definitions, a lack of appropriate, or any, comparison groups, among other things, in studies looking at the incidence, prevalence, and control of the condition-;epidemiology-;have distorted the risks, say the researchers.