Level of agreement between DIS and UM-CIDI alcohol and marijuana use disorders: Implications for research

Howard D. Chilcoat, Victoria C. Lucia, Naomi Breslau

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


By and large, recent epidemiological investigations of alcohol and drug use disorders are based on structured interviews, such as the Diagnostic Interview Schedule (DIS) and the Composite International Diagnostic Interview (CIDI), which are designed to be administered by trained lay interviewers. Little is known about the level of agreement between these instruments. To fill this gap, we compared responses from: 971 young adults from an epidemiological sample, who were administered both the DIS and the University of Michigan CIDI (UM'CIDl) in the same interview session. Level of agreement was estimated for responses to individual items, and for DSM-IIIR abuse/dependence (A/D) diagnoses and symptoms for alcohol and marijuana. DIS/UM-CIDI agreement for drug use disorders in the year preceding the interview was moderate for each disorder (alcohol A/D, kappa - 0.58; marijuana A/D, kappa = 0.68), although one-year prevalences of alcohol and marijuana AID were approximately 40% higher for the DIS compared to the UM-CIDI. Despite the discrepancies, levels of comorbidity with other psychiatric disorders and associations with risk factors were similar for the DIS and LM-CIDI.

Original languageEnglish (US)
Pages (from-to)83-101
Number of pages19
JournalInternational Journal of Methods in Psychiatric Research
Issue number2
StatePublished - 1999
Externally publishedYes


  • Agreement
  • Comorbidity
  • DIS
  • Interview

ASJC Scopus subject areas

  • Psychiatry and Mental health


Dive into the research topics of 'Level of agreement between DIS and UM-CIDI alcohol and marijuana use disorders: Implications for research'. Together they form a unique fingerprint.

Cite this