Use of the BASIS-32
In Reply: As we described in our paper, we did indeed use clinician ratings as the basis for our ratings in the interview condition and, as we noted, we achieved very good interrater reliability. Unpublished data from a generalizability analysis revealed that very little variance in test scores was due to rater variance: 15 percent for relation to self and others, 2 percent for depression-anxiety, 4 percent for daily living and role functioning, 1 percent for impulsive and addictive behaviors, and 6 percent for psychosis.
We should also note that the internal consistency coefficients for our interview version, which were significantly lower than the self-report version, were quite acceptable for two subscales (relation to self and others and daily living and role functioning) and were borderline acceptable for depression-anxiety. The other two subscales—impulsive and addictive behaviors and psychosis—have also had borderline to low reliability in most of the other published studies (1,2,3), including our self-report version and Eisen's original report (1).
1. Eisen SV, Dill DL, Grob MC: Reliability and validity of a brief patient-report instrument for psychiatric outcome evaluation. Hospital and Community Psychiatry 45:242-247, 1994Abstract, Google Scholar
2. Russo J, Roy-Byrne P, Jaffe C, et al: The relationship of patient-administered outcome assessments to quality of life and physician ratings: validity of the BASIS-32. Journal of Mental Health Administration 24:200-214, 1997Crossref, Medline, Google Scholar
3. Hoffman F, Capelli K, Mastrianni X: Measuring treatment outcome for adults and adolescents: reliability and validity of BASIS-32. Journal of Mental Health Administration 24:316-331, 1997Crossref, Medline, Google Scholar