Functional response models for intraclass correlation coefficients

Comparison of Posttraumatic Stress Disorder Checklist Instruments From Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition vs Fifth Edition in a Large Cohort of US Military Service Members and Veterans

IMPORTANCE

The definition of posttraumatic stress disorder (PTSD) changed markedly between the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) and DSM-5, creating challenges for studies and in medical settings spanning this transition.

OBJECTIVE

To evaluate the ability to compare and assess PTSD, based on DSM-IV and DSM-5 criteria, using PTSD Checklists (PCLs).

DESIGN, SETTING, AND PARTICIPANTS

This diagnostic study was conducted with survey data collected in October 2019, from the Millennium Cohort Study, a population-based US military cohort study. The population for the present study was restricted to a subset of initial web responders of the 2019 survey cycle, randomly assigned to 1 of 4 survey groups.

EXPOSURES

Each group received the DSM-IV and DSM-5 PCL (PCL-Civilian [PCL-C] version and PCL for DSM-5 [PCL-5]). PCL instruments were counterbalanced to control for order effects.

MAIN OUTCOMES AND MEASURES

Survey data were used to assess PTSD (using the PCL-C and PCL-5), major depressive disorder (using the Patient Health Questionnaire), generalized anxiety (using the Generalized Anxiety Disorder scale), and problem drinking (using the Patient Health Questionnaire). Demographic and military characteristics included age, sex, race/ethnicity, marital status, education, service branch, pay grade, enrollment panel, and military service status.

RESULTS

Among the 1921 participants (mean [SD] age, 50.1 [12.5] years), 1358 (70.7%) were men, 1638 (85.3%) were non-Hispanic White individuals, 1440 (75.0%) were married, and 1190 (61.9%) had at least a bachelor's degree; 295 (15.4%) had probable PTSD according to DSM-IV criteria with PCL-C compared with 286 (14.9%) using DSM-5 criteria with PCL-5 (κ = 0.77). There was substantial agreement between PCLs for probable PTSD based on DSM-IV criteria (295 [15.4%] with PCL-C; 316 [16.4%] with PCL-5; κ = 0.80) and DSM-5 criteria (286 [14.9%] with PCL-5; 258 [13.4%] with PCL-C; κ = 0.77). Estimated PTSD sum scores showed excellent agreement with observed scores. Using an established crosswalk, PCL-5 sum scores estimated with the PCL-C were similar to observed PCL-5 scores. Of the 17 corresponding items between the 2 instruments, 16 had substantial agreement. Appending 2 additional PCL-C items to the PCL-5 did not significantly alter estimates. The PCL-C and PCL-5 had nearly identical associations with comorbid conditions.

CONCLUSIONS AND RELEVANCE

The findings of this diagnostic study suggest that PTSD can be successfully assessed and compared over time with either PCL instrument in veteran and military populations.

A semiparametric marginalized zero-inflated model for analyzing healthcare utilization panel data with missingness

Zero-inflated count outcomes arise quite often in research and practice. Parametric models such as the zero-inflated Poisson and zero-inflated negative binomial are widely used to model such responses. However, interpretations of those models focus on the at-risk subpopulation of a two-component population mixture and fail to provide direct inference about marginal effects for the overall population. Recently, new approaches have been proposed to facilitate such marginal inferences for count responses with excess zeros. However, they are likelihood based and impose strong assumptions on data distributions. In this paper, we propose a new distribution-free, or semiparametric, alternative to provide robust inference for marginal effects when population mixtures are defined by zero-inflated count outcomes. The proposed method also applies to longitudinal studies with missing data following the general missing at random mechanism. The proposed approach is illustrated with both simulated and real study data.

Distribution-free models for latent mixed population responses in a longitudinal setting with missing data

Many biomedical and psychosocial studies involve population mixtures, which consist of multiple latent subpopulations. Because group membership cannot be observed, standard methods do not apply when differential treatment effects need to be studied across subgroups. We consider a two-group mixture in which membership of latent subgroups is determined by structural zeroes of a zero-inflated count variable and propose a new approach to model treatment differences between latent subgroups in a longitudinal setting. It has also been incorporated with the inverse probability weighted method to address data missingness. As the approach builds on the distribution-free functional response models, it requires no parametric distribution model and thereby provides a robust inference. We illustrate the approach with both real and simulated data.

Variable selection for distribution-free models for longitudinal zero-inflated count responses

Zero-inflated count outcomes arise quite often in research and practice. Parametric models such as the zero-inflated Poisson and zero-inflated negative binomial are widely used to model such responses. Like most parametric models, they are quite sensitive to departures from assumed distributions. Recently, new approaches have been proposed to provide distribution-free, or semi-parametric, alternatives. These methods extend the generalized estimating equations to provide robust inference for population mixtures defined by zero-inflated count outcomes. In this paper, we propose methods to extend smoothly clipped absolute deviation (SCAD)-based variable selection methods to these new models. Variable selection has been gaining popularity in modern clinical research studies, as determining differential treatment effects of interventions for different subgroups has become the norm, rather the exception, in the era of patent-centered outcome research. Such moderation analysis in general creates many explanatory variables in regression analysis, and the advantages of SCAD-based methods over their traditional counterparts render them a great choice for addressing this important and timely issues in clinical research. We illustrate the proposed approach with both simulated and real study data. Copyright © 2016 John Wiley & Sons, Ltd.