Parallel Process Growth Mixture Model Research Articles

Joint Trajectories of Depression and Rumination: Experiential Predictors and Risk of Nonsuicidal Self-Injury

Nonsuicidal self-injury (NSSI) is common in adolescence. Rumination is a key risk factor and often co-occurs with depressive symptoms. This is the first study to examine the joint longitudinal trajectories of rumination and depressive symptoms as predictors of NSSI, and the adverse experiences associated with these trajectories. A community sample of 1,835 adolescents (55.9% male participants, 12.3 ± 0.5 years of age) completed questionnaires to assess adverse childhood experiences, rumination, depressive symptoms, and NSSI. Assessments were made 4 times over 18 months. A parallel process growth mixture model showed that youth with high trajectories of rumination but low trajectories of depression had moderately increased odds of NSSI (2.43-fold, 95% CI 1.53-3.91) compared with adolescents with low trajectories of both rumination and depression. Odds ratios (ORs) in adolescents with low trajectories of rumination but increasing or high trajectories of depression were similarly elevated, suggesting that high trajectories of rumination or depression were risk factors in isolation. However, odds were 10.06-fold greater (95% CI 5.68-18.02) when high trajectories of rumination occurred in tandem with high trajectories of depression. Multinomial logistic regression showed that male sex (OR 10.54, 95% CI 5.66-19.63), peer victimization (OR 2.25, 95% CI 1.72-2.96), and parental alienation (OR 1.94, 95% CI 1.46-2.57) were key determinants of membership in the highest risk group. Risk for NSSI is markedly increased in adolescents with high longitudinal trajectories of depression and rumination. Reducing exposure to peer victimization, cyber victimization, emotional abuse, parental alienation, and interparental conflict may reduce risk.

Parallel trajectories of vaping and smoking cannabis and their associations with mental and physical well-being among young adults

BackgroundVaping and smoking are common modes of using cannabis (THC) among young adults, but little is known about how patterns of cannabis vaping and smoking unfold over time or how using one or both types of products may differently affect mental and physical well-being. This study examines parallel processes of cannabis vaping and smoking over 5 years and mental and physical outcomes in a sample of young adults. MethodsAnnual surveys were conducted between 2016 and 2022 with a mostly California-based cohort of 2428 young adults. Parallel process growth mixture models examined trajectories of past-month frequency of cannabis vaping and smoking from ages 20 – 25. Classes were extracted based on parallel trajectories of vaped and smoked product use. Models assessed differences in self-reported mental (anxiety, depression) and physical (ailments, subjective overall) well-being outcomes in young adulthood across classes, adjusting for demographic characteristics and mental and physical well-being at pre-baseline (average age 19). ResultsFour cannabis vaping/smoking classes emerged: low use of cannabis (84.7%), decreasing smoking, low-moderate vaping (7.1%), stable moderate smoking, decreasing vaping (4.6%), and rapid increasing dual use (3.4%). Classes were similar on physical well-being indicators in young adulthood. The rapid increasing dual use class showed higher anxiety and depressive symptoms compared to other classes. ConclusionProgression to higher frequency of both vaping and smoking cannabis in young adulthood may contribute to poorer mental well-being compared to other use patterns. Targeted efforts to reduce dual vaping and smoking in young people who use cannabis may be needed.

Maturing out of alcohol and cannabis co-use: A test of patterns and personality predictors.

Reductions in substance involvement into adulthood are thought to represent a normative maturing out of substance use. However, patterns and predictors of maturing out of alcohol and cannabis co-use remain largely unstudied. Therefore, the current study tested developmental trajectories of alcohol and cannabis use from late adolescence into adulthood and whether late adolescent personality traits predicted trajectory class membership. Data come from a longitudinal study of family history of alcohol disorder (N=458). Age bands were created to model trajectories of drinking quantity, negative alcohol consequences, and cannabis use frequency from late adolescence (age 18-22) to young adulthood (age 23-28) and adulthood (age 29-36). Participants reported on their sensation seeking, conscientiousness, and neuroticism during late adolescence and their typical drinking quantity, negative alcohol consequences, and cannabis use frequency at each age band. Three trajectory classes were derived from an initial Parallel Process Growth Mixture Model: (1) low-risk maturing out of alcohol-only use, (2) high-risk maturing out of co-use, and (3) high-risk switchers who increased their cannabis use into adulthood. Late adolescent sensation seeking was associated with higher odds of being in both co-use trajectories, whereas a lack of conscientiousness was associated with higher odds of being a co-use switcher. We identified heterogeneity in trajectories of co-use, which suggests that a lack of maturing out of alcohol involvement may be accompanied by increased cannabis use. Moreover, late adolescent personality traits may predispose individuals toward riskier developmental trajectories of substance use into adulthood.

