Soeun Kim

Psychometric Properties of the UCLA PTSD Reaction Index: Part I

This article presents psychometric characteristics of the UCLA PTSD Reaction Index for DSM IV (PTSD-RI) derived from a large sample of children and adolescents (N = 6,291) evaluated at National Child Traumatic Stress Network centers. Overall mean total PTSD-RI score for girls was significantly higher as compared with boys. Age-related differences were found in that overall mean total PTSD-RI scores and within sex groups were higher among those aged 7-9 years and 16-18 years. There were no significant differences in mean total PTSD-RI scores across racial/ethnic groups. The PTSD-RI total scale displayed good to excellent internal consistency reliability across age ranges, sex, and racial/ethnic groups (α = .88-.91). Correlations of PTSD-RI scores with PTS subscale scores on the TSCC-A for the entire sample and within sex, age, and ethnic/racial groups provided evidence of convergent validity, although not discriminant validity. In contradistinction to previously reported 4-factor models, an exploratory factor analysis revealed 3 factors that mostly reflected the underlying dimensions of PTSD in DSM IV. PTSD-RI scores were associated with increased odds ratios for functional/behavior problems (odds ratio [OR] = 1-1.80). These findings are striking in light of the wide range of trauma exposures, age, and race/ethnicity among subjects.

Randomized, double-blind, placebo-controlled trial of modafinil for the treatment of methamphetamine dependence.

OBJECTIVE To compare modafinil to placebo for reducing methamphetamine (MA) use, improving retention, and reducing depressive symptoms and MA cravings. Rates of adverse events and cigarette smoking with modafinil versus placebo were also compared. METHODS Following a 2-week, non-medication lead-in period, 71 treatment-seeking MA-dependent participants were randomly assigned to modafinil (400mg once daily; N=34) or placebo (once daily; N=37) for 12 weeks under double-blind conditions. Participants attended clinic thrice-weekly to provide urine samples analyzed for MA-metabolite, to complete research assessments, and to receive contingency management and weekly cognitive behavioral therapy (CBT) sessions. RESULTS There were no statistically significant effects for modafinil on MA use, retention, depressive symptoms, or MA cravings in pre-planned analyses. Outcomes for retention and MA use favored modafinil in a post hoc analysis among participants with low CBT attendance and among participants with baseline high-frequency of MA use (MA use on >18 of past 30 days), but did not reach statistical significance in these small subgroups. Modafinil was safe and well tolerated and did not increase cigarette smoking. CONCLUSIONS Modafinil was no more effective than placebo at 400mg daily in a general sample of MA users. A post hoc analysis showing a trend favoring modafinil among subgroups with baseline high-frequency MA use and low CBT attendance suggests that further evaluation of modafinil in MA users is warranted.

Trauma history and psychopathology in war-affected refugee children referred for trauma-related mental health services in the United States.

There is an increasing need to deliver effective mental health services to refugee children and adolescents across the United States; however, the evidence base needed to guide the design and delivery of services is nascent. We investigated the trauma history profiles, psychopathology, and associated behavioral and functional indicators among war-affected refugee children presenting for psychological treatment. From the National Child Traumatic Stress Networks Core Data Set, 60 war-affected refugee children were identified (51.7% males, mean age = 13.1 years, SD = 4.13). Clinical assessments indicated high rates of probable posttraumatic stress disorder (30.4%), generalized anxiety (26.8%), somatization (26.8%), traumatic grief (21.4%), and general behavioral problems (21.4%). Exposure to war or political violence frequently co-occurred with forced displacement; traumatic loss; bereavement or separation; exposure to community violence; and exposure to domestic violence. Academic problems and behavioral difficulties were prevalent (53.6% and 44.6%, respectively); however, criminal activity, alcohol/drug use, and self-harm were rare (all < 5.45%). These findings highlight the complex trauma profiles, comorbid conditions, and functional problems that are important to consider in providing mental health interventions for refugee children and adolescents. Given the difficulties associated with access to mental health services for refugees, both preventive and community-based interventions within family, school, and peer systems hold particular promise.

