Maladaptive coping among military-connected adolescents: Examining combined risk using QCA

Introduction

Military-connected students in public schools face a unique set of stressors that may impact their wellbeing and academic functioning.

Methods

Twenty-four youth in the 7th to 12th grades who had an active-duty parent (mother or father) serving in the U.S. Armed Forces were interviewed. Participants completed a qualitative interview while actively completing a Life History Calendar (LHC) to mark deployment and family military service milestones and discuss how they impacted the youth respondent. This study used Qualitative Comparative Analysis (QCA) to explore the interplay and combination of specific stressors related to relocation and deployment experiences among adolescents, and to determine key factors associated with maladaptive outcomes.

Results

The results of the QCA analysis identified bullying experiences and negative experiences with other military-connected youth as conditions that are associated with maladaptive coping.

Discussion

Chronic and acute stressors in adolescence are established risk factors for mental, emotional, and behavioral problems in the short and long-term including suicidality, substance use and abuse, and substance use disorders. Through qualitative inquiry we were able to identify specific contextual details related to maladaptive coping that can be used to further refine areas of focus for research, prevention, and interventions for military-connected adolescents.

Adolescent Relationship Abuse, Gender Equitable Attitudes, Condom and Contraception Use Self-Efficacy Among Adolescent Girls

Introduction: Experiencing adolescent relationship abuse (ARA) negatively impacts sexual health and influences risk behaviors of adolescent girls. ARA may be associated with more inequitable gender attitudes among girls, a potentially modifiable factor in violence prevention. This study examines the association among gender equitable attitudes, experiences of ARA, and sexual behaviors among girls participating in Sisterhood 2.0, a community-based violence prevention program implemented in low resource neighborhoods. Methods: Data were from baseline surveys collected for Sisterhood 2.0 implemented in Pittsburgh, PA. Participant demographics, gender equitable attitudes, self-efficacy to use condoms with partners, and self-efficacy to select appropriate contraception were assessed. A latent class analysis (LCA) estimated probability of responses to nine indicators, including sexual behavior self-efficacy and violence. Multigroup LCA by grade (9-12) was also estimated and analyses were performed with SAS V9.4. Results: Female-identified adolescents ages 13-19 (n = 246) were primarily Black (75%) and evenly distributed across grade in school. Sixty-five percent reported emotional relationship abuse and 31% reported physical abuse within the previous nine months. A three-class solution was best fitting for the LCA. Experiences of violence were related to less equitable gender attitudes, being sexually active, and lower condom and contraception self-efficacy. Younger participants who were sexual minorities with less educated heads of household had more experiences with ARA and less equitable gender attitudes. Discussion: Gender equitable attitudes were lower in adolescent girls with greater experiences of ARA and worse condom and contraception self-efficacy. Integrating discussions about healthy sexual relationships and gender equity may be salient factors in violence prevention.

Racial/Ethnic Inequities in Adverse Childhood Experiences and Selected Health-Related Behaviors and Problems Among Maryland Adolescents

We examined racial/ethnic inequities in the prevalence of adverse childhood experiences (ACEs) and examined the association between ACEs and selected health-related behaviors and problems. Data for this cross-sectional study come from the 2018 Maryland Youth Risk Behavior Survey/Youth Tobacco Survey, a statewide survey of high school students (n = 40,188). ACEs included caregiver verbal abuse and household food insecurity, substance use or gambling, mental illness, and involvement with the criminal justice system. We estimated the prevalence of ACEs overall and by race/ethnicity, and then used multiple logistic regression to determine associations between ACEs and emotional/behavioral problems, adjusting for race/ethnicity. Outcome variables included emotional distress, poor school performance, suicidal ideation, fighting, alcohol use, and marijuana use. More than one fifth of students reported each individual ACE. Differences in the prevalence of ACEs by race/ethnicity were statistically significant (p < .001). More than one fourth (25.8%) reported one of the five ACEs, 15.1% reported two, and 15.4% reported three or more. For each ACE, reporting having experienced it (vs. not) was associated with a >30% higher prevalence for each of the outcome variables. Among students who reported three or more ACEs (relative to none), the odds of emotional distress and suicidal ideation were more than 8 times greater. Among Maryland adolescents, ACEs are common, are inequitably distributed by race/ethnicity, and are strongly linked to behavioral health. Findings suggest the need to monitor ACEs as a routine component of adolescent health surveillance and to refocus assessment and intervention toward "upstream" factors that shape adolescent health.

Identifying individuals with multiple non-communicable disease risk factors in Kenya: a latent class analysis

OBJECTIVES

Non-communicable diseases (NCDs) are the leading causes of death globally. In Kenya, the number of deaths resulting from NCDs is projected to surpass malaria and tuberculosis by 2030. Studies in Kenya show increasing NCDs; the aim of the present study is to examine the clustering of NCDs and risk factors in Kenya.

STUDY DESIGN

This is a cross-sectional study using data from the 2015 Kenya STEPwise Survey.

METHODS

This study examined relationships between NCDs (e.g. obesity, hypertension and diabetes) and health behaviours (e.g. sedentary activity, and fruit and vegetable consumption). Survey probability weights, which estimated the sampling design effect, were applied to consider the sampling units, and stratifications were used during sampling so that the results could be generalisable to the national adult Kenyan population. In total, 4350 adults were included in the study sample.

RESULTS

Overall, 24.43% of participants were classified as having hypertension, 1.88% as having type 2 diabetes, and 27.94% were classified as being overweight or obese. The best-fit model was a four-class solution. Class 1 is best described as 'young with high NCD risk' and had the highest sedentary activity. Class 2 is best described as 'poor rural with lower NCD risk' with a high chance of smoking and alcohol consumption. Class 3 is best described as 'rural with high NCD risk' and had the highest fruit and vegetable consumption. Class 4 is best described as 'wealthy young urban dwellers with high NCD risk' with a high chance of alcohol consumption and smoking. Individuals in Class 4 had the highest chance (40%) of being overweight/obese, a 2% chance of type 2 diabetes and a 23% chance of having hypertension.

CONCLUSIONS

NCDs are clustered in groups with high-risk behaviours. The group with the highest chance of having NCDs also had the highest chance of engaging in high-risk behaviours. The findings of this study suggest that smoking and alcohol consumption increase NCD risk in rural areas. Tailored and targeted interventions are needed to curb the increasing NCD prevalence in Kenya.