Parent problem drinking trajectory classes predict anxiety in adolescence and emerging adulthood

BackgroundThis study identified latent trajectory classes for maternal problem drinking and paternal problem drinking and examined the associations between these trajectory classes and offspring anxiety symptoms during adolescence and emerging adulthood. MethodsParticipants (n = 870; 54% female; 59% non-Hispanic White; Mage = 16.10, SD = 0.71) were administered surveys during the spring of 2007, 2008, and 2009, and 2014. ResultsFit indices from parallel process growth mixture models suggested three dual trajectory classes: (1) Low initial levels of maternal problem drinking and paternal problem drinking that increased over time (Low-Both); (2) Low initial levels of maternal problem drinking that increased over time and high initial levels of paternal problem drinking that increased slightly over time (Low-Mom/High-Dad); (3) High initial levels of maternal problem drinking that increased slightly over time and low initial levels of paternal problem drinking that remained relatively stable over time (High-Mom/Low-Dad). Girls were more likely than boys to be classified in the Low-Mom/High-Dad and High-Mom/Low-Dad classes, relative to the Low-Both trajectory class. In addition, adolescents in the High-Mom/Low-Dad trajectory class reported the most anxiety symptoms during adolescence and emerging adulthood. LimitationsLimitations include the reliance on one informant (the adolescent/emerging adult) and the geographically limited sample (northeastern United States). ConclusionsPrevention and intervention programs aimed at reducing anxiety should consider changes in alcohol use in both the father and the mother over time. Moreover, special attention should be paid to maternal problem drinking given that it appears to be a salient risk factor for anxiety during adolescence and emerging adulthood.

Latent Change and Co-Occurrence of Overactive and Underactive Behavior Problems in American Early Education

This research examined growth and co-occurrence of latent developmental patterns of classroom behavior problems among young, economically disadvantaged children. A nationally representative sample (N = 3,827) of children was assessed for overactive and underactive behavior problems through two years of prekindergarten, kindergarten and first grade. Latent growth mixture modeling revealed three distinct subpopulations of change in overactivity and two in underactivity. A parallel-process growth mixture model showed that well-adjusted patterns of overactivity and underactivity co-occurred in the majority of the sample, whereas co-occurrence of marginal overactivity and underactivity existed in approximately 7% of the sample. Results are discussed in light of theory and implications for practice. Impact Statement Young children’s classroom behavior develops in different ways and at different rates, but some patterns of development are more common than others. In looking at children’s overactive and underactive behavior, we found that most children have fewer behavior problems from preschool through 1st grade, some have relatively moderate overactive problems that show little change over time, and a small proportion have increasingly difficult problems with overactivity and with underactivity. Even though overactivity and underactivity are distinct aspects of classroom behavior, children with increasing patterns of overactive behavior problems were more likely than expected to also show more underactivity in the classroom, and these children may require more support from parents and teachers during the early school years. Supplemental data for this article can be accessed online at https://doi.org/10.1080/2372966X.2021.2000842.

Simultaneous trajectories of alcohol and cannabis use from adolescence to emerging adulthood: Associations with role transitions and functional outcomes.

Alcohol and cannabis are the most widely used substances among young people. This study identifies alcohol and cannabis (AC) trajectory classes, and examines class differences on adult role transitions and functioning during emerging adulthood. Participants (n = 3,265) completed six annual surveys (ages 16 to 22) on frequency of past month alcohol and cannabis use. Parallel process growth mixture models (PP-GMM) were used to model simultaneous trajectories of use and identify latent trajectory classes based on AC co-use. Trajectory classes were compared on age 23 outcomes. Analyses identified four trajectory classes of co-use: Initial AC Abstainer with Increasing AC Use (85%); Initial Low AC Use with Increasing Drinking (5%); Increasing Cannabis Use with Steady Drinking (3%); and Decreasing Cannabis Use with Steady Drinking (7%). The two trajectory classes with early AC co-use were less likely to have made certain role transitions, and reported poorer functioning in emerging adulthood, but in distinct ways. For example, the Decreasing Cannabis Use with Steady Drinking class was less likely to graduate from college, and reported more AC consequences and related problems, whereas the Initial Low AC Use with Increasing Drinking class was less likely to be employed full-time, and reported poorer mental health. The Increasing Cannabis Use with Steady Drinking class tended to fall between the Initial AC Abstainer class and the two early AC co-use trajectory classes on outcomes. Results indicate the ways in which AC co-use can affect functional outcomes and the timing of role transitions during the transition to young adulthood. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Correlates of dual trajectories of physical activity and sedentary time in youth: The UP & DOWN longitudinal study.

