Highly active antiretroviral therapy (HAART) reduces the incidence of human immunodeficiency virus (HIV) dementia (HAD), whereas the overall prevalence appears to have increased. Recent changes to diagnostic nosology have emphasized the presence of neurocognitive deficits. Uniform methods of ascertaining neuropsychological impairment and excluding confounding causes are critical to between-study comparison. We conducted a systematic review on all studies that use single-cohort prospective treatment effect design that reported on the neurocognitive or neuropsychological profile of individuals commencing HAART. Fifteen 15 relevant studies were included. A large number of studies using observational or cross-sectional designs were excluded, as these do not allow for a within-subject description of pre- and post-HAART predictive factors. Eleven studies reported a significant improvement in neurocognitive status or neuropsychological profile over an average study period of 6 months. Variable or nonreporting of HAART regimens in these studies did not allow for an analysis of individual agent or regimen effectiveness. The results show that although HAART does improve cognition, it does not appear to fully eradicate impairments. The methods used in this research differ widely and therefore comparison across studies is difficult. Studies examining the long-term effects of HAART on HIV-associated neurocognitive disorders (HANDs) using uniform methods of data collection are needed, together with clear reporting of HAART regimens.