International OR
A Life or Death Question
Allocating scarce healthcare resources:
In AIDS-stricken Africa, strategic models of operations research offer hope
By Steffen Flessa
Many young people dream of becoming a hero. Some join the army and rescue one of their fellow soldiers. Others study medicine and give 50 people their eyesight back. Still others join the priesthood and establish orphanages for 100 poor children.
Impressed? Of course. But there is another way to become a hero and save the lives of thousands, perhaps even of millions, of people: become a management scientist. Management scientists develop models of healthcare resource allocation that improve the efficiency of healthcare management in developing countries and thus make the difference between death and life for millions.
Undeveloped countries are suffering from an enormous degree of illness and unnecessary death. The infant mortality rate (the number of newborns who will not survive their first year of life) reaches up to 170 per 1,000 in parts of the world. In those same countries, the child mortality rate (the number of newborns who will die before they reach the age of five) can be as high as 250 per 1,000. Adults in developing countries are far more likely to die from infectious diseases before they see their own children grow up than adults in other parts of the world. In some countries life expectancy is less than 40 years. This is an enormous human tragedy - and an economic catastrophe. Poor health induces a weak economy, and a weak economy does not provide sufficient resources to improve the health status. In Africa, where most of the poorest countries of the world are located, the two most devastating diseases, malaria and AIDS, are the biggest threats to economic development, and, of course, to human lives.
Health economists and politicians from these nations have to decide on the allocation of healthcare resources. Which sector of the healthcare system should get which share of the available resources? Market forces cannot satisfactorily answer this question in developing countries because the majority of poor people would be without access to any modern healthcare if the market was left uncontrolled by the government. Thus, politicians have to allocate the scarce resources.
Curative and preventive care, primary and hospital-based care, programs fighting childhood and adult diseases all need a sufficient budget. In developing countries the budget is never adequate. If resources are allocated to an inefficient institution or program, these scarce resources are missing elsewhere, which will result in additional human suffering and death. Efficient allocation, therefore, becomes a question of life or death. Management scientists can save millions of lives and ease the suffering of entire nations if they can show, with their models, where healthcare resources should be allocated.
AIDS
The Acquired Immune Deficiency Syndrome is rapidly changing the nature of healthcare in developing countries. In some regions in Africa, AIDS is already the leading cause of death. At the end of 2000, about 36.1 million people were living with AIDS or an HIV-infection worldwide [UNAIDS 2000, UNAIDS/WHO 2001]. About 5.3 million of those people had been infected during the past year, and three million died due to AIDS. An estimated 22 million people worldwide have lost the fight against AIDS since the epidemic was identified two decades ago.
The region most severely affected is Sub-Saharan Africa, where 70 percent of the worldwide HIV/AIDS cases have occurred, and the annual increase exceeds all other parts of this world combined. In Swaziland, Malawi, Sambia, Zimbabwe, Namibia and Botswana, up to 36 percent of adults have been infected. In order to allocate healthcare resources to different curative and preventive services, politicians must be able to predict the spread of AIDS and assess the impact of different prevention strategies. Therefore, we developed a system dynamics model to simulate the AIDS-epidemic in a severely affected region in Eastern Africa. The model consists of 19,300 sub-populations and five Africa-specific paths of infection. The simulation is performed on a personal computer.
Map 1: Expectancy of life.
Source: Worldbank
Map 2: Child mortality.
Source: WHO, Worldbank
Map 3: Infant mortality.
Source: WHO, Worldbank
Map 4: HIV-prevalence rates worldwide.
Source: Worldbank
Map 5: HIV-prevalence in Africa, 1982-1997.
Source: UNAIDS
Basic Projection
Charts 1 and 2 show the predicted spread of AIDS in the region. Although many people become infected with HIV, and even die from AIDS, it takes a long time until the demographic impact of AIDS can be measured on population totals. If you compare the population for the scenario with and without AIDS, hardly any difference can be detected for many years. In 1990, for example, the population in the AIDS scenario was only 0.25 percent lower than that of the non-AIDS scenario. Politicians could not appreciate the tiny difference. Even by the year 2000, the difference between the population in the case with and in the case without AIDS was merely 7.63 percent. Some politicians might even have been happy about the reduced speed of population growth.
Chart 1: Population.
Chart 2: AIDS-Cases and mortality in the region.
The long incubation period (up to 15 years) makes it extremely difficult for those responsible for planning and allocating healthcare resources to assess the real extent of the AIDS tragedy. The number of healthy people in the region reached its maximum by the year 1996 and has steadily declined since. The number of infected will peak by 2011, the number of AIDS patients will peak by 2016, and the number of AIDS-related deaths will peak by 2017. This "slow plague" [Gould 1993] comes so silent and is delayed by so many years, that other diseases might appear more urgent for health politicians. They might allocate insufficient resources to fight AIDS, only because the various impacts of this disease are not perceptible right now.
