Significant Predictors Of Class Membership Research Articles

Aging together: Dyadic profiles of older couples’ marital quality, psychological well-being, and physical health

The current study examined couple-level profiles to reveal systematic patterns of health and well-being in older couples. Using latent profile analysis, dyadic profiles were constructed in a national sample of 535 older, different-sex couples based on couples’ marital quality, psychological well-being, and physical health. Results revealed 4 distinct profiles: Maritally Satisfied and Healthy (63.5%), Maritally Dissatisfied and Moderately Healthy (14.7%), Maritally Satisfied with Unhealthy Wives (12.3%), and Maritally Satisfied with Unhealthy Husbands (9.3%). Husband age and education, wife education, and marriage length were significant predictors of class membership. Within-class partner concordance analyses confirmed that the latter two profiles featured the largest discordance between partners’ health and well-being. Most couples in this sample of older adults belonged to a class who demonstrated successful aging in all domains. A smaller group maintained high marital quality despite physical and psychological difficulties, and other couples were moderately healthy but maritally dissatisfied.

Men's preconception health and the social determinants of health: What are we missing?

BackgroundLife course perspectives suggest that optimizing men's health before conception is requisite to equitably improve population health, an area of increasing public health focus. Although scholarship on the social determinants of health (SDOH) suggests that men's health and health behaviors do not occur in a vacuum, preconception health studies have not explicitly examined how these factors influence men's preconception health.ObjectiveTo identify latent classes of men's preconception health and the role of the SDOHs in predicting class membership.MethodsPooled data from the 2011–2019 male file of the National Survey of Family Growth were analyzed (n = 10,223). Latent class analysis (LCA) was used to identify distinct classes of men's preconception health. Eight manifest variables were used to fit latent class models. A classify-analyze approach was subsequently used to create a preconception health phenotype (PhP) outcome variable. SDOHs (exposure variable) were assessed in four domains (rural/urban residence, health access, socioeconomic status, and minority/immigrant status) to predict class membership. Survey-weighted multinomial regression models were fitted to examine the association between the exposure and the outcome.ResultsThree unique PhPs were identified (lowest risk (69%), substance users (22.9%), and sexual risk-takers (8.1%) from the LCA model. Health access, socioeconomic status, and minority/immigrant status were significant predictors of class membership but not rural/urban residence. Sexual risk takers were more likely to be uninsured (aOR: 1.25, 95% CI 1.02, 1.52), college-educated (aOR: 1.94 95% CI: 1.34, 2.79), and non-Hispanic Black (aOR: 1.99 95% CI: 1.55, 2.54) while substance users were more likely to have unstable employment (aOR: 1.23 95% CI:1.04, 1.45) and have a high school degree or higher (aOR 1.48 95% CI: 1.15, 1.90) than men in the lowest risk category.ConclusionSocial determinants may impact men's preconception health in ways that are not conventionally understood. These findings raise important questions about how preconception health interventions should be created, tailored, and/or retooled. Specifically, studies that examine the sociocultural and political contexts underpinning the relationship between social class, masculinity, and men's preconception health are needed to provide nuanced insights on factors that shape these outcomes.

Exploration of Test Anxiety Profiles in U.S. Undergraduate Students

The present study was conducted using latent profile analysis to determine whether homogeneous test anxiety groups could be identified among 625 undergraduates and to determine whether gender, self-critical perfectionism, and rigid perfectionism were significant predictors of class membership in the identified test anxiety groups. The students completed two multidimensional instruments, measures of test anxiety and perfectionism, online. Results of the latent profile analysis identified a test anxiety profile with three homogeneous groups as the best model. The three homogeneous groups were labeled low, medium, and high. Moreover, gender, self-critical perfectionism, and rigid perfectionism were found to predict class membership differentially in the three homogeneous groups. Implications of the results are discussed in relation to prevention and intervention in the fields of perfectionism and test anxiety.

