Introduction: South Africa has embarked on a massive roll out of ARVs to more than 1.4 million people living with HIV/AIDS. Provision of ARVs to people living with HIV/AIDS encounters many challenges associated with adherence. Properly taken ARVs have been shown to reduce viral loads to undetectable levels and increase the CD4 count. This in turn leads to a drop in opportunistic infections and better health outcomes but the requirements for adherence are high. Several patient-related factors have been reported to affect adherence rates. Nonadherence on the other hand has been reported to lead to the development of drug resistant strains of HIV. It recognised that the resistance to ARVs can quickly lead to build up of highly resistant strains in the blood due to one week of missed medication. Aims and objectives: This study set out to identify factors which affect adherence to HAART among adults on HAART in two health facilities in Gauteng province in 2006.The main objectives were to assess the patient adherence using viral load response and self-report data. Secondly, the study was to determine factors that facilitate adherence and finally barriers to adherence at the two sites. Materials and methods: A cross sectional study was done at the two ARV facilities in Gauteng from July to November 2006. Two physiological methods -CD4 counts and plasma viral load, and one subjective-3 day recall self- report methods were used to asses adherence. Exit interviews and record reviews were done to collect data. Virologic outcome was the preferred surrogate marker for adherence. Univariate and bivariate analyses were done to determine measures of association. Measures of association (Chi square) at a 95% significance level for factors affecting adherence were then determined and results obtained. Results: The mean age was 36.9 years (range 18-70 years) and 73.5% were women. Self-report data (n=343) indicated 98.4% in the higher adherence category (taken 100% of their doses). Viral load data (n=343) showed that 88.8% were in the adherence lower category (<400 RNA copies). Viral load outcome (“adherence”) was significantly associated with the length on treatment (p<0.05) and patients who had been on treatment for 12-24 months had lower viral load than those who had been treatment for a shorter time (<12 months) or longer (>24months). However, gender (p=1.000), age (p=0.223), level of education (p=0.697) and access to social grants (p=0.057) were not associated with “adherence”. Socio-economic status was significantly associated with viral load outcome (p<0.01) as well as cost (n=185; p<0.05). Individuals who incurred the highest costs (>R25) were the least likely to adhere followed by those facing average costs (R15-25) compared to the reference group (< R15). Conclusion: Adherence rates of 88.8% suggest that respondents from both facilities can optimally adhere to their medication when they have been on ARVs for longer than a year. These are minimum adherence rates. There were factors that still hinder adherence at both the individual patient level. There is still a need for more targeted interventions especially towards men who were noted to have a relatively low uptake of HAART within the two sites.