Trajectories of physical activity and sedentary time (SED) may differ between subgroups of youth. The aim of this study was to identify group-based dual trajectories of physical activity and SED and explore individual, social, and environmental correlates of these trajectories. Longitudinal data (three time points, baseline 2011-2012) of Spanish youth (n=1597, mean age=11.94±2.52, 50.9% boys) were used. Moderate-to-vigorous physical activity (MVPA) and SED were assessed objectively at each time point, and 21 potential correlates were self-reported at baseline. Parallel process growth mixture models identified shared categorical latent groups, adjusting for school and age. Multinomial logistic regression models identified baseline correlates of a given trajectory. Four shared categorical latent groups were identified: (1) stable MVPA and decreasing SED (4%); (2) stable MVPA and increasing SED (3%); (3) consistently higher MVPA (18%); and (4) stable low MVPA and slight increase in SED (75%). Multinomial logistic regression models with group 3 as reference found: negative affect (RRR=0.90, 95% CI 0.84-0.97), parental screen-time rules (RRR=1.15, 95% CI 1.00-1.33), and household media equipment (RRR=1.17, 95% CI 1.05-1.30) predicted likelihood of group 1 membership; cons of reducing SED (RRR=2.70, 95% CI 1.77-4.10) predicted likelihood of group 2 membership; and co-participation in physical activity with friends (RRR=0.80, 95% CI 0.69-0.94), fathers' modeling of TV viewing (RRR=1.22, 95% CI 1.02-1.47), and household media equipment (RRR=1.16, 95% CI 1.02-1.31) predicted likelihood of group 4 membership. Results suggest that strategies to improve MVPA and SED behaviors among youth may need to be multifaceted, targeting all levels of influence.

Multiple developmental pathways underlying conduct problems: A multitrajectory framework.

In the past decades, there has been an overemphasis of a descriptive/behavioral approach to study conduct disorder. In an equifinal perspective, we aimed to examine the developmental multitrajectory groups of psychological features (irritability, interpersonal callousness, hyperactivity/impulsivity, and depressive-anxiety symptoms) and their associations with conduct problems. In a population-based cohort (n = 1,309 participants followed from 5 months to 17 years old), latent-class growth analysis was performed for each psychological feature to identify a two-trajectory model (from ages 6 to 12 years). Based on parameter estimates of the two-trajectory models for each of the four psychological features, a parallel process growth mixture model identified eight significant developmental patterns that were subsequently compared with typically developing children. Furthermore, we observed that while interpersonal callousness conferred an increased risk for childhood and adolescence conduct problems, its co-occurrence with hyperactivity/impulsivity, irritability, and/or depressive-anxiety symptoms heightened the general risk, but also predicted distinct subtypes of conduct problems (i.e., aggressive and rule-breaking behaviors). Thus, by studying complex developmental combinations of psychological features, we observed qualitatively distinct pathways towards conduct problems. A multitrajectory framework of psychological features should be considered as a significant step towards unveiling the multiple etiological pathways leading to conduct disorder and its substantial clinical heterogeneity.

Investigating convergence of cardiac and behavioral indicators of distress during routine vaccinations over the second year of life.

There is considerable variability regarding the convergence between behavioral and biological aspects of distress responses in toddlerhood, and little research has investigated the convergence of these measures in high distress. The aim of the current study was to describe patterns of distress responses to vaccinations as indexed by both pain-related behavioral distress and heart rate (HR) at 12 and 18months. Caregiver-toddler dyads were part of an ongoing longitudinal cohort observed during 12- (N=158) and 18-month (N=122) well-baby vaccinations. Parallel-process growth mixture models discerned two distinct groups at 12months and three distinct groups at 18months. All groups had comparable pain-related behavioral distress and HR responses post-vaccination, with most participants displaying high arousal and regulation to baseline levels following the vaccination. However, at 18months, an important minority had a blunted response or did not regulate to a low level of distress by 3min post-needle. Post hoc analyses revealed that higher baseline pain-related behavioral distress predicted membership in the majority groups at 12 and 18months. These results highlight the developmental differences and variability in behavioral and cardiac indicators of distress regulation across the second year of life.