Mental Health Outcomes Associated with Risk and Resilience among Military-Connected Youth

The present study aimed to describe patterns of risk and protective factors affecting U.S. Army families and their association with mental health diagnoses among military-connected children. Wartime military service is associated with increased adverse outcomes for military-connected youth, but few studies have explored the impact of concurrent risk and access to protective factors. Using big data methods to link existing datasets, protective factors (e.g., marital and family functioning) were drawn from a voluntary survey completed by 1,630 US Army spouses. Risk factors (e.g., parent mental health, family moves, deployment) were drawn from Department of Defense (DoD) archival data. Rates of mental health diagnoses among youth were derived from DoD healthcare records. Using the three-step method of latent profile analysis, five profiles emerged with variability across risk and protective factors. The largest group (40% of the sample) had considerable protective factors and limited risk exposure. Statistically significant differences in the prevalence of mental health diagnoses among military-connected youth were observed across profiles (χ²  = 30.067, df = 4, p < .001), with the highest rates (31.1% and 30.5%) observed in the two profiles with the lowest protective factors. Findings suggest most military families are faring well and highlight the importance of a thorough assessment that evaluates both the stressors military families face and the strengths they possess.

Mental health outcomes associated with profiles of risk and resilience among U.S. Army spouses

The current study examined patterns of risk and protective factors among military families and associations with mental health diagnoses among U.S. Army spouses. Spouses (N = 3,036) completed a survey of family psychosocial fitness, which informed protective factors including coping, family cohesion, and social support. Survey results were linked with Department of Defense archival data, which provided information on military-specific risks, including relocation, deployments, and reunification, as well as mental health care diagnoses. The three-step method of latent profile analysis identified six profiles, suggesting significant heterogeneity in military families with respect to their access to resources and exposure to risk. The largest profile of families (40.48% of the sample) had limited risk exposure and considerable strengths. Variability in risk and protection across profiles was associated with statistically significant differences in the prevalence of mental health diagnoses among spouses (χ² = 108.968, df = 5, p < .001). The highest prevalence of mental health diagnoses among Army spouses (41.2%) was observed in the profile with the lowest levels of protective factors. Findings point to the importance of evaluating both concurrent risk and protective factors. Increasing access to resources may be a fruitful avenue for prevention among military families that are struggling. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Polytobacco use among a nationally-representative sample of black high school students

Background

Studies of the patterns of polytobacco use have increased. However, understanding the patterns of using multiple tobacco products among Black adolescents is minimal. This study identified the patterns of polytobacco use among U.S. Black adolescents.

Methods

Latent class analysis (LCA) was used to identify patterns of adolescent polytobacco use among a representative sample of Black youth from the 2017 Youth Risk Behavior Survey (n = 2782). Ever and recent (past 30 day) use of cigarettes, electronic cigarettes, cigars, and dip or chewing tobacco were used as latent class indicators. Multinomial regression was conducted to identify the association if smoking adjusting for sex, age, grade, and marijuana use.

Results

Most students were in the 9th grade (29%), e-cigarette users (21%) and were current marijuana users (25%). Three profiles of tobacco use were identified: Class 1: Non-smokers (81%), Class 2: E-cigarette Users (14%), and Class 3: Polytobacco Users (5%). Black adolescent Polytobacco users were the smallest class, but had the highest conditional probabilities of recent cigarette use, e-cigarette use, ever smoking cigars or chewing tobacco. Ever and current use of marijuana were associated with increased odds of being in the e-cigarette user versus non-smoker group, and current marijuana use was associated with increased odds of polytobacco use (aOR = 24.61, CI = 6.95–87.11).

Conclusions

Findings suggests the need for targeted interventions for reducing tobacco use and examining the unique effects of polytobacco use on Black adolescents. Findings confirm a significant association of marijuana use with tobacco use.

Preliminary Psychometrics and Potential Big Data Uses of the U.S. Army Family Global Assessment Tool

The purpose of the present study is to explore the psychometric properties of the U.S. Army's Family Global Assessment Tool (GAT), which assesses the psychosocial fitness of Army families. With data from 1,692 Army spouses, we examined the structure, reliability and validity of the GAT, using confirmatory factor analysis (CFA) and two validity studies. Fifty-three items and 9 factors were retained following CFA. This model provided a good fit, and scales demonstrated strong internal consistency. Bivariate correlations and results from a theoretically driven model provide preliminary evidence of validity. Findings support the usefulness of the GAT for measuring psychosocial fitness of Army spouses.

Latent classes of partner-seeking venues and sexual risk among men who have sex with men in Paris, France

In France, men who have sex with men (MSM) experience the highest HIV incidence. MSM who use multiple venues such as bars and smartphone apps to find sexual partners have greater HIV and sexually transmitted infection risk than those who use a single physical or online venue. To explore the role of venues and sexual behavior on HIV risk, we used latent class analysis to highlight underlying profiles of 580 MSM in Paris, France. Three latent classes emerged: Multi-venue Users (19%), Non-serosorting App Users (49%), and Serosorting App Users (32%). Multi-venue Users had the highest probabilities of condomless receptive anal intercourse (CRAI) and condomless insertive anal intercourse (CIAI), group sex, and meeting partners at venues such as bars or online. Non-serosorting App Users had the lowest probabilities of CRAI, CIAI, and condomless serosorting. Serosorting App Users had an 87% chance of condomless serosorting. MSM reporting a pre-exposure prophylaxis history had 14 and 5 times the odds of being classified as Multi-venue Users and Serosorting App Users than Non-serosorting App Users. MSM were 3% more likely to be Multi-venue Users than Non-serosorting App Users for every one-year increase in age. Interventions addressing sexual risk behaviors among MSM in France should consider tailored risk-reduction and prevention messaging.

Understanding Community-based HIV/AIDS Service Organizations: An Invaluable Source of HPV-related Cancer Information for At-risk Populations

To describe non-clinical HIV service providers (NCHSPs) as surrogate seekers and health information mavens for people living with HIV (PLWH), men who have sex with men (MSM), and other vulnerable populations.In May/June 2016, we recruited 30 NCHSPs from three community-based HIV/AIDS service organizations. NCHSPs completed a 118-item self-administered, paper-and-pencil survey about HPV, cancer, and health communication. Data were analyzed using Stata/SE 14.1.Almost all (97%) NCHSPs were surrogate seekers and had looked for HIV/AIDS (97%), STD (97%), and cancer (93%) information. Most (60%) cancer information seekers had looked for information about HPV. The Internet (97%) and healthcare providers (97%) were health information sources almost all NCHSPs trusted. Nearly all NCHSPs (93%) were completely or very confident about their ability to find health information. The mean health information mavenism score (17.4 ± 2.1) was significantly higher than the scale's high-score cutoff (15.0) (p < 0 .001).NCHSPs look for and share health information with the vulnerable populations (e.g., PLWH, MSM) they serve. More research is needed to understand what NCHSPs' know and think about the health information they are sharing with vulnerable populations.

