ABSTRACT
<p><b>OBJECTIVE</b>To evaluate HIV risk behaviors among heroin drug users who were treated in methadone maintenance treatment (MMT) clinics.</p><p><b>METHODS</b>A prospective cohort study recruited and followed up clients of eight MMT clinics treated no more than two and half months in Guizhou province, China. Through face-to-face interviews, the baseline and following up informations were collected. And the baseline information included both demographic information and HIV risk behaviors, the following up informations included only HIV risk behaviors. The baseline investigation started in June, 2006 and the following up investigation finished in June, 2007. A total of 1003 heroin drug users were recruited at baseline, among them 666 (66.4%) were still at treatment by the end of follow up and 469 (70.4%) clients participated in the follow up interview. Wilcoxon two sample test and McNemar test were used to test for changes in HIV risk behaviors between baseline and following up investigation.</p><p><b>RESULTS</b>Among the 469 MMT clients who were followed up, the average days of reported heroin use decreased from 26.4 to 0.9 in the past 30 days (Z = 27.21, P < 0.05). Average days of alcohol use at baseline were 3.3 but 3.7 at follow up (Z = 0.45, P = 0.96). Needle-sharing behavior reported in the past 30 days decreased from 1.3% at baseline to 0.2% at follow up (χ(2) = 5.00, P = 0.025). At baseline, 5.5% (26/469) subjects reported having multiple sex partners in the past 30 days compared to 3.4% (16/469) at following up (χ(2) = 3.18, P = 0.08).6.4% (30/469) subjects reported casual sex with non-regular sex partners in the past 30 days at baseline compared to 5.1% (24/469) at following up (χ(2) = 0.95, P = 0.33). Of those who reported having casual sex relationship in the past 30 days 56.7% (17/30) reported using condoms at baseline but 58.3% (14/24) reported using condoms at follow up (χ(2) = 1.96, P = 0.16).</p><p><b>CONCLUSION</b>MMT was observed to decrease needle-sharing HIV risk behavior. However, decreased HIV sexual risk behaviors were not observed at statistical significant level.</p>