The average age of the sample was 37 years (SD=11). The sample was almost evenly split between women and men (52% male, n=94). Almost a third of participants were married (32%, n= 58). The mean years of school completed was 13 (SD=3.2). The mean Hollingshead index was 52 (SD=15) which suggests the average SES of participants fell in the middle-lower class range; 48% (n=87) of the sample were currently employed. Thirty-one percent (n=56) reported military service. Participants largely identified themselves as African-American (62%, n=114) or Caucasian (31%, n=57) with a minority identifying as either Asian, Native American, or multi-racial (5%, n=10).
Trauma exposure was highly prevalent in the sample. Results from the TLEQ indicated that almost all participants (98%, n=181) had been exposed to one or more traumatic events that resulted in “fear, helplessness, or horror.” Forty-eight percent (n=79) of the sample reported exposure to some form of childhood physical assault and 26% (n=43) reported experiencing childhood sexual assault. The most commonly endorsed TLEQ trauma categories included learning of the unexpected death of close friend/loved one or survival of a life-threatening accident or illness (82%, n=137), exposure to a natural disaster or serious accident (66%, n=111), and exposure to adult physical violence (66%, n=110). Fifteen percent (n=31) of the sample endorsed adult sexual trauma and 24% (n=31) reported service in a warzone.
A categorization of each individual’s index traumatic events is provided in . Consistent with the results from the TLEQ, almost all participants (98%, n=181) described an event that met DSM-IV Criterion A1. Of these, almost all (97%, n=175) also met DSM-IV Criterion A2 resulting in 95% (n=175) of the sample endorsing a DSM-IV Criterion A traumatic event. Each participant’s traumatic event was rated for whether it would qualify as a DSM-5 Criterion A event. In total, 7% (n=12) of those who met DSM-IV Criterion A reported events that would no longer qualify under the new stressor criterion. Only one individual (a veteran who reported combat trauma but denied fear, helplessness or horror) who did not meet DSM-IV Criterion A met the new DSM-5 criterion. Thus, a total of 89% of sample reported an event that met DSM-5 Criterion A. Results of the CAPS interview revealed that approximately half (n=93) of participants met DSM-IV criteria for PTSD resulting in a base rate of 50%. As shown in , only 3 individuals who met full DSM-IV criteria for PTSD, did not meet DSM-5 Criterion A while meeting DSM-IV criteria B, C, and D.
Trauma exposure reported for the CAPS interview and DSM-IV PTSD prevalence.
Among those with PTSD, depression and other anxiety disorders were a common comorbidity. As many as 33% (n=31) met criteria for comorbid major depressive disorder and 29% met criteria for another DSM-IV anxiety disorder. Psychopathology was less prevalent among those without PTSD, where only 13% (n=11) met criteria for any Axis I disorder. Anxiety disorders accounted for the majority of psychopathology among non-PTSD participants (82%, n=8).
displays the rate of endorsement of each of the DSM-5 symptoms and the correlations of each symptom with DSM-5 symptom clusters and total PTSD symptom score. Overall, each symptom was more highly correlated with its purported symptom cluster than other symptom clusters. The majority of symptoms were highly correlated (21
) with their respective symptom clusters, however, symptoms D1 (inability to recall important aspects of the trauma) and E2 (reckless or self-destructive behavior) only showed moderate item-cluster correlations (see ).
DSM-5 Symptom endorsement and Item-Cluster correlations
The proposed DSM-5 criteria require indication of active avoidance. Almost all participants who were diagnosed with DSM-IV PTSD endorsed active avoidance and thus met DSM-5 Criterion C (99%, n=92). displays the observed prevalence rates for various DSM-5 diagnostic algorithms where the number of symptoms required for Criterion D and E were varied simultaneously while Criteria A, B, C, F, G, and H were fixed as currently proposed. Results indicate that several of the permutations resulted in good concordance (AUCs > .90, Kappa > .8) between classification methods (see ). The currently suggested algorithm of three or more D symptoms and three or more E symptoms resulted in an acceptable AUC of .91 but was characterized by a sensitivity of only .85 and resulted in a change in the prevalence rate from the DSM-IV observed 50% down to 44% Two other algorithms produced more balanced sensitivity and specificity and slightly better AUC and kappa values. For example, while it was not statistically significantly different from the algorithm requiring three D and three E symptoms (z = 0.61, n.s.), the algorithm requiring only 2 D and E symptoms produced an AUC of .93 (sensitivity = .95, specificity = .91) and was associated with a slight increase in the prevalence rate (i.e., 52% vs 50% observed using DSM-IV). Similarly, the algorithm requiring three D symptoms and 2 E symptoms resulted in an AUC of .93, provided a nice balance between sensitivity (.92) and specificity (.93), and a less than 1% decrease in observed prevalence compared to the DSM-IV base rate of 50% (see ).
Prevalence and diagnostic efficiency indices of proposed DSM-5 PTSD criteria