Out-of-pocket costs of AIDS care in China: are free antiretroviral drugs enough?

Abstract

Financial access to HIV care and treatment can be difficult for many people in China, where the government provides free antiretroviral drugs but does not cover the cost of other medically necessary components, such as lab tests and drugs for opportunistic infections. This article estimates out-of-pocket costs for treatment and care that a person living with HIV/AIDS in China might face over the course of one year. Data comes from two treatment projects run by Médecins Sans Frontières in Nanning, Guangxi Province and Xiangfan, Hubei Province. Based on the national treatment guidelines, we estimated costs for seven different patient profiles ranging from WHO Clinical Stages I through IV. We found that patients face significant financial barriers to even qualify for the free ARV program. For those who do, HIV care and treatment can be a catastrophic health expenditure, with cumulative patient contributions ranging from approximately US$200–3939/year in Nanning and US$13–1179/year in Xiangfan, depending on the patient's clinical stage of HIV infection. In Nanning, these expenses translate as up to 340% of an urban resident's annual income or 1200% for rural residents; in Xiangfan, expenses rise to 116% of annual income for city dwellers and 295% in rural areas. While providing ARV drugs free of charge is an important step, the costs of other components of care constitute important financial barriers that may exclude patients from accessing appropriate care. Such barriers can also lead to undesirable outcomes in the future, such as impoverishment of AIDS-affected households, higher ARV drug-resistance rates and greater need for complex, expensive second-line antiretroviral drugs.

Keywords:

• cost of AIDS care
• access to treatment
• China
• MSF
• out-of-pocket costs
• AIDS

Introduction

Financial access to healthcare is a difficult problem for many people in China, where there is a widespread fee-for-service system but an estimated 70% of the population has no health insurance (Blumenthal & Hsiao, 2005, p. 1167; Development Research Center of the State Council, 2005). For those living with an impoverishing chronic illness such as HIV, this challenge is even more daunting. In 2003, the government began providing free antiretroviral (ARV) drugs (among other services1), to the rural and urban poor living with HIV. However, adequate HIV-related care and treatment requires many other components beyond the ARV drugs themselves – such as lab tests and hospitalizations – and the resulting costs are unaffordable for many.

This article quantifies the out-of-pocket costs for treatment and care that persons living with HIV/AIDS (PLWHA) in China might face over the course of one year, depending on their WHO clinical stage of HIV infection. It does not estimate the costs to the health system, but rather, quantifies the costs that an individual patient might pay. Data is drawn from two HIV/AIDS treatment projects run by the aid organization Médecins Sans Frontières (MSF) in collaboration with the local Chinese Centers for Disease Control (CDC), in Xiangfan City, Hubei Province and Nanning City, Guangxi Province. Data from these two projects reflect a seemingly widespread problem – that cost poses a persistent and serious barrier to accessing medical care. While data from two sites cannot represent the breadth of China, they do provide important evidence that the cost of HIV care can quickly rise beyond affordable levels.

Background

AIDS in China and the "Four Frees, One Care" policy

HIV-related illness is widely recognized by both national authorities and the international community as a serious and growing problem in China. The government estimates that 650,000 people were living with HIV, 75,000 of whom are in the final clinical stage of infection (AIDS) as of the end of 2005 (Ministry of Health, UNAIDS & WHO, 2006, p.1). The numbers of new HIV infections and deaths have both been increasing and authorities fear it is spreading from high risk groups to the general population (Xinhua News Agency, 2006a).

The government began rolling out antiretroviral therapy (ART) in 2003 under the Four Frees, One Care policy; 28,757 PLWHAs were receiving free ARV drugs as of the end of 2006 (Xinhua News Agency, 2006b). Furthermore, in July 2004, the Ministries of Health and Finance, recognizing that opportunistic infections (OIs) also created a heavy financial burden for patients, directed provincial and local authorities in high HIV-prevalence areas to provide OI drugs at “no or reduced charge” for “AIDS patients in economic hardship” (Ministries of Health and Finance, 2004). A number of provinces, including Henan, Anhui, Hubei and the relatively wealthy municipalities of Beijing and Guangzhou, have adopted policies subsidizing various aspects of care, from providing some OI drugs free of charge to covering a fixed amount of hospitalization costs per year (Guangzhou City, 2005; Hubei Province, 2005). However, implementation of these policies can be uneven within provinces. In addition, a person must have hukou (an official residency registration card) to be eligible for these benefits, thus excluding many migrant workers. Furthermore, in the majority of the country not covered by such initiatives, cost is even more likely to be a barrier to accessing care.

