The aim of this study was to compare the PC-PTSD and GHQ-12 in detecting new cases of PTSD among primary care patients. Data on the PC-PTSD, GHQ-12 and psychiatric diagnoses was extracted from clinical databases for 11,230 VA primary care patients. Signal detection analyses and likelihood ratios were used to compare screens. Logistic regression analysis was used to examine the prediction of PTSD by the PC-PTSD after controlling for the GHQ. The PC-PTSD had a higher positive predictive value than the GHQ (41 percent vs. 31 percent). Combining positive results on the two screens in predicting PTSD yielded the highest likelihood ratio (LR=17.3) compared to a positive result on the PC-PTSD only (LR=8.3) or the GHQ only (LR=4.6). The PC-PTSD performed slightly better than the GHQ and provided unique information in identifying PTSD, suggesting that disorder specific screens are important to use in primary care settings.