. 2024 May 10;33(Suppl 1):e2013. doi: 10.1002/mpr.2013

TABLE 7.

Individual‐level concordance of PTSD diagnoses based on the original and recalibrated PCL‐5 short‐form with diagnoses based on the DSM5 PTSD lifetime in the 10% probability subsample of WMHQ respondents who were administered the short‐form followed by the remaining items to assess DSM‐5 lifetime disorder (n = 66). ^a

Original threshold

Recalibrated threshold

Est

(SE)

Est

(SE)

Prevalence ^b

59.5

(8.7)

40.4

(7.5)

90.0

(4.7)

77.3

(8.7)

61.9

(10.8)

85.6

(5.7)

PPV

62.5

(8.5)

79.1

(7.1)

NPV

89.7

(5.4)

84.2

(7.0)

TCA

73.5

(6.4)

82.1

(5.1)

AUC

0.76

0.81

χ_{1}^{2}

610.5 ^c

2.5

Abbreviations: AUC, area under the receiver operator characteristic curve; NPV, negative predictive value; PPV, positive predictive value; SE, standard error of Est; SN, sensitivity; SP, specificity; TCA, total classification accuracy.

^{^a}

The calculations are based on a weighted version of the 10% WMHQ subsample that administered the full 20‐item PCL‐5. As this was in the Part II sample, the weight adjusted for the over‐sampling of respondents who met CIDI criteria for one or more Part I disorders as well as for the fact that the Part II respondents who were administered the remaining 14 questions in the full PCL‐5 over‐sampled those whose screening scale scores based on the 6 PCL‐5 items in the screening scale (which were administered first) were over the recommended diagnostic threshold.

^{^b}

Full PCL‐5 prevalence is 41.4% (7.6).

^{^c}

Significant at the 0.05 level, two‐sided test.