Handling Missing Covariates in Conditional Mixture Models Under Missing at Random Assumptions

Pervasive versus situational childhood ADHD: latent classes and their clinical characteristics, based on parent and teacher ratings in a large longitudinal population sample

Attention-deficit/hyperactivity disorder (ADHD) diagnoses require symptoms to be present in two settings. Low levels of concordance between symptoms rated at home and school raise questions regarding this approach. The aim was to examine whether there are sub-groups with context-specific expressions of ADHD symptoms (i.e., at home or school only) with clinically significant problems sufficient to support a new diagnostic formulation. We applied latent class transition analysis to parent and teacher data (N = 10,476) from the Avon Longitudinal Study of Parents and Children (ALSPAC), collected at ages 8, 10, and 20 years. We examined the short-term stability of emergent classes and their childhood and adult-associated risk profiles. In addition to an Unaffected class (~ 45%), there was a Pervasive Combined class with elevated inattentive and hyperactive/impulsive symptoms at both home and school (~ 11%) and three classes with situational expressions; School Combined (~ 9%), Home Combined (~ 18%), and School Inattentive (~ 16%). Stability ranged from 0.27 to 0.78. The Pervasive Combined class was most symptomatic and impaired. School inattentive also displayed clinical symptom levels, whereas the School and Home Combined classes displayed subclinical levels. Different profiles regarding sex, cognition, conduct problems, and substance use emerged for the three situational classes. Distinct groupings of pervasive and situational ADHD expressions are identifiable in the general population. The isolation of a stable and burdensome Pervasive Combined class lends support to the current diagnostic approach. However, there are indications of situational expressions of ADHD with clinical symptom levels and associated difficulties.

Severe Pandemic Learning Loss and the Promise of Remotely Delivered Intervention in Students With Comorbid Reading and Mathematics Learning Difficulty

Analyses were conducted with second graders, drawn from an ongoing multi-cohort randomized controlled trial (RCT), who had been identified for RCT entry based on comorbid reading comprehension and word-problem solving difficulty. To estimate pandemic learning loss, we contrasted fall performance for 3 cohorts: fall of 2019 (pre-pandemic; n = 47), 2020 (early pandemic, when performance was affected by the truncated preceding school year; n = 35), and 2021 (later pandemic, when performance was affected by the truncated 2019 to 2020 school year plus the subsequent year's ongoing interruptions; n = 75). Across the 2 years, declines (standard deviations below expected growth) were approximately 3 times larger than those reported for the general population and for students in high-poverty schools. To estimate the promise of structured remote intervention for addressing such learning loss during extended school closures, we contrasted effects in the RCT's 2018 to 2019 cohort (entirely in-person intervention delivery; n = 66) against the same intervention's effects in the 2020 to 2021 cohort (alternating periods of remote and in-person delivery; n = 29). Large intervention effects were not moderated by pandemic status, suggesting potential for structured remote intervention to address student needs during extended school closures.

Intimate partner violence perpetration and sexual risk behaviour: Identifying shared determinants among young men in Tanzania

In this study we examined factors that predict involvement in different patterns of sexual risk behavior and IPV perpetration among young men in Tanzania (n = 979), with a focus on identifying factors that distinguish men who engage in both behaviours from those who do not. Risk factors were drawn from three domains thought to be upstream drivers of both IPV and sexual risk: poverty, adverse childhood experiences, and inequitable gender norms. A three-step latent class analysis was used to assess whether and how factors from each domain distinguished subgroups of men whose behaviour patterns were characterised as comorbid (involvement in IPV and sexual risk behaviour), IPV-only, sexual risk only, and normative (low risk). Consistent with expectations, greater food insecurity, adverse childhood experiences, and inequitable gender norms related to violence and sexual behaviour predicted increased risk of membership in the comorbid group compared to other sub-groups. Findings support the promise of integrated prevention programmes targeting the common causes of IPV perpetration and sexual risk behaviour.

Explorations of Individual Change Processes and Their Determinants: A Novel Approach and Remaining Challenges

Over the past 40 years there have been great advances in the analysis of individual change and the analyses of between-person differences in change. While conditional growth models are the dominant approach, exploratory models, such as growth mixture models and structural equation modeling trees, allow for greater flexibility in the modeling of between-person differences in change. We continue to push for greater flexibility in the modeling of individual change and its determinants by combining growth mixture modeling with structural equation modeling trees to evaluate how measured covariates predict class membership using a recursive partitioning algorithm. This approach, referred to as growth mixture modeling with membership trees, is illustrated with longitudinal reading data from the Early Childhood Longitudinal Study with the MplusTrees package in R.

