HIV will inevitably be more prevalent in some groups of people than others because of the history of the epidemic.

This will mean that the same act will carry different levels of risk in different groups of the population and different places:

The prevalence in different core groups varies from place to place and over time, e.g. between gay and bisexual men in the USA and the UK. For instance, prevalence in both gay men and heterosexual women in at least twice as high in London as it is in the rest of the UK. However this is for different reasons; while prevalence in London gay men as a whole is higher, it is only so in women because more African women live in London. So it would be misleading to say that prevalence in women in general as a population group is higher in London.

• Statistical scientific knowledge about prevalence rates in groups of the population cannot always be safely translated into practical knowledge of individuals' risk factors in particular cases.
• Groups defined by a common sense of identity are not functionally the same as groups defined by their actual behaviour. Thus, many people have mistakenly believed themselves not to be at risk because they saw themselves as heterosexual even though they were also engaging in unsafe homosexual activity in practice.

Identification and avoidance of people with HIV is not an effective or desirable line of defence against HIV for individuals who engage in activities which can transmit the virus. This is because, at any time the majority of people with HIV cannot be identified by external physical signs. Moreover, any attempt to fit individuals into broad and blunt risk categories (although useful in the planning of public health services) is likely to fail on an individual level most of the time. It is not always possible to detect the sexual or drug using history of partners with sufficient accuracy for this to offer sufficient protection.