Injection and Sexual Behavior Profiles among People Who Inject Drugs in Miami, Florida

Background The dynamics of injection drug use and higher-risk sexual practices compound the risk of HIV and HCV acquisition. Published literature on people who inject drugs (PWID) has examined risk of infection assuming homogeneity of cohort behavior. Categorizing subgroups by injection and sexual risk can inform a more equitable approach to how syringe services programs (SSPs) adapt harm reduction resources and implementation of evidence-based interventions. We explored injection and sexual risk profiles among PWID to determine significant predictors of class membership. Methods Data were collected from 1,272 participants at an SSP in Miami-Dade County. Latent Class Analysis (LCA) examined how 10 injection/sexual behavior indicators cluster together to create profiles. Model fit statistics and multivariable multinomial latent class regression identified the optimal class structure and significant predictors of class membership. We assessed SSP visits, naloxone access, HIV/HCV testing and prevalence, and incidence of self-reported wounds. Results Three distinct profiles of injection/sexual risk were determined: Low Injection/High Sexual (LIHS) (9.4%); High Injection/Moderate Sexual (HIMS) (18.9%); and Low Injection/Low Sexual (LILS) (71.7%). Participants reporting gay/bisexual orientation and methamphetamine injection more likely belonged to the LIHS class. LIHS class members had higher prevalence of HIV, while those of HIMS reported increased hepatitis C prevalence. Compared to members of LILS, those of HIMS more likely experienced unstable housing, gay/bisexual orientation, heroin or speedball injection, and identifying as women. HIMS cohort members had more SSP visits, naloxone accessed, and higher wound incidence than those of LILS. Conclusions Understanding PWID subgroups amplifies the importance of implementing evidencebased interventions such as PrEP for those engaging in highest risk behavior, with focused interventions of antiretroviral management and access to condoms for members of the LIHS class and HCV screening with wound care for those belonging to HIMS.

Sexual Behavior Patterns of Black Young Adults in Georgia: Results from a Latent Class Analysis.

The state of Georgia ranks first in the rate of new diagnosis of HIV/AIDS, yet studies focused on understanding the drivers of this epidemic in the most at-risk groups outside the Atlanta Metropolitan Statistical Area (MSA) remain sparse. Among the many consequences of this is the inability to effectively design tailored intervention programs that appropriately address the reduction of HIV/AIDS and other STDS in the state. Reducing disparities in HIV/AIDS remains an important goal in the National HIV/AIDS Strategy. Given the disproportionate burden of the HIV/AIDS epidemic among Blacks in the state of Georgia, there is a significant need to focus research attention on the sexual behavior of young people that predisposes them to increased risk of HIV/AIDS infection. The current study used latent class analysis (LCA) to examine the sexual behavior patterns of Black young adults living in Georgia. LCA results revealed a two-class fit with distinct sexual behavior patterns. Gender and educational attainment emerged as significant predictors of class membership. Our findings contribute to the vast literature showing that sexual behavior patterns of Black young adults are not homogenous, and underscore the significance of socio-demographic and contextual factors in shaping sexual risk behavior.

Assessing neuropsychological phenotypes of pediatric brain tumor survivors.

Pediatric brain tumor survivors (PBTS) are at risk for both neurocognitive impairments and psychological difficulties, yet these two domains have historically been discretely examined, with assessment of psychosocial outcomes rarely included in studies of cognitive outcomes. Taking a person-centered approach, the current study aimed to more comprehensively evaluate PBTS late effect profiles, including both neurocognitive and psychological sequelae, and predictors of these profiles. PBTS (N=89) were assessed in a pediatric neuropsychological clinic between May 2009 and May 2018, diagnosed at least 1year prior, and off-treatment for at least 3months (Mage =6.57years, SD=4.53; 46.1% female). Parent- and teacher-report of psychological symptoms, and performance-based measures of neurocognitive functioning were examined using latent profile analysis. The R3STEP procedure identified predictors of class membership. The optimal model identified four classes characterized by: (1) average functioning across all measures ("Average," n=47), (2) average psychosocial functioning and impaired neurocognitive functioning ("Cognitive Deficit," n=25), (3) elevated social problems and significant neurocognitive impairments ("Social/Cognitive Deficit," n=9), and (4) impaired visual planning and problem-solving and elevated parent-reported psychosocial problems, but average processing speed, working memory, and teacher-reported psychosocial outcomes ("Discrepant," n=8). Ethnicity, race, radiation treatment, and diagnoses of neurofibromatosis 1, hydrocephalus, and cerebellar mutism syndrome were significant predictors of class membership (ps<0.05). The present study identified distinct phenotypes with unique patterns of relations among neurocognitive and psychological domains. These findings are a vital first step toward identifying those at highest risk for poor outcomes and informing interventions that effectively address interrelated treatment targets for specific groups.