Codevelopment of Delinquency, Alcohol Use, and Aggression Toward Peers and Dates: Multitrajectory Patterns and Predictors

We aimed to characterize developmental patterns of involvement in alcohol use, delinquency, and interpersonal aggression in a normative sample of adolescents by applying multitrajectory group-based modeling. Using seven waves of data from a cohort sequential study spanning the 6th to 12th grades (n = 2,825; 50% girls), we identified four distinct trajectory groups: low risk (33%), declining peer aggressors (44%), peer and dating aggressors (13%), and multidomain high risk (10%). Across all comparisons, girls were more likely than boys to be members of the peer and dating aggressor group and less likely to be members of the multidomain high-risk group. Moreover, individual (self-control, negative emotionality), family (family violence, parental monitoring), and peer (substance use norms) distinguished class membership.

Time of day effects on the relationship between daily sleep and anxiety: An ecological momentary assessment approach

Previous research has linked sleep disturbance to anxiety. However, evidence for this relation has been inconsistent, largely limited to retrospective reports that do not account for daily variability, and silent on when the association is most pronounced. Thus, the present study utilized ecological momentary assessment (EMA) to examine the effects of daily deviations in total sleep time (TST) and person-average TST on anxiety and whether these effects varied as a function of time of day in a sample of unselected adults (N = 138). Results indicate that the amount of TST on a given night, relative to personal average TST, negatively predicted anxiety, and this relation was significant in the morning and afternoon, but not evening. In contrast, person-average TST was unrelated to average anxiety. Relations between TST and anxiety did not differ across objective (e.g., actigraphy) and subjective (e.g., sleep diary) measures. Furthermore, the pattern of results remained the same when controlling for previous day's anxiety and were not bidirectional. These findings suggest that getting less sleep than is typical for the individual predicts subsequent anxiety, and this effect is particularly strong in the morning. Average sleep duration may be less important to the experience of anxiety than deviations from that average. These findings highlight the importance of EMA to examine how and when variability in sleep confers vulnerability for anxiety symptoms.

Consequences of Involvement in Distinct Patterns of Adolescent Peer and Dating Violence

Typological theoretical perspectives suggest that the consequences of involvement in peer and dating violence may depend on the particular pattern of violent behaviors that youth experience and/or engage in. Yet few studies have examined whether distinct patterns of dating and peer violence involvement differentially predict developmental outcomes. Using two waves of data, the current study examined the prospective associations between distinct patterns of peer and dating aggression and victimization, identified using latent class analysis, and a range of potential developmental outcomes in a general population sample of adolescents in the 8th to 10th grades (n = 3068; 46% female, 58% White, 31% Black, 11% other race/ethnicity). The findings suggest that, compared to youth involved in other patterns of violence, youth involved in peer and dating violence as aggressors and victims are at greatest risk for negative sequelae, although results differed considerably for girls and boys and on the outcome variable and comparison groups being examined.

Patterns of adolescent aggression and victimization: Sex differences and correlates

This study identified heterogenous patterns of peer and dating aggression and victimization among boys and girls and examined their relation to risk and protective correlates. Girls (n=1648) and boys (n=1420) in grades 8-10 completed surveys assessing 14 indicators of violence involvement. Latent class analyses indicated a four-class solution, though a test of measurement invariance indicated the nature of the classes differed by sex. Among boys and girls, three classes emerged: Uninvolved (45% of girls, 61% of boys), Peer Aggressor-Victims (23% of girls, 21% of boys), and Cross-Context Aggressor-Victims (12% of girls, 5% of boys). Those in the Peer Aggressor-Victims class were likely to report involvement in peer aggression only; however, girls in this class were likely to be involved only in moderate violence, whereas boys were likely to be involved in moderate and severe violence. Those in the Cross-Context Aggressor-Victims class were likely to report involvement in all forms of violence except sexual and controlling aggression, which was likely only among boys. Among girls, but not boys, a Verbal Dating Aggressor-Victims class (21% of girls) emerged that was characterized by involvement in occasional verbal dating aggression only. Among boys, but not girls, a Cross-Context Physical Victims class (13% of boys) emerged that was characterized by being only a victim of moderate physical peer and dating violence. Unique and shared risk and protective factors distinguished class membership for girls and boys. Findings suggest the pathways leading to violence may differ by sex and result in different patterns of violence involvement.

