Diomandé, Fabien V. K.a; Bissagnéné, Emmanuelb; Nkengasong, John N.a,c; Maurice, Chantala; Monga, Bena,c; Laga, Mariea,d; Nolan, Monica L.a,c
With 42 million people currently living with HIV and 5 million new infections annually, in 2002 the HIV pandemic has surpassed the most pessimistic forecast . Although a global problem, the brunt of the epidemic is borne in sub-Saharan Africa where 70% of all HIV-infected persons live among less than 10% of the world's population. In sub-Saharan Africa, AIDS is the leading cause of morbidity and mortality, and places an increasing burden on already overstretched social and health care systems which further reduces their capacities to respond to other public health and development challenges . Expanded access to comprehensive, effective HIV prevention and care interventions are a prerequisite for African development and stability.
Côte d'Ivoire has a severe, mature and generalized HIV epidemic with an antenatal prevalence of approximately 10% . In 1998, in recognition of this chronic humanitarian emergency, the national government partnered with UNAIDS to establish the ‘Côte d'Ivoire Drug Access Initiative’ (DAI–CI) to improve access to comprehensive HIV care including subsidized antiretroviral therapy (ART). The cost of treatment fell precipitously through DAI–CI price negotiations with the pharmaceutical industry, subsidies from national and international solidarity funds and the provision of free laboratory tests from the Centers for Disease Control and Prevention (CDC). The 2-year pilot programme was consistent with standards of care in industrialized countries including individualized regimens, sophisticated laboratory tests at baseline to assess eligibility for treatment, and ongoing clinical and laboratory monitoring to assess response to treatment . Special measures were also put in place to increase access for women and children.
The evaluation of the 2-year DAI–CI pilot programme demonstrated a substantial proportion of persons with advanced disease had an initial visit but never began ART. Nevertheless by December 2001, 1011 men, 878 women and 126 children – a total of 2015 persons had commenced ART in the capital city, Abidjan. Overall, ART was successful; among those starting ART in the pilot period, toxicity, and immunologic and virologic outcomes were similar to those reported from populations in industrialized countries [5,6].
The government of Côte d'Ivoire and partners are now planning to decentralize antiretroviral services to increase access to care outside the capital city. The previous screening algorithm to define biologic eligibility for antiretroviral therapy included CD4 cell count and viral load quantification. However, if these sophisticated laboratory facilities and tests are prerequisites for care, they would represent a critical barrier to care for millions of HIV-infected Africans. As in other African countries, the current public health laboratory capacity in Côte d'Ivoire is grossly insufficient to process multiple samples from the millions of persons in need of life-saving treatment [7–12].
Our objective was to examine empirical data from the DAI–CI pilot programme in Côte d'Ivoire to describe the contribution of CD4 cell count and viral load screening tests as part of the screening strategy to identify persons eligible for ART. We also model current biologic eligibility recommendations and a range of test costs to determine if these would have a substantial impact on the cost per identified patient eligible for ART. These should be used to inform policy rapidly to expand access to life-saving antiretroviral therapy in resource-poor countries.
We used empirical data from the DAI–CI pilot programme to determine the direct costs of performing internationally accepted laboratory screening tests, as well as the proportion of clients identified to be eligible for subsidized ART by clinical staging, by CD4 cell count criteria, and by viral load criteria in a serial (sequential) screening algorithm.
Criteria for initiation of subsidized antiretroviral therapy in the DAI–CI programme
Eligibility criteria were set by a national advisory committee and were consistent with international standards of that time. These included:
* HIV infection (HIV-1, HIV-2 or dual HIV-1/HIV-2 infection)
* HIV-related symptoms – CDC stage B or C 
* CD4 cell count < 500 × 106 cells μl (for asymptomatic patients)
* HIV-1 plasma viral load > 10 000 copies/ml (for asymptomatic patients with CD4 cell count > 500 × 106 cells μl)
We then used the same population to model costs with different eligibility criteria, with parallel as well as sequential testing, and with different test prices: we modeled current US and WHO treatment eligibility criteria [8,14] in place of the DAI–CI criteria; and we modeled costs across a range of CD4 cell count and viral load test prices. We considered ‘benefit’ to be the outcome of being defined ‘eligible for subsidized ART’ when meeting clinical criteria, CD4 cell count criteria in the absence of clinical eligibility or viral load criteria alone.
The data were analyzed from the project RETRO-CI maintained DAI–CI database. Project RETRO-CI is a collaboration between the Centers for Disease Control and Prevention (CDC) and the Ivorian Ministry of Health and provides free laboratory testing and data management support to the clinical centers participating in the DAI–CI. The data are primarily collected for patient care and for programme evaluation, however the protocol and written consent forms have also been reviewed and approved by the CDC and Côte d'Ivoire ethics committees.
Patients were either self-referred or referred from other clinics and screened to determine if they were biologically eligible for subsidized ART and socio-economic criteria were used to determine the level of financial subsidy for drugs . At the first screening visit a clinical examination was performed to categorize the stage of the patient's disease according to CDC staging criteria  and blood was drawn for CD4 cell count and hematological (hemoglobin, platelets, total white cell count and profile) and biochemical tests (liver function enzymes: alanine transaminase (ALT), alkaline phosphatase (ALP) and gamma-glutamyltransaminase (GGT); bilirubin, creatinine, urea, glucose and amylase). Patients were reviewed at a second visit, and if they were not eligible by clinical and/or CD4 cell count a further blood draw was performed to determine if the patient was eligible through viral load criteria. Eligible patients without significant hematological or biochemical abnormalities were prescribed subsidized ART. After ART initiation, patients were reviewed 1 month later and then quarterly for clinical and laboratory monitoring including CD4 cell count and viral load testing.
Definition of outcomes of interest and cost parameters
Patients enrolled in the DAI–CI were identified as eligible for subsidized ART if they met clinical or CD4 cell count criteria (< 500 × 106 cells μl). If they did not meet either clinical or CD4 cell count criteria, they were also eligible if they had elevated viral loads (> 10 000 copies/ml). With the data collected, we were able to define each of the following parameters:
a. Each eligible person as eligible for subsidized therapy in the Drugs Access Initiative through meeting:
1. Clinical criteria
2. CD4 cell count criteria (in absence of clinical criteria)
3. Viral load criteria (in absence of clinical and CD4 cell count criteria)As 2–4% of clients were ineligible due to clinical or CD4 cell count criteria and did not have viral load results we also estimated the number of clients eligible through viral load criteria and calculated the minimum (actual) and maximum estimates.
b. The programme cost reflects the costs associated with providing CD4 cell count and viral load tests for screening purposes. We used the Project RETRO-CI direct laboratory test costs to calculate the costs associated with performing these tests to identify eligible patients. The additional costs associated with providing CD4 cell count and viral load tests as part of the eligibility screening algorithm (serial or parallel) were calculated.
c. The programme cost per eligible person identified reflects the direct laboratory costs required to identify an eligible person for each screening component (HIV, CD4 cell count and viral load tests, respectively).
All costs are shown in US dollars. Project RETRO-CI laboratory costs reflect only the direct costs for reagents, test kits, consumables, technician salaries and overheads. Shipping, supervision and infrastructures costs are not included and thus represent non-profit, partially subsidized costs. As all patients require confirmation of their HIV-infected status, HIV screening costs are included in the costs for clinical criteria alone. These estimates reflect current project RETRO-CI costs for HIV rapid tests able to differentiate HIV-1 