The authors translated a CIDI-Alcohol, a highly structured diagnostic interview, into Korean
and examined the validity and reliability of the Korean version of CIDI-Alcohol.
In order to examine the validity and reliability of CIDI-Alcohol, the authors calculated the
degree of agreement of CIDI-Alcohol diagnoses between the two lay interviewers and sensitivity,
specificity, kappa coefficient between clinical and lay CIDI-Alcohol diagnoses.
The results were as follows :
1) Agreement rates of CIDI-Alcohol diagnoses between the two lay interviewers ranged
from 97% to 100%. The kappa coe伍cient for each diagnosis was quite high(0.96 - 1.0).
2) The kappa value of agreement between clinical and lay CIDI-Alcohol diagnoses ranged
from 0.43 to 0.55,which was lower than those of other studies.
3) With changing the diagnostic criteria of CIDI-Alcohol, The highest agreement rate
between clinical and CIDI-Alcohol diagnoses was acquired when the authors used two or more
diagnostic criteria of ICD-10/harmful use and four or more diagnostic criteria of DSM- ID -R/
alcohol dependence.