Eating Disorders
Early Development of Maladaptive Eating Symptom Networks among Mexican American Children
Juan C. Hernandez, M.A.
Graduate Student Researcher
Arizona State University
San Antonio, Texas
Claire Cusack, M.A.
Graduate Student
University of Louisville
Louisville, Kentucky
Linda Luecken, Psy.D.
Professor
Arizona State University
Tempe, Arizona
Cheri Levinson, Ph.D. (she/her/hers)
Associate Professor
The University of Louisville
Louisville, Kentucky
Marisol Perez, Ph.D.
Professor and Associate Dean
Arizona State University
Tempe, Arizona
Objective: Maladaptive eating behaviors commonly emerge during childhood and elevate risk for future physical and mental health impairment, diminished health related quality of life, and eating disorder occurrence. Exploring maladaptive eating among diverse groups of children can optimize prevention efforts by identifying culturally indicated sensitive periods and symptom targets along a developmental trajectory. Mexican American children offer one such diverse group, uniquely characterized by a dietary context that often includes culturally limited nutrition guidelines and complex bicultural dynamics.
The current study leverages symptom network analysis, an emerging psychometric approach in behavioral sciences, to 1) identify sensitive periods in the development of Mexican American children’s maladaptive eating by exploring longitudinal changes in the strength and structure of syndromic interconnectivity, and 2) identify symptom-level influences on broader symptom activation from early- to middle-childhood.
Impact: Our results show that while strength of network connectivity remained stable, specific affective experiences may vary across development (e.g., eating less when distressed; eating too much if allowed). Furthermore, leaving food on the plate may be a key practice to promote as it may potentially decrease activation of less desirable symptoms (as demonstrated through strong negative EI effect). Future research should test if interventions targeting central symptoms cultivate healthy eating behaviors in Mexican American children transitioning from early- to middle-childhood.
Method: The current sample includes 322 low-income mother–child dyads (53.7% of children assigned female at birth). At enrollment, on average mothers were 27.8 years old and completed 10 years of education. Eighty-six percent were born in Mexico, 82% spoke Spanish as their primary language, and 83.5% were unemployed. Demographics and maladaptive eating behaviors (nine maternal-report items from the Child Eating Behavior Questionnaire) were assessed when children were 6, 7.5, and 9 years old. We estimated three cross sectional regularized partial correlation networks. Because networks had some negative edges between items (i.e., activation of one symptom diminished activation of another symptom), we calculated expected influence (EI) which appropriately estimates symptom centrality in networks with both positive and negative edges. We then computed network comparison tests (NCTs) to examine changes in network strength and structure across time.
Results: The items with the highest EI at age 6 were leaving food on the plate (EI=-1.83) and eating less when feeling upset (EDI=1.02); at age 7.5 were leaving food on the plate (EI=-1.38) and enjoying eating (EI =-1.31); and at age 9 were enjoying eating (EI=-1.49), eating too much if allowed (EI =1.23) and leaving food on the plate (EI=-1.23). NCTs revealed that at age 6 the network structure significantly differed from the 7.5- and 9-year-old network structures (ps < .001); the network structures did not differ between 7.5 and 9 (p = .081). Further, networks did not vary in global strength across time points (ps = .392–.892), meaning networks had similar strength regardless of age.