Health Care System Issues
Somsak Chunharas, MD
Health System Research Institute, Nonthaburi, Thailand
In response to the HIV drugs and vaccines in Thailand, especially for pregnant women, what should be the valid issues for consideration in terms of public health policy and practice? The three relevant issues for the health care system and health technologies are: (1) efficacy and safety of the drugs and vaccines; (2) effectiveness with respect to patient and provider compliance; and, (3) efficiency and cost-effectiveness (i.e., what is the best, given the resources and constraints?). These issues can take on an optimistic, critical or over-optimistic perspective.
Optimistic View
Assuming that HIV drugs and vaccines are made available, the question now is how to deliver the technologies to those who need them.
First, the service packages need to be clearly defined. How should the drugs and vaccines be used? What other related services and care are needed? Secondly, ways to reach the target group should be identified. What level of health services is required? What types of health personnel are needed? What are the roles of the family, community, and NGOs?
Thirdly, the health care system has to be reoriented and its capacity built up technically (i.e., how to deal with drugs and vaccines), socially (i.e., perception and sociocultural aspect), and managerially (i.e., how to ensure continuity and proper use of resources). Lastly, it will become necessary to mobilize the family, community and NGOs and empower the partner along with interactive learning.
Critical View
What if nobody sees the drugs as public goods and leaves it to the free market? The government has no public health program and does not pay for the cost of the drugs and vaccines, even for the poor. Health insurance does not cover this medical treatment due to the high cost and its emphasis on primary preventive measures.
Over-Optimistic View
Proponents of the over-optimistic view take on a technology bias and believe that there is no need to worry. More resources are allocated for drugs and vaccines than for other interventions. The health sector dominates the control effort, and there is public demand.
By looking at public health and policy implications, two issues arise. First, with the prospect or possibility of the HIV drugs and vaccines being made available, whether for partial control or eradication of the HIV virus, it is still more important to place emphasis on AIDS prevention, behavioral advocacy, and the value of avoidance rather than becoming too optimistic about the drug or vaccine. Second, proper policies about investment in research for high cost technology such as drug and vaccine for HIV might help to make the pricing issues more manageable and affordable for countries.
In conclusion, with the availability of this technology, there exists a need to look at how the existing health care system can be reoriented to cope more efficiently with the problems of HIV infection which will require more practical types of services.
Impact of AIDS on Health Sector Costs in Developing Countries
Donald S. Shepard, Ph.D.
Brandeis University, Waltham, Massachusetts, USA
Cost of AIDS Care in Thailand
In Thailand, the total health resources devoted to HIV and AIDS prevention, treatment, and mitigation were 2.39 billion Baht in 1994. About a third-848 million Baht or approximately US$34 million-was devoted to treatment of patients living with HIV and AIDS. The public and private sectors respectively finance 60% (70 million Baht) and 40% (379 million Baht) of the of national cost of AIDS treatment care (Table 1). Hospital-based services represent the bulk of HIV-related treatment expenditures. Hospital care includes inpatient and outpatient treatment directly related to AIDS, antiretroviral therapy (ARV), and treatment of other opportunistic infections (OI). Home care is little used and accounts for a small share of total costs.
Thailand spends more on AIDS prevention than curative care at 64% versus 36% of AIDS expenditures (Table 2). The total AIDS-related amount is a modest proportion (6%) of the overall public health budget. The total public expenditures are modest overall in Thailand.
A Comparative Study on the Cost of AIDS Care
The Joint United Nations Program on HIV/AIDS (UNAIDS) estimated that, as of the end of 1996, 29 million people worldwide had been infected with HIV and 6.4 million had already died as a result of the epidemic. About 90% of those living with HIV/AIDS reside in developing countries. To help national governments and donor agencies reassess how best to manage the worldwide epidemic, the World Bank and the European Commission commissioned a brief study of the costs of HIV to the health system of developing countries.
