Some authors argue that there are no predictors of adherence to medication, while others proffer indicators of likely adherence. These contextual factors are wide-ranging and may be interdependent. There are few studies of adherence in resource-poor settings. Of these, many were linked to particular trials whose conditions were highly controlled. For instance, the number of participants in the trial was small, the participants were strictly selected and they were offered much adherence support. These conditions will largely fall away with wider access to antiretroviral drugs (ARVs) such as South Africa is undertaking. Poor adherence remains one of the challenges to scaling-up access to ARVs, and, as such, the likely barriers to adherence need to be understood and countered. This study explores contextual factors — such as disclosure, preparation for treatment, treatment meaning and treatment acceptance — as aspects of lifestyle that can bear on adherence. Those who had disclosed were largely supported at household level, and after initial problems with adherence, participants adjusted and were able to include ARV treatment in their daily repertoire. Although the anticipation of stigma remains an obstacle to wide disclosure, antiretroviral therapy offers new hope and has the potential to impact on stigma.