Effects of the child–perpetrator relationship on mental health outcomes of child abuse: It's (not) all relative

The present study was conducted to better understand the influence of the child-perpetrator relationship on responses to child sexual and physical trauma for a relatively large, ethnically diverse sample of children and youth presenting for clinical evaluation and treatment at child mental health centers across the United States. This referred sample includes 2,133 youth with sexual or physical trauma as their primary treatment focus. Analyses were conducted to ascertain whether outcomes were dependent on the perpetrators status as a caregiver vs. non-caregiver. Outcome measures included psychiatric symptom and behavior problem rating scales. For sexual trauma, victimization by a non-caregiver was associated with higher posttraumatic stress, internalizing and externalizing behavior problems, depression, and dissociation compared to youth victimized by a caregiver. For physical trauma, victimization by a non-caregiver was also associated with higher posttraumatic symptoms and internalizing behavior problems. The total number of trauma types experienced and age of physical or sexual trauma onset also predicted several outcomes for both groups, although in disparate ways. These findings are consistent with other recent studies demonstrating that perpetration of abuse by caregivers results in fewer symptoms and problems than abuse perpetrated by a non-caregiving relative. Thus, clinicians should not make a priori assumptions that children and adolescents who are traumatized by a parent/caregiver would have more severe symptoms than youth who are traumatized by a non-caregiver. Further exploration of the role of the perpetrator and other trauma characteristics associated with the perpetrator role is needed to advance our understanding of these findings and their implications for clinical practice.

Does McNemar’s test compare the sensitivities and specificities of two diagnostic tests?

McNemar’s test is often used in practice to compare the sensitivities and specificities for the evaluation of two diagnostic tests. For correct evaluation of accuracy, an intuitive recommendation is to test the diseased and the non-diseased groups separately so that the sensitivities can be compared among the diseased, and specificities can be compared among the healthy group of people. This paper provides a rigorous theoretical framework for this argument and study the validity of McNemar’s test regardless of the conditional independence assumption. We derive McNemar’s test statistic under the null hypothesis considering both assumptions of conditional independence and conditional dependence. We then perform power analyses to show how the result is affected by the amount of the conditional dependence under alternative hypothesis.

Risk factors associated with psychiatric readmission.

Abstract The present study focused on identifying risk factors for early readmission of patients discharged from an urban community hospital. Retrospective chart reviews were conducted on 207 consecutive inpatient psychiatric admissions that included patients who were readmitted within 15 days, within 3 to 6 months, and not admitted for at least 12 months post-discharge. Findings indicated that a diagnosis of schizophrenia/schizoaffective disorder (OR = 18; 95% CI 2.70–117.7; p < 0.05), history of alcohol abuse (OR = 9; 95% CI 1.80–40.60; p < 0.05), number of previous psychiatric hospitalizations (OR = 2; 95% CI 1.28–3.73; p < 0.05), and type of residence at initial admission (e.g., homeless, OR = 29; 95% CI 3.99–217; p < 0.05) were significant risk factors for early readmission, where OR compares readmission group 1 versus group 3 in the multinomial logistic regression. Initial positive urine drug screen, history of drug abuse or incarceration, and legal status at initial admission did not predict early readmission. Reducing the risk factors associated with psychiatric readmissions has the potential to lead to the identification and development of preventative intervention strategies that can significantly improve patient safety, quality of care, well-being, and contain health care expenditures.