Médecins Sans Frontières treatment projects in China

Médecins Sans Frontières has been working in China since 1988, carrying out both emergency interventions and long-term projects. In 2003, MSF opened two joint projects, with the Hubei Province Public Health Bureau and Guangxi CDC, respectively, to provide free, comprehensive HIV/AIDS care. In both projects, patients come from a defined geographical catchment area, generally within three hours’ travel time from the clinic site. Patients are either referred from area hospitals, voluntary counseling and testing centers (VCT) or self-referred. Though some patients travelled more than three hours to the project sites, patients are generally not excluded or denied treatment once they present at the clinic. At the time of data collection, in Xiangfan, about 70% of patients were farmers, 49% of patients were women, and the cohort had a mean age of 38. In Nanning, about 64% of patients were male (possibly because over one quarter of the patients were injecting drug users who are almost all male) and the cohort had a mean age of 32.4.

As of May 2007, the projects had cumulatively followed 1263 patients, 599 of whom were continuing to receive ART, including 54 children. Patients present relatively late at the clinics, with about half of patients in WHO Stage III (advanced disease) or IV (severe disease). Services provided include pre- and post-test counseling, consultations, diagnosis and management of OIs, ARV treatment preparedness and counseling, ARV treatment, prevention of mother-to-child transmission (PMTCT), hospitalization and long-term follow-up.

Methods

Because MSF provides services free of charge, the projects pay for many costs that the patient would otherwise bear, such as lab tests and OI drugs, thereby generating a significant amount of price data not publicly available elsewhere. We have used the prices paid by MSF for various components of care to construct out-of-pocket non-ARV expenses that a patient would face over the course of one year, without the financial support of an outside actor such as the local government or a non-governmental organisation. Costs were calculated based on the national treatment guidelines (Chinese Center for Disease Control and Prevention, 2005, p.10) and based on data collected between October 2005 and March 2006. Data come from official policies regarding the price of HIV testing and medical consultations and prices paid by MSF for commonly-used drugs, hospitalizations, Caesarean section delivery,2 infant formula and transport3 (lodging excluded4). All prices are provided in US dollars (US$), although calculations were done in Chinese renminbi (RMB); we use an exchange rate of 8 RMB/US$, the approximate rate at the time of data collection. For the remainder of this article, the term “cost” refers specifically to out-of-pocket costs to the patient.

Limitations of data

For most of the data categories listed above, it is reasonable to assume that individual patients would face similar prices as MSF, with the exceptions of OI drugs and hospitalization.

Opportunistic infections drugs

OI drug prices vary widely by producer, distributor or other factors; thus patients’ prices may be significantly different from MSF's. Prices that patients pay in pharmacies or hospitals may be higher than the wholesale prices paid by MSF, due to overhead and profit margins. Mark-ups on drug prices may be very high when a health facility relies on drug sales to cover salaries or other overhead costs (Liu, 2004, p. 532) or when there are perverse incentives to over-prescribe drugs (Wagstaff, 2005, p. 14). According to the Development Research Center of the State Council (2005), too often in China “hospitals and doctors choose and use medicine in terms of maximizing their own profit” (p. 15).

Hospitalization

Similarly, for hospitalization, prices charged to MSF could systematically differ from those charged to patients. On the one hand, since it is a large international organization, MSF may be perceived as having greater capacity to pay and therefore receive generally higher bills than individual patients. On the other hand, patients sometimes face inflated hospital bills because medically unnecessary items are recommended or provided (Zhang, Pan, Yu, Wen, & Zhao, 2005, p. 879); without expert medical knowledge, patients are in a weak negotiating position to contest costs. Liu & Mills (1999) found that nearly 20% of expenditure for appendicitis and pneumonia in China was clinically unnecessary and that the length of hospitalization could have been reduced by 10–16% without any adverse effects. Because each project site employs HIV specialist physicians, MSF is well-placed to contest unnecessary services and thus may face lower overall hospital bills.

However, the data on both OI drug prices and hospitalization are insufficient to quantify what effects these factors may have or whether they cancel each other out.

