66. Screening for PTSD in Female VA Patients: Validation of the PTSD Checklist
DJ Dobie, MIRECC, VA Puget Sound Health Care System and Univ of Washington School of Medicine; DR Kivlahan, CESATE, VA PSHCS and U W School of Medicine; M McFall, MIRECC, VA PSHCS and UW School of Medicine; KR Bush, HSRDS, VAPSHCS; C Maynard, HSRDS, VAPSHCS and UW School of Medicine; AJ Epler, HSRDS, VAPSHCS; KA Bradley, HSRDS, VAPSHCS and UW School of Medicine
Objectives: Posttraumatic Stress Disorder (PTSD) is associated with increased physical health problems and diminished quality of life. Women are at greater risk than men for developing PTSD. Female veterans who receive VA care report high levels of lifetime trauma exposure, and preliminary data suggest correspondingly high rates of PTSD. However, efficient screening instruments for PTSD have not been validated in veteran women. In this study we evaluated the diagnostic utility of a self-report measure to screen for PTSD in female VA patients.
Methods: As part of the longitudinal Veteran Women’s Alcohol Problems (VWAP) study, all women who received care at the VA Puget Sound Health Care system from October 1996- January 2000 were invited by mail to complete an annual women’s health survey and to participate in a face-to-face research interview. Recruitment materials described the interview as concerning women's health habits, drug and alcohol use, pain, and mental health. Invitations were randomly staggered to allow for prompt scheduling, and participants received $35.00 upon completion. Three hundred-and-one respondents completed the 17-item self-report PTSD Checklist (PCL), followed by a structured diagnostic interview for PTSD, the CAPS (Clinician Administered PTSD Scale). Clinician interviewers were blind to PCL self-report results. Receiver Operating Characteristic (ROC) analysis was used to determine the optimal diagnostic cutpoint for the PCL in this sample.
Results: Preliminary analysis of the data revealed that of the 301 women interviewed, 104 (34%) met diagnostic criteria for current PTSD by the CAPS interview. The area under the ROC curve for the PCL compared to the CAPS was 0.88 (95% CI 0.84-0.92). A PCL cutpoint score of 38 provided optimal diagnostic efficiency (sensitivity = 0.81, specificity = 0.81) with a kappa of 0.64. Lower scores yielded more sensitive screening with a substantial decline in specificity (e.g. score of 30, sensitivity = 0.85, specificity = 0.68). Using the PCL cutpoint of 38, screening prevalence for PTSD was 30 % among Year 1 respondents to the previously mailed VWAP women’s health survey (N=1259, mean age = 46). Interview recruitment bias will be evaluated by comparing clinical and utilization features of interviewed women to all eligible women.
Conclusions: The PCL performed well as a screening measure for the detection of PTSD in female VA outpatients. The optimal screening cutpoint of 38 in this population is lower than thresholds determined in trauma treatment settings (e.g. optimal score of 50 in combat veterans), but only slightly lower than the cutpoint of 42 proposed for male VA ambulatory patients in the Veterans Life Experiences Study (Spiro et al, 2000). This lower cutpoint is also consistent with thresholds derived in non-VA primary care samples.
Impact: Female veterans are a rapidly growing and highly trauma-exposed population in VHA. The observed 30% screening prevalence of PTSD in female veteran patients indicates that the impact of PTSD on the health of veteran women warrants further study. The implementation of an efficient screening measure for PTSD may improve recognition and treatment of PTSD among female veteran patients.