Longitudinal transitions in patterns of post‐traumatic stress disorder symptoms and psychological distress among South Korean adults with traumatic experiences

This study aimed to investigate transitions in patterns of post-traumatic stress disorder (PTSD) symptoms and psychological distress among South Koreans with traumatic experiences. This study consists of two parts: In Part 1 the time since trauma ranged between 0 and 41 years, while in Part 2 participants experienced at least one traumatic event within the previous three years. We identified subgroups and transitions between classes over a one-year period using latent profile analysis and latent transition analysis. First, in Part 1, four classes were identified at T1 and T2. Second, we explored whether age, sex, time since trauma, and time perspective (TP) predicted class membership at T1. Age, past-negative (PN), present-fatalistic, and future TPs were significant predictors of class membership at T1. Finally, sex, PN, and future TPs were identified as significant predictors of class transition over time. The same analysis was conducted in Part 2 with individuals chosen from the Part 1 participants. Differential effects of TP on class membership and transitions in PTSD symptoms and psychological distress indicated the need for intervention programs that consider the TP profiles of individuals with traumatic experiences.

A Latent Class IRT Approach to Defining and Measuring Language Proficiency

English language learner (EL) status has high stakes implications for determining when and how ELs should be evaluated for academic achievement. In the US, students designated as English learners are assessed annually for English language proficiency (ELP), a complex construct whose conceptualization has evolved in recent years to reflect more precisely the language demands of content area achievement as reflected in the standards of individual states and state language assessment consortia, such as WIDA and ELPA21. The goal of this paper was to examine the possible role for and utility of using content area assessments to validate language proficiency mastery criteria. Specifically, we applied mixture item response models to identify two classes of EL students: (1) ELs for whom English language arts and math achievement test items have similar difficulty and discrimination parameters as they do for non-ELs and (2) ELs for whom the test items function differently. We used latent class IRT methods to identify the two groups of ELs and to evaluate the effects of different subscales of ELP (reading, writing, listening, and speaking) on group membership. Only reading and writing were significant predictors of class membership. Cut-scores based on summary scores of ELP were imperfect predictors of class membership and indicated the need for finer differentiation within the top proficiency category. This study demonstrates the importance of linking definitions of ELP to the context for which ELP is used and suggests the possible value of psychometric analyses when language proficiency standards are linked to the language requirements for content area achievement.

Typologies of coping in young adults in the context of the COVID-19 pandemic

The COVID-19 pandemic has created major upheavals in the lives of people worldwide. The virus has mostly affected elderly populations, but there may be corollary effects on young adults’ psychosocial adjustment due to educational, economic, and occupational disruptions. Using latent class analysis, we examined unique typologies of coping in response to the pandemic among young adults. We used an expanded set of indicators including traditional measures of problem- and emotion-focused coping as well as measures of resilience and coping flexibility. We also examined whether class membership could be predicted by demographics, stress appraisal, and psychosocial characteristics including catastrophic thinking and impulsivity. The sample of 1,391 young adults (ages 18–35) was recruited via Amazon’s Mechanical Turk (MTurk) and snowball methods from late-April to early-May 2020. Six classes were identified: (1) Resilient Flexible Problem-Focused Copers, (2) Resilient Inflexible Problem-Focused Copers, (3) Non-Resilient Flexible Problem-Focused Venters, (4) Non-Resilient Flexible Problem-Focused Copers, (5) Non-Resilient Flexible Non-Copers, and (6) Non-Resilient Inflexible Non-Copers. Using Class 1 as the reference class, we found perceived centrality and uncontrollability of the pandemic as well as catastrophic thinking and impulsivity were significant predictors of class membership. The mean levels of stress appraisal and psychosocial characteristics varied significantly between the classes, reinforcing the structural validity of these classes. The findings suggest the importance of training young adults to develop resilience and flexibility as well as specific coping skills that can help offset the psychological effects of dramatic lifestyle changes that may result from pandemics or other health crises in the future.

Who are the young poker players? A latent class analysis of high school teenagers.

For the last 20 years, poker has been one of the most popular forms of gambling for adults. Although various studies have demonstrated the specific characteristics of these players, few studies have focused on teenagers playing poker. However, a better understanding of this activity among this vulnerable group would help to develop more effective strategies for preventing gambling problems. Thus, this study aims to identify, through latent class analysis, subgroups in that population and to assess co-occurrence across various characteristics typically associated with gambling behavior. The sample was constituted of 759 adolescents (70.8% boys; M age = 15.44 years, range = 14-19) recruited in high schools and who had played poker in the last year. The statistical fit indices revealed a four-class solution. Class 1 almost exclusively played simulated poker. Class 2 played poker exclusively in the school context. Class 3 played poker almost exclusively at home. Class 4 showed a very diversified pattern regarding their modalities of poker playing. Results of the logistic regression suggested that gambling related variables (e.g., time spent playing, reading about gambling strategies and diversity of gambling funding) were significant predictors of class membership. This study shows that there is a variety of profiles among young poker players. Although one profile has few risk factors, others have more factors associated with adults' gambling problems. These profiles suggest that specific prevention strategies are probably appropriate to reach these different groups of young people. (PsycINFO Database Record (c) 2020 APA, all rights reserved).

