The validity of the interviewer severity ratings in groups of ASI interviewers with varying training

Aims. This study evaluated the hypothesis that the subjective interviewer severity rating (ISR) summary indices of the intake Addiction Severity Index (ASI) of less trained interviewers are less valid than those of more highly trained interviewers. Design. Baseline ASIs from three completed studies whose interviewers varied in degree of initial ASI training and subsequent quality assurance monitoring were examined. Associations between baseline ISRs and three other sets of ASI summary indices not based on interviewer ratings-composite scores, clinical indices and evaluation indices-were compared for three groups of interviewers with varying amounts of training. The assumption underlying these analyses was that more reliable ISRs, found in more trained interviewers, would be more highly associated with the other more objective indices, Setting. Methadone maintenance patients in the Philadelphia and New York City areas. Participants. Thirty-five interviewers with the most intense training who administered 295 interviews; 10 interviewers with an intermediate level of training who administered 763 interviews; and eight identified (and other unidentified) least trained interviewers who administered a total of 276 interviews. Measurements and methods. Four sets of summary indices from the above ASIs. Both bivariate and multivariate analyses were performed. Findings. The study found that the validity of the validity of ISRs was greater in more trained interviewers. Conclusions. Greater training and subsequent monitoring of ASI interviewers generally appears to be associated with increased ISR validity.