The Measurement Equivalence and Differential Predictive Validity of Scores On the Primary Care Posttraumatic Stress Disorder Screen Across Gender in a Sample of Military Veterans
Purpose: Previous research indicates that as many as 8.7% of military personnel exposed to combat experiences in Iraq and Afghanistan later received a new onset posttraumatic stress disorder (PTSD) diagnosis or were identified as having PTSD symptoms. Within the Department of Defense, the Primary Care Posttraumatic Stress Disorder (PC-PTSD) screen is the main screening instrument used to identify military members who are at increased risk of PTSD. Since 2005, this screen has been offered to all military members returning from deployment. Once identified, those at increased risk for PTSD can be referred for further evaluation and, as appropriate, offered treatment. However, validation studies of PC-PTSD scores from military samples have seldom employed a significant number of female subjects and no published studies have examined this instrument for gender bias. Ruling out bias is important because routine under-identification of PTSD risk in any group could result in hindered access to needed assessment and/or care. With the current proportion of military females historically high, it is imperative that the PC-PTSD be analyzed to ensure measurement equivalence across gender. This study examined two questions: 1) Do scores on the PC-PTSD exhibit gender bias at the item level? 2) Do scores on the scale differentially predict a clinical PTSD diagnosis for males and females?
Methods: This non-experimental measurement study used a large sample that included all female and male US Air Force veterans who completed the PC-PTSD as part of a routine post-deployment health screening during 2008 and 2009 within 90 to 180 days after returning from Iraq and Afghanistan. The first analysis of PC-PTSD scores involved an item response theory (IRT) differential item functioning (DIF) analysis across the male (N = 8381) and female (N = 1425) subgroups. Then, using a clinical diagnosis as the criterion, both logistic regression and diagnostic likelihood ratio methods were employed to assess for differential predictive validity by sex. This second analysis used a subset of the original sample consisting of 5785 men and 1214 women, all of whom had at least one military medical diagnostic visit within the 90 days after taking the PC-PTSD screening. All data was collected from existing US Air Force post-deployment screening and medical databases.
Results: Results of the IRT DIF analysis revealed no statistically significant gender-related DIF. At the scale level, the logistic regression and diagnostic likelihood ratio analyses found no significant differential prediction of later PTSD diagnosis by PC-PTSD scores.
Implications: The results supported the continued use of the PC-PTSD with both female and male veterans. Also, the US Air Force scoring regime for the PC-PTSD, which differs somewhat from the scoring recommended in the literature, was confirmed as valid. Future research should focus on establishing measurement equivalence across both gender and ethnicity for scores on more extensive PTSD instruments used as diagnostic aids.