With only two months available for field work, the assignment was approached through case studies in five countries, spanning different levels of economic development and parts of the world: Tanzania and Côte d'lvoire (Ivory Coast) in Africa, Thailand in Asia, and Mexico and Brazil in the Americas. The studies were national in scope in the first four countries. In Brazil, the largest of the five countries, the study focused only on the state of São Paulo, which had the majority of the country's AIDS cases. A collaborative effort among researchers and public health officials in each country and in the United States allowed the study to proceed simultaneously in all regions and also strengthened local capacity to address future questions.
Knowing that no one source of data would be sufficient, the study team used analyses of government budgets and reimbursements, studies from hospitals with careful records, comprehensive health sector analyses where available, and estimates by informed experts. Official epidemiological tabulations of declared AIDS cases understate the burden of disease. HIV infections that have not progressed to AIDS, cases not known to health workers, or suspected cases not completely diagnosed are not counted in official reports. The studies spanned both preventive and curative services, as well as all three sources of funding for these services (government, donors, and private payments).
To elicit expert opinion in the most systematic way, the team consulted experts with diverse perspectives. In Côte d'lvoire, the team and the National AIDS Control Program convened a workshop with researchers, clinicians from the tertiary and smaller hospitals, and a traditional healer who cared for AIDS patients. Workshop participants noted that patterns and sites of care varied according to both the part of the country in which the patient resided and the financial means to buy drugs and private care. Even among those without insurance, a patient living in the country's largest city of Abidjan was estimated to receive more services because of his proximity to hospitals and physicians than one living in the interior. By knowing the cost per day or per consultation, the study team derived the total cost of care.
To compare costs among countries, the team first converted local currencies to United States dollars at market exchange rates. It then converted to "international dollars" using 1994 Purchasing Power Parities, which standardized for international differences in the costs of basic commodities. The resulting per capita expenditures on HIV-related prevention and treatment ranged from US$1.48 in Mexico to US$8.69 in São Paulo, Brazil (Figure 1). The overall average across the five countries was US$4.36. As São Paulo had the highest per capita GDP of all five areas (US$5,600 in 1994 international dollars), its high absolute expenditures per person were understandable. However, although São Paulo spent a lot on health care overall, it actually devoted the lowest share of its health expenditures on HIV illness of all five countries (1.4%).
Tanzania, the poorest of the five countries, devoted 13.7% of its health expenditures to HIV. Surprisingly, the high share did not reflect the burden of curative care but impressive preventive efforts. Eighty-five percent of Tanzania's AIDS expenditures were directed toward prevention, supported largely by international donor contributions. Among the other countries, the share of health expenditures devoted to HIV/AIDS was 1.5% in Thailand, 1.8% in Mexico, and 6.7% in Côte d'lvoire. The overall average was 5.0%.
Most policy makers feel that an adequate prevention program is key to stemming the growth in HIV infections. As part of an international update on the AIDS pandemic (AIDS in the World II) researcher J. Broomberg, estimated the cost of minimal program adequate to control the growth of HIV/AIDS infections. Such a program, for example, would cost US$0.63 per capita in international dollars. Among the five countries, only three-Tanzania, Thailand, and São Paulo, Brazil-achieved this level of expenditures.
Although the short term successes of protease inhibitors and combined drug therapies may hold promise for a few AIDS patients, the cost of their widespread use would be prohibitive. Even without these latest drugs, the lifetime treatment of an average AIDS case in São Paulo, the richest of the case study areas, already cost three times the state's per capita GDP. Thailand is the only one of the five countries in which the majority of public funds are devoted to prevention.
The study identified several promising approaches that could and should be widely replicated. In a 1995 article in the Lancet, H. Grosskurth and others reported on the impressive results of a randomized trial in Mwanza, Tanzania. Through a vigorous program of treating sexually transmitted diseases, which included an enhanced supply of drugs and education and condoms for the patients and their contacts, HIV incidence was reduced by 42%.
Under the direction of the National AIDS Committee chaired by its Prime Minister, Thailand has implemented a policy of universal condom use in commercial sex. The clients themselves sometimes pay for the condoms, but the government monitors the brothels to ensure compliance.
