Sedentary Behavior and Dementia

To the Editor A recent study on physical activity and risk of dementia has several strengths, including the large population and an objective measure of physical activity. There is an important limitation that the authors do not appear to fully address, however: ascertainment bias. As the diagnosis of dementia depended on either having a death certificate or hospitalization (in which inpatient diagnostic codes would be assigned), the outcome could be ascertained only in participants with these health outcomes. Participants without either of these 2 health events could not be classified as having dementia regardless of actual dementia status. Because of the strong links between physical activity and overall health, it seems likely that the association between physical activity and dementia with this design could represent, in part, higher likelihood of hospitalization or death (ascertainment opportunity) due to health conditions other than dementia. It seems that potential ascertainment bias could be addressed, in part, by performing an analysis limited to participants who died or were hospitalized or by controlling for number of hospitalizations (opportunities for outcome ascertainment) and death.
Source: JAMA - Category: General Medicine Source Type: research