Costs by stage of illness: seven patient profiles

Because costs depend very much on how ill an HIV-infected person is, cost of treatment for an ‘average’ patient is not a useful concept. Instead, we calculated costs for six patient profiles in various clinical stages of HIV-related illness, plus one PMTCT intervention, with results summarized in Table 3. The seven patient profiles have been simplified to capture the most significant and/or quantifiable expenses; therefore, real costs may well be higher than those calculated here.

Patient 1: No ART, good health (WHO Stage I: asymptomatic)

In the lowest-cost scenario, the patient is HIV-positive but in relatively good health and requires little medical attention. CD4 count is greater than 350 cells/mL and patient has no OIs; needs to attend clinic once every six months for check-up and CD4 test; no ART.

Patient 2: No ART, declining health (WHO Stage I: asymptomatic)

CD4 count between 200 and 350 and no OIs; CD4 gets checked every three months; no ART.

Patient 3: ART, lower-cost estimate (WHO Stage II: mild disease)

ART necessary on the basis of CD4 count below 200. Expenses begin to rise. Assume patient has no OIs; no need for hospitalization; and no additional trips to the clinic beyond the minimum for monitoring tests.

Patient 4: ART, lower-middle cost estimate (WHO Stage III: advanced disease)

Patient 4 is similar to Patient 3, but experiences two of the most common OIs, pulmonary tuberculosis (TB)5 and oral candidiasis, both of which are relatively inexpensive to treat. Assume no need for hospitalization.

Patient 5: ART, upper-middle cost (WHO Stage IV: severe disease)

Patient 5 similar to Patient 4, but experiences three different, relatively common HIV-related illnesses in one year: herpes zoster (shingles), vulvovaginal candidiasis and pneumocystis carinii (or jirovecii) pneumonia (requiring a 21 day hospitalization). Hospitalization dramatically increases costs: hospitalization costs (summarized in Table 1) calculated based on mean cost per day of hospitalization in each project.

Table 1. Cost of hospitalization.

Hospital costs	Nanning	Xiangfan
Number of hospitalizations 2005	131	123
Mean cost/hospitalization: US$ (SD)*	1338 (1268)	402 (508)
Mean cost per day of hospitalization: US$ (SD)	64 (179)	30 (23)
Mean length of hospitalization: days (SD)	21 (22)	14 (11)
*SD: standard deviation.

Patient 6: ART, high-cost estimate (WHO Stage IV: severe)

Patient 6 has multiple OIs, including the costliest to treat: penicilliosis (an OI specific to Nanning6), atypical mycobacteria infection and oesophageal candidiasis. This type of patient also frequently requires lengthy hospitalization. We chose the 75% quartile of hospital bills paid in order to represent above-average costs (1824 US$ Nanning, 469 US$ Xiangfan).

Patient 7: PMTCT costs

Pregnant woman requiring PMTCT; has no OIs; and has been initiated on ART, which the government has provided free of charge in line with official policy. We add to the regular costs of ART the cost of delivery by Caesarean section and six months of infant formula.

Results

Costs pose an important barrier to entering the free ARV program

There are a number of financial barriers to entering the free ARV program in Xiangfan and even greater ones in Nanning. There are several reasons for the major differences in cost between the two sites.

According to the national treatment guidelines, a patient must first be screened for HIV infection with an ELISA test, followed by confirmation with a Western blot (WB) assay and baseline laboratory tests (costs detailed in Table 2) before being able to access free ART.

Table 2. Costs of entering free ARV program.

Component	Description	Nanning (US$)	Xiangfan (US$)
ELISA HIV test	Test at VCT (free if patient provides name and identification) or in hospital [hospital cost]	Free or [3.75]	Free or [10]
Western blot	Confirmation HIV test	62.5	Free
Basic lab tests	9 lab tests required by national guidelines	62	22
Transport	Cost for median round-trip to clinic×3 trips	8.75×3 = 26.25	2.5×3 = 7.5
	Total	150.75 US$	29.5 US$

The ELISA screening test is available at specially appointed VCTs; if patients provide their name and identification number the test is free, otherwise they must pay to be tested. Patients who do not want to give their name and ID number, live far from a VCT centre and/or are unaware of such services, may pay for an ELISA at a local hospital.

