The Homeless Supplement to the Diagnostic Interview Schedule: test-retest analyses.

This study sought to extend previous work on reliability of self-reported residential history in a homeless population with high rates of drug abuse. The latest version of the Homeless Supplement to the Diagnostic Interview Schedule (DIS/HS) was used to achieve reliability on homelessness experience, use of shelters, transience, and recent residential patterns. Homeless study volunteers were recruited for a test-retest study from a drop-in day centre for mentally ill homeless people (N = 25) and a substance abuse day programme (N = 30). They were administered the instrument approximately one to two days apart. Kappa and intraclass correlation analyses were performed to assess reliability. Overall, the reliabilities of most variables were acceptable, ranging from fair to excellent. Six items were reconstructed to achieve reliability and two were dropped. Substance dependence and adult antisocial behaviour patterns did not affect reliability on most items. This study has developed a reliable self report instrument for measuring residential history that can be used with homeless and drug abusing populations. Replication is needed in larger, more representative samples and comparison of reliability with other psychiatric and cognitive characteristics.