In order to allocate the proper share of the AIDS-prevention budget to different programs, it is important to realize that the significance of different paths of infection is changing with time. Until 1991, the majority of people were infected by promiscuous contacts. Prostitutes in urban centers were a major source of infection. Afterwards, the virus had already spread into the rural population so that sexual intercourse between married partners became the main source of infection; between 50 percent and 60 percent of infections are now attributable to this path.
Chart 3: Share of paths of infection.
The increasing number of prenatal infections, i.e., from mother to child before or just after childbirth, is due to the fact that more and more women in rural places are becoming infected. As long as AIDS was merely a problem of the urban population and a few traveling men, prenatal infections were rare. This has now changed. Since HIV-infected children have no chance to survive in developing countries, AIDS will roll back the reduction in the childhood mortality rate that was achieved during the last decade. According to our simulations, the rate would have been 109.67 per 1,000 newborns, without AIDS. The impact of AIDS increases the rate to 181.23.
The number of studies on the costs of AIDS in developing countries is very limited. Our own research in Tanzanian hospitals retrieved annual costs of curative care at $67.35 per patient. Chart 4 demonstrates the consequences: fewer and fewer healthy people have to pay for the treatment of more and more AIDS patients. In some African countries almost 70 percent of the healthcare budget will have to be spent for treating AIDS patients. It is difficult to estimate the loss of gross national product due to AIDS. If we assume with Cuddington [1993] that the national product is a function of age, than we have to realize that the national income per capita will shrink from the year 2000 to the year 2020 by 17.62 percent. Decision-makers who have to allocate healthcare resources to different programs must keep this in mind.
Chart 4: Direct annual costs of AIDS [U.S. dollars].
Prevention
So what should we do? How should we fight back against AIDS? Or, to put it more precisely: to which programs should scarce AIDS-prevention funds be allocated? The most successful intervention would be an effective vaccination. However, the impact of vaccination programs is not as easy to understand as one might expect. Chart 5 shows the results for different scenarios. The "Standard" scenario assumes that no vaccination is possible. The scenario "Vacc" assumes that an effective, life-long vaccination is available on January 1, 2001, and that it is easy to vaccinate everybody. Another simulation shows the consequences of a vaccine that has only a limited efficacy of 50 percent ("Half"). Another scenario ("Short") assumes that the vaccination can be applied in 2001, but will last only for five years. No re-vaccination is planned. Finally the last simulation ("Delayed") is similar to the "Vacc" scenario, only that the vaccine will not be available before January 2006.
Chart 5: Vaccination.
As expected, delayed vaccination programs as well as vaccines with limited efficacy induce increasing numbers of infected. The "Short" scenario demonstrates the most important result. As expected, the graphs for "Short" and "Vacc" are identical for the first five years. Afterwards, the number of newly infected increases so much that by the year 2020 there will be more people suffering from AIDS than for the case if no vaccination had ever been attempted! This is due to the fact that human beings who were protected against AIDS for five years have an additional risk afterward of being infected. This means that a vaccine that does not provide life-long protection against AIDS might accelerate the spread of AIDS in developing countries. Many examples show that disease control programs are only funded for a limited number of years. Unless these programs are sustained for decades, such a vaccine will worsen AIDS!
It is difficult to assess the economies of a vaccine against AIDS since no vaccine is currently available and no costs can be estimated. Table 1 shows the maximum amount a vaccine could cost in order to be cost-effective. This amount depends on the discounting factor of future health gains. This demonstrates that the higher the discounting factor is, the less cost-effective the vaccination program. In other words, nations with a low "future orientation" will not spend much money on prevention. Health economists and politicians must be aware of the influence of time preference rates on healthcare allocation decisions.
|
Discounting Rate |
SCENARIO |
|||
|
Vacc [US$] |
Half [US$] |
Short [US$] |
Delayed [US$] |
|
|
0 % |
39,00 |
5,56 |
15,09 |
22,21 |
|
1 % |
34,41 |
4,94 |
13,82 |
19,18 |
|
2 % |
30,45 |
4,41 |
12,68 |
16,61 |
|
3 % |
27,03 |
3,95 |
11,66 |
14,42 |
|
4 % |
24,07 |
3,54 |
10,74 |
12,54 |
|
5 % |
21,50 |
3,19 |
9,91 |
10,94 |
|
6 % |
19,26 |
2,88 |
9,16 |
9,56 |
|
7 % |
17,30 |
2,61 |
8,48 |
8,38 |
|
8 % |
15,58 |
2,36 |
7,88 |
7,35 |
|
9 % |
14,08 |
2,15 |
7,32 |
6,47 |
|
10 % |
12,75 |
1,96 |
6,81 |
5,71 |
|
20 % |
5,427 |
0,89 |
3,60 |
1,81 |
|
30 % |
2,82 |
0,48 |
2,16 |
0,68 |
|
40 % |
1,70 |
0,30 |
1,42 |
0,29 |
|
50 % |
1,13 |
0,20 |
1,00 |
0,14 |
Table 1: Maximum costs per person vaccinated so that vaccination is cost effective.
