Abstract:
|
We previously reported that age-specific mortality of participants in the National Health and Nutrition Examination Survey (NHANES) 1999-2010 differed by survey cycle (better survival for later survey cycles), which was attributed to the longer followup time for early survey cycles rather than a true secular trend. In the current work, using Cox regression, we explore alternative methods to adjust for the followup time-mortality association, which was found to be more pronounced for older age during followup, in the analysis of blood lead and cardiovascular disease (CVD) mortality. We examined whether time-related factors including followup time (and its functions interacted with age during followup), age at followup start, calendar time at followup start, age during followup, and calendar time during followup) operate as effect modifiers of the blood lead-CVD mortality association. Our preliminary analyses indicate that the time-related factors are not effect modifiers of blood lead-CVD mortality association.
|