Developmental multi-trajectory of irritability, anxiety, and hyperactivity as psychological markers of heterogeneity in childhood aggression.

A growing body of evidence suggests that child aggression is likely to be driven by multiple developmental pathways. However, little is known about the complex interactions between developmental trajectories of child psychological factors (such as anxiety, irritability, and hyperactivity/impulsivity dimensions) and their associations with aggression from childhood to adolescence. Therefore, the current study aimed to identify clusters of individuals with different developmental multi-trajectory, investigate their early risk factors, and describe their longitudinal associations with physical aggression. The sample comprised 4898 children derived from the Fragile Families and Child Wellbeing Study. A parallel process growth mixture model was used to identify developmental multi-trajectory groups at 5, 9 and 15 years old. Associations between multi-trajectory group membership and physical aggression were examined with Generalized Estimating Equations models. Finally, multinomial logistic regression was performed to assess perinatal and early risk factors for multi-trajectory groups. Multi-trajectory groups differed in the magnitude of risk for exhibiting physical aggression, compared to typically developing children. The risk for physical aggression was the most prominent in children who were hyperactive/impulsive and irritable [odds ratio (OR) 6.47; 95% confidence interval (CI) 5.44-7.70] and hyperactive/impulsive, irritable, and anxious (OR 7.68; CI 6.62-8.91). Furthermore, maternal cigarette and alcohol use during pregnancy and maternal depression consistently predicted multi-trajectory groups characterized by problematic levels of at least two co-occurrent psychological symptoms. Identified combinations of developmental trajectories of psychological characteristics were associated with different magnitude in risk for exhibiting physical aggression. These results may highlight the heterogeneity of developmental trajectories associated with childhood aggression.

Developmental Trajectories of Tobacco/Nicotine and Cannabis Use and Patterns of Product Co-use in Young Adulthood.

Introduction:Concurrent co-use of tobacco/nicotine and cannabis (T/C) products is common among young people and may increase risks for negative health and psychosocial outcomes, but little is known about developmental patterns of T/C co-use. This study aimed to identify distinct trajectory classes of concurrent T/C co-use from ages 16 to 21 and compare groups on T/C co-use behaviors in young adulthood.Methods:Participants (n = 2497) reported T/C use on annual online surveys from 2015 to 2019 (ages 16-22). We used parallel process growth mixture models to model simultaneous trajectories of past-month cigarette, e-cigarette, smokeless tobacco, and cannabis use and identify latent classes of T/C trajectories. Classes were then compared on types and number of T/C products used and types of T/C co-use in young adulthood.Results:Models revealed 4 T/C classes: Low/No T/C Use, Early Concurrent T/C Co-use, Late Concurrent T/C Co-use, and Tobacco Quitters/Cannabis Maintainers. Compared to other classes, the Early Concurrent T/C Co-use group—individuals with rapid progression to concurrent T/C co-use during adolescence—were more likely to report poly-tobacco use, poly-cannabis use, same-occasion sequential T/C co-use and T/C co-administration (ie, mixing T/C) of both combustible and vaping products in young adulthood.Conclusion:Early progression to concurrent T/C co-use in adolescence is prospectively linked to poly-product use and co-use of T/C products in young adulthood. Prevention efforts targeting co-use of T/C products in adolescence may help to reduce riskier patterns of T/C use and co-use in young adulthood.

Adolescent Psychological Functioning and Membership in Latent Adolescent-Parent Communication Dual Trajectory Classes.

The primary goal of this study was to examine the relationship between adolescents' psychological functioning (as indicated by depressive symptoms) and substance use (alcohol and drug use) and membership in adolescent-parent communication trajectory subgroups in a large, diverse community sample of adolescents from the United States (n=1,057; 53% female; 51% Caucasian; Age: M=16.15, SD=.75). Adolescents completed questionnaires at three annual assessments. Fit indices from parallel process growth mixture models suggested three dual trajectory classes: (1) Average communication with both parents (Average-Both); (2) Good adolescent-mother and poor adolescent-father communication (Good-Mom/Poor-Dad); and (3) Poor adolescent-mother and good adolescent-father communication (Poor-Mom/Good-Dad). The trajectory classes differed by gender. In addition, psychological functioning and substance use were differentially related to the trajectory classes.

