Advancing the Study of Adolescent Substance Use Through the Use of Integrative Data Analysis

Advancing Science to Prevent Firearm Violence in Communities: A Process for Harmonizing Studies to Develop Research Infrastructure

The Community Firearm Violence Prevention Network (CFVP Network), funded by the National Institutes of Health (NIH), supports a network of research projects that develop and test interventions through collaborations with community partners to prevent firearm violence, injury, and mortality. The CFVP Network presents a unique opportunity to accelerate the science of preventing firearm injuries. The data harmonization workgroup of the CFVP Network led the process of aligning studies across the three unique inaugural network projects, with particular attention to how the CFVP Network could address current gaps in the science. The goal of the data harmonization workgroup was to align study measures, assessment timelines, and data management and archival processes across projects to enable robust cross-project analyses that accelerate the science of preventing firearm injuries. To accomplish this goal, the workgroup established the infrastructure to facilitate cross-project data collection, data sharing and archiving, and analyses. Among the three inaugural network projects, the workgroup's process resulted in harmonizing two assessment timepoints (baseline and one year post-implementation) and 60 constructs (with 31 identical standardized constructs). These harmonized products provide opportunities for novel analyses across the network projects. We expect that the harmonized study infrastructure developed through this process will catalyze future research focused on preventing firearm injury, including and extending beyond CFVP Network projects. The CFVP data harmonization workgroup's process can serve as a model for future networks that seek to build the science in a particular area.

Brief Alcohol Interventions are Effective through 6 Months: Findings from Marginalized Zero-inflated Poisson and Negative Binomial Models in a Two-step IPD Meta-analysis

To evaluate and optimize brief alcohol interventions (BAIs), it is critical to have a credible overall effect size estimate as a benchmark. Estimating such an effect size has been challenging because alcohol outcomes often represent responses from a mixture of individuals: those at high risk for alcohol misuse, occasional nondrinkers, and abstainers. Moreover, some BAIs exclusively focus on heavy drinkers, whereas others take a universal prevention approach. Depending on sample characteristics, the outcome distribution might have many zeros or very few zeros and overdispersion; consequently, the most appropriate statistical model may differ across studies. We synthesized individual participant data (IPD) from 19 studies in Project INTEGRATE (Mun et al., 2015b) that randomly allocated participants to intervention and control groups (N = 7,704 participants, 38.4% men, 74.7% White, 58.5% first-year students). We sequentially estimated marginalized zero-inflated Poisson (Long et al., 2014) or negative binomial regression models to obtain covariate-adjusted, study-specific intervention effect estimates in the first step, which were subsequently combined in a random-effects meta-analysis model in the second step. BAIs produced a statistically significant 8% advantage in the mean number of drinks at both 1-3 months (RR = 0.92, 95% CI = [0.85, 0.98]) and 6 months (RR = 0.92, 95% CI = [0.85, 0.99]) compared to controls. At 9-12 months, there was no statistically significant difference in the mean number of drinks between BAIs and controls. In conclusion, BAIs are effective at reducing the mean number of drinks through at least 6 months post intervention. IPD can play a critical role in deriving findings that could not be obtained in original individual studies or standard aggregate data meta-analyses.

Harmonizing Social, Emotional, and Behavioral Constructs in Prevention Science: Digging into the Weeds of Aligning Disparate Measures

While integrative data analysis (IDA) presents great opportunity, it also necessitates a myriad of methodological decisions related to harmonizing disparate measures collected across multiple studies. There is a lack of step-by-step methodological guidance for harmonizing disparate measures of latent constructs differently conceptualized or operationalized across studies, such as social, emotional, and behavioral constructs often utilized in prevention science. The current paper addressed this gap by providing methodological guidance and a case illustration focused on harmonizing measures of disparately conceptualized and operationalized constructs. We do so by outlining a five-phased harmonization approach paired with an illustrative example of the approach as applied to harmonization of broadband latent emotional and behavioral health constructs assessed with different measures across studies. This approach builds on and expands upon procedures currently recommended in the IDA literature with parallels to best practices in test development procedures. The illustrative example of our phased approach is drawn from an IDA study of 11 randomized controlled trials of Coping Power (Lochman & Wells, 2004), an evidence-based preventive intervention. We demonstrate the harmonization of two constructs, internalizing and externalizing problems, as harmonized across the teacher-reported scales of the Achenbach System of Empirically Based Assessment (Achenbach, 1991a) and the Behavior Assessment System for Children (Reynolds & Kamphaus, 2004). Finally, we consider the potential strengths and limitations of this phased approach, underscoring areas for future methodological research and conclude with some recommendations.