While national policy requires that a WB confirmatory test be carried out, it does not stipulate who should pay for the test, so policies and prices vary by province. In Hubei province (Xiangfan site), the Global Fund for AIDS, Tuberculosis and Malaria has provided funds so that WB tests can be given free of charge; however, in Guangxi province (Nanning site), patients had to pay 62.5 US$ at the time of data collection (though government policy has since reduced the price to about 12.5 US$).

These are the first major financial barriers a patient may face and these fees may have blocked many from accessing the free ARV program.7

Next, the national guidelines require the following lab tests before starting ARVs: complete blood count and differential; blood sugar level; creatinine; CD4 cell count; hepatitis B antigen test; hepatitis C antibody test; chest X-ray; alanine aminotransferase (ALT); and amylase (Chinese Center for Disease Control and Prevention, 2005, p. 10). There are no nationally-determined prices for lab tests; we found that lab tests and transport are generally costlier in Nanning, which is a provincial capital, than in Xiangfan, which is a prefectural capital and smaller city.

Once the three steps above have been completed, a patient could enter the free ARV program, having spent 150.75 US$ in Nanning and 29.5 US$ in Xiangfan, including estimated transport costs.

For patients who can afford to enter the free ARV program, costs begin to rise steeply as the illness progresses

Results for the seven patient profiles described above are given in Table 3. Increasing costs are mainly due to the high cost of treating or preventing certain OIs, such as cryptococcal meningitis or penicilliosis, and the very sizeable expense of hospitalization. Figures 1 and 2 indicate the share of costs due to each category of care, including testing (Western blot, CD4 and all other lab tests), OI (outpatient), hospitalization and transport.

Figure 1.  Cost by component – Nanning.

Figure 2.  Cost by component – Xiangfan.

Table 3. Summary of costs by patient profile.*

	Patient profile	WHO Stage	Nan-ning (%)	Xiang-fan (%)	Description	Year	Nanning (US$)	Xiangfan (US$)
1	Non-ART Good health	I	28 n = 144	22 n = 80	CD4 > 350; HIV Elisa, Western blot, CD4 test each 6 months.	1^st year:	200	12.5
						Following years:	80	5
2	Non-ART Declining health	I	28 n=144	22 n = 80	CD4 250-350; HIV Elisa, Western blot, CD4 test each 3 months.	1^st year:	280	17.5
						Following years:	160	10
3	ART Lower cost	II	13 n = 67	28 n = 101	Starts ART after baseline blood tests; monitoring blood tests; no OI; no hospitalization	1^st year:	359	81
						Following years:	104	27.5
4	ART Lower-middle cost	III	33 n = 170	35 n = 127	Starts ART, necessary blood testing; 2 common and relatively inexpensive OIs; no hospitalization	1^st year:**	395	140
5	ART Upper-middle cost	IV	25 n = 129	15 n = 54	Starts ART; blood testing; 3 common HIV-related conditions, including a costly one requiring a 21-day hospitalization	1^st year:**	1723	746
6	ART High cost	IV	25 n = 129	15 n = 54	Starts ART; blood testing; 3 costliest OIs; above-average hospitalization cost	1^st year:**	3939	1179
7	PMTCT	n/a	n/a	n/a	Starts ART; no OIs; C-section delivery; infant formula	1^st year:**	1498	818
*Approximate costs for first year and subsequent years of HIV care, estimated for various stages of illness. We approximate the proportion of patients that may fall into each category by indicating in columns four and five the percentage of each project's cohort that presented in each WHO stage of illness (column three).
**Since it is unlikely the same OIs would occur year after year, we do not provide an estimate for each additional year's costs for Patients 4,5,6 or 7.

Generalizability

How common are each of these patient profiles in China? Publicly available data is scarce regarding disease progression, OI incidence and need for hospitalization in China. Liu et al. (2002) estimated that PLWHA require an average of 1.2 hospitalizations per person/year, but does not provide any data on WHO stage or occurrence of OIs (p. 9). Thus, we have made an approximate estimation of the percentage of patients facing the various cost levels by associating each patient profile with the WHO clinical staging system for HIV infection. While a direct extrapolation would not be realistic, this method of analysis can provide an approximate idea of how applicable each profile is. In Table 3, each patient profile number is followed by a WHO stage and the percentage of patients in Nanning and Xiangfan that first presented at the clinic in the relevant stage. At least 50% of the cohort presents in WHO Stage III or IV, where costs are already very high.