Latent classes of course in Alzheimer's disease and predictors: the Cache County Dementia Progression Study.

Several longitudinal studies of Alzheimer's disease (AD) report heterogeneity in progression. We sought to identify groups (classes) of progression trajectories in the population-based Cache County Dementia Progression Study (N = 328) and to identify baseline predictors of membership for each group. We used parallel-process growth mixture models to identify latent classes of trajectories on the basis of Mini-Mental State Exam (MMSE) and Clinical Dementia Rating sum of boxes scores over time. We then used bias-corrected multinomial logistic regression to model baseline predictors of latent class membership. We constructed receiver operating characteristic curves to demonstrate relative predictive utility of successive sets of predictors. We fit four latent classes; class 1 was the largest (72%) and had the slowest progression. Classes 2 (8%), 3 (11%), and 4 (8%) had more rapid worsening. In univariate analyses, longer dementia duration, presence of psychosis, and worse baseline MMSE and Clinical Dementia Rating sum of boxes were associated with membership in class 2, relative to class 1. Lower education was associated with membership in class 3. In the multivariate model, only MMSE remained a statistically significant predictor of class membership. Receiver operating characteristic areas under the curve were 0.98, 0.88, and 0.67, for classes 2, 3, and 4 relative to class 1. Heterogeneity in AD course can be usefully characterized using growth mixture models. The majority belonged to a class characterized by slower decline than is typically reported in clinical samples. Class membership could be predicted using baseline covariates. Further study may advance our prediction of AD course at the population level and in turn shed light on the pathophysiology of progression.

Trajectories of Depression and Anxiety Symptoms in Adolescent Girls: A Comparison of Parallel Trajectory Approaches

Longitudinal mixture models have become popular in the literature. However, modest attention has been paid to whether these models provide a better fit to the data than growth models. Here, we compared longitudinal mixture models to growth models in the context of changes in depression and anxiety symptoms in a community sample of girls from age 10 to 17. Model comparisons found that the preferred solution was a 5-class parallel process growth mixture model that differed in the course of depression and anxiety symptoms reflecting both ordering of symptoms and qualitative group differences. Comparisons between classes revealed substantive differences on a number of outcomes using this solution. Findings are discussed in the context of clinical assessment and implementation of growth mixture models.

Trajectories of Body Mass and Self-Concept in Black and White Girls

As a stigmatizing condition, obesity may lead to the internalization of devalued labels and threats to self-concept. Modified labeling theory suggests that the effects of stigma may outlive direct manifestations of the discredited characteristic itself. This article considers whether obesity's effects on self-concept linger when obese youth enter the normal body mass range. Using longitudinal data from the National Growth and Health Study on 2,206 black and white girls, we estimated a parallel-process growth mixture model of body mass linked to growth models of body image discrepancy and self-esteem. We found that discrepancy was higher and self-esteem lower in formerly obese girls compared to girls always in the normal range and comparable to chronically obese girls. Neither body image discrepancy nor self-esteem rebounded in white girls despite reduction in body mass, suggesting that the effects of stigma linger. Self-esteem, but not discrepancy, did rebound in black girls.

A parallel process growth mixture model of conduct problems and substance use with risky sexual behavior

Conduct problems, substance use, and risky sexual behavior have been shown to coexist among adolescents, which may lead to significant health problems. The current study was designed to examine relations among these problem behaviors in a community sample of children at high risk for conduct disorder. A latent growth model of childhood conduct problems showed a decreasing trend from grades K to 5. During adolescence, four concurrent conduct problem and substance use trajectory classes were identified (high conduct problems and high substance use, increasing conduct problems and increasing substance use, minimal conduct problems and increasing substance use, and minimal conduct problems and minimal substance use) using a parallel process growth mixture model. Across all substances (tobacco, binge drinking, and marijuana use), higher levels of childhood conduct problems during kindergarten predicted a greater probability of classification into more problematic adolescent trajectory classes relative to less problematic classes. For tobacco and binge drinking models, increases in childhood conduct problems over time also predicted a greater probability of classification into more problematic classes. For all models, individuals classified into more problematic classes showed higher proportions of early sexual intercourse, infrequent condom use, receiving money for sexual services, and ever contracting an STD. Specifically, tobacco use and binge drinking during early adolescence predicted higher levels of sexual risk taking into late adolescence. Results highlight the importance of studying the conjoint relations among conduct problems, substance use, and risky sexual behavior in a unified model.