Evaluating model-based imputation methods for missing covariates in regression models with interactions

Imputation strategies are widely used in settings that involve inference with incomplete data. However, implementation of a particular approach always rests on assumptions, and subtle distinctions between methods can have an impact on subsequent analyses. In this research article, we are concerned with regression models in which the true underlying relationship includes interaction terms. We focus in particular on a linear model with one fully observed continuous predictor, a second partially observed continuous predictor, and their interaction. We derive the conditional distribution of the missing covariate and interaction term given the observed covariate and the outcome variable, and examine the performance of a multiple imputation procedure based on this distribution. We also investigate several alternative procedures that can be implemented by adapting multivariate normal multiple imputation software in ways that might be expected to perform well despite incompatibilities between model assumptions and true underlying relationships among the variables. The methods are compared in terms of bias, coverage, and CI width. As expected, the procedure based on the correct conditional distribution performs well across all scenarios. Just as importantly for general practitioners, several of the approaches based on multivariate normality perform comparably with the correct conditional distribution in a number of circumstances, although interestingly, procedures that seek to preserve the multiplicative relationship between the interaction term and the main-effects are found to be substantially less reliable. For illustration, the various procedures are applied to an analysis of post-traumatic stress disorder symptoms in a study of childhood trauma.

Predictors of Functional Decline in Early Parkinson’s Disease: NET-PD LS1 Cohort

BACKGROUND Data on predictors of decline in PD are largely based on de-novo populations and limited to the use of motor outcomes that fail to capture the full scope of disease. OBJECTIVE Determine the clinical predictors of decline in early treated PD using a novel multi-domain measure. METHODS Data from NINDS Exploratory Trials in PD Long-Term Study 1 (NET-PD LS1), a multicenter Phase 3 study of creatine in early treated PD, were analyzed. Functional decline was defined by a global outcome metric (GO) that consisted of: Schwab and England ADL scale, PD 39-item Questionnaire, Unified PD Rating Scale, Ambulatory Capacity Score, Symbol Digit Modalities Test, and Modified Rankin Scale. Univariate and multivariate models were used to test the association of predictors of interest with a standardized rank-sum of the GO. RESULTS 765 of 1741 participants completed five-year assessments and were included. Older age at disease onset (p <  0.0001), higher baseline levodopa equivalent dose (p = 0.01), and worse Scales for Outcomes of Parkinsons Disease Cognition score (p = 0.001) at baseline were the strongest predictors of functional decline in multivariate analysis. PD symptom subtype was not a significant predictor of outcome (p = 0.42). The full model was only a modest predictor of change in GO (R2 = 0.186). CONCLUSIONS This is the largest study to systematically assess predictors of functional decline in early treated PD over several years, and the first to use a multi-domain outcome measure of decline. Older age at disease onset and worse cognition, and not PD subtype, were predictors of decline.

Multiple imputation with non-additively related variables: Joint-modeling and approximations

This paper investigates multiple imputation methods for regression models with interacting continuous and binary predictors when continuous variable may be missing. Usual implementations for parametric multiple imputation assume a multivariate normal structure for the variables, which is not satisfied for a binary variable nor its interaction with a continuous variable. To accommodate interactions, missing covariates are multiply imputed from conditional distribution in a manner consistent with the joint model. Alternative imputation methods under multivariate normal assumptions are also considered as candidate approximations and evaluated in a simulation study. The results suggest that the joint modeling procedure performs generally well across a wide range of scenarios and so do the approximation methods that incorporate interactions in the model appropriately by stratification. It is critical to include interactions in the imputation model as failure to do so may result in low coverage and bias. We apply the joint modeling approach and approximation methods in the study of childhood trauma with gender × trauma interaction.

Sensitivity analysis on the ecological bias for Seoul tuberculosis data

In ecological studies, researchers often try to convey the analysis results to individual level based on aggregate data. In order to do this correctly, the possibility of ecological bias should be studied and addressed. One of the key ideas used to address the ecological bias issue is to derive the ecological model from the individual model and to check whether the parameter of interest in the individual model is identifiable in the ecological model. However, the procedure depends on unverifiable assumptions, and we recommend checking how sensitive the results are to these unverifiable assumptions. We analyzed the tuberculosis data that was collected in Seoul in 2005 using a spatial ecological regression model for the aggregate count data with spatial correlation, and found that the deprivation index is likely to have a small positive effect on the occurrence risk of tuberculosis in individual level in Seoul. We considered this finding in various aspects by performing in depth sensitivity analyses. In particular, our findings are shown to be robust to the distribution assumptions for the individual exposure and missing binary covariate across various scenarios.