Integrating time into stigma and health research

Stigma changes over time: it waxes and wanes through history, is manifested within humans who develop over time and is tied to statuses (such as attributes, illnesses and identities) that have varying courses. Despite the inherent fluidity of stigma, theories, research and interventions typically treat associations between stigma and health as stagnant. Consequently, the literature provides little insight into when experiences of stigma are most harmful to health and when stigma interventions should be implemented. In this Perspective, we argue that integrating time into stigma research can accelerate progress towards understanding and intervening in associations between stigma and health inequities. We situate time in relation to key concepts in stigma research, identify three timescales that are relevant for understanding stigma (historical context, human development and status course), and outline a time-based research agenda to improve scientists’ ability to understand and address stigma to improve health.

Associations between stigma and health are typically treated as stagnant. In this Perspective, Earnshaw et al. argue that considering stigma in relation to historical, human development and status course timescales can advance progress in understanding and addressing stigma to improve health.

The Role of Social Position Within Peer Groups in Distress-Motivated Smoking Among Adolescents

OBJECTIVE

The relationship between smoking and adolescents' peer relationships is complex, with studies showing increased risk of smoking for adolescents of both very high and very low social position. A key question is whether the impact of social position on smoking depends on an adolescent's level of coping motives (i.e., their desire to use smoking to mitigate negative affect).

METHOD

We assessed how social position predicts nicotine dependence in a longitudinal sample (N = 3,717; 44.8% male; mean age = 13.41 years) of adolescent lifetime smokers measured between 6th and 12th grades. Using both social network analysis and multilevel modeling, we assessed this question at the between-person and within-person level, hypothesizing that within-person decreases in social position would lead to increased risk of nicotine dependence among those with high levels of coping motives.

RESULTS

In contrast to our hypotheses, only interactions with the between-person measures of social position were found, with a slight negative relationship at low levels of coping motives. In addition, the main effect of coping motives was considerably stronger than that of social position at the between-person level, and social position had no significant within-person main effect on nicotine dependence risk.

CONCLUSIONS

These results suggest that adolescents with higher overall levels of social position among their peers may have slightly decreased risk for nicotine dependence, but only when coping motives are low. Counter to expectations, higher levels of nicotine dependence risk were not linked to fluctuations in social position.

Relationship between stress and alexithymia, emotional processing and negative/positive affect in medical staff working amid the COVID-19 pandemic

The psychological burden of the COVID-19 pandemic may have a lasting effect on emotional well-being of healthcare workers. Medical personnel working at the time of the pandemic may experience elevated occupational stress due to the uncontrollability of the virus, high perceived risk of infection, poor understanding of the novel virus transmission routes and unavailability of effective antiviral agents. This study used path analysis to analyze the relationship between stress and alexithymia, emotional processing and negative/positive affect in healthcare workers. The sample included 167 nurses, 65 physicians and 53 paramedics. Sixty-two (21.75 %) respondents worked in COVID-19-designated hospitals. Respondents were administered the Toronto Alexithymia Scale-20, Cohen's Perceived Stress Scale, Emotional Processing Scale, and the Positive and Negative Affect Schedule. The model showed excellent fit indices (χ2 (2)=2.642, p=0.267; CFI=0.999, RMSEA=0.034, SRMR=0.015). Multiple group path analysis demonstrated physicians differed from nurses and paramedics at the model level (X2diff (7)=14.155, p<0.05 and X2diff (7)=18.642, p<0.01, respectively). The relationship between alexithymia and emotional processing was stronger in nurses than in physicians (difference in beta=0.27; p<0.05). Individual path χ2 tests also revealed significantly different paths across these groups. The results of the study may be used to develop evidence-based intervention programs promoting healthcare workers' mental health and well-being.

