BACKGROUND: There is paucity of information on the cost-effectiveness of delivery strategies to retain patients on antiretroviral therapy (ART) and this study tries to fill this gap. METHODS: The analysis is based on a representative sample of 2835 patients attending 32 ART sites in KwaZulu-Natal (KZN), South Africa. Extended Cox regression and Kaplan Meier were used to estimate the transition probabilities to remain on ART among patients who attended sites with different staff and workload profiles. Annual costs per patient-year of observation for these delivery profiles were estimated. Probabilistic sensitivity analysis took into account parameters' uncertainty. RESULTS: The delivery sites with a full-time doctor and a full-time senior professional nurse and an intake of less than 200 new patients per doctor per year were the most cost-effective in retaining patients on ART. If 1000 new patients were followed up by this type of site, 724 patients would still be on ART after 10 years at a discounted cost of US$8.41 million at 2006 value with an incremental cost-effectiveness ratio of US$12,271 per extra retained patient over the second not dominated site profile. CONCLUSIONS: The results could be used to estimate the human resources needed for a sustainable scaling up of ART in KZN.
BACKGROUND: There is paucity of information on the cost-effectiveness of delivery strategies to retain patients on antiretroviral therapy (ART) and this study tries to fill this gap. METHODS: The analysis is based on a representative sample of 2835 patients attending 32 ART sites in KwaZulu-Natal (KZN), South Africa. Extended Cox regression and Kaplan Meier were used to estimate the transition probabilities to remain on ART among patients who attended sites with different staff and workload profiles. Annual costs per patient-year of observation for these delivery profiles were estimated. Probabilistic sensitivity analysis took into account parameters' uncertainty. RESULTS: The delivery sites with a full-time doctor and a full-time senior professional nurse and an intake of less than 200 new patients per doctor per year were the most cost-effective in retaining patients on ART. If 1000 new patients were followed up by this type of site, 724 patients would still be on ART after 10 years at a discounted cost of US$8.41 million at 2006 value with an incremental cost-effectiveness ratio of US$12,271 per extra retained patient over the second not dominated site profile. CONCLUSIONS: The results could be used to estimate the human resources needed for a sustainable scaling up of ART in KZN.