As it is not very likely that an effective vaccine will be developed soon, direct prevention is the only way of fighting AIDS. The use of condoms is effective, but much more expensive than expected. Chart 6 compares the costs of treating AIDS patients with the costs of providing condoms for different scenarios. The first scenario ("Promis") assumes that condoms are used in all promiscuous sexual contacts. The coitus frequency in long-term partnerships is unchanged. The scenario "Part" calculates the costs for the case that 30 percent of sexual intercourse in partnerships and 50 percent in promiscuous relations are protected. Finally we analyze the impact of a prevention strategy that provides condoms only to prostitutes ("Prost").
Chart 6: Cost saving of condoms.
Providing condoms for everybody is not cost-effective. Donating condoms for prostitutes, however, pays back many times over. In the years 2011 and 2012, every dollar invested in condoms for prostitutes saves $5.80 of treatment costs of AIDS patients. If we provide condoms for all promiscuous sexual intercourse, we have a ratio of up to 14:1. Thus, condoms must be made available for risk groups soon.
The last two years have witnessed a strong decline of prices of anti-retroviral drugs that can be used to prevent the infection of the child by the mother before or during childbirth (prenatal transmission). Thus, preventing prenatal transmission might be cost-effective in developing countries as well as in the industrialized world [Marseille, Kahn & Saba, 1998; Shaffer, 1999; Guay, 1999; Marseille, 1999]. Two components contribute to the cost of this measure. First of all, all pregnant females must be screened for HIV. Our own study in Tanzania showed that this costs about $5 per mother. Secondly, if a mother is detected to be HIV-positive, she must be given a dose of the anti-retroviral drug. Marseille [1999] shows that Nevirapine has the best cost-benefit ratio, as it reduces the risk of a prenatal transmission by 50 percent [Guay] and costs about $4 per woman treated.
For any sensible discounting rate, the present value of the treatment costs saved by screening all pregnant women and providing Nevirapine to all those found HIV positive is less then the present value of the costs of this prevention strategy. Thus, mass screening for HIV among pregnant women is not cost-effective. Nevertheless, this must not remain so forever: Screening pregnant women in all African countries and providing Nevirapine to all those found HIV positive will definitely reduce the unit costs of screening kits and drugs. Thus, it may become cost-effective soon.
In conclusion, it's clear that allocating healthcare resources in developing countries presents a real challenge to the management scientist. The decision-making process is complicated and requires deep knowledge of modeling as well as of particular diseases. The effort, however, is well worth it. Modeling helps developing countries properly allocate precious health-care resources and save thousands, if not millions, of poor, vulnerable people. Misallocating funds will cause their death. The stakes are high. The developing world needs more heroes. It needs more management scientists.
|
References
- Cuddington, J. T., (1993), "Modelling the macroeconomic affects of AIDS, with an application to Tanzania," World Bank Economic Review, Vol. 7, No. 2, p. 173-189.
- Gould, P., (1993), "The slow plague," Cambridge, Oxford.
- Guay, L. A., et. al., (1999), "Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda. HIVNET 012 randomised trial," Lancet, Vol. 354, No. 9181, p. 795-802
- Marseille, E., et. al., (1999), "Cost effectiveness of single-dose nevirapine regimen for mothers and babies to decrease vertical HIV-1 transmission in sub-Saharan Africa," Lancet, Vol. 354, No. 9181, p. 803-809
- Marseille, E., Kahn, J. G., Saba, J., (1999), "Cost-effectiveness of the female condom in preventing HIV, STDs and pregnancy in urban sub-Saharan Africa," International AIDS Economics Network.
- Shaffer, N., et. al., (1999), "Short-course zidovudine for prenatal HIV-1 transmission in Bangkok, Thailand: a randomised controlled trial," Lancet, Vol. 353, p. 773-780.
- UNAIDS (1998), "Spread of HIV over time in sub-Saharan Africa, 1982-1997," Internet: WWW.UNAIDS.ORG, 98036-E-12, July 1998.
- UNAIDS (2000), "Report on the global HIV/AIDS epidemic," June 2000.New York.
- UNAIDS/WHO (2001), "AIDS epidemic update," January 2000.
- WHO (1998b), "The world health report 1998. Life in the 21st century. A vision for all."
- Worldbank (1997), "Confronting AIDS," Oxford.
- Worldbank (2000a), World development indicators, Washington, D.C.
- Worldbank (2000b), World Bank atlas, Washington, D.C.
Steffen Flessa is professor of healthcare management with the Evangelical University of Applied Sciences-Nuremberg in Germany. After completing a master's degree in business administration, he was a lecturer on healthcare management during a four-year stay in Tanzania. He received his Ph.D. in 1996. He has worked as a healthcare management consultant with the Lutheran Church of Tanzania and as a research fellow at the University Erlangen-Nuermberg's Department of Management Science and Operations Research. He has held a professor's position at the Evangelical University of Applied Sciences-Nuremberg since 1998. A frequent consultant in the Third World, he is currently in the Democratic Republic Congo to assist the Lutheran Church to sustain their health services.
OR/MS Today copyright © 2001 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Street, Suite 220, Marietta, GA 30060, USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com
Web Site © Copyright 2001 by Lionheart Publishing, Inc. All rights reserved.