Bethany C. Bray

Latent Class Analysis With Distal Outcomes: A Flexible Model-Based Approach

Although prediction of class membership from observed variables in latent class analysis is well understood, predicting an observed distal outcome from latent class membership is more complicated. A flexible model-based approach is proposed to empirically derive and summarize the class-dependent density functions of distal outcomes with categorical, continuous, or count distributions. A Monte Carlo simulation study is conducted to compare the performance of the new technique to 2 commonly used classify-analyze techniques: maximum-probability assignment and multiple pseudoclass draws. Simulation results show that the model-based approach produces substantially less biased estimates of the effect compared to either classify-analyze technique, particularly when the association between the latent class variable and the distal outcome is strong. In addition, we show that only the model-based approach is consistent. The approach is demonstrated empirically: latent classes of adolescent depression are used to predict smoking, grades, and delinquency. SAS syntax for implementing this approach using PROC LCA and a corresponding macro are provided.

College students on the autism spectrum: Prevalence and associated problems

As more young people are identified with autism spectrum diagnoses without co-occurring intellectual disability (i.e. high-functioning autism spectrum disorder; HFASD), it is imperative that we begin to study the needs of this population. We sought to gain a preliminary estimate of the scope of the problem and to examine psychiatric risks associated HFASD symptoms in university students. In a large sample (n = 667), we examined prevalence of ASD in students at a single university both diagnostically and dimensionally, and surveyed students on other behavioral and psychiatric problems. Dependent upon the ascertainment method, between .7 per cent and 1.9 per cent of college students could meet criteria for HFASD. Of special interest, none of the students who were found to meet diagnostic criteria (n = 5) formally for HFASD in this study had been previously diagnosed. From a dimensional perspective, those students scoring above the clinical threshold for symptoms of autism (n = 13) self-reported more problems with social anxiety than a matched comparison group of students with lower autism severity scores. In addition, symptoms of HFASD were significantly correlated with symptoms of social anxiety, as well as depression and aggression. Findings demonstrate the importance of screening for autism-related impairment among university students.

Transitions into and out of light and intermittent smoking during emerging adulthood

INTRODUCTION The purpose of this study was to examine transitions in smoking from adolescence into emerging adulthood and to identify factors that might influence these transitions, specifically, movement into and out of light and intermittent smoking. METHODS This study used Markov models to examine movement across three stages of smoking (nonsmoking, light and intermittent smoking, and heavy smoking) from adolescence into emerging adulthood. Biannual data were collected from 990 young men and women from the 12th grade until 2 years after high school. RESULTS At each timepoint, most youth were nonsmokers. Those who were heavy smokers in 12th grade had a 79% chance of also being heavy smokers 2 years after high school. Between 17% and 21% of participants were light and intermittent smokers at each timepoint, and the likelihood of remaining so at the next timepoint ranged from 56% to 72%. Less than one-half of the 12th-grade light and intermittent smokers were light and intermittent smokers 2 years later, and 3% of the sample were light and intermittent smokers across all assessments. Prevalence and transition rates did not differ by gender. College attendees reported less smoking than nonattendees before and after their transition to college, and attendees compared with nonattendees who smoked were less likely to transition from light and intermittent to heavy smoking and remain heavy smokers. Binge drinking was significantly related to 12th-grade smoking stage and to transitions from nonsmoking to smoking. Overall, few emerging adults maintained light and intermittent smoking consistently over time. DISCUSSION Light and intermittent smoking during emerging adulthood may not be the same phenomenon as light and intermittent smoking in adulthood.

Eliminating Bias in Classify-Analyze Approaches for Latent Class Analysis

Despite recent methodological advances in latent class analysis (LCA) and a rapid increase in its application in behavioral research, complex research questions that include latent class variables often must be addressed by classifying individuals into latent classes and treating class membership as known in a subsequent analysis. Traditional approaches to classifying individuals based on posterior probabilities are known to produce attenuated estimates in the analytic model. We propose the use of a more inclusive LCA to generate posterior probabilities; this LCA includes additional variables present in the analytic model. A motivating empirical demonstration is presented, followed by a simulation study to assess the performance of the proposed strategy. Results show that with sufficient measurement quality or sample size, the proposed strategy reduces or eliminates bias.

Examining shared and unique aspects of Social Anxiety Disorder and Autism Spectrum Disorder using factor analysis.

Social Anxiety Disorder (SAD) and Autism Spectrum Disorders (ASD) are fairly common psychiatric conditions that impair the functioning of otherwise healthy young adults. Given that the two conditions frequently co-occur, measurement of the characteristics unique to each condition is critical. This study evaluated the structure and construct overlap of two screening measures of SAD and ASD. Results from 623 young adults indicated that separable, though highly correlated, factors can be derived from the two measures related to social anxiety and social difficulties. The ASD screening measure also taps unique factors related to restricted interests and attention to details, theory of mind deficits, and a preference for routine. Recommendations are provided for accurate screening of symptoms of both SAD and ASD.

Modeling Relations among Discrete Developmental Processes: A General Approach to Associative Latent Transition Analysis

To understand one developmental process, it is often helpful to investigate its relations with other developmental processes. Statistical methods that model development in multiple processes simultaneously over time include latent growth curve models with time-varying covariates, multivariate latent growth curve models, and dual trajectory models. These models are designed for growth represented by continuous, unidimensional trajectories. The purpose of this article is to present a flexible approach to modeling relations in development among two or more discrete, multidimensional latent variables based on the general framework of loglinear modeling with latent variables called associative latent transition analysis (ALTA). Focus is given to the substantive interpretation of different associative latent transition models, and exactly what hypotheses are expressed in each model. An empirical demonstration of ALTA is presented to examine the association between the development of alcohol use and sexual risk behavior during adolescence.

