Citation

Silove D, Manicavasagar V, Mollica R, Thai M, Khiek D, Lavelle J, Tor S. J. Nerv. Ment. Dis. 2007; 195(2): 152-157.

Affiliation

School of Psychiatry, University of New South Wales, and Centre for Population Mental Health Research, Sydney South West Area Health Service, Sydney, Australia.

Copyright

(Copyright © 2007, Lippincott Williams and Wilkins)

DOI

10.1097/01.nmd.0000254747.03333.70

PMID

17299303

Abstract

The Hopkins Symptom Checklist depression scale (HSCL-D) and the Harvard Trauma Questionnaire (HTQ) have been used extensively in international studies, particularly among refugees and conflict-affected populations. Like many other screening measures, however, the HSCL-D and HTQ have not been subjected to extensive epidemiologic testing, particularly among communities not affected by war. The present study examined the psychometric properties of the HSCL-D and HTQ by comparing the measures with the Structured Clinical Interview for DSM-IV (SCID) among ethnic Cambodians living in Surin, Thailand, a community that was spared the long period of mass violence that affected Cambodia proper. The PTSD data of the HTQ were then contrasted with those obtained from a refugee clinic. The main finding was that in the Surin study, the screening measures showed greater agreement with the SCID in identifying noncases (negative prediction) than cases (positive prediction). In contrast, in the earlier clinic study, the HTQ showed high positive and moderately low negative prediction. The results support previous observations that clinical interviews such as the SCID may be more conservative in identifying cases. We raise the possibility, however, that structured interviews may perform differently across clinic and community populations. In particular, clinicians may be more accurate in identifying cases than noncases in highly symptomatic clinic populations, with the converse applying in low prevalence community populations. These issues warrant further investigation to specify more clearly the value of using the two approaches to case identification.

Language: en