Crosswalk Analysis of Antiretroviral Therapy Costing Models And Their Policy Impact

The World Health Organization (WHO) reports that antiretroviral therapy (ART) is now available to approximately 37 percent of individuals living in sub-Saharan Africa who require it . This increase in access to treatment is evidence of impressive scale-up over the past decade. However, much works remains to be done to achieve the goal of universal access. Although providing universal access to ART and mitigating the impact of the HIV epidemic is central to local government planning, many lack sufficient capacity to develop and manage their national strategic responses.

Mathematical models can assist policymakers in estimating the resources that are currently being spent on existing HIV programs, as well as plan for the scale of increase that will be necessary in the future. Furthermore, models can be used to help policymakers adjust current allocations in order to more effectively address the epidemic. Through the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) and other initiatives, several different packages of modeling software have been developed to address different aspects of resource allocation. Focusing on software commonly in use, this crosswalk analysis evaluates and compares nine ART costing models, including model utility, data output, impact on policy decisions, and other pertinent information to inform HIV policymakers.

Download the crosswalk analysis below or view summaries online by clicking the links in the summary.

Link to Crosswalk Analysis of ART Costing Models and Their Policy Impact (PDF, 583 KB)
Link to ART Costing Crosswalk Analysis Table (PDF, 34 KB)

Published and unpublished literature on ART costing analysis was reviewed. The models themselves, when publicly available, were accessed and analyzed. Telephone interviews were conducted with key informants who had firsthand experience implementing the models. When possible, more than one key informant per model was interviewed in order to glean information from multiple sources. Following the telephone interviews, the ART Costing Crosswalk Analysis was sent to each of the key informants for their input and validation.


Nine models were found to be commonly implemented in sub-Saharan Africa in cost HIV treatment programs and were included in this crosswalk analysis. Click on the name of each model for a summary page:

These models have the ability to address slightly different needs—from costing operational plans, to identifying program bottlenecks, to estimating the resource needs for HIV drugs. The majority of the models measure financial costs, which can be defined as the amount paid by a program for its inputs. The costing data included in the models is mostly recurrent, meaning that the inputs consumed in the past year are repurchased in the following years. Most of the models also use a bottom-up costing approach in which the researchers estimate the unit costs of each input and then multiply this cost by the number of units delivered. The notable exception is PACM, which uses a top-down, economic approach that measures the total program expenditures to determine program costs and includes the value of inputs not directly paid for such as volunteer time, travel costs for service providers, and so forth (e.g., economic costs). More information on model methodology is detailed in Appendix A of the PDF above.

While each model was created for a specific purpose, each can also be used to answer a variety of questions. For example, the Goals model is used to calculate the relative cost-effectiveness of varying HIV interventions, but it also can answer the question, “What will be the impact of cuts in current levels of funding?” Likewise, HAPSAT was created to analyze the sustainability of a country’s comprehensive national portfolio of HIV interventions. It analyzes the financial and human resource requirements for achievement of various HIV policy targets over a five-year time horizon and evaluates the resulting resource gaps. However, HAPSAT can also be used to answer the question, “What are the key bottlenecks and when do they appear over time?” A list of the questions that could potentially be answered by each of the models is included in Appendix B of the PDF above.

PACM, SIMCLIN, and Peds 2010 cost only those program areas related to HIV treatment. However, several of the models cost a variety of HIV program areas, not solely the cost of ART. For example, MBB was designed to look at implementation constraints for all the Millennium Development Goals. ASAP, Goals Model, HAPSAT, RNM, and Spectrum include program areas related to HIV prevention, care, and support; orphans and vulnerable children (OVC); and nutritional support in addition to treatment. See Table 1, and Appendix C for a full list of program areas and data outputs included in each model.

These models have been used to influence policy throughout the world. RNM has been utilized in numerous countries, including Kenya and Honduras, to assist with costing national HIV strategic plans. UNAIDS currently uses RNM to develop its Global Resource Needs Estimates. These estimates provide UNAIDS with information on the financial gap between the available and the needed resources for ART. In addition, Spectrum is the foundation of epidemiological projections in 135 countries. This model analyzes HIV information to estimate the need for ART, as well as the cost of HIV care and treatment for a variety of opportunistic infections (OIs).