Drugs to treat HIV-related conditions

We identified the drugs needed to treat all outpatient HIV-related conditions that affected at least 1% of the cohort and found just eleven different medicines; of these, six cost less than $1.25 per treatment course, but five were particularly costly: acyclovir, clarithromycin, ethambutol, fluconazole and itraconazole. For these drugs, cost ranged from a low of 8.75 US$ for a seven-day treatment course of acyclovir to as high as 1290 US$ for one year of fluconazole prophylaxis.

Hospitalization

Costs of hospitalization vary greatly, depending on the medical problem, services available, level of hospital, the individual patient and other factors. Nevertheless, even median costs per hospitalization can be catastrophic for the average patient, equaling 66% of annual per capita income for a farmer in Xiangfan and 26% for an urban resident. In Nanning, median costs per hospitalization represent 273% of a farmer's per capita income and 79% for an urban resident. Hospital bills can also often be inflated with unnecessary services. According to Chinese Health Minister, Gao Qiang, “Presently, hospitals are paying too much attention to making profit by selling medicines and ordering unnecessary tests instead of caring about the interests of patients” (as cited in Feng, 2006, p.1).

Large differences in hospitalization costs between the two sites may be due to differences in costs of overhead, equipment and labor between a provincial capital city (Nanning) and prefectural hospital (Xiangfan) and longer median and mean lengths of stay in Nanning than in Xiangfan. These differences correspond to the findings of Liu et al. (2002) that AIDS patients spend a mean of 24.43 days in the hospital at 61 US$/day in “economically developed areas” but only 14.7 days at 39 US$/day in a provincial hospital of an “economically underdeveloped area” (p. 9).

Discussion

The findings illustrate that the cost of medical care and treatment for those infected with HIV are significant, even when there is access to free ARV drugs. Sometimes, treatment costs exceed the cost of the ARV drugs themselves.8 Non-ARV costs of care begin to match or exceed the (government procurement) cost of ARV drugs at the lower-middle cost level (Patient 4) in Nanning and upper-middle cost level (Patient 5) in Xiangfan.

Such costs can quickly rise beyond the reach of many patients (see Figures 3 and 4). An estimated 70% of PLWHA in China live below the poverty line (Liu et al., 2005, p.18). Average annual per capita income of farmers is 335 US$ in Nanning (Annual Statistics Report of Nanning City, 2005) and 399 US$ in Xiangfan (Annual Statistics Report of Xiangfan City, 2005). Thus, in Nanning, even a lower-cost ART patient's annual medical fees (Patient 3, year 1) will practically equal annual income.

Figure 3.  Costs as percentage of rural and urban annual per capita incomes – Nanning.

Figure 4.  Costs as percentage of rural and urban annual per capita incomes – Xiangfan.

Income for HIV-affected households may be even lower: in Xiangfan, 82% of MSF patients are farmers, with a median self-reported annual income of 313 US$. Infection with HIV results in impoverishment for many reasons – it is not only the costs of medical care that wear down a family's resources, but also that people living with HIV/AIDS become less able to work if they fall ill and/or if stigma and discrimination hamper their ability to work (Russell, 2003, p. 26). Other family members may have to give up income-earning opportunities to stay home and provide care to a PLWHA who is ill (Meessen, 2003, p. 581). In their study covering Beijing and eight provinces, Liu et al. (2002) found that when one family member developed AIDS, the family lost 30% of its annual income (p. 9).

For those living in urban areas, average annual per capita income is much higher at 1150 US$ in Nanning and 1018 US$ in Xiangfan. Still, the lower-middle cost patient's medical expenses (Patient 4) equal 34% of the annual income of an urban resident in Nanning and 14% in Xiangfan. Upper-middle cost (Patient 5) and high cost (Patient 6) figures far exceed annual urban and rural incomes in Nanning and rural incomes in Xiangfan (see Figure 1). These expenses could easily be considered “catastrophic” health expenditures, defined as anywhere from 5–40% of a household's income after subsistence needs are met (Ke et al., 2003, p. 111) (see Figures 3 and 4).

