Our aim was to determine if a comprehensive adherence package improved self reported adherence to antiretroviral therapy. The adherence package included an education programme, individualized planning of regimens, and the opportunity for a patient to choose from a number of adherence aids and reminder devices. A randomized step wedge design was used. Forty-three individuals were randomized to begin the intervention over a five-month period. There was a substantial fall in the number of missed doses reported for the last four days (0.76 to 0.38, P =0.03) and last seven days (1.5 to 0.74, P =0.005) but not for the last 28 days (2.5 to 2.5, P =0.63). There was no statistical difference in the viral load or CD4 lymphocyte count in the period before or after the intervention. The Morisky score during the pre and post intervention periods was significantly different (P =0.006), 2.9 (SD 0.9) and 3.3 (SD 0.8) respectively. This adherence package improved self reported adherence during the last four and seven days.