An Exploration of Cognitive Subgroups in Alzheimer's Disease

Heterogeneity is observed in the patterns of cognition in Alzheimers disease (AD). Such heterogeneity might suggest the involvement of different etiological pathways or different host responses to pathology. A total of 627 subjects with mild/moderate AD underwent cognitive assessment with the Mini-Mental State Examination (MMSE) and the Dementia Rating Scale-2 (DRS-2). Latent class analysis (LCA) was performed on cognition subscale data to identify and characterize cognitive subgroups. Clinical, demographic, and genetic factors were explored for association with class membership. LCA suggested the existence of four subgroups; one group with mild and another with severe global impairment across the cognitive domains, one group with primary impairments in attention and construction, and another group with primary deficits in memory and orientation. Education, disease duration, age, Apolipoprotein E-epsilon4 (APOE epsilon4) status, gender, presence of grasp reflex, white matter changes, and early or prominent visuospatial impairment were all associated with class membership. Our results support the existence of heterogeneity in patterns of cognitive impairment in AD. Our observation of classes characterized by predominant deficits in attention/construction and memory respectively deserves further exploration as does the association between membership in the attention/construction class and APOE epsilon4 negative status.

Assessing the Total Effect of Time-varying Predictors in Prevention Research

Observational data are often used to address prevention questions such as, “If alcohol initiation could be delayed, would that in turn cause a delay in marijuana initiation?” This question is concerned with the total causal effect of the timing of alcohol initiation on the timing of marijuana initiation. Unfortunately, when observational data are used to address a question such as the above, alternative explanations for the observed relationship between the predictor, here timing of alcohol initiation, and the response abound. These alternative explanations are due to the presence of confounders. Adjusting for confounders when using observational data is a particularly challenging problem when the predictor and confounders are time-varying. When time-varying confounders are present, the standard method of adjusting for confounders may fail to reduce bias and indeed can increase bias. In this paper, an intuitive and accessible graphical approach is used to illustrate how the standard method of controlling for confounders may result in biased total causal effect estimates. The graphical approach also provides an intuitive justification for an alternate method proposed by James Robins [Robins, J. M. (1998). 1997 Proceedings of the American Statistical Association, section on Bayesian statistical science (pp. 1–10). Retrieved from http://www.biostat.harvard.edu/robins/research.html; Robins, J. M., Hernán, M., & Brumback, B. (2000). Epidemiology, 11(5), 550–560]. The above two methods are illustrated by addressing the motivating question. Implications for prevention researchers who wish to estimate total causal effects using longitudinal observational data are discussed.

Failure to sustain prepulse inhibition in adolescent marijuana users

BACKGROUND Marijuana use is typically initiated during adolescence, which is a critical period for neural development. Studies have reported reductions in prepulse inhibition (PPI) among adults who use marijuana chronically, although no human studies have been conducted during the critical adolescent period. METHODS This study tested PPI of acoustic startle among adolescents who were either frequent marijuana users or naïve to the drug (Controls). Adolescents were tested using two intensities of prepulses (70 and 85 dB) combined with a 105 dB startle stimulus, delivered across two testing blocks. RESULTS There was a significant interaction of group by block for PPI; marijuana users experienced a greater decline in the PPI across the testing session than Controls. The change in PPI of response magnitude for users was predicted by change in urine THC/creatinine after at least 18 h of abstinence, the number of joints used during the previous week before testing, as well as self-reported DSM-IV symptoms of marijuana tolerance, and time spent using marijuana rather than participating in other activities. CONCLUSIONS These outcomes suggest that adolescents who are frequent marijuana users have problems maintaining prepulse inhibition, possibly due to lower quality of information processing or sustained attention, both of may contribute to continued marijuana use as well as attrition from marijuana treatment.

Behavioral Impulsivity and Risk‐Taking Trajectories Across Early Adolescence in Youths With and Without Family Histories of Alcohol and Other Drug Use Disorders

BACKGROUND Youths with family histories of alcohol and other drug use disorders (FH+) are at increased susceptibility for developing substance use disorders relative to those without such histories (FH-). This vulnerability may be related to impaired adolescent development of impulse control and elevated risk-taking. However, no previous studies have prospectively examined impulse control and risk-taking in FH+ youth across adolescence. METHODS A total of 386 pre-adolescents (305 FH+, 81 FH-; aged 10 to 12) with no histories of regular alcohol or other drug use were compared on behavioral measures of impulsivity including delay discounting, response initiation (Immediate Memory Task), response inhibition impulsivity (GoStop Impulsivity Paradigm), and risk-taking (Balloon Analogue Risk Task-Youth). Youths completed these laboratory tasks every 6 months, allowing for the examination of 10- to 15-year-olds. Hierarchical linear modeling was used to characterize the development of impulse control and risk-taking as shown in performance of these tasks throughout adolescence. RESULTS We found that (i) FH+ youths had increased levels of delay discounting and response inhibition impulsivity at study entry; (ii) regardless of FH status, all youths had relatively stable delay discounting across time, improvements in response inhibition and response initiation impulsivity, and increased risk-taking; and (iii) although FH+ youths had increased response inhibition impulsivity at pre-adolescence, these differences were negligible by mid-adolescence. CONCLUSIONS Heightened delay discounting in FH+ pre-adolescents coupled with normal adolescent increases in risk-taking may contribute to their increased susceptibility toward problem substance use in adolescence.