Children’s Health Insurance Status and Emergency Room Utilization: An Examination of Complex Survey Data

Since the Children’s Health Insurance Program’s passage into law in 1997, the program has increased in cost to over $15 billion in recent years. Emergency room usage has also increased throughout the United States, leading to nationwide issues in overcrowding and surges in service costs. This study seeks to examine emergency room utilization of children insured under Children’s Health Insurance Program to determine if Children’s Health Insurance Program enrollees use the emergency room more or less frequently than their privately insured counterparts. The data used in this study were from the 2017 National Health Interview Survey. SAS statistical software was used to conduct a multinomial regression assessing the relationship between insurance type (private v. Children’s Health Insurance Program) and frequency of emergency room utilization over the last 12 months. The analysis results indicate no statistically significant difference between Children’s Health Insurance Program insured and privately insured children in terms of frequency of emergency room utilization and suggest a need to explore other factors that more directly influence Children’s Health Insurance Program costs.

Social determinants of hypertension and type-2 diabetes in Kenya: A latent class analysis of a nationally representative sample

Introduction

Cardiovascular disease is among the leading causes of death in Kenya and type II diabetes (T2D) is a growing chronic health concern in the country. However, a gap exists in examining how demographic and social characteristics coalesce to identify individuals at high risk for hypertension and/or T2D in Kenya. The current study examined demographic typologies associated with self-report diagnoses.

Methods

Nationally representative cross-sectional study using 43,898 individuals from the Kenya Demographic and Health Survey 2014. Main Outcome Measures were self-reported Hypertension and Type 2 Diabetes diagnosis. Descriptive analyses were conducted using STATA 14. Latent class analysis (LCA) was conducted using Mplus 7.4.

Results

Approximately 5% reported hypertension and 1% reported T2D. Latent class analysis suggested a 4-class solution. The class with the highest likelihood to report previous diagnosis of hypertension (10.4%), consisted of high proportion of married adult women. The second highest prevalence of previous diagnosis of hypertension (4.4%) consisted of a high proportion of married middle aged men with high probability of being smokers. The results suggest that Kenyan women over 30 years may be at increased risk of hypertension compared to men. Future studies should include additional socio-demographic and behavioral characteristics to better understand gender differences in correlates for hypertension to be used for targeted and tailored health promotion-interventions.

Identifying Syndemics for Sexually Transmitted Infections Among Young Adults in the United States: A Latent Class Analysis

PURPOSE

Syndemic theory suggests that the convergence of social, environmental, and ecological factors can interact to exacerbate behavioral health problems and are often intensified by social conditions and disparities. This study used latent class analysis (LCA) to determine gender and racial/ethnic specific classes for sexually transmitted infection (STI) risk.

METHODS

LCA included 18 measured socioeconomic, depression, substance use, and sexual behavioral variables from 1,664 young adults ages 18-25 in the NHANES. Models were stratified by gender and then by race/ethnicity. Logistic regression determined associations between latent class membership and testing positive for one or more STIs (Chlamydia trachomatis, HIV or herpes simplex virus-II). For each stratified analysis, classes with the lowest probability of reported risk factors in the LCA were the reference groups.

RESULTS

Class 3 in females (highest probability of reporting both socioeconomic and behavioral factors) and class 3 in males (majority behavioral factors) had increased odds of STI (females: OR = 2.7, 95% CI 1.6-4.5; males: OR 2.5, 95% CI 1.3-4.6). By race for females, depression (highest in Hispanics), poverty, and less educated households (highest in blacks and Hispanics) were evident in classes associated with STI. Class 1 black males (majority behavioral factors) had a higher odds of STI compared with low risk white males (OR = 16.4 95% CI 3.7-72.0) However, no other associations were observed among males.

CONCLUSIONS

Risk patterns for STI differed by gender and race/ethnicity. Consistent with syndemic theory, effective STI interventions need to address socioeconomic factors and mental health rather than individual behaviors, particularly for minority women.

Psychological Distress Behavioral Patterns Among Latinos: We Don't See Ourselves as Worthless

A cross-sectional study of 4921 Latinos from the National Health Interview Survey was conducted to examine behavioral patterns of psychological distress among Latinos. Latent class analysis was used to ascertain psychological distress behavioral profiles among Latinos. Analysis revealed four latent classes of distress, which included moderate psychological distress (13.6%); mild sadness, nervousness, and restlessness (13.0%); high psychological distress (2.8%); and no psychological distress (70.7%). Worthlessness, a widely-accepted dimension of distress, was not a significant behavioral trait. Results from the present study suggest that underlying cultural elements affect the subjective interpretations of symptoms reported by Latinos. These findings highlight distress profiles among Latinos and the possibility of overlooking behaviors that are uniquely indicative of distress, potentially leading to the underreporting of serious psychological distress in this population. Furthermore, these findings shed light on other paradoxical issues impacting the mental health of Latinos.

A latent class analysis of the co-occurrence of sexual violence, substance use, and mental health in youth

Purpose: The purpose of this study was to understand co-occurrence of sexual violence, substance use, and mental health in youth and see if there are differences in experiences among sexual identities, races, and genders. Methods: The 2017 Youth Risk Behavior Survey sampled 14,638 high school students. Latent class analysis was used to understand the relationship between incidence of sexual violence, substance use, and depression and suicidality in a nationally representative sample. The model controlled for demographic and bullying covariates. Results: A four-class solution provided optimal fit. Classes were low risk (51.8%); low violence, frequent substance use, and high depression and suicidal ideation (29.7%); experienced violence and moderate mental health issues(5.1%); and some violence, high substance use, and high depression and suicidal ideation (13.4%). Youth who identified as sexual minority youth were almost 2.5-5 times more likely to be in the three higher risk classes than their heterosexual peers. Those who identified as "I don't know" were almost 3 times more likely to be in the three higher risk classes than heterosexual youth. Multiracial youth had higher likelihood of being in the three riskier classes compared to Caucasian youth. Females were 2.5 and 3 times more likely to be in the low violence and some violence classes compared to males. Conclusion: The study's finding distinct classes of co-occurring behavioral and mental health outcomes fill gaps in the current literature and informs practitioners that LGB youth and bi/multiracial youth were at an elevated risk of co-occurring victimization, substance use, and mental health problems.