The World Health Organization has developed treatment guidelines for managing AIDS cases. By ensuring that health workers in clinics and dispensaries are trained and supplied with diagnostic materials and basic drugs, such as Bactrim, many opportunistic infections can be treated at a reasonable cost at peripheral health facilities.
Both Thailand and Brazil are starting to expand access to antiretroviral drugs, such as zidovudine (ZVD). Following the finding by the AIDS Clinical Trials Group 076 that ZVD reduces maternal-infant infections by 66%, Thailand has begun to make ZVDavailable for deliveries in a few hospitals. This time-limited use of ZVD appears to be cost-effective and affordable. In a population in which 10% of women are infected, this protocol would add about $30 to the cost of each delivery. Controlled tests of the original and shorter regimens, which would be suitable for more widespread use, are underway in Thailand, Côte d'lvoire, and other countries.
Experience to date advises considerable caution in using public financing to provide antiretrovirals to AIDS patients other than pregnant women. To be effective, the drug must be taken regularly over many months. Drug side effects and stigma compound the problems of taking any medication regularly. Where patient compliance has been assessed, fewer than half of the patients returned after their initial treatment.
The study team found that the HIV epidemic is most advanced in sub-Saharan Africa, where over half of the medical patients in many tertiary hospitals are HIV-infected. In Côte d'lvoire, for example, one person in 20 is HIV-infected, and an HIV-infected patient occupies one hospital bed in five.
Despite the HIV burden, the health systems of the five countries have managed so far to cope with the AIDS epidemic, since AIDS patients receiving expensive treatment still represent a limited share of the population. It takes five to ten years for an HIV infection to progress to clinical AIDS.
For the future, however, the 20 million persons already infected with HIV in developing countries pose a sobering challenge. The highest priorities for future expenditure are based on prevention (particularly ensuring that persons with many sexual partners always use condoms) and first line treatment (diagnosis and antibiotic treatment of sexually transmitted diseases and uncomplicated opportunistic infections). Policy makers must balance investing resources for antiretroviral therapy and other sophisticated treatment of AIDS with the growing demands for treating other chronic diseases, such as cancer and cardiovascular disease.
Acknowledgments
The members of the study team were: Justine Agness-Soumahoro, MD; Richard N. Bail, MD, MPH; Charles S. M. Cameron, MPH, MBA; Antonio C. C. Campino, PhD; Roberto F. Iunes, MS; José Antonio Izazola, MD, ScD; Tijcoura Konj, PhD; Sukhontha Kongsin, PhD; Phare Mujinja, MS; Jeffrey Prottas, PhD; Jorge Saavedra, MD, MPH; Donald S. Shepard, PhD; Adple Siluj, MS; Laksami Suebsaeng; Paula Tibandebage, PhD; and Samuel Wangwe, PhD.
Beyond Clinical Trials: Operations Research
Donald S. Shepard, PhD
Brandeis University, Waltham, Massachusetts, USA
Operations research is a set of techniques to systematically analyze current knowledge and uncertainties to improve the operation of program or services. The goals are to implement the best feasible program currently and to gain knowledge for future improvements. The goal of implementing the best program gives operations research a population perspective, rather than a clinical one. Thus, it is concerned with addressing the needs of all those who have a problem or need, and not just those served by current programs. Applied to perinatal AIDS, operations research is concerned with obtaining the best pregnancy outcomes for HIV-positive women and not just patients enrolled in clinical trials. Finally, operations research is action-oriented (i.e., it accepts innovation and learns from experience).
Conceptual Framework
The conceptual framework of operations research entails four components:
Objective. The objective is to operate successful programs with high impact which could be effective in reducing perinatal transmission and have high coverage; and high quality that would have high compliance and quality, be affordable, minimize drug resistance, be ethical and equitable, and be accepted by patients and providers. If successful, Thai programs could be a possible model for other regions or countries, particularly Southeast Asia.
