Drinking alcohol is part of many social and ceremonial activities in diverse cultures. The manufacture and sale of alcohol produce employment and income for individuals and government revenues through taxes. However, alcohol abuse has a broad negative impact on the social, economic and health status of those involved and on society in general. One of the possible unforeseen consequences is the spread of HIV/STDs in vulnerable populations.

Alcohol use and abuse may be of particular concern among persons who live in situations of high risk for HIV/STD infection in developing countries. They include women living in households where alcohol abuse is common (and negotiating safer sex may be particularly difficult), commercial sex workers, runaway and homeless youth, and men in occupations that require them to travel over long distances such as migrant labourers, building tradesmen, truck drivers or the military.

Research on alcohol abuse prevention has tended to stress control measures that change the availability of alcohol (e.g., restriction on underage drinking, drug and driving laws, taxation). While these measures may indirectly affect risk for HIV/STD infection, they do not directly address alcoholrelated risks associated with HIV or STDs.

A harm reduction approach is different from, but complementary to, traditional prevention approaches. This approach focuses on avoiding problems directly associated with HIV infection while drinking. These problems include unsafe sex or injecting drugs while intoxicated or, in the case of a person living with HIV/AIDS, failure to adhere to medication regimens for controlling progression of HIV infection.

Heavy drinking occasions can form the primary focus of interventions. The goal is to shift the pattern of alcohol consumption and subsequently change the circumstances in which HIV/STD risk behaviours take place. Attention can be paid to different levels, for example, the individual level (e.g., an individual believes he or she must be drunk to have sex) or the economic level (e.g., more alcohol is sold when commercial sex workers promote drinking to patrons). Some harm reduction interventions have focused on changing social or physical characteristics of bars and other drinking settings. Strategies used have included:

• identifying opinion leaders in bar settings and changing group norms to endorse safer sex
• scheduling interventions at risky time periods, such as when the bar closes and
• patrons leave
• providing condoms at the bar
• promoting safer behaviour through posters and other media
• training servers to identify patrons at risk.

These types of intervention are successful when they reduce the chance of adverse behaviour (unsafe sex). The fact that drinking will occur is accepted. However, the drinker is held accountable for his or her actions. Most of these harm reduction interventions are relatively new; the role of alcohol has only been indirectly evaluated. Nevertheless, initial results suggest that a wide range of HIVrisk behaviours can be reduced after intervention and at followup.

In general, support for controlling alcohol availability is declining in many parts of the world. Policy-makers should give careful thought to the need for regulating alcohol among those vulnerable to HIV/STD infection and for programmes that minimize the consequences of drinking in vulnerable and infected populations. Some communities have taken action in this area with interventions focusing on:

• providing alternatives to drinking and visiting brothels
• changing drinking patterns among commercial sex workers
• changing the availability of liquor and expectancies for safer sex behaviour in social
• clubs and brothels
• providing HIV education to partners of alcohol users
• providing informal treatment settings, such as drop-in centres, for individuals with alcohol problems.

As the production and sale of alcohol is a major source of income for some women in developing countries, this must be taken into account. Prevention programmes can enlist their support in promoting safer sex, e.g., as community-based condom distributors.

Preventive interventions should be initiated and implemented by researchers themselves for the specific purpose of testing the effects of their strategies. Interventions may also occur naturally through the actions of public and private organizations (e.g., reducing the availability and accessibility of alcohol, increasing condom distribution at bars, health promotion campaigns that highlight linkages between alcohol use and AIDS). Careful evaluation of these interventions is also needed.

Investigatorinitiated interventions may be nested within the context of naturally occurring interventions, permitting the effects of both types of interventions to be studied simultaneously. Examples include developing or augmenting HIV/STD interventions within clinical settings, such as hospitals and public health clinics, to address alcoholrelated problems (e.g., improving adherence of alcohol abusers to therapeutic drug regimens). Ongoing alcoholproblem studies can be supplemented by including HIV-infected or atrisk populations and adapting the intervention to address HIV/STD issues in these sub-groups.

A wide range of contexts may be appropriate for intervention studies. They may be aimed at individuals, social networks, institutions, and specific settings such as bars and clubs, to change alcoholrelated sexual expectancies, behavioural norms, and risktaking behaviours. Some examples:

Hardtoreach populations: research on homeless and runaway youth indicates a high rate of cooccurring alcohol abuse and unsafe sexual behaviour, often resulting in the spread of STDs. Alcohol abusers often delay seeking medical treatment and may be difficult to retain in controlled clinical trials. New interventions should attract and retain individuals at extremely high risk for alcohol abuse and HIV/STD infection, e.g., by including more informal and culturally relevant dropin clinics. New research designs, such as casecontrol or casebased analyses, should be developed to test the effects of these interventions on such factors as HIV/STD exposure, alcohol abuse, and retention in clinical trials.

Healthcare systems: increasing attention is being paid to the role of healthcare systems and professionals in preventing alcoholrelated problems, facilitating early detection of alcoholrelated highrisk behaviours, and providing appropriate treatment. Experimental designs may be used within healthcare settings to test the efficacy of prevention strategies. These may include risk assessment, provision of brief and/or more extensive advice, case monitoring, and improved linkages to services for alcoholics in treatment or for HIVinfected individuals with alcoholproblems.

Media/communications: ongoing research is needed to assess the efficacy of media strategies, often combined with other strategies, to prevent alcoholrelated risky sexual behaviour. Promising media messages, new communications technologies and special media for cultural sub-groups need to be developed and tested to determine the most effective approaches for varied target audiences. Of particular interest are communication strategies that reach audiences at highest risk for alcohol abuse and HIV/STD infection.

Family studies: research suggests that family involvement, broadly defined, can enhance the effectiveness of school and clinicbased alcohol abuse prevention programmes among youth at risk. Research should be expanded in this area to develop effective interventions among family members to reduce the risk for HIV/STD infection.