Latent Classes of Sexual Positioning Practices and Sexual Risk Among Men Who Have Sex with Men in Paris, France

HIV/STI risk varies by sexual positioning practices; however, limited data have characterized the behavioral profiles of men who have sex with men (MSM) in France. This study used latent class analysis (LCA) to explore sexual risk profiles among MSM in Paris, France. LCA was used to classify sexual positioning and serosorting profiles among MSM in Paris (n = 496). Age, HIV status, relationship status, substance use, group sex, and PrEP history were used in a multinomial regression model predicting class membership. Three latent classes were identified: majority top/serosorters, versatile/low partners, and majority bottom/some serosorters. Majority top/serosorters had the highest probability of condomless serosorting; majority bottom/some serosorters had the highest mean number of partners (~ 12 partners) for condomless receptive anal intercourse. HIV-positive MSM were more likely to be classified as majority bottom/some serosorters than versatile/low partners (AOR 7.61; 95% CI 2.28, 25.3). Findings support the need for tailored and targeted interventions for highest-risk individuals.

HIV Testing and Healthcare Utilization Among U.S. African-American Women

A severe HIV/AIDS disparity exists for U.S. African-American women. Although HIV testing and healthcare engagement are paramount to HIV prevention and community health, many African-American women do not test for HIV or optimize healthcare services. To unpack the relationship between HIV testing and healthcare utilization among African-American women, latent class analysis (LCA) was used among a national, representative sample from the 2014 National Health Interview Survey to assess subgroup profiles of African-American women regarding HIV testing and healthcare utilization behaviors. This study also explored the covariates predicting latent class membership: age, income, cost of living worry, and healthcare service satisfaction. LCA identified three subgroups of U.S. African-American women: (a) Moderate HIV testing/Low healthcare utilization, (b) Moderate HIV testing/High healthcare utilization, and (c) No HIV testing/High healthcare utilization. Future HIV prevention interventions should be tailored to specific subgroups of African- American women based on HIV testing behaviors and healthcare use and experiences.

The association of parentification indicators with substance use patterns among military-connected adolescents

Military-connected youths are a vulnerable population exposed to a unique set of stressors, which may put them at increased risk for adverse mental and behavioral health outcomes, compared to their civilian counterparts. Among military-connected adolescents, emotional and instrumental parentification are mechanisms hypothesized to account for negative outcomes, including substance use. However, parentification may be protective in some cultures and has not been examined in the military population. Data were drawn from 1,441 7th-, 9th-, and 11th-graders who had a parent in the military and completed the 2013 California Healthy Kids Survey. Latent class analysis was used to examine patterns of instrumental parentification indicators among military-connected youths and associated patterns of substance use. Military-connected students in the high parentification class were significantly less likely to be in the frequent polysubstance using class (odds ratio = 0.376, 95% confidence interval [0.180, 0.782]). Findings suggest that experiences of instrumental parentification, including having more responsibilities at home, being more independent, and being able to solve problems better than peers, may be protective. For military-connected youths, these skills, developed in the context of their families, may translate to better coping with other stressful situations. Professionals may be able to build on these strengths and develop adaptive coping strategies to support military-connected youths in managing the stressors of wartime military life. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Depression and Alcohol Use in a National Sample of Hispanic Adolescents

BACKGROUND

Underage alcohol use and depression remain public health concerns for Hispanic adolescents nationwide.

OBJECTIVES

The study purpose was to identify the profiles of depression among Hispanic adolescents who reported experiencing depressive symptoms in their lifetime and classify them into groups based on their symptoms. Based on classifications, we examined the relationship between past year alcohol use and severity of depressive symptoms while controlling for sex and age.

METHODS

A secondary analysis of the 2013 NSDUH was conducted among Hispanic adolescents from 12 to 17 years of age (n = 585) who reported experiencing depressive symptoms. Latent class analysis was used to identify latent classes of depressive symptoms among Hispanic adolescents. A zero-inflated negative-binomial regression model was used to examine the relationship between alcohol use and depressive symptoms.

RESULTS

"High depressive" and "moderate depressive" classes were formed. The items that highly differentiated among the groups were felt worthless nearly every day, others noticed they were restless or lethargic, and had changes in appetite or weight. There was a significant difference (p = 0.03) between the classes based on alcohol use; those in the moderate depressive class were 1.71 times more likely to be identified as not reporting past alcohol use. Results indicated the high depressive class was estimated to have 1.62 more days of past year alcohol use than those in the moderate depressive class for adolescents who used alcohol (p < 0.001). Conclusions/Importance: Study findings can be used to address these significant public health issues impacting Hispanic adolescents. Recommendations are included.

School Protective Factors and Substance Use Among Lesbian, Gay, and Bisexual Adolescents in California Public Schools

PURPOSE

The majority of studies examining substance use among lesbian, gay, and bisexual (LGB) youth have focused on a wide array of risk factors (e.g., victimization). Few studies have explored the protective role of schools. This study aims to fill this gap in the literature and inform programs aimed at reducing substance use among LGB youth. More specifically, this study explores the extent to which school connectedness and support from teachers and other adults at school are associated with substance use among LGB youth in school and within the past 30 days.

METHODS

A secondary analysis of the 2013-2015 California Healthy Kids Survey (CHKS) was conducted to examine associations between school protective factors (i.e., school connectedness and adult support) and substance use among LGB youth, above and beyond a key risk factor, school victimization. The study outcomes were past 30-day and in-school use of cigarettes, alcohol, marijuana, inhalants, prescription pain medication, and other illegal drugs.

RESULTS

Overall, school connectedness and school adult support were associated with lower odds of substance use. For example, higher levels of school connectedness were associated with 22% decreased odds of past 30-day inhalant use (adjusted odds ratio [AOR] = 0.78; 95% confidence interval [CI] = 0.72-0.86), and 25% decreased odds of past 30-day prescription pain medication use (AOR = 0.75; 95% CI = 0.69-0.82). Higher levels of adult support in school were also associated with 17% decreased odds of marijuana use on school property in the past 30 days (AOR = 0.83; 95% CI = 0.77-0.91).

CONCLUSION

The results indicate a need for substance use prevention programs that integrate school connectedness and adult support in school.

Substance Use Among Transgender Students in California Public Middle and High Schools

BACKGROUND

Transgender adolescents face tremendous social stress in families and schools, which often leads to behavioral health disparities. This study assessed whether rates of substance use were higher among transgender adolescents when compared to nontransgender adolescents.

METHODS

This study is a secondary data analysis of the 2013-2015 California Healthy Kids Survey (CHKS) that examines whether rates of substance use are higher among transgender youth when compared to nontransgender youth. Participants included 4778 transgender and 630,200 nontransgender students in middle and high schools in nearly all school districts in California. The study outcomes were lifetime, recent, and in-school use of cigarettes, tobacco, alcohol, marijuana, cocaine, and ecstasy as well as nonmedical use of prescription painkillers, diet pills, Ritalin or Adderall, and cold medicine.