Complex Posttraumatic Stress Disorder in Patients Exposed to Emotional Neglect and Traumatic Events: Latent Class Analysis

The inclusion of a complex posttraumatic stress disorder (CPTSD) diagnosis in the 11th revision of the International Classification of Diseases reflects growing evidence that a subgroup of individuals with PTSD also suffer from disturbances in emotion regulation, interpersonal skills, and self-concept, which together are termed "disturbances in self-organization" (DSO). Although CPTSD is assumed to result from exposure to complex traumatic events, emotional neglect may be an important contributor. This study investigated the presence of CPTSD, defined by endorsement of PTSD and DSO symptoms in a clinical postwar generation sample. The sample consisted of 218 patients who had been exposed to emotional neglect in childhood, a subgroup of whom had also been exposed to potentially traumatic events. Using items from the Harvard Trauma Questionnaire and the Brief Symptom Inventory, a latent class analysis revealed two classes: high endorsement of almost all CPTSD symptoms (n = 83; 38.1%) and low endorsement of all CPTSD symptoms (n = 135; 61.9%). Contrary to our hypothesis, no DSO-only class was found. The R3step method showed gender and number of traumatic events to be significant predictors of class membership. Compared to the low endorsement class, individuals in the CPTSD class were more likely to be female, p = .013, and to report a higher number of traumatic experiences, p < .001. The potential intermediary role of emotional neglect in the development of DSO and CPTSD is discussed.

Exploring patterns of personal growth initiative and posttraumatic stress: A latent profile analysis

ABSTRACTObjective: This study examined the inconsistent relationship found between personal growth initiative (PGI) and posttraumatic stress symptoms (PTSS) by exploring potential subgroups. In addition, after identifying the subgroups, potential predictors of these subgroups were examined. Participants: Participants were 534 undergraduate students who experienced a potentially traumatic event (PTE). Data collection occurred from March 2013 to December 2014. Methods: Because this was a preliminary study, archival data was used. Participants consisted of students who experienced a PTE within the past three years. Results: Latent profile analysis indicated a five-class model as the best-fitting model. Direct exposure and intentionality of the event were significant predictors of class membership; length of time since event and gender did not predict class membership. Conclusions: The current study provides a potential explanation of the inconsistent results between PGI and PTSS and also highlights the importance of acknowledging subgroups when utilizing PGI after a PTE.

Co-occurring addictive behaviors: an analysis of risk profiles among university students

Objectives: The first goal of this study was to empirically identify, among university students, groups with varying levels of risk based on indicators of gambling and substance use. The second goal was to compare the identified groups on various demographic characteristics.Methods: The study comprised of 2139 full-time undergraduate students, representative of university students in Montreal, Canada. A 3-step latent class logistic regression analysis was performed to identify groups and compare them on demographic characteristics.Results: The statistical fit indices of the latent class analysis revealed a four-class solution. Class 1 (30.1% of the sample) included non-gamblers with low probabilities of substance use. Class 2 (11.2% of the sample) grouped non-gamblers with high-risk patterns of consumption. Class 3 (36.42% of the sample) included gamblers who are low-risk substance users, and individuals in Class 4 (22.25% of the sample) reported risky patterns of gambling and substance-using behaviors. Results of the logistic regression suggested that gender, being born in Canada, and working full- or part-time are significant predictors of class membership. Participants in both groups labeled at-risk (with and without gambling) were also more likely to report psychological distress and to live outside the family environment.Conclusions: This study raises important questions regarding the choice of preventive models and feeds into the long-standing debate around universal versus high-risk approaches.

Empirically derived subtypes of serious emotional disturbance in a large adolescent sample.

