Background: Efforts to increase access to life-saving treatment, including antiretroviral therapy (ART), for people living with HIV/AIDS in resource-limited settings has been the growing focus of international efforts. One of the greatest challenges to scaling up will be the limited supply of adequately trained human resources for health, including doctors, nurses, pharmacists and other skilled providers. As national treatment programmes are planned, better estimates of human resource needs and improved approaches to assessing the impact of different staffing models are critically needed. However there have been few systematic assessments of staffing patterns in existing programmes or of the estimates being used in planning larger programmes.
Methods: We reviewed the published literature and selected plans and scaling-up proposals, interviewed experts and collected data on staffing patterns at existing treatment sites through a structured survey and site visits.
Results: We found a wide range of staffing patterns and patient-provider ratios in existing and planned treatment programmes. Many factors influenced health workforce needs, including task assignments, delivery models, other staff responsibilities and programme size. Overall, the number of health care workers required to provide ART to 1000 patients included 1-2 physicians, 2-7 nurses, <1 to 3 pharmacy staff, and a much wider range of counsellors and treatment supporters. We estimate from these data that the equivalent of 20,000 to 100,000 physicians, nurses, pharmacists and other core clinical staff will be needed to meet the WHO target of treating 3 million people by the end of 2005. The total number of staff, including counsellors, administrators and other cadres, could be substantially higher.
Discussion: These data are consistent with other estimates of human resource requirements for antiretroviral therapy, but highlight the considerable variability of current staffing models and the importance of a broad range of factors in determining personnel needs. Few outcome or cost data are currently available to assess the effectiveness and efficiency of different staffing models, and it will be important to develop improved methods for gathering this information as treatment programmes are scaled up.