Evaluation of the Psychometric Properties and Utility of Abbreviated Forms of the Truama Symptom Checklist (TSC-40) Among a Sample of Women in Treatment for Substance Use

Background & Purpose: Trauma symptoms limit functioning and influence outcomes across domains of living. It is important for research and intervention to utilize valid and current measures. This study examined the Trauma Symptom Checklist-40 (TSC-40), a self-report assessment of symptomatology in adults associated with traumatic experiences. The purpose of this investigation was to evaluate the factor structure, identify and examine the utility of an abbreviated form, and assess the content and predictive validity of the TSC-40 using advanced statistical methods.

Methods: This analysis included 361 participants in a NIDA-funded longitudinal study examining personal support networks among women in treatment for a substance use disorder. Frequency of trauma symptomatology over the previous two months was assessed at one week post-treatment intake using a 4-point scale (0 = never to 3 = often) with a possible range of 0-120. The TSC-40 in its original form has six sub-scales: Dissociation (“spacing out—or going away in your mind”), Anxiety (“feeling tense all the time”), Depression (“uncontrollable crying”), Sexual Abuse Trauma Index (“Flashbacks”), Sexual Problems (“Low sex drive”), and Sleep Disturbances (“Insomnia”). Exploratory factor analyses (EFA) using principal axis factoring (oblique rotation) were conducted to identify the underlying factor structure—number of factors was determined based on a scree plot, magnitude of factor loadings (≥ .40), and theory. Confirmatory factor analyses (CFA) were conducted using Mplus 7.11 to evaluate the model fit of the factor structure derived from EFA. Items selected for inclusion in the abbreviated form were based on factor loadings and relevant theory. Validity was examined using structural equation modeling (SEM) among these variables as criteria, and controlling for relevant covariates: substance use 12-months post-treatment intake, co-morbid diagnoses, abstinence support and exposure to violence (ETV).

Results: The sample had a mean age of 36.5 (SD=10.3), and was predominantly African American (60%), never married (65%), and reliant on government assistance (73%), with 40% reporting less than a high school education. A majority of women indicated receiving previous treatment for substance abuse (73%). EFA identified 6 subscales from the full 40 items: Dissociation, Anxiety, Depression, Abuse, Sleep Disturbance, and Sexual Problems. CFA using the 6 composite sub-scale mean scores yielded a parsimonious model: χ2=8.75, df=4, p=.07; CFI=.995; RMSEA=.06. The abbreviated form utilized 20 items on 5 subscales ([subscale1], [subscale2], [subscale3], [subscale4], [subscale5], 4 items drawn from each sub-scale) with a CFA model indicating excellent fit: χ2=7.67, df=5, p=.18; CFI=.995; RMSEA=.04. SEM results indicated that both content and predictive validity for the multiple models of the TSC were supported; trauma was associated with the presence of a dual disorder, ETV and substance use post-treatment intake, with acceptable model fit χ2=48.73, df=33, p=.04; CFI=.96; RMSEA=.04, WRMR=.68.

Implications: Results support the use of an abbreviated measure of trauma symptomatology, as well as the use of composite sub-scale scores in future analyses. The utilization of a brief measure of trauma symptomatology could improve research through the reduction of administration time, measurement error and missing data. Furture research may further refine item reduction.