The relationship between alcohol consumption and HIV has spawned a growing body of literature, with mounting support suggesting not only that the two are related [1–5] but that they should be treated as intertwined public health issues. The “fuel to the fire” nature of the relationship between alcohol consumption and HIV is evidenced throughout sub-Saharan Africa, which has 67% of the world’s HIV infections,  as well as countries with high rates of risky drinking patterns, such as drinking until intoxicated and heavy episodic drinking . Risky drinking is a predictor of sexual risk behavior in sub-Saharan Africa [5, 8] and is common throughout the region. The World Health Organization (WHO) estimates heavy episodic drinking, commonly defined as consuming five or more drinks in one sitting, to occur at least once weekly among 25% of drinkers in Africa, which is double the worldwide average of 11.5% .
Understanding the relationship between alcohol consumption and HIV is a crucial step in addressing the HIV epidemic. While the biological consequences of alcohol consumption, such as a weakening of the immune system  and an increase in an HIV positive individual’s viral load,  have been noted as important avenues of transmission, the social consequences may be even more important. That is, alcohol’s facilitation of unprotected sex may increase the potential for HIV-infection. Several mechanisms have been theorized to explain this association, including the physiological effects of alcohol on decision making [11, 12], expectations regarding alcohol intake and condom use [13, 14], as well as certain personality types being more likely to engage in risky behaviors [15–17].
The vast majority of evidence supporting an association between alcohol consumption and unprotected sex is based on either global associations, which correlate self-reports of overall frequency of alcohol consumption or situational-overlap associations, which analyze the frequency of drinking during sex with self-reported frequency of unprotected sex [1, 18]. These methodologies do not permit the researcher to determine if drinking during a specific sexual intercourse event is specifically associated with failure to use a condom during that sexual event. This limitation can be addressed by using event-level analysis, which requires pairing alcohol consumption during a specific instance of sexual intercourse and whether or not a condom was used.
Earlier literature reviews of studies based in the US and other developed nations found no association at the event-level, [19, 20] however, more recent meta-analyses that included both developed and developing nations have found support for an event-level association between alcohol use prior to or during sex with risk of HIV [1, 21] and unprotected sex . Furthermore, if data is collected on which partner is consuming alcohol, as well as on the quantity of alcohol consumed or level of intoxication, using event-level analyses allows the researcher to determine gender differences and if alcohol exerts effects only at certain levels of intoxication. For example, Kiene et al.  found that drinking alcohol before sex by the female, the male, or by both partners, increased the proportion and number of subsequent unprotected sex events, but only when the quantity of alcohol consumed was at a moderate or higher risk drinking level. Despite strong support for the use of event-level analysis, little research exists that looks at alcohol use and risky sexual behavior at the event-level and even less that have been conducted in sub-Saharan Africa. Of the studies that meet this description to date [22–24], a significant association between alcohol use prior to or during sex and unprotected sex has been demonstrated.
The present analysis extends the relatively limited literature, none of which derives from population-based data, on the event-level association between alcohol consumption during sex and unprotected sex by testing this association across sub-Saharan African countries using data from Demographic and Health Survey (DHS) and AIDS Indicator Survey (AIS) population-based surveys. We hypothesize that drunkenness at last sex will be associated with lower odds of condom use, especially when sex is with a casual partner.