RESULTS

Transgender students were about 2-1/2 times more likely as nontransgender students to use cocaine/methamphetamine in their lifetime (odds ratio [OR] = 2.53; 95% confidence interval [CI] = 2.18-2.95) and about 2.8 times as likely to report past 30-day inhalant use (OR = 2.80; 95% CI = 2.41-3.26). Transgender students were more than twice as likely to report past 30-day prescription pain medication use (OR = 2.19; 95% CI = 1.90-2.53) and more than 3 times as likely to use cigarettes in school (OR = 3.37; 95% CI = 2.84-3.99).

CONCLUSIONS

The study's findings indicate a need for community- and school-based interventions that reduce substance use among transgender youth.

HIV Testing and Health Care Utilization Behaviors Among Men in the United States: A Latent Class Analysis

Emphasis has been placed on HIV testing and health care engagement, but little is known about how testing and engagement intersect, especially for men. We used latent class analysis to explore underlying profiles of U.S. men regarding HIV testing and health care utilization using data from the 2014 National Health Interview Survey. Multinomial regression was used to predict class membership in four classes: (a) Low HIV Testing/No Health Care Utilization, (b) Some HIV Testing/Low Health Care Utilization, (c) No HIV Testing/Some Health Care Utilization, and (d) High HIV Testing/High Health Care Utilization. Most men were in the No HIV Testing/Some Health Care Utilization class (46%), with a 0% chance of ever having had an HIV test but an 89% chance of seeing a general practitioner in the previous year. Research should include qualitative measures to capture information on facilitators and barriers to HIV testing for men who see general practitioners.

E-cigarettes, Cigarettes, and the Prevalence of Adolescent Tobacco Use

BACKGROUND

Adolescent e-cigarette use has increased rapidly in recent years, but it is unclear whether e-cigarettes are merely substituting for cigarettes or whether e-cigarettes are being used by those who would not otherwise have smoked. To understand the role of e-cigarettes in overall tobacco product use, we examine prevalence rates from Southern California adolescents over 2 decades.

METHODS

The Children's Health Study is a longitudinal study of cohorts reaching 12th grade in 1995, 1998, 2001, 2004, and 2014. Cohorts were enrolled from entire classrooms in schools in selected communities and followed prospectively through completion of secondary school. Analyses used data from grades 11 and 12 of each cohort (N = 5490).

RESULTS

Among 12th-grade students, the combined adjusted prevalence of current cigarette or e-cigarette use in 2014 was 13.7%. This was substantially greater than the 9.0% adjusted prevalence of current cigarette use in 2004, before e-cigarettes were available (P = .003) and only slightly less than the 14.7% adjusted prevalence of smoking in 2001 (P = .54). Similar patterns were observed for prevalence rates in 11th grade, for rates of ever use, and among both male and female adolescents and both Hispanic and Non-Hispanic White adolescents.

CONCLUSIONS

Smoking prevalence among Southern California adolescents has declined over 2 decades, but the high prevalence of combined e-cigarette or cigarette use in 2014, compared with historical Southern California smoking prevalence, suggests that e-cigarettes are not merely substituting for cigarettes and indicates that e-cigarette use is occurring in adolescents who would not otherwise have used tobacco products.

Associations between school violence, military connection, and gang membership in California secondary schools

Recent studies have found that military-connected students confront many challenges-such as secondary traumatization-that may stem from a parent's deployment and frequent relocations. It is possible that multiple moves and deployments of family service members are associated with military-connected students' gang membership and involvement with school violence behaviors. In this study, a total of 13,484 students completed the core and military modules of the California Healthy Kids Survey. Logistic regressions examined the odds of a student being a member of a gang given their grade, gender, race/ethnicity, school violence behaviors, military-connectedness, changes in schools, and familial deployments. Results indicated that of the nearly 8% of students sampled who reported being in a gang, those with a parent or sibling currently serving in the military reported a higher prevalence rate of gang membership than students with no military connection. Students who reported being in fights or carrying weapons to school were at least twice more likely to be a gang member than students who reported not having been in fights or carrying weapons. Changing schools 4 or more times in a 5-year period and experiencing at least 1 familial deployment were also associated with an increased likelihood of gang membership. The findings of this study offer incentive to further explicate the gang and school violence experiences of military-connected students. This study supports schools in understanding the characteristics of the military-connected students and families they serve so they can implement appropriate interventions to curb gang and school violence behaviors. (PsycINFO Database Record

The E-cigarette Social Environment, E-cigarette Use, and Susceptibility to Cigarette Smoking

PURPOSE

One concern regarding the recent increase in adolescent e-cigarette use is the possibility that electronic (e-) cigarettes may be used by those who might not otherwise have used cigarettes, and that dual use, or transition to cigarette use alone, may follow.

METHODS

Questionnaire data were obtained in 2014 from 11th/12th grade students attending schools in 12 communities included in the Southern California Children's Health Study. We evaluated the cross-sectional association between e-cigarette use, the social environment (family and friends' use and approval of e-cigarettes and cigarettes), and susceptibility to future cigarette use among never cigarette smokers (N = 1,694), using previously validated measures based on reported absence of a definitive commitment not to smoke.

RESULTS

Among adolescents who had never used cigarettes, 31.8% of past e-cigarette users and 34.6% of current (past 30-day) e-cigarette users indicated susceptibility to cigarette use, compared with 21.0% of never e-cigarette users. The odds of indicating susceptibility to cigarette use were two times higher for current e-cigarette users compared with never users (odds ratio = 1.97; 95% confidence interval: 1.21-3.22). A social environment favorable to e-cigarettes (friends' use of and positive attitudes toward the use of e-cigarettes) was also associated with greater likelihood of susceptibility to cigarette use, independent of an individual's e-cigarette use.

CONCLUSIONS

E-cigarette use in adolescence, and a pro-e-cigarette social environment, may put adolescents at risk for future use of cigarettes. E-cigarettes may contribute to subsequent cigarette use via nicotine addiction or social normalization of smoking behaviors.

Patterns of Alternative Tobacco Product Use: Emergence of Hookah and E-cigarettes as Preferred Products Amongst Youth

PURPOSE

There is a growing public health concern related to the rapid increase in the use of multiple tobacco products among adolescents. This study examined patterns of adolescent use of cigarettes, e-cigarettes, cigars/cigarillo, hookah/waterpipe, and smokeless/dip/chewing tobacco in a population of southern California adolescents.

