Identifying Differences in Predictive Variables for Change in Self-Rated Health by Generational Cohorts

Abstract

Background: Self-rated health (SRH) is an individual's assessment of overall health measured as the response to a question such as general, would you say your health is: excellent, very good, good, fair, or poor? In the literature, there is a consistent link between a low SRH and mortality in a dose-dependent response when controlling for a number of covariates. However, there has not been much research describing how SRH changes over time and how this change might differ between members of different generational cohorts. Methods: This study aims to identify and describe these differences through using the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a large nationally representative data set with two waves of data. Generational cohorts were split into five levels; the oldest three levels were analyzed: the Baby Boomers (born from 1946-1964), the Silent Generation (born from 1925-1964), and the Pre-Silent Generation (born before 1925). An individual's change of SRH was defined by their SRH at wave 1 and at wave 2 as no change in SRH, an increase in SRH, or a decrease in SRH. Models were constructed using a multi-tiered backwards selection process for each generation and initial SRH with demographic, behavioral, and health variables known to be linked to SRH. Results: The oldest generation had the worst initial SRH and the worst change in SRH. The generation groups differed in the frequency distributions of wave 1 variables: the oldest generation was more female, more Non-Hispanic White, smoked and drank less, and had higher levels of disability and pain. However, the proportion of individuals with depressive symptoms was similar between the generational groups. The logistic models were consistent with the literature except not all variables were significant predictors of a change in SRH at every level of initial SRH and generation, particularly for the Pre-Silent Generation. Conclusion: The Pre-Silent Generation consistently had the fewest number of significant predictors at each level of initial SRH in the logistic regression models, even with variables known to be strongly linked with SRH. The possibility that cohort-specific differences are causing the discrepancies in how SRH changes between the generations remains a viable explanation.%%%%M.P.H., Public Health – Drexel University, 2014