These figures may even be an underestimate, as Liu et al. (2002) found that patients spent on average 2188 US$/year for care without ART and 10,250–13,000 US$ with ART (p. 9). These figures are relatively high compared to the Nanning and Xiangfan figures, possibly because they were calculated before the free ARV policy was announced in 2003. Since then, important changes such as the roll-out of ART and widespread use of generic ARVs would have reduced costs. Nevertheless, Liu et al. (2002) also reported that 75.5% of AIDS patients paid out-of-pocket for medical expenses, while only 10.3% were covered by the public system and 7.9% by insurance companies. While the percentage covered by the public system would have increased since then, the level of insurance coverage for this population is strikingly low.

This analysis has assumed that ARVs are provided continuously and free of charge to patients, in line with official policy. However, MSF has been contacted numerous times regarding stock-outs of drugs, because the healthcare facility has run out, never stocked the drug or has insufficient quantities to meet patient needs. As ARV treatment should not be interrupted, the patient must often turn to the private market to buy ARV drugs. There is no publicly available national data on stock ruptures, however, MSF has received requests to assist with stockouts of two drugs in particular: lamivudine (3TC) and efavirenz (EFV), both included in the free national ARV list. One month's supply of 3TC (150 mg, 60 tablets) costs 163 US$ and one month's supply of EFV (600mg, 30 tablets) costs 77 US$ in the private market (MSF project data, 2006). Furthermore, for those who are unable to access the free ARV program at all, the cost of ARV drugs will remain a very considerable expense. Table 4 indicates annual wholesale prices of various first-line ARV regimens in the private sector.

Table 4. Sample costs of first-line ARV regimens (one year, patient > 60kgs).

Triple-combination regimens	Cost/year (US$)
D4T + 3TC + NVP	751
AZT + 3TC + NVP	1240
D4T + 3TC + EFV	1456
AZT + 3TC + EFV	1944
D4T (stavudine, 80mg/day), Chinese generic; AZT (zidovudine, 600mg/day), Chinese generic; 3TC (lamivudine, 300mg/day), Brand-name import; NVP (nevirapine, 400mg/day), Chinese generic; EFV (efavirenz, 600mg/day), Brand-name import.

Studies in other countries have shown that charging user fees for components of HIV care can have important consequences, including: delayed entry into programs, negative impacts on adherence and regularity of clinic attendance and higher mortality rates (Desclaux et al., 2002). Kumarasamy et al. (2006) found that in southern India, cost was the most frequently cited reason for discontinuing therapy (cited by 64% of discontinued patients) and the second-most cited reason for modifying therapy (19% of patients, after secondary effects). In a review of publications on ART in developing countries, Ivers, Kendrick, and Doucette (2005) found that providing medicines free of charge was associated with a 29–31% higher probability of achieving undetectable viral load at six and 12 months than if the patient had to pay part or all of the cost of ART. Similarly, a study covering 18 ART programs in low-income countries found that mortality was higher in programs that charged user fees (Dabis, Egger, & Schechter, 2006, p.817). In short, user fees have been associated with factors that lead to worse health outcomes and a greater risk of developing resistance in the future due to decreased adherence.

Transport

Although representative transport costs are difficult to ascertain, they may pose significant financial barriers. In an MSF HIV project in Malawi, Zachariah et al. (2006) found that patients who had to pay less than 0.50 US$ per one-way trip to the clinic were four times as likely to agree to initiate ART as those who had to pay over 1 US$.

Specific patient populations

Specific patient populations such as injecting drug users (IDUs) and pregnant women may require services that imply greater costs. For example, there is a methadone maintenance clinic in Nanning for IDUs, but methadone costs 1.25 US$/day, totaling 445 US$ per year (assuming continuous use). For pregnant women, the cost of PMTCT implies an additional expenditure of 1135 US$ in Nanning and 708 US$ in Xiangfan. While the government policy is to provide ARVs for PMTCT (zidovudine or nevirapine) the cost of these drugs is but a small percentage of the total cost of preventing HIV transmission.

Future considerations

In addition to the expenses that patients already face, the future may bring increased costs as the need for access to second-line ARVs grows. Over time, increasing numbers of patients will become resistant to first-line drugs and require more expensive second-line regimens. Key second-line drugs (lopinavir/ritonavir, tenofovir) were not available in China at the time of data collection, so country-specific prices did not exist; however, in other low- and middle-income countries, second-line regimens cost 7–28 times more than first-line regimens (MSF, 2005, p. 1). Even if the government provides these drugs free of charge through the national program, it is not clear that the tools needed to initiate and monitor second-line therapy will be covered, particularly viral load tests. Currently, viral load testing is only available in some large cities. Médecins Sans Frontières recently analyzed viral loads for 65 patients in Xiangfan and paid 175 US$ per test at the national reference lab. Resistance testing, which can guide physicians on which second line drugs to choose for patients, is also important but is available in few sites.

