The nominative technique: a new method of estimating heroin prevalence.

Over the years, nominative estimates of heroin prevalence have been consistently higher than self-reports of heroin use. During this time, nominative data have generally followed mainstream patterns of drug use: nominative estimates for young adults and for males are higher than nominative estimates for older persons, youth, and females; moreover, the recent downward trends in drug use have been replicated by the nominative heroin data. Thus, the overall picture presented by the nominative data--similar patterns but higher levels of prevalence--seems to support the validity of the new approach. Nevertheless, considerable caution should be exercised in interpreting nominative data. This is chiefly because a substantial minority of nominators cannot report the number of other close friends of the heroin user who also "know." While missing data has been handled by a conservative imputation rule, the fact that so many persons are unable to provide an answer to this key question casts doubt on the accuracy of the answers that were given. In fact, the nominative approach might tend to produce over-estimates, because of the potential for undercounts of the numbers of others who "know." Additional tests of validity should be performed, such as application of the nominative approach to nonsensitive behaviors or minimally sensitive behaviors, such as marijuana use or perhaps cocaine use. Certainly, the overall validity of the nominative heroin data would be supported if in future surveys new nominative heroin estimates for relatively unstigmatized forms of drug use proved to be similar to self-reported levels of use, thus pointing to the unique difference in estimates that might be observed for heroin. Finally, in interpreting the heroin estimates presented here, it should be remembered that both the nominative and self-report estimates refer to heroin use in the household population of the United States. Thus, many heroin addicts and other users who reside in various unconventional living arrangements would not be included in the counts presented here. Among the excluded groups are transients residing in rooming houses or "crashing" in the home of one "friend" after another or who are incarcerated in jails or confined to residential drug treatment centers. This is a caution for interpreting the estimates presented in this paper, not a criticism of the nominative technique itself.(ABSTRACT TRUNCATED AT 400 WORDS)