Abstract
OBJECTIVE: To estimate associations of retirement with self-reported frequency and duration of naps. DESIGN: Prospective cohort study. SETTING: Population-based. PARTICIPANTS: 1359 current and former Wisconsin state employees, aged 54-69. MEASUREMENTS: Four annual surveys mailed between 2010 and 2014 elicited employment status and nap characteristics. Changes in employment status and nap characteristics were identified from survey pairs measured 1 year apart (up to 3 survey pairs per subject). General linear mixed models with repeated measures were used to estimate changes in minutes napped per week (MNPW), weekly nap frequency, and individual nap duration as predicted by retirement transitions vs stable employment status. All models were adjusted for demographic characteristics, self-rated health, medical diagnoses, sleep problems, circadian preference, and change in nocturnal sleep duration. RESULTS: There were 3101 survey pairs in the analytic sample. Full retirement (transition from working ≥35 h/wk to not working for pay) over a 1-year period predicted a statistically significantly larger mean change in MNPW than stable employment status: mean (95% confidence interval) = +48 (+16, +80) MNPW. Associations between staged retirement transitions (from full-time to part-time work, or from part-time work to full retirement) and 1-year changes in MNPW were not statistically significant. The MNPW changes associated with full retirement were attributable to nap frequency increase of +0.4 (+0.1, +0.8) d/wk; nap duration did not change significantly. CONCLUSIONS: Compared with stable employment status, full retirement is associated with an average 1-year increase of +48 MNPW. This change is attributable to a frequency gain of 0.4 d/wk napped. Changes in nap duration were negligible.