METHODS

Data from 2,097 11th- and 12th-grade participants in the Southern California Children's Health Study were collected via self-report in 2014. Study participants were asked about lifetime and current (past 30 days) use of cigarettes, cigars/cigarillos/little cigars, e-cigarettes, hookah/waterpipe, and smokeless/dip/chewing tobacco. Latent class analysis (LCA) was used to identify patterns of tobacco use.

RESULTS

Hookah/waterpipe tobacco use had the highest current prevalence (10.7%) followed by e-cigarettes (9.6%). The prevalence of use of smokeless/dip/chewing tobacco was lowest, with 2.2% of adolescents reporting current use. The LCA suggested four distinct classes, comprising nonusers (72.3% of the sample), polytobacco experimenters (13.9%), e-cigarette/hookah users (8.2%), and polytobacco users (5.6%). Multinomial logistic regression based on these four classes found that males had double the odds to be polytobacco users relative to nonusers compared to females (odds ratio, 2.3; 95% confidence interval, 1.26-4.25).

CONCLUSIONS

By identifying naturally occurring configurations of tobacco product use in teens, these findings may be useful to practitioners and policymakers to identify the need for tobacco control interventions that address specific tobacco products and particular combinations of polytobacco use. LCA can be used to identify segments of the population overrepresented among certain tobacco use classes (e.g., boys) that may benefit most from targeted polyproduct intervention approaches.

Suicidality among military-connected adolescents in California schools

Previous research indicates that suicidal ideation is higher among military-connected youth than non military-connected youth. This study extends prior work by examining suicidal ideation, plans, and attempts in military-connected and non military-connected adolescents. Data were gathered from 390,028 9th and 11th grade students who completed the 2012-2013 California Healthy Kids Survey. Bivariate comparisons and multivariate logistic analyses were conducted to examine differences in suicidal ideation, plans, attempts, and attempts requiring medical attention between military and not military-connected youth. In multivariate logistic analyses, military-connected youth were at increased risk for suicidal ideation (OR = 1.43, 95 % CI = 1.37-1.49), making a plan to harm themselves (OR = 1.19, CI = 1.06-1.34), attempting suicide (OR = 1.67, CI = 1.43-1.95), and an attempted suicide which required medical treatment (OR = 1.71, CI = 1.34-2.16). These results indicate that military-connected youth statewide are at a higher risk for suicidal ideation, plans, attempts, and attempts requiring medical care because of suicidal behaviors. It is suggested that policies be implemented to increase awareness and screening among primary care providers, school personnel, and military organizations that serve military-connected youth.

Substance Abuse and Other Adverse Outcomes for Military-Connected Youth in California: Results From a Large-Scale Normative Population Survey

IMPORTANCE

Military families and military-connected youth exhibit significant strengths; however, a sizeable proportion of these families appear to be struggling in the face of war-related stressors. Understanding the consequences of war is critical as a public health concern and because additional resources may be needed to support military families.

OBJECTIVE

To determine whether rates of adverse outcomes are higher for military-connected adolescents during war compared with nonmilitary peers.

DESIGN, SETTING, AND PARTICIPANTS

This study is a secondary data analysis of a large, normative, and geographically comprehensive administrative data set (2013 California Healthy Kids Survey) to determine whether military-connected youth are at risk for adverse outcomes, including substance use, experiencing violence and harassment, and weapon carrying, during wartime. These outcomes are of particular concern because they affect socioemotional adjustment and academic success. Data were collected in March and April 2013 and participants included 54,679 military-connected and 634,034 nonmilitary-connected secondary school students from public civilian schools in every county and almost all school districts in California.

MAIN OUTCOMES AND MEASURES

Outcomes included lifetime and recent use of alcohol, tobacco, marijuana, prescription medications, and other drugs, as well as experiences of physical and nonphysical violence and harassment and weapon carrying during the last year.

RESULTS

Multivariable logistic regression models indicated that military-connected youth had greater odds of substance use, experience of physical violence and nonphysical harassment, and weapon carrying. For example, military-connected youth had 73% greater odds of recent other drug use (eg, cocaine and lysergic acid diethylamide; odds ratio [OR], 1.73; 95% CI, 1.66-1.80) and twice the odds of bringing a gun to school (OR, 2.20; 95% CI, 2.10-2.30) compared with nonmilitary-connected peers. Their odds of being threatened with a weapon or being in a fight were also significantly higher than their civilian counterparts (OR, 1.87; 95% CI, 1.80-1.95 and OR, 1.67; 95% CI, 1.62-1.71, respectively).

CONCLUSIONS AND RELEVANCE

Most military-connected youth demonstrate resilience. However, results suggest that during wartime, military-connected youth are at increased risk for adverse outcomes. Further, when compared with data from 2011, the rates of these negative outcomes appear to be increasing. These findings suggest a need to identify and intervene with military-connected adolescents and reflect a larger concern regarding the well-being of military families during wartime.

Understanding Consistent Condom Use Among Homeless Men Who Have Sex with Women and Engage in Multiple Sexual Partnerships: A Path Analysis

Consistent condom use is the main strategy aimed at preventing individuals from acquiring HIV through sexual intercourse. The mechanism of consistent condom use among a high-risk homeless subpopulation-homeless men who have sex with women and also engage in multiple sexual partnerships-remains unclear. This study identified 182 homeless men who engaged in multiple sexual partnerships from a representative sample of homeless men, who self-identified as heterosexual, using meal line services in Downtown Los Angeles' Skid Row area. Information such as participants' condom use psychosocial correlates, sexual risk behaviors, and social network characteristics were collected. Results suggested that condom efficacy is a potential intervening mechanism through which condom attitudes (β = -0.199; p = 0.005) and depression (β = -0.156; p = 0.029) are associated with an individual's consistent condom use. Having more network members with whom participants talked about HIV prevention (β = 0.051; p = 0.006) was also found to be associated with an individual's consistent condom use. HIV prevention programs should focus on increasing their condom efficacy to help reduce HIV risks among this vulnerable homeless subpopulation.

Psychosocial Factors Associated With Adolescent Electronic Cigarette and Cigarette Use

BACKGROUND

Use of electronic cigarettes (e-cigarettes) among adolescents has increased since their introduction into the US market in 2007. Little is known about the role of e-cigarette psychosocial factors on risk of e-cigarette or cigarette use in adolescence.

METHODS

Information on e-cigarette and cigarette psychosocial factors (use and attitudes about use in the home and among friends) was collected from 11th- and 12th-grade participants in the Southern California Children's Health Study during the spring of 2014.