Alternatives. The major alternatives for women during pregnancy are deciding whether to test for HIV infection; whether to offer ZDV during pregnancy and labor; and if so, how much ZDV and whether or not it should be administered for a long or short duration. The option for infected women after pregnancy is to continue treatment after delivery if symptomatic, or resume treatment if they become symptomatic. The options for the infected baby are to: (1) give ZDV treatment after birth (choosing the amount and duration); (2) encourage bottle feeding by providing and improving the distribution of formula milk and monitoring hygiene; and (3) perhaps provide treatment if the child is symptomatic at one or two years of age. Educational and monitoring services are needed to strengthen diarrhea management, strengthen social acceptance of people living with HIV/AIDS (Persons With AIDS, or PWAs), and provide counseling regarding enrollment and compliance during prenatal care.
Constraints. The constraints include cost (limited budget for drugs and limited resources for general public health programs), compliance among patients and providers, and drug resistance.
Relationships. Relationships pertain to how the alternatives relate to the objectives of controlling perinatal transmission. The impact of a public health program depends on the product of two factors, which are effectiveness and coverage. The components of effectiveness include: (1) diagnostic accuracy or how well screening identifies the person with the problem; (2) efficacy, or how well the intervention when properly applied addresses the problem; (3) provider compliance or how well providers do what they are supposed to do; and (4) patient compliance or how well patients do what they are supposed to do. Coverage refers to the share of patients having the problem who actually receives treatment.
Empirical Base
There are about 5,000 HIV-positive pregnancies per year in North Thailand; about 2.1% of the region's pregnant women in North Thailand are HIV-positive. Approximately 1,000 pregnant women per year can be enrolled in the North Thailand Perinatal HIV Prevention Trial, but 4,000 non-enrolled pregnant women will need protection.
ZDV has been proven to be efficacious in industrialized countries by ACTG 076 with 66% efficacy in 1994, and in Thailand by the Chulalongkorn trial with about 50% efficac in 1997. Moreover, ZDV was well-tolerated by Thai women based on the outcome of the Chulalongkorn and CDC trials in Bangkok. However, compliance is poor (<50%) in adult antiretroviral trials not related to perinatal transmission. Table 1 gives the projected costs of alternative AIDS programs.
There is a strong need to demonstrate the efficacy of a short course treatment because, even if the standard or long-term ZDV treatment works as shown in the Chulalongkorn trial, there are financial constraints of extending it to other women. The long regimen would add a third onto the 860 million Baht budget for AIDS treatment.
Another way of understanding the relationship between treatment and outcome is a cost-effectiveness analysis. Treatment cost is the cost per pregnant woman in a cohort: net public sector costs (including the costs of screening and intervention) less present value of avoided treatment costs for antiretroviral drugs, medical services, laboratory tests, etc., for the infected child ($1000). Effectiveness is the disability adjusted life years (DALYs), which is a single scale coined by the World Bank. It is an extension of the Quality Adjusted Life Year (QALY), extensively used in industrialized countries. It adjusts life expectancy for timing, quality of life, and the value of a year of life at various ages. Averting an infant death (with a 65-year life expectancy) gains 34 DALYs. Based on this analysis, possible alternatives were classified as either favorable or unfavorable (Table 2). Favorable alternatives yield a substantial number of QALYs in relation to their costs and merit as wide application as possible. Unfavorable alternatives generate few QALYs in relation to their costs and should receive limited or no use to make the best use of scarce funding.
Recommendations
This brief application of an operations research framework suggests three types of recommendations:
Expand treatment under current guidelines: Thai health authorities currently recommend the following treatment for HIV infected women: offer ZDV to pregnant women beginning at 36 weeks gestational age; give a brief treatment to infants until discharge; encourage mothers to use bottle feeding; and provide subsidized formula milk, charged on a sliding scale based on income; and refer women to NGOs to receive emotional support and potentially serve as future peer counselors.
Monitor coverage: The second recommendation is to monitor progress by monitoring coverage (e.g., percent of deliveries with prenatal care), percent of pregnant women counseled and tested for HIV, and percent of HIV-positive women receiving ZDV. Compliance can be monitored by the percent of HIV-positive women who start treatment at 36 weeks of gestation, percent of HIV-positive women who receive an adequate dose and actually deliver, and percent of babies who receive an adequate dose.