Genetic susceptibility for major depressive disorder associates with trajectories of depressive symptoms across childhood and adolescence

BACKGROUND

Early-onset depression during childhood and adolescence is associated with a worse course of illness and outcome than adult onset. However, the genetic factors that influence risk for early-onset depression remain mostly unknown. Using data collected over 13 years, we examined whether polygenic risk scores (PRS) that capture genetic risk for depression were associated with depressive symptom trajectories assessed from childhood to adolescence.

METHODS

Data came from the Avon Longitudinal Study of Parents and Children, a prospective, longitudinal birth cohort (analytic sample = 7,308 youth). We analyzed the relationship between genetic susceptibility to depression and three time-dependent measures of depressive symptoms trajectories spanning 4-16.5 years of age (class, onset, and cumulative burden). Trajectories were constructed using a growth mixture model with structured residuals. PRS were generated from the summary statistics of a genome-wide association study of depression risk using data from the Psychiatric Genomics Consortium, UK Biobank, and 23andMe, Inc. We used MAGMA to identify gene-level associations with these measures.

RESULTS

Youth were classified into six classes of depressive symptom trajectories: high/renitent (27.9% of youth), high/reversing (9.1%), childhood decrease (7.3%), late childhood peak (3.3%), adolescent spike (2.5%), and minimal symptoms (49.9%). PRS discriminated between youth in the late childhood peak, high/reversing, and high/renitent classes compared to the minimal symptoms and childhood decrease classes. No significant associations were detected at the gene level.

CONCLUSIONS

This study highlights differences in polygenic loading for depressive symptoms across childhood and adolescence, particularly among youths with high symptoms in early adolescence, regardless of age-independent patterns.

Capturing Environmental Dimensions of Adversity and Resources in the Context of Poverty Across Infancy Through Early Adolescence: A Moderated Nonlinear Factor Model

Income, education, and cumulative-risk indices likely obscure meaningful heterogeneity in the mechanisms through which poverty impacts child outcomes. This study draws from contemporary theory to specify multiple dimensions of poverty-related adversity and resources, with the aim of better capturing these nuances. Using data from the Family Life Project (N = 1,292), we leveraged moderated nonlinear factor analysis (Bauer, 2017) to establish group- and longitudinally invariant environmental measures from infancy to early adolescence. Results indicated three latent factors-material deprivation, psychosocial threat, and sociocognitive resources-were distinct from each other and from family income. Each was largely invariant across site, racial group, and development and showed convergent and discriminant relations with age-twelve criterion measures. Implications for ensuring socioculturally valid measurements of poverty are discussed.

Using Integrative Data Analysis to Investigate School Climate Across Multiple Informants

Survey research frequently involves the collection of data from multiple informants. Results, however, are usually analyzed by informant group, potentially ignoring important relationships across groups. When the same construct(s) are measured, integrative data analysis (IDA) allows pooling of data from multiple sources into one data set to examine information from multiple perspectives within the same analysis. Here, the IDA procedure is demonstrated via the examination of pooled data from student and teacher school climate surveys. This study contributes to the sparse literature regarding IDA applications in the social sciences, specifically in education. It also lays the groundwork for future educational researchers interested in the practical applications of the IDA framework to empirical data sets with complex model structures.

A Latent Variable Approach to Measuring Social Dynamics in Adolescence

In the study of adolescent health, it is useful to derive indices of social dynamics from sociometric data, and to use these indices as predictors of health risk behaviors. In this manuscript, we introduce a flexible latent variable model as a novel way of obtaining estimates of social integration and social status from school-based sociometric data. Such scores provide the flexibility of a regression-based approach while accounting for measurement error in sociometric indicators. We demonstrate the utility of these factor scores in testing complex hypotheses through a combination of structural equation modeling and survival models, showing that deviance mediates the relationship between social status and smoking onset hazard at the transition to high school.

