To compare the situational characteristics of protected and unprotected sexual encounters that involved alcohol use 2 hours prior with ones that did not.
Data were collected between December 2002 and December 2005 as part of enrollment in a prospective cohort study designed to identify HIV seroconversion risk factors among women bar and hotel workers in Northern Tanzania. A total of 608 (37.3%) of the women who were inconsistent condom users were asked a set-matched questions concerning situational characteristics surrounding their last protected and unprotected sexual encounter including whether they had been drinking within 2 hours of sex. The associations between drinking 2 hours before sex (yes/no), condom use (protected/unprotected), and their interaction with the situational descriptors were examined with a 2 × 2 model for paired categorical data after controlling for time since the last type of encounter.
Condom failure was 5 times more likely if someone (woman, man, or both partners) had been drinking in advance of the encounter (OR, 5.19; 95% CI, 2.05–15.46) and was especially likely to occur if only the woman had been drinking before sex (OR, 14.05; 95% CI, 4.03–50.41). Alcohol use before sex was associated with sexual contacts where the woman was having sex with her partner for the first time, their relationship was casual or transitory or sex was transactional, the location was unfamiliar and less under her control, and the partner had been drinking or using drugs before having sex. Condom use was more frequent in precisely the same types of encounters. Interestingly, there were no significant interactions between alcohol use before sex and condom use, suggesting that drinking before sex and use of condom are distinct and not contingent risk factors.
Alcohol use before sex is associated with an increased likelihood of condom failures and with high-risk sexual encounters, ones that have consistent situational characteristics regardless of whether condoms are used or not.
Alcohol use before sex is associated with an increased likelihood of condom failures and with high risk sexual encounters, ones that have consistent situational characteristics regardless of whether condoms are used or not.
From the *Interdata, Inc., Sanibel, FL; and †London School of Hygiene and Tropical Medicine, Mwanza, TZ
Supported by the Rockefeller Foundation, New York (Grant Number 2002 HE 036).
The Rockefeller Foundation was not involved in the study design, data collection and analysis, and writing of this report.
Correspondence: Joseph C. Fisher, PhD, MPH, Interdata, Inc., P.O. Box 129, Sanibel, FL 33957. E-mail: firstname.lastname@example.org.
Received for publication July 14, 2009, and accepted February 22, 2010.