RESULTS

Of 2084 participants, 499 (24.0%) had used an e-cigarette, including 200 (9.6%) current users (past 30 days); 390 participants (18.7%) had smoked a combustible cigarette, and 119 (5.7%) were current cigarette smokers. Cigarette and e-cigarette use were correlated. Nevertheless, 40.5% (n = 81) of current e-cigarette users had never smoked a cigarette. Psychosocial factors (home use of each product, friends' use of and positive attitudes toward e-cigarettes and cigarettes) and participant perception of the harm of e-cigarettes were strongly positively associated both with e-cigarette and cigarette use. Most youth who reported e-cigarette use had friends who used e-cigarettes, and almost half of current users reported that they did not believe there were health risks associated with e-cigarette use.

CONCLUSIONS

Longitudinal studies of adolescents are needed to determine whether the strong association of e-cigarette psychosocial factors with both e-cigarette and cigarette use will lead to increased cigarette use or dual use of cigarettes and e-cigarettes, or whether e-cigarettes will serve as a gateway to cigarette use.

Fighting With Siblings and With Peers Among Urban High School Students

Understanding the determinants of fighting is important for prevention efforts. Unfortunately, there is little research on how sibling fighting is related to peer fighting. Therefore, the aim of this study was to evaluate the association between sibling fighting and peer fighting. Data are from the Boston Youth Survey 2008, a school-based sample of youth in Boston, MA. To estimate the association between sibling fighting and peer fighting, we ran four multivariate regression models and estimated adjusted prevalence ratios and 95% confidence intervals. We fit generalized estimating equation models to account for the fact that students were clustered within schools. Controlling for school clustering, race/ethnicity, sex, school failure, substance use, and caregiver aggression, youth who fought with siblings were 2.49 times more likely to have reported fighting with peers. To the extent that we can confirm that sibling violence is associated with aggressive behavior, we should incorporate it into violence prevention programming.

Perceived risk and condom use among adolescents in sub-Saharan Africa: a latent class analysis

While studies have examined factors associated with condom use behaviors, few have assessed risk perception and condom use among SSA adolescents. This work sought to fill this gap. Data for this study are derived from the National Survey of Adolescents (2004-2005), a survey of 12-19 year olds in Burkina Faso, Ghana, Malawi, and Uganda. Latent class analysis was conducted. Two risk profiles emerged in the data. Percent of adolescents in the high perceived risk sub-groups ranged from 49.6% - 90.8% and moderate risk from 29.1-83.8%. Those in the high perceived risk class were 2.26 to 5.10 times more likely to report condom use. The findings provide additional information about the factors that influence condom use among adolescents in these four countries. Because an individual's development is influenced by his/her settings and systems, expanded views of behavior must be included in prevention efforts.

Prevalence and correlates of victimization and weapon carrying among military- and nonmilitary-connected youth in Southern California

OBJECTIVES

The present analysis sought to explore the normative rates and correlates of school victimization and weapon carrying among military-connected and nonmilitary-connected youth in public schools in Southern California.

METHODS

Data are from a sub-sample of the 2011 California Healthy Kids Survey (N=14,512). Items to assess victimization and weapon carrying were separated into three categories: physical acts (e.g., being pushed or shoved), nonphysical acts (e.g., having rumors spread about them) and weapon carrying.

RESULTS

The bivariate results indicate that youth with a military-connected parent had higher rates of physical victimization (56.8%), nonphysical victimization (68.1%), and weapon carrying (14.4%) compared to those with siblings serving (55.2%, 65.2%, and 11.4%, respectively) and nonmilitary-connected (50.3%, 61.6%, and 8.9%, respectively) youth. Having a parent in the military increased the odds of weapon carrying by 29% (Odds Ratio=1.29, 95% confidence interval=1.02-1.65). Changing schools and a larger number of family member deployments in the past 10years were associated with significant increases in the likelihood of victimization and weapon carrying.

CONCLUSIONS

The results of this analysis warrant a focus on school supports for youth experiencing parental military service, multiple relocations and deployments of a family member.

Substance Use Among Adolescents in California: A Latent Class Analysis

Data from the California Healthy Kids Survey of 7th, 9th, and 11th graders were used to identify latent classes/clusters of alcohol, tobacco, and marijuana use (N = 418,702). Analyses revealed four latent classes of substance use, which included nonusers (61.1%), alcohol experimenters (some recent alcohol use; 22.8%), mild polysubstance users (lifetime use of all substances with less than 3 days of recent use; 9.2%), and frequent polysubstance users (used all substances three or more times in the past month; 6.9%). The results revealed that alcohol and marijuana use are salient to California adolescents. This information can be used to target and tailor school-based prevention efforts.

Gang membership of California middle school students: behaviors and attitudes as mediators of school violence

Empirical evidence examining how risk and protective behaviors may possibly mediate the association between gang membership and school violence is limited. This study utilizes a statewide representative sample of 152 023 Latino, Black and White seventh graders from California to examine a theoretical model of how school risk (e.g. truancy, school substance use and risky peer approval) and protective (e.g. connectedness, support and safety) behaviors and attitudes mediate the effects of gang membership on school violence behaviors. The dataset was collected in the 2005-2006 and 2006-2007 academic school years using the ongoing large-scale California Healthy Kids Survey conducted by WestEd for the State of California. Approximately 9.5% of the sample considered themselves to be a member of a gang. The findings indicate that school risk behaviors and attitudes mediate the association between gang membership and school violence behaviors. Although the direct negative association between gang membership and school violence perpetration is weak, the positive indirect effect mediated by school risks behaviors and attitudes is strong. This indicates that when gang members engage in school risk behaviors, they are much more likely to be school violence perpetrators. Implications for further research, theory and practice for both gang and school violence researchers are discussed.

Influence of caregiver network support and caregiver psychopathology on child mental health need and service use in the LONGSCAN study

Using structural equation modeling, this study examined the relationship of caregiver network support on caregiver and child mental health need, as well as child mental health service use among 1075 8-year-old children participating in the LONGSCAN study. The final model showed acceptable fit (χ(2) = 301.476, df = 136, p<0.001; RMSEA = 0.052; CFI = 0.95). Caregiver and child mental health needs were positively related. As predicted, caregiver network support exerted a protective effect, with greater levels of caregiver network support predictive of lower caregiver and child need. Contrary to prediction, however, caregiver network support was not directly related to child service use. Higher child need was directly related to child service use, especially among children whose caregivers had mental health problems. The findings appear to indicate that lower levels of caregiver network support may exert its impact on child service use indirectly by increasing caregiver and child need, rather than by directly increasing the likelihood of receiving services, especially for African American children.