Conclusion

This study demonstrates the unaffordable costs that many PLWHA in China may face in trying to access adequate care and treatment, despite the availability of free ARVs through the government program. The high cost of HIV tests and other lab tests may prevent some patients from even entering the free ARV program; for those who do enter, many are likely unable to afford a medically adequate package of care. Further research is needed to understand how patients and their families cope with such financial barriers, the quality of care they receive and the ultimate impact on health outcomes and family finances.

We urge policy-makers to consider seriously providing a free minimum package of HIV care that goes beyond provision of ARV drugs to include HIV tests, consultations, laboratory testing, hospitalizations, prophylaxis and treatment of common OIs, as others have also advocated elsewhere.9 Evidence from MSF projects, as well as other studies, indicate – not surprisingly – that health outcomes suffer when patients cannot afford the care they need. Beyond the individual patient, decreased adherence also implies more widespread resistance, which could lead to even greater technical and financial challenges in the future. Policy-makers should think beyond ARV drugs when funding ART programs, both for the health of the patient and the long-term viability of treatment programs.

Acknowledgements

We wish to thank two anonymous referees for their helpful comments and suggestions, the staff and PLWHA of the Xiangfan and Nanning projects and other PLWHA who have brought these issues to our attention. All errors remain our own.

Notes

1. The Four Frees, One Care policy also stipulates “…(b) free voluntary counselling and testing (VCT); (c) free drugs to HIV-infected pregnant women to prevent mother-to-child transmission, and HIV testing of newborn babies; free schooling for children orphaned by AIDS; and (d) care and economic assistance to the households of people living with HIV/AIDS.”

2. Delivery by Caesarean section is standard practice in China for PMTCT. However, abortion is sometimes recommended as the first option, before PMTCT. For example, Article 21 of the Hubei province policy states, “If either the husband or the wife, or both of the couple, is HIV-positive, the pregnancy should be terminated; if they don't want to terminate the pregnancy, they must adopt measures to prevent mother-to-child-transmission of HIV.” (Hubei Province, 2004). We do not include price information on abortion for lack of data.

3. Median amount per round-trip that MSF reimbursed to patients, extrapolated to estimate annual transport costs. Médecins Sans Frontières reimburses transport costs to selected patients based on needs assessments made by counselors; transport costs also vary widely depending not only on where patients live, but also on the type of transport, catchment area of the program, availability of services, geographical factors and other issues. Thus, transport figures are not representative and should only be used as a point of reference.

4. Patients frequently mention lodging as an additional cost. However, MSF does not provide financial support for lodging as another NGO has established free shelters in both locations where patients coming to the clinics can stay. Thus, lodging costs are excluded from this calculation for lack of data.

5. As of 2004, national policy states that TB drugs should be free for patients who are sputum positive or negative (with clinical diagnosis). Many TB/HIV-co-infected patients are sputum negative (including 100% of TB-diagnosed patients in the Xiangfan project as of May 2006); yet, the majority of patients seen in the Xiangfan clinic did not receive free drugs for various reasons, indicating that the new policy was not yet implemented everywhere.

6. Penicilliosis is one of the most common OIs seen in Nanning (17%) and is also prevalent in other southern provinces such as Yunnan, Guizhou and Guangdong. However, it is not seen in central China and, therefore, the cost of treatment is not included in the Xiangfan calculation.

7. After a number of patients requested financial assistance, MSF began paying the Western blot fee for all patients in June 2005 in the Nanning project.

8. The government price for a generic ARV triple-combination of stavudine or zidovudine, didanosine and nevirapine is about 438 US$ per year (Bureau of Economic Operations, 2005). There is no publicly available information regarding the government price paid for a 3TC-based combination, although this is the recommended first-line regimen, as the terms of supply from GlaxoSmithKline for 3TC are kept confidential.

9. See Free by 5: Economists’, Public Health Experts’ & Policy Makers’ Declaration on Free Treatment for HIV/AIDS, (2005).