Drug testing was used as an adjunct to a general population household drug use survey administered via audio computer-assisted self-interview. Participants, ages 18–40 years residing in Chicago, were recruited to participate in three different biological tests (hair, oral fluid, and urine) presented in random order subsequent to completing an interview. Subjects had the option of participating in zero to three different tests. We examined participation/refusal in tests, reaction to testing requests, as well as variables associated with participation and reaction. Subjects were randomly assigned to a low (US$ 10 per test) or high (US$ 20 per test) incentive condition. Over 90% of the sample participated in at least one test, usually the oral fluid test. Associations between refusal status and two variables, socioeconomic status (SES) and the presence of children in the household, provided partial support for the notion that drug test participation parallels the survey response process in general. The incentive level did not directly increase drug test participation. Reporting of recent illicit drug use was associated with participation in only one procedure, hair testing. Type of test offered and individual differences in willingness to be drug tested were important predictors of drug test refusal and subject reaction to testing requests. Compared with urine and hair testing, oral fluid testing had lower refusal rates and was generally more acceptable to respondents in a general population survey. The findings support the feasibility of incorporating multiple drug tests with modest incentives into general population household surveys on drug abuse.
Biological testing for drugs of abuse has long been accepted as an essential means of gathering information about recent substance use in treatment settings. For example, in the United States, federal regulations require the use of drug testing in programs providing treatment for opiate dependence Department of Health and Human. Until recently, the use of biological testing as a tool for gathering data about drug use in research studies has been limited to studies of “special” populations, such as prisoners, arrestees, and those in treatment settings. Despite a growing body of research suggesting limits in the validity of drug surveys, drug testing has been less commonly used in population-based epidemiological research.
While treatment-oriented and special population studies commonly employ urine testing, a limited number of published epidemiological studies employing drug testing as an adjunct to population surveys of illicit drug abuse have employed hair testing. Saliva or “oral fluid” testing has been extensively employed in epidemiological, and community surveys focused on confirming self-reports of tobacco use.