The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility

Full text

The PTSD Checklist (PCL): Reliability, Validity, and Diagno stic
Utility
Frank W. Weathers, Brett T. Litz,
Debra S. Herman, Jennifer A. Huska, & Terence M. Keane
The PTSD Checklist (PCL) is a new self- repo r t rating sc al e for as ses sin g post-
trau ma tic stress disord er (PTSD). The PCL consists of 17 items which corre s po nd to th e
DSM-III-R sympt om s of PTSD. Exami n ee s are instruct ed to indicat e how much they have
been both ere d by eac h sy mpt o m in th e past mo nth using a 5-point (1- 5) scal e. The
anchor s for the severity ratin gs ar e identical to thos e used on the SCL-90- R (Deroga t is,
19 8 3), and range from "Not at all" to "Extre me ly." There ar e two versions of th e PCL:
the PCL-M, with reex per i en cing sym pto ms written specifically for military expe ri en c es, and
the PCL-C, with reexp erie nci ng symp to ms writt en gene rically to apply to any trau ma tic
ev ent . The PCL can be modified easily to fit specific ass es sm en t need s. For exa m pl e, in
order to ass es s symp to m severity repea te dl y in the cont ext of a treat men t protoc ol, the
time frame of one mont h can be chan g e d to "the past we ek" instea d of "the pa s t mont h."
Similarly, the wording of the items can be modified to reflect a specific trau m atic ev ent. In
the secon d study desc rib ed below the phras e "your military ex p erie nc es " was cha n ge d to
"your ODS exp erie n ce s. "
The PCL is useful in a variety of clinical and rese a rc h ass e ss me nt conte xts ,
especi ally when inform atio n ab o ut PTSD sympt om s is de si r ed but ad minist ering a
struct u re d inte r vi ew is not feasibl e . The PCL can be us ed as a continuo us mea su re of PTSD
sy mpt o m severity by sum min g scor es across the 17 ite m s. It also can be us e d to deriv e a
PTSD diagn osis by con si d erin g a score of 3 (Modera tel y) or gre at e r as a symp t om , then
followin g the DSM-III-R di a g no s tic rule (1 B sympt om , 3 C sym pto ms, and 2 D sy mpt o ms).
STUDY 1
Method Subject s for the first st u d y were 123 male Vietna m the at er veter an s who cont a ct ed
the Natio nal Cent er for PTSD either for clinical servi c es or rese a rc h particip a tion.
Dem ogr ap hic inform atio n for this samp l e is shown in Table 1. In three separ at e session s,
two to thre e days apart, subj e cts compl e te d the PCL twice, comple t ed sev eral oth e r
qu esti o nn aire mea s ur es of PTSD and gener al psycho pa th ol o gy, and were adminis ter ed the
PTSD modul e of the Structur ed Clinical Interview for DSM-III-R (SCID). The preval en c e of
curren t PTSD was 54%. Sev eral an alyse s were conduc te d to evalua te the psycho m etri c
prop e rtie s of the PCL.
Results
1. Mean PCL score s were 63. 6 (SD=14 .1) for PTSD subje cts and 34. 4 (SD =14 .1) for non-
PTSD su bj ects.
2. Test- retes t reli a bility was .96.
3. Inter nal consiste nc y (alpha coefficien t) wa s .93 for B sympt om s, .92 for C symp t om s, .92
for D sym p to ms , and .97 for all 17 sy mpt om s. Item- scal e total correlatio ns rang e d fro m .
62 - .87.
4. Conver g en t validity was demo ns tr at ed by strong correlatio ns betw e en the PCL and: the
Mississippi Scale (.93), the PK scale of the MMPI-2 (.77), th e Impact of Event Scal e (.90), and
the Comb at Expos ure Scale (.46).
5. Diagnos t i c utility was deter min ed by usin g PCL scor e s to predict PTSD diagn o ses deriv ed
fro m the SCID. The opti mally efficient cut off scor e was 50, which yielde d a sensitivity of .
82, a sp ecificity of .83, and a kappa of .64.
STUDY 2