Conduct systematic experimentation: The third recommendation to test interventions entails undertaking a program of experimentation with the goal of finding the best way to improve coverage and compliance. This goal can be reached by working with NGOs to develop outreach strategies as well as with health professionals in teaching and supervising health workers to educate and counsel patients and measuring performance. Results can be evaluated with an experimental randomized design at a health facility or clinic or with health care workers. Likewise, a qualitative experimental design can be used (e.g., before and after comparison at a health facility, or cost-effectiveness comparison of the intervention implemented at a health facility versus control at other health facilities). Using the same strategy as coverage, compliance strategies can be improved by working with health care workers to visit homes, remind patients at each visit, or teaching families to provide support. The results of such work should be evaluated and the successes replicated.
In summary, the recommendations are intended to explore strategies that are working now and can be implemented on a wide basis. What is undertaken should be affordable and feasible within a set of constraints. Monitoring current performance involves testing a series of program components that can lead to better implementation in the future and a replication of success.
International Cooperation
Richard Marlink, MD
Harvard AIDS Institute, Boston, USA
The logistics of clinical and laboratory collaboration in Africa are the following: (1) training each other on how to get the job done; (2) "a Sengalese project" done through the public health system in Senegal and determined by the Sengalese involved; (3) multi-national input; and (4) expenses or determining individual expenditures in the country.
The public health issues on how to do international collaboration and set up long-term public health efforts are that policy is created at the top of the pyramid and decisions are made for the population. A national AIDS committee can eventually become a policy-making mechanism. As discussion is essential, it can be facilitated by creating a forum for discussion and dissertation and involving international and/or private agencies.
In terms of ethics according to the Helsinki Declaration, an international forum for ethical issues is important. However, decisions should be made locally or "in-country" with multidisciplinary input at the local level. Overall, whenever possible, a national or specific policy should be established.
Although lacking experience with AIDS, top down efforts were initiated with collaborators to create awareness of the epidemic, resulting in Sengalese landmarks early in the epidemic (Table 1). HIV-2 in Senegal has been studied as a model for long-term non-progression. A study of disease-free survival in seroincident women where the endpoint is CD4+ count < 200 cells/mm3 found that HIV-1-positives survived fewer years since seroconversion than HIV-2-positives. This demonstrated a different natural history of types and subtypes in the comparison between HIV-1 and HIV-2 (Table 2).
With respect to the public health impact of AIDS in Africa, the issues are complex and there is no simple solution. A multidisciplinary approach needs to be taken, while commitment and collaboration must be long-term. Solutions to problems in Africa can best be found by African researchers and public health officials in place. Helping and studying the epidemic in Africa will help both Africa and the rest of the world.
International collaboration has helped foster diagnostic testing for HIV in laboratory; treatment; and telecommunications. Starting with the lack of a regional testing laboratory, a non-commercial testing analysis (immunoblot) which was much less expensive than commercial kits was developed for testing and confirming HIV-1 and HIV-2 (US$3.13 versus US$21.05 for HIV-1 in 1994). In addition to dramatic savings, the new diagnostic technique was put in place to screen and confirm large numbers of patients.
Secondly, the individual health needs of female sex workers were addressed by providing them with better treatment in the clinic and setting up extra physicians and nurses who were not previously available. This treatment also offered education and information; access to care and medications; access to condoms; education for male STD patients; and expenses.
Thirdly, the Program for Collaboration Against AIDS and Related Epidemics (ProCAARE) was established as an ongoing e-mail conference on AIDS via satellite. ProCAARE contains up-to-date abstracted literature on AIDS and tuberculosis, STDs, and other retrovirals, and represents a cost-effective alternative to traveling to a conference or placing a telephone call. These three examples illustrate that new ways can be found to get the job done, whether through testing in laboratory, better treatment in a clinic, or better communications through telecommunications.
