South Africa is the country with the worldwide largest number of HIV-infected individuals (5.7 million in December 2007). Antiretroviral treatment (ART) did not become available in the public sector in South Africa until late 2004, a delay that caused large numbers of avoidable deaths. Since then, however, scale up has been rapid. By the middle of 2008, more than half a million people were receiving ART in South Africa. Despite this enormous treatment effort – funded largely through public budget allocation to the Department of Health with support from external agencies, such as the Presidential Emergency Fund for AIDS Relief (PEPFAR) and Médecins San Frontières (MSF) – about three-fifths of HIV-infected people in need of ART in South Africa currently do not receive it.
Research questions
One of the major challenges in the continued ART scale-up – the number and types of human resources that will be required to achieve universal coverage – deserves attention from new perspectives. We investigate
· Feedback effects from current ART coverage to future human resources need and HIV incidence (“top-down models”)
· ART task time distributions and health worker numbers and types required under different models of ART delivery (“bottom-up models”)
Fit
The modeling projects fit in with the emerging health systems research agenda at the Africa Centre and the interest in ensuring access to high-quality ART in the Africa Centre community.
Data sources and methods
For the “top-down models” we used secondary data from a variety of sources to investigate feedback effects in deterministic Markov models and microsimulations. For the “bottom-up models” we collected data on task time distributions in three primary care clinics within the DoH-Africa Centre antiretroviral treatment and care programme in a time-motion study. We are currently coding the data for use in predictive models and scenario analyses of future HRH needs in this rural, decentralized programme of ART delivery.
Findings
Analysis of the data from the time-motion study in “bottom-up models” is currently under way. In “top-down models” we demonstrate that estimates of human resource needs that ignore the feedback effect significantly underestimate the future human resources required to achieve universal ART coverage. For instance, while estimating the size of the ART workforce required for providing universal coverage in sub-Saharan Africa within 10 years, we compare a model with feedback to a model without feedback. Our results show that feedback alone implied that 2.45 times more health workers (compared to the non-feedback case) will need to be added to the system every year to achieve universal coverage in 10 years, assuming that health worker productivity and ART delivery models remain unchanged.
Policy implications
Our “top-down models” have demonstrated the large and growing need of health workers to provide ART in sub-Saharan Africa and South Africa. The need will most likely far exceed the health worker education capacity in the region. Thus, it is imperative that health worker productivity in ART programmes increases and new models of ART delivery are established, where health worker-patient contact times are optimized and tasks are shifted from highly skilled to lower skilled health worker categories. Our “bottom-up models” will contribute to our understanding how ART delivery can be re-organized, so that HRH supply can meet HRH need in future years.