Method Subject s for the secon d st u dy were 1006 nation a lly surveye d male (88%) and
fem ale (12 %) vetera ns of the Persian Gulf the at er. Dem ogr ap hic infor m ati o n for this
sampl e also is shown in Tabl e 1.
The survey instru me nt include d the PCL and the Mississippi Scale, both of which
were modified to b e specific to the Persian gulf th eat er , and a mea su re of com ba t
ex p osu re . Approxima te ly half of the subjects com plet ed the surv ey wh e n the y wer e
seeking services (primarily psychot he ra py or be n efits couns eling) at a vet cente r , and half
com plet ed it at an Army base or throu g h their National Gu ar d or Reserv e unit.
Results
1. Mean PCL score s were 64. 2 (SD=9.1 ) for PTSD subje c ts and 29. 4 (SD =11 .5) for non- PTSD
su bj ects .
2. Inter nal consiste nc y (alpha coefficien t) wa s .90 for B sympt om s, .89 for C symp t om s, .91
for D sym p to ms , and .96 for all 17 sy mpt om s. Item- scal e total correlatio ns rang e d fro m .
52 - .80.
3. Conver g en t validity was demo ns tr at ed by a stro ng correlatio n betw ee n th e PCL and the
Mississippi Scale (.85).
4. A principal co mpo n e nts analysis with varima x rotatio n yielded one large factor
acc o un ting for 59% of the varian ce, and one smaller fact o r acco unt i ng for 7% of the
varianc e. Loading most highly on the first factor were item s mea su ring reex pe rien cin g (B1-
B4), effortful avoidan ce (C1- C2), and hypera ro u sal (D4- D6). Loading most highly on the
second factor wer e item s mea su ring nu m bin g of resp o ns i v en ess (C4-C7) and hyper aro us al
(D1- D3).
CONCLUSIONS
1. The PCL is an easily administ er e d self-report rating scale for ass es sin g the 17 DSM-III-R
sy mpt o ms of PTSD.
2. The PCL h a s excell ent test- ret e st reliability over a 2- 3 day period.
3. Inter nal consiste nc y is very hig h for each of the thre e gro ups of items corresp on din g to
the DSM-III-R symp to m clust ers as well as for the full 17- item scale.
4. The PCL correlate s strongly with other me a su re s of PTSD, such as the Mississippi Scale,
the PK scale of the MMPI-2, and the Impact of Event Scale, and also correlat es mode ra t el y
with level of com ba t expos ur e.
5. Used as a contin uou s me as ur e, the PCL has good diag n os tic utility. In Vietna m com ba t
vet era ns a cutoff of 50 on th e PCL is a good predictor of a PTSD diag nos i s bas ed on th e
SCID PTSD modul e.
6. Princi p al comp on en ts an alysis rev eal ed on e large factor, consistin g prim arily of
re e xpe rie n cin g and hyp e rar ou sa l items , and on e much smaller factor, consisting prim a rily
of em o tion al numb i ng ite ms.

Table 1 -- Demo gr a ph ic and Psych o met ric Informa tion
Variable Vietna m
(n = 12 3) Persia n Gulf
(n = 10 0 6)
Age 43 .74 (2.69) 30 .46 (7.98)
Gend er
Men 10 0 .0% 82 .8%
Wom en 0.0% 12 .8%
Educatio n
< High School 10 .6% 1.8%
High School Degre e 24 .4% 56 .2%
So m e Colleg e 35 .0% 28 .3%
College 7 .3% 8.7%
> College 7.4% 4.5%
Marital
Single 26 .0% 33 .6%
Married 28 .4% 48 .2%
Divorced 27 .6% 9.7%
Separ a ted 13 .8% 6.1%
Widow e d 2.4% .9%
Branch
Army 48 .0% 63 .4%
Navy 13 .8% 9.3%
Air Force 7 .3% 15 .1%
Marines 29 .3% 10 .0%
Race
Whit e 73 .2% 64 .1%
Black 22 .8% 17 .0%
Hisp a nic 2.4% 5.8%
Other 0.0% 5.8%
Mississippi Scale 106 .63 (27.59) 74 .13 (24. 2 4)
PTSD Checklist 50 .58 (20. 2 4) 34 .77 (16.8 3)