Trends in injection risk behaviors in a sample of New York City injection drug users: 1992-1995.

Objectives: This study assessed the relation between year of recruitment into an AIDS prevention project and likelihood of engaging (yes/no) in injection risk behaviors.

Methods: In total, 834 injection drug users were recruited over a 44-month period (January 1992-August 1995) in New York City. Logistic regression was used to examine trends in three behaviors, across four successive annual cohorts: using needles/syringes that were previously used by another person, using injection supplies (e.g., cookers, cotton, rinse water) that had been previously used, and giving or lending of used needles/syringes to another person.

Results: Significant (p < .0001) decreasing trends occurred in two behaviors: giving or lending of used needles/syringes to another person and use of injection supplies that had been used by another person. Sample characteristics were generally consistent over time and did not obviate significant injection risk trends.

Conclusions: Decreasing trends in injection risk behaviors could not be explained by changing sample characteristics. Even though some injectors report engaging in risky injection practices, the confluence of numerous AIDS prevention efforts in the 1990s has contributed to an overall reduction in the likelihood of such behaviors.