Abstract

Objectives: To descriptively document birth cohort differences in sleeping patterns; self-reported age-specific sleep duration and insomnia symptoms among adults aged 50+ from the National Health Interview Survey (NHIS) and the Health and Retirement Study (HRS).

Methods: We analyzed respondents aged 50+ (born 1920 to 1969) from the 2006-2018 NHIS (n=162,400) and HRS (n=28,918). We fit multinomial models among the NHIS sample predicting age-specific optimal sleep duration (optimal for age versus short for age, and optimal for age versus long for age). For the HRS sample, we fit growth curve models predicting age-based insomnia symptom trajectories. The models for both samples adjusted for age, gender, race/ethnicity, and educational attainment.

Results: Results regarding sleep duration in the NHIS, suggested that cohorts born in the 1950's and 1960's had significantly higher odds of reporting short sleep duration than cohorts before them. Results from the HRS similarly illustrated that cohorts born in the 1950's and 1960's had significantly higher levels of insomnia symptoms than those born before them. The worsening sleep among cohorts entering midlife was consistent regardless of alternative cohort specification, when age groups or periods were analyzed, and when more extensive covariates were modeled.

Discussion: We observe a pronounced decline in healthy sleeping patterns among American cohorts in midlife, with consistent and striking results across datasets, methods, and measures. These findings have important implications for the well-being and longevity of Americans who have entered midlife in the 21st century.