Multivariate Higher-Order IRT Model and MCMC Algorithm for Linking Individual Participant Data From Multiple Studies

Many clinical and psychological constructs are conceptualized to have multivariate higher-order constructs that give rise to multidimensional lower-order traits. Although recent measurement models and computing algorithms can accommodate item response data with a higher-order structure, there are few measurement models and computing techniques that can be employed in the context of complex research synthesis, such as meta-analysis of individual participant data or integrative data analysis. The current study was aimed at modeling complex item responses that can arise when underlying domain-specific, lower-order traits are hierarchically related to multiple higher-order traits for individual participant data from multiple studies. We formulated a multi-group, multivariate higher-order item response theory (HO-IRT) model from a Bayesian perspective and developed a new Markov chain Monte Carlo (MCMC) algorithm to simultaneously estimate the (a) structural parameters of the first- and second-order latent traits across multiple groups and (b) item parameters of the model. Results from a simulation study support the feasibility of the MCMC algorithm. From the analysis of real data, we found that a bivariate HO-IRT model with different correlation/covariance structures for different studies fit the data best, compared to a univariate HO-IRT model or other alternate models with unreasonable assumptions (i.e., the same means and covariances across studies). Although more work is needed to further develop the method and to disseminate it, the multi-group multivariate HO-IRT model holds promise to derive a common metric for individual participant data from multiple studies in research synthesis studies for robust inference and for new discoveries.

Approaches for creating comparable measures of alcohol use symptoms: Harmonization with eight studies of criminal justice populations

BACKGROUND

With increasing data archives comprised of studies with similar measurement, optimal methods for data harmonization and measurement scoring are a pressing need. We compare three methods for harmonizing and scoring the AUDIT as administered with minimal variation across 11 samples from eight study sites within the STTR (Seek-Test-Treat-Retain) Research Harmonization Initiative. Descriptive statistics and predictive validity results for cut-scores, sum scores, and Moderated Nonlinear Factor Analysis scores (MNLFA; a psychometric harmonization method) are presented.

METHODS

Across the eight study sites, sample sizes ranged from 50 to 2405 and target populations varied based on sampling frame, location, and inclusion/exclusion criteria. The pooled sample included 4667 participants (82% male, 52% Black, 24% White, 13% Hispanic, and 8% Asian/ Pacific Islander; mean age of 38.9 years). Participants completed the AUDIT at baseline in all studies.

RESULTS

After logical harmonization of items, we scored the AUDIT using three methods: published cut-scores, sum scores, and MNLFA. We found greater variation, fewer floor effects, and the ability to directly address missing data in MNLFA scores as compared to cut-scores and sum scores. MNLFA scores showed stronger associations with binge drinking and clearer study differences than did other scores.

CONCLUSIONS

MNLFA scores are a promising tool for data harmonization and scoring in pooled data analysis. Model complexity with large multi-study applications, however, may require new statistical advances to fully realize the benefits of this approach.

An Algorithm for Creating Virtual Controls Using Integrated and Harmonized Longitudinal Data

We introduce a strategy for creating virtual control groups-cases generated through computer algorithms that, when aggregated, may serve as experimental comparators where live controls are difficult to recruit, such as when programs are widely disseminated and randomization is not feasible. We integrated and harmonized data from eight archived longitudinal adolescent-focused data sets spanning the decades from 1980 to 2010. Collectively, these studies examined numerous psychosocial variables and assessed past 30-day alcohol, cigarette, and marijuana use. Additional treatment and control group data from two archived randomized control trials were used to test the virtual control algorithm. Both randomized controlled trials (RCTs) assessed intentions, normative beliefs, and values as well as past 30-day alcohol, cigarette, and marijuana use. We developed an algorithm that used percentile scores from the integrated data set to create age- and gender-specific latent psychosocial scores. The algorithm matched treatment case observed psychosocial scores at pretest to create a virtual control case that figuratively "matured" based on age-related changes, holding the virtual case's percentile constant. Virtual controls matched treatment case occurrence, eliminating differential attrition as a threat to validity. Virtual case substance use was estimated from the virtual case's latent psychosocial score using logistic regression coefficients derived from analyzing the treatment group. Averaging across virtual cases created group estimates of prevalence. Two criteria were established to evaluate the adequacy of virtual control cases: (1) virtual control group pretest drug prevalence rates should match those of the treatment group and (2) virtual control group patterns of drug prevalence over time should match live controls. The algorithm successfully matched pretest prevalence for both RCTs. Increases in prevalence were observed, although there were discrepancies between live and virtual control outcomes. This study provides an initial framework for creating virtual controls using a step-by-step procedure that can now be revised and validated using other prevention trial data.