The heterogeneity of serious emotional disturbance has been thoroughly documented among adolescents with nationally representative data derived from structured interviews, although use of these interviews may not be feasible within the context of brief and self-administered school surveys. This study seeks to identify distinct subtypes of serious emotional disturbance in a large school-based sample. A total of 108,736 students fully completed the K6 scale that was included on the 2012 Kentucky Incentives for Prevention Survey. Latent class analysis was used to derive subtypes of serious emotional disturbance among students receiving a positive screen (n = 15,147). To determine significant predictors of class membership, adjusted rate ratios and 95 % confidence intervals were calculated using multinomial logistic regression. A four-class model was the most parsimonious, with four distinct subtypes emerging that varied by both symptom type and severity: comorbid moderate severity, comorbid high severity, anxious moderate severity, and depressed high severity. Age, gender, race/ethnicity, family structure, substance use, antisocial behavior, role impairments, and peer victimization were significant predictors of class membership, although the magnitude of these effects was stronger for the two high severity groups. Our results suggest heterogeneity of serious emotional disturbance by both symptom type and severity. Prevention programs may benefit by shifting focus from specific disorders to the core features of serious emotional disturbance, including psychological distress, high comorbidity, and role impairments.

Patterns and predictors of sleep quality before, during, and after hospitalization in older adults.

The impact of hospitalization on sleep in late-life is underexplored. The current study examined patterns of sleep quality before, during, and following hospitalization, investigated predictors of sleep quality patterns, and examined predictors of classification discordance between two suggested clinical cutoffs used to demarcate poor/good sleep. This study included older adults (n = 163; mean age 79.7 ± 6.9 years, 31% female) undergoing inpatient post-acute rehabilitation. Upon admission to inpatient post-acute rehabilitation, patients completed the Pittsburgh Sleep Quality Index (PSQI) retrospectively regarding their sleep prior to hospitalization. They subsequently completed the PSQI at discharge, and 3 months, 6 months, 9 months, and 1 year post discharge. Patient demographic and clinical characteristics (pain, depression, cognition, comorbidity) were collected upon admission. Using latent class analysis methods, older adults could be classified into (1) Consistently Good Sleepers and (2) Chronically Poor Sleepers based on patterns of self-reported sleep quality pre-illness, during, and up to 1 year following inpatient rehabilitation. This pattern was maintained regardless of the clinical cutoff employed (> 5 or > 8). Logistic regression analyses indicated that higher pain and depressive symptoms were consistently associated with an increased likelihood of being classified as a chronic poor sleeper. While there was substantial classification discordance based on clinical cutoff employed, no significant predictors of this discordance emerged. Clinicians should exercise caution in assessing sleep quality in inpatient settings. Alterations in the cutoffs employed may result in discordant clinical classifications of older adults. Pain and depression warrant detailed considerations when working with older adults on inpatient units when poor sleep is a concern.

Latent class analysis of need descriptors within an Irish youth mental health early intervention program toward a typology of need.

Significant overlap and comorbidity has been demonstrated among young people with mental health problems. This paper examined demographic characteristics, heterogeneity of need descriptors and services provided among young people (12-25 years) engaging in brief interventions at Jigsaw in the Republic of Ireland. Between 1 January 2013 and 31 December 2013, a total of 2571 young people sought help from 1 of 10 Jigsaw sites. Of these, 1247 engaged in goal-focused brief interventions, typically consisting of one to six face-to-face sessions. Descriptive statistics were used to summarize social and demographic factors. Latent class analysis was used to cluster young people into relevant typologies of presenting issues. Multinomial logistic regression was then performed to determine significant predictors of class membership. The most common age of young people was 16. More women (59.6%) than men engaged in brief interventions, 56% attended school, 74% lived with their family of origin or with one parent, and 54.2% came from families where parents were married. Using established fit criteria, four relevant typologies emerged: Developmental (26.8%), Comorbid (15.8%), Anxious (42.7%) and Externalising (14.6%). Predictors varied by class membership, but general family problems and lack of adult support emerged as the strongest predictors for all classes. This study demonstrated that the mental health needs of young people in Ireland are significant and diverse. Because Jigsaw favours a more descriptive approach to problem identification, the four typologies suggest a need to determine program capacity in engaging youth with heterogeneous presenting issues and to tailor brief interventions to each group's clinical profiles.

An exploratory study of multiple distinct gambling trajectories in emerging adults

This study uses data from a four-wave longitudinal survey of emerging adults (18–20 years of age in Wave 1) to examine whether there are multiple distinct trajectories of problem gambling risk severity and whether membership in these trajectory classes can be predicted by certain risk and protective factors. Four trajectory classes of gambling risk severity were identified – nonproblem-diminishing (73.9%), low-risk-stable (16.8%), marginal/nongambler-diminishing (7.1%), moderate-risk-increasing (2.2%) – with most youths' gambling involvement remaining stable or diminishing across the years and only the smallest most at-risk group showing a slight increase in severity across this transitional period. Three risk factors were significant predictors of class membership – being male, scoring higher on alcohol dependence, and escape-avoidance coping were all associated with increased probability of being in one of the more gambling involved trajectory classes, while lower alcohol dependence scores were associated with increased likelihood of being in the marginal/nongambling class.

Patterns of sleep quality during and after postacute rehabilitation in older adults: a latent class analysis approach

Sleep quality is related to emotional, physical, psychological and cognitive functioning and functional independence in later life. After acute health events, older adults are likely to utilize postacute rehabilitation services to improve functioning and facilitate return to independent living. Patterns of how sleep changes with postacute rehabilitation, and predictors of such patterns, are unknown. The current investigation employed latent class analysis (LCA) methods to classify older adults (n=233) into groups based on patterns of self-reported sleep quality pre-illness, during postacute rehabilitation and up to 1year following postacute rehabilitation. Using LCA, older adults were grouped into (1) consistently good sleepers (46%), (2) good sleepers who transitioned into poor sleepers (34%), (3) consistently poor sleepers (14%) and (4) poor sleepers who transitioned into good sleepers (6%). In three planned analyses, pain was an independent predictor of membership in classes 1 or 2 (good pre-illness sleep quality) versus classes 3 or 4 (poor pre-illness sleep quality), and of membership in class 1 (consistently good sleep) versus class 2 (good sleep that transitioned to poor sleep). A lower Mini-Mental State Examination score was a predictor of membership in class 1 versus class 2. There were no statistically significant predictors of membership in class 3 versus class 4. Demographics, comorbidities and depressive symptoms were not significant predictors of class membership. These findings have implications for identification of older adults at risk for developing poor sleep associated with changes in health and postacute rehabilitation. The findings also suggest that pain symptoms should be targeted to improve sleep during postacute rehabilitation.

Differential Susceptibility to the Effects of Child Temperament on Maternal Warmth and Responsiveness

ABSTRACT A child's difficult temperament can elicit negative parenting and inhibit positive parenting behavior. However, mothers appear to be differentially susceptible to child temperament. The author examined the differential susceptibility to the effects of a child's temperament on the mother–child interaction style (i.e., maternal warmth and responsiveness) as well as plausible reasons for these differences. With 2,130 mothers of 14-month-old infants (51% male) as subjects, a regression mixture analysis identified three latent classes with varying associations between the child's temperament and mother–child interactions: nonsusceptible class, susceptible–high class, and susceptible–low class. Mother-reported depression was most predictive of class membership. Latent class differences in the maternal self-efficacy, marital conflict, and coparenting alliance were also found. On the other hand, family income, maternal employment, and the child's gender were not significant predictors of class membership when individual and contextual resources were considered. Overall, mothers with abundant individual and family resources (i.e., less depressed, highly self-efficacious, few marital conflicts, and high coparenting alliance with their spouse) showed that their interaction style with a child would vary according to the child's temperament, whereas mothers with slender resources interacted with their children in a less warm and responsive manner, regardless of the child's temperament. The implications of these findings are also discussed.

Trajectories of adolescent alcohol use after brief treatment in an Emergency Department

ObjectiveThe primary aim of this study was to identify distinct classes of trajectories of adolescent substance use following a brief motivational interviewing (MI) intervention in an Emergency Department (ED). The secondary aim was to identify predictors of class membership. MethodsLatent growth mixture modeling was used with 177 adolescents who participated in two randomized clinical trials evaluating MI for an alcohol-related event. ResultsThree classes were identified: (1) moderate use, decreasers consisting of 56.8% of participants; (2) heavy use, decreasers, consisting of 10.5% of participants, and (3) heavy use sustainers, consisting of 32.7% of participants. Hispanic ethnicity, parental monitoring, and days of high-volume drinking were significant predictors of class membership. Hispanic ethnic status and high levels of parental monitoring were associated with decreased likelihood of belonging to either of the two heavy use classes. More frequent high-volume drinking at baseline was associated with increased likelihood of belonging to the heavy use, sustainer class, and decreased likelihood of belonging to the heavy use, decreaser class. Across all three classes, being female and having frequent high-volume drinking at baseline were associated with worse response to the intervention. ConclusionsThese findings have important implications for identifying adolescents who may benefit from different or additional intervention, and for anticipating and informing families of adolescents’ potential drinking course following treatment.