Tobacco use and suicidality: latent patterns of co-occurrence among black adolescents

INTRODUCTION

Tobacco use is responsible for a considerable portion of the health disparities experienced by Blacks. In addition to its physiological effects, tobacco use has consistently been shown to be associated with suicidality among adolescents. The purpose of the present study is to describe the association between specific patterns of tobacco use behavior and those of suicidality among a nationally representative sample of Black high school students.

METHODS

Responses from Black adolescents (N = 2,931) who completed the 2007 Youth Risk Behavior Surveillance Survey were included. Latent class analysis was utilized to identify typologies based on two domains: (a) lifetime and recent tobacco use and (b) suicidality. The association between tobacco use class membership and suicidality class membership as well as the direct effects of age and gender on class membership was also investigated.

RESULTS

A significant proportion of Black youth reported smoking and suicidal behaviors. A 4-class model provided the best overall fit to the data for tobacco use behavior (nonsmokers, former smokers, light current smokers, and frequent current smokers); a 3-class model provided the best overall fit to the data for suicidality (not suicidal, mild suicidality, suicidal). Smoking status was associated with suicidality class membership, with more intense patterns of smoking related to increased probability of identification with both mild suicidality and being classified as suicidal compared with not suicidal.

CONCLUSIONS

The results of this study indicate that any current smoking status increases the likelihood of suicidality at least 5-fold and provides incentive to target this specific portion of the population of Black adolescent smokers.

Social-ecological influences on patterns of substance use among non-metropolitan high school students

Patterns of substance use are examined in a sample of over 1,200 youth in a non-metropolitan region of New England. Self-reported history and frequency of alcohol, tobacco, marijuana, inhalants, pain medications, and other hard drug use was assessed for 9th and 10th grade students. Latent class analyses identified four patterns of substance use: non-users (22%), alcohol experimenters (38%), occasional polysubstance users (29%), and frequent polysubstance users (10%). Contextual risk and protective factors in the individual, family, peer, and community domains predicted substance use patterns. Youth report of peer substance use had the largest effects on substance use class membership. Other individual characteristics (e.g., gender, antisocial behavior, academic performance, perceived harm from use), family characteristics (e.g., parental drinking, parental disapproval of youth use), and community characteristics (e.g., availability of substances) demonstrated consistent effects on substance use classes. Implications for prevention are discussed from a social-ecological perspective.

A multiprocess latent class analysis of the co-occurrence of substance use and sexual risk behavior among adolescents

OBJECTIVE

To assess the co-occurrence of patterns of adolescent substance use and sexual behavior and test for potential moderating effects of gender.

METHOD

The 2005 Youth Risk Behavior Survey sample of 13,953 high school students was used in this study. Latent class analyses were conducted to examine the relation between patterns of substance use and sexual risk behavior in a nationally representative adolescent sample. The final model controlled for demographic covariates and an interaction between gender and substance-use classes.

RESULTS

Four class solutions to each behavior provided optimal fit. Substance-use classes were nonusers (27%), alcohol experimenters (38%), occasional polysubstance users (23%), and frequent polysubstance users (13%). Sexual risk classes were abstainers (53%), monogamous (15%), low-frequency multipartner (18%), and high-frequency multipartner (14%). Substance-use class had a strong association with sexual behavior even when controlling for demographic covariates. African Americans were less likely to use substances but more likely to report sexual risk behaviors. interaction effects indicated that substance use among females was associated with increased sexual risk behavior compared with males.

CONCLUSIONS

The results of this study provide impetus for the development of preventive interventions targeting change across multiple risk behaviors, and also for targeting particular high-risk groups or patterns of behavior.

Associations between maternal closeness, suicidal ideation, and risk behaviors in Cape Town

Health compromising behaviors, such as smoking and other risk behaviors tend to co-occur, and contribute to the leading causes of preventable death, disease, and disability among adolescents and young adults worldwide. The present study assesses a model of the direct and indirect effects of maternal closeness with suicidal ideation on smoking and risky behaviors. The sample consisted of 657 South African youth assessed over two time points with comparison of effects between males and females. Maternal closeness had a significant effect on suicidal ideation among females. Suicidal ideation had a significant effect on risky behaviors among males and lifetime smoking among females. A significant indirect effect was observed from maternal closeness to lifetime smoking among females. These results indicate that suicidal ideation is associated with lifetime smoking and risky behaviors separately among male and female adolescents and highlight the need to focus on the development of mental health preventive interventions for both genders.

How many imputations are really needed? Some practical clarifications of multiple imputation theory

Multiple imputation (MI) and full information maximum likelihood (FIML) are the two most common approaches to missing data analysis. In theory, MI and FIML are equivalent when identical models are tested using the same variables, and when m, the number of imputations performed with MI, approaches infinity. However, it is important to know how many imputations are necessary before MI and FIML are sufficiently equivalent in ways that are important to prevention scientists. MI theory suggests that small values of m, even on the order of three to five imputations, yield excellent results. Previous guidelines for sufficient m are based on relative efficiency, which involves the fraction of missing information (gamma) for the parameter being estimated, and m. In the present study, we used a Monte Carlo simulation to test MI models across several scenarios in which gamma and m were varied. Standard errors and p-values for the regression coefficient of interest varied as a function of m, but not at the same rate as relative efficiency. Most importantly, statistical power for small effect sizes diminished as m became smaller, and the rate of this power falloff was much greater than predicted by changes in relative efficiency. Based our findings, we recommend that researchers using MI should perform many more imputations than previously considered sufficient. These recommendations are based on gamma, and take into consideration one's tolerance for a preventable power falloff (compared to FIML) due to using too few imputations.

How many imputations are really needed? Some practical clarifications of multiple imputation theory

Multiple imputation (MI) and full information maximum likelihood (FIML) are the two most common approaches to missing data analysis. In theory, MI and FIML are equivalent when identical models are tested using the same variables, and when m, the number of imputations performed with MI, approaches infinity. However, it is important to know how many imputations are necessary before MI and FIML are sufficiently equivalent in ways that are important to prevention scientists. MI theory suggests that small values of m, even on the order of three to five imputations, yield excellent results. Previous guidelines for sufficient m are based on relative efficiency, which involves the fraction of missing information (gamma) for the parameter being estimated, and m. In the present study, we used a Monte Carlo simulation to test MI models across several scenarios in which gamma and m were varied. Standard errors and p-values for the regression coefficient of interest varied as a function of m, but not at the same rate as relative efficiency. Most importantly, statistical power for small effect sizes diminished as m became smaller, and the rate of this power falloff was much greater than predicted by changes in relative efficiency. Based our findings, we recommend that researchers using MI should perform many more imputations than previously considered sufficient. These recommendations are based on gamma, and take into consideration one's tolerance for a preventable power falloff (compared to FIML) due to using too few imputations.