Body mass index, body dissatisfaction and adolescent smoking initiation

BACKGROUND

Smoking influences body weight, but there is little evidence as to whether body mass index (BMI) and body dissatisfaction increase smoking initiation in adolescents.

METHODS

We evaluated the association between measured BMI, body dissatisfaction and latent classes of smoking initiation (never smokers, experimenters, late onset regular smokers, early onset regular smokers) in the Avon Longitudinal Study of Parents and Children. In observational analyses we used BMI (N=3754) and body dissatisfaction at age 10.5 years (N=3349). In Mendelian randomisation (MR) analysis, we used a BMI genetic risk score of 76 single nucleotide polymorphisms (N=4017).

RESULTS

In females, higher BMI was associated with increased odds of early onset regular smoking (OR: 1.11, 95% CI: 1.04, 1.18) compared to being a never smoker, but not clearly associated with experimenting with smoking (OR: 1.04, 95% CI: 0.99, 1.10) or late onset regular smoking (OR: 1.01, 95% CI: 0.94, 1.09). No clear evidence was found for associations between BMI and smoking initiation classes in males (p-value for sex interaction≤0.001). Body dissatisfaction was associated with increased odds of late-onset regular smoking (OR: 1.71, 95% CI: 1.32, 1.99) in males and females combined (P-value for sex interaction=0.32). There was no clear evidence for an association between the BMI genetic risk score and smoking latent classes in males or females but estimates were imprecise.

CONCLUSIONS

BMI in females and body dissatisfaction in males and females are associated with increased odds of smoking initiation, highlighting these as potentially important factors for consideration in smoking prevention strategies.

Patterns of cannabis use during adolescence and their association with harmful substance use behaviour: findings from a UK birth cohort

BACKGROUND

Evidence on the role of cannabis as a gateway drug is inconsistent. We characterise patterns of cannabis use among UK teenagers aged 13-18 years, and assess their influence on problematic substance use at age 21 years.

METHODS

We used longitudinal latent class analysis to derive trajectories of cannabis use from self-report measures in a UK birth cohort. We investigated (1) factors associated with latent class membership and (2) whether latent class membership predicted subsequent nicotine dependence, harmful alcohol use and recent use of other illicit drugs at age 21 years.

RESULTS

5315 adolescents had three or more measures of cannabis use from age 13 to 18 years. Cannabis use patterns were captured as four latent classes corresponding to 'non-users' (80.1%), 'late-onset occasional' (14.2%), 'early-onset occasional' (2.3%) and 'regular' users (3.4%). Sex, mother's substance use, and child's tobacco use, alcohol consumption and conduct problems were strongly associated with cannabis use. At age 21 years, compared with the non-user class, late-onset occasional, early-onset occasional and regular cannabis user classes had higher odds of nicotine dependence (OR=3.5, 95% CI 0.7 to 17.9; OR=12.1, 95% CI 1.0 to 150.3; and OR=37.2, 95% CI 9.5 to 144.8, respectively); harmful alcohol consumption (OR=2.6, 95% CI 1.5 to 4.3; OR=5.0, 95% CI 2.1 to 12.1; and OR=2.6, 95% CI 1.0 to 7.1, respectively); and other illicit drug use (OR=22.7, 95% CI 11.3 to 45.7; OR=15.9, 95% CI 3.9 to 64.4; and OR=47.9, 95% CI 47.9 to 337.0, respectively).

CONCLUSIONS

One-fifth of the adolescents in our sample followed a pattern of occasional or regular cannabis use, and these young people were more likely to progress to harmful substance use behaviours in early adulthood.

A Comparison of Methods for Uncovering Sample Heterogeneity: Structural Equation Model Trees and Finite Mixture Models

Although finite mixture models have received considerable attention, particularly in the social and behavioral sciences, an alternative method for creating homogeneous groups, structural equation model trees (Brandmaier, von Oertzen, McArdle, & Lindenberger, 2013), is a recent development that has received much less application and consideration. It is our aim to compare and contrast these methods for uncovering sample heterogeneity. We illustrate the use of these methods with longitudinal reading achievement data collected as part of the Early Childhood Longitudinal Study-Kindergarten Cohort. We present the use of structural equation model trees as an alternative framework that does not assume the classes are latent and uses observed covariates to derive their structure. We consider these methods as complementary and discuss their respective strengths and limitations for creating homogeneous groups.