Perceived importance of condom use among African Americans using drugs

American Journal of Health Studies, Spring-Summer, 2003 by Joseph Schumacher, Min Qi Wang, Sacia Dear, Sapna Marepally, Foster Cook

To determine the probability sample needed per zip code, we first estimated the size of the target population of users of drugs by taking the Alabama drug use prevalence rate (11%) and applying it to each zip code. This calculation provided an estimate of 8,200 users of drugs in these sites. This was most likely a conservative estimate because it was based on state prevalence rates, while the prevalence rates within the target zip codes were believed to be higher. Next, we determined the minimum probability sample size needed for a 95% confidence interval and 3% error rate which yielded a target sample of 401 users of drugs (Wang, Fitzhugh, & Westerfield, 1995).

The completed interviews achieved this statistical power with a total of 1,630 completed interviews. For the purpose of this study only individuals who used marijuana and crack (cocaine) during the past three months were included (n=718), exceeding the required sample size of 401 with a 95% confidence interval and a 3% error rate.

In order to make the sample more representative of the people residing within each zip code area, individuals were interviewed at selected interview sites that covered all zip codes. Interviewers were given a set number of people to talk to before moving on to the next site. One question, included in the survey, asked if the respondent lived in the zip code area. The number of people per zip code was divided by the number of sites to give a per-site quota.

For several days prior to surveying, the teams visited the sites and distributed literature about the project and free condoms in order to make the area residents feel more comfortable with their presence. Recruitment was facilitated by approaching individuals and offering them an incentive (a $5 McDonald's gift certificate). The estimated refusal rate was about 15%.

SURVEY INSTRUMENT AND INTERVIEWS

Street interviews about drug use, sex, condom use, and reasons for condom use were conducted almost daily from year 2000 to 2002. Interviews lasted from 15 to 20 minutes initially, but were slightly shorter after the team gained experience. Interviewers asked respondents about their alcohol, marijuana and cocaine (including crack) use in the past 30 days. Respondents were also asked if they had sex without a condom in the past 30 days with their main and/or other sex partners. Respondents were also asked about the reasons for condom use with their partners. There were items indicating reasons for condom use and respondents rated the importance of each statement for their main partner and other partners. Consequently, for each statement, there were two questions. For example, if the first statement was: "You think your partner is safe", then the two questions were: "How important is this for using condoms with your main partner?" and "How important is this for using condoms with your other partner?" The response categories were on a 5-point Likert scale ranging from "Not Important" (score 1) to "Very Important" (score 5). There were a total of 15 statements. The complete list of these statements is in the Appendix. The alpha reliability coefficients of reasons for condom use were .96 for main partners, and .95 for other partners.