Do type, timing and duration of life course non-employment differentially predict dementia risk? An application of sequence analysis.

Periods out of employment may influence dementia, but characterizing lifecourse employment is difficult and prior research is sparse. This study used sequence and cluster analysis to characterize type, timing, and duration of lifecourse work gaps and estimate associations with dementia risk. Life History Mail Survey supplement to the U.S. Health Retirement Study participants (N = 5,945, 13.6 % of the Health and Retirement Study sample) reported lifecourse employment (full time or part time) and reasons and age of work gaps (unemployment, schooling, caregiving, or disability). Our exposure was gender-stratified employment trajectories from age 18-65, characterized using sequence analysis and cluster analysis. Our outcomes were algorithmically defined dementia probability scores and memory scores. We estimated the association between employment trajectories and dementia risk using generalized estimating equations and memory decline using linear mixed effect models, adjusted for age, gender, birthplace, and childhood socioeconomic status. We identified 11 employment trajectories for women (including predominant work, disability, unemployment, caregiving, retirement) and 10 for men (similar, but no caregiving). Compared to "predominant work", "disability" and "unemployment" trajectories were associated with higher dementia risk for men and women (e.g., disability among women: OR = 3.62; 95 % CI = 2.25, 5.81). Among women who cared for family, those who did not re-enter the labor force full-time had higher dementia risk (e.g. "family gap, go back part time": OR = 1.79; 95 % CI = 1.15, 2.79) compared to the predominant work cluster. Women who cared for family and returned to full-time work had similar cognitive outcomes to those in the predominant work cluster. Men who had long spells of part-time work also had elevated dementia risk (e.g. part time earlier: OR = 1.64; 95 % CI = 1.16, 2.57). Finally, women and men with long periods of unreported employment status had higher dementia risk than those in the predominant work trajectory. Results suggest the type, timing and duration of work gaps are differentially associated with dementia risk. Work gaps due to disability, unemployment or unreported employment status predicted higher dementia risk. Permanently leaving full-time work for caregiving predicted worse cognitive outcomes but temporary caregiving-related work interruptions did not.