2024 journal article
Allostatic Load in Childhood, Adolescence, and Young Adulthood: Are Assumptions of Measurement Invariance Warranted?
Psychosomatic Medicine.
UN Sustainable Development Goal Categories
1. No Poverty
(OpenAlex)
3. Good Health and Well-being
(Web of Science)
Source: ORCID
Added: April 8, 2024
ABSTRACT Overview Allostatic load represents the cumulative toll of chronic mobilization of the body’s stress response systems, as indexed by biomarkers. Higher levels of stress and disadvantage predict higher levels of allostatic load, which, in turn, predict poorer physical and mental health outcomes. To maximize the efficacy of prevention efforts, screening for stress- and disadvantage-associated health conditions must occur before middle age—that is, during childhood, adolescence, and young adulthood. However, this requires that models of allostatic load display properties of measurement invariance across age groups. Because most research on allostatic load has featured older adults, it is unclear if these requirements can be met. Methods To address this question, we fit a series of exploratory and confirmatory analytic models to data on eight biomarkers using a nationally representative sample of N = 4260 children, adolescents, and young adults drawn from the National Health and Nutrition Examination Survey dataset. Results Exploratory and confirmatory models indicated that, consistent with allostatic load theory, a unidimensional model was a good fit to the data. However, this model did not display properties of measurement invariance; post-hoc analyses suggested that the biomarkers included in the final confirmatory model were most strongly intercorrelated among young adults and most weakly intercorrelated among adolescents. Conclusions These results underscore the importance of testing assumptions about measurement invariance in allostatic load before drawing substantive conclusions about stress, disadvantage, and health by directly comparing levels of allostatic load across different stages of development, while underscoring the need to expand investigations of measurement invariance to samples of longitudinal data.