click to return to the Heller School Ph.D. in Social Policy homepage click to go to the Heller School homepage click to go to  Brandeis University website

Resource Needs and Performance of National HIV/AIDS Programs in Low- and Middle-Income Countries

A Dissertation Presented to the Faculty of the Heller School for Social Policy and Management, Brandeis University, Waltham, Massachusetts.

Written by Wu Zeng, MD, MS

The pandemic of HIV/AIDS adversely affects many developing counties, particularly those in Sub-Saharan Africa. In the past decade, international donors have allocated billions of dollars to address HIV/AIDS in the most affected countries. Despite the unprecedented worldwide commitment, available funding remains insufficient to meet the compelling goals of universal access to HIV prevention, treatment, care and support by 2010 (2010 HIV/AIDS targets) and Millennium Development Goals (MDGs) on HIV/AIDS. Given the limits on available resources, it is critical to make available funding work most efficiently, thereby maximizing benefits to the target populations.  This study is designed to help policy makers meet this need through research on the efficiency of national AIDS programs in low- and middle-income countries.

This study first employs the economic theory of production to evaluate national HIV/AIDS programs using data envelopment analysis (DEA). Then, the analytical frameworks of governance and absorptive capacity are used to identify determinants of a country’s efficiency in delivering HIV/AIDS services.  The relationships between the efficiency and its determinants are estimated with a random-effects Tobit model. Lastly, an extended DEA approach is used to project the resource needs of HIV/AIDS to meet the 2010 HIV/AIDS targets, and to decompose each country’s performance gap into a resource gap and an efficiency gap. The resource gap and efficiency gap refer to the amount of shortfall due to lack of financial resources and that due to less than optimal use of available resources respectively.

The study constructed a data set of inputs and outputs of national HIV/AIDS programs based on data collected jointly by research groups at Brandeis University and UNAID through publications, reports and data bases.  It analyzed trend patterns to impute missing values.  Altogether, the data set contained 68 countries and 151 country years over the period of 2002 through 2007.  

The analysis showed that these countries achieved moderate efficiency (48.3%) in implementing HIV/AIDS programs, but with enormous variation. The average efficiency of countries in each of four quartiles of observations was 12.4%, 33.3%, 53.9% and 95.1% respectively, from the least efficient to the most efficient quartile. The countries, on average, could have doubled the volume of the services with current spending, or could have saved about half of their current spending without reducing their outputs, if they performed efficiently.

The study found that governance, financing mechanisms, and demographic characteristics of a country all influence the efficiency of HIV/AIDS service delivery. If the countries in the worst quartile of “voice and accountability” could improve this characteristic up to the average level in the sample, the efficiency of the countries’ HIV/AIDS programs would have been increased by 42.5%. In the countries with the lowest rates of law enforcement, corruption was found to serve as a second best option to move the HIV/AIDS program forward. In addition, higher shares of external sources of health expenditure and high prevalence of HIV/AIDS were positively associated with efficiency of HIV/AIDS programs, but per capita gross national income was not associated with such efficiency.

The DEA analysis projected resource needs for 45 countries for 2006 at $5.9 billion to meet the 2010 HIV/AIDS targets. This projection represented a gap of $1.5 billion compared with available funding in 2006, if all countries performed as efficiently as their counterparts on the production frontier. These estimates are smaller than ones from resource needs model (RNM) used by UNAIDS. The further decomposition of performance gap for 25 countries requiring additional funding to scale up services to meet the 2010 HIV/AIDS targets from the projection indicates, on average, that 62% of the performance gap is a resource gap and the remaining 38% is an efficiency gap. The share due to the efficiency gap increases when all observations are included in the analysis because the remaining 20 countries do not require additional funding if they perform efficiently.

This analysis suggests: 1) There is a substantial scope for improving HIV/AIDS programs, particularly in the countries in the lowest quartile of efficiency in the analysis; 2) Closing the wide gap between resources needed and funds available requires not only obtaining more funding from donors and national sources, but also improving countries' ability to use available resources efficiently to reach global goals in HIV/AIDS control; 3) It is critical to strengthen health care system by addressing governance (particularly both rule of law and control of corruption simultaneously) and financing mechanisms; 4) Low-income countries with high HIV prevalence remain the targets for controlling the spread of HIV/AIDS, and continued efforts from the international community is required; and 5) DEA approach can serve as an complementary tool to the resource needs model to project the resource needs to reach internationally agreed HIV/AIDS goals under the scenario in which all countries perform optimally.


  • Donald S. Shepard, PhD, Chair
  • Jon Chilingerian, PhD
  • Joan Kaufman, ScD
  • John Stover, MA
    President, Futures Institute

PhD Dissertation Abstract