Heroin incidence: a trend comparison between National Household Survey data and indicator data.

Because of the small proportion of the population reporting ever having used heroin, the year of first use data from NIDA's National Household Surveys on Drug Abuse conducted in 1977, 1979, and 1982 were pooled to show the number of new users in the household population by year of first use. In addition, the data were "smoothed" by using a 2-year moving average. The early 1970s and the mid-1970s epidemics were evident. These epidemic periods occurred at the same periods reported by high school seniors and by heroin users in treatment. The household self-report data trends based on age and frequency of use were also consistent with the trends in periods of initiation reported by heroin users in treatment as noted in drug abuse treatment admission data for year of first heroin use. Trends in indicators of heroin epidemics were compared with trends based on self-report data from the National Household Surveys. The trends in hepatitis B cases, heroin-related emergency room visits, heroin-related deaths, and the average retail heroin purity were consistent with the epidemic periods suggested by the household data. This consistency among the three sources of self-reported data on trends in year of first heroin use combined with the consistency of these self-reported data with the trends based on the indicators of heroin epidemics offers some validation to the use of retrospective direct questions concerning age of first use of heroin to monitor heroin incidence in the household population.