Test-retest reliability of psychoactive substance abuse and dependence diagnoses in telephone interviews using a modified Diagnostic Interview Schedule-Substance Abuse Module

Abstract

The test-retest reliability of lifetime substance abuse and dependence diagnoses obtained by telephone interviewers was investigated. Trained personnel administered two identical interviews based on a modified Diagnostic Interview Schedule-Substance Abuse Module (DISSAM) approximately a week apart for 100 respondents, of whom 55 were receiving alcohol or other drug treatment and 45 and randomly selected from residential households in one Michigan county. The uncorrected agreement for all lifetime dependence diagnoses exceeded 93% for all six categories assessed and the more conservative chance corrected agreement (Cohen's Kappa coefficient kappa) was .92 (alcohol),.76 (marijuana),.87 (cocaine), and .71 (other opiates). Kappa coefficients for hallucinogens and heroin dependence could not be calculated due to low (i.e., 5% or less) base rates. Likewise, kappa was calculated only for a single abuse diagnosis, alcohol, with kappa = .42 and 95% agreement. In the interpretation of kappa, the standard applied was: kappa ranging from .41 to .60 represented moderate agreement, kappa ranging from .61 to .80 represented substantial agreement, and kappa ranging from .81 to 1.00 represented excellent agreement. Thus, test-retest reliability was excellent for lifetime alcohol and cocaine dependence and was substantial for lifetime marijuana and other opiates dependence. These results indicate that lifetime psychoactive substance abuse diagnoses can be obtained fairly reliably over the telephone using trained lay interviewers.