Beyond The Sum Score: A Multidimensional Examination Of Allostatic Load Using Principal Component Analysis And Latent Profile Analysis In Previously Institutionalized Youth

Abstract

Allostatic Load (AL), the stress-related physiological toll on the body, has been significantly associated with long-term health in humans. In this dissertation study, I examined the impact of early institution care on allostatic load in adolescents, using 13 biomarker data collected from 97 previously institutionalized youth who were adopted internationally from orphanage-liked institutions (PI; 61 females and 36 males; age range 12.02 to 21.39 years; Mean age = 16.31, SD age= 2.4 years) and 96 youth born and raised by birth families (Non-adopted, NA; 50 females and 46 males; age range 12.11 to 21.82 years; Mean age = 15.24, SD age= 2.35 years). Three distinct statistical approaches, namely, the Group Allostatic Load Index (GALI), the Principal Component Analysis (PCA), and the Latent Profile Analysis (LPA) were used to characterize the allostatic load. ANCOVA, Structural Equation Modeling (SEM), and multinomial logistic regression were utilized to examine how AL was associated with the experience of early institutional care. Our findings revealed a significant association between early institutional care and elevated allostatic load in adolescents. The sex difference was also found, such that males were more vulnerable to elevated AL levels than females. Moreover, age was positively associated with allostatic load. Furthermore, the construct of AL transcended beyond a single sum score, as evidenced by the detected multi-dimensions of AL and various underlying AL profiles among the studied individuals. Finally, three methods in the present study each offered unique perspectives that collectively enriched our understanding of AL. Taken together, this dissertation study underlined early health risks in youths who experienced early institutional care, highlighting the need for early, targeted, and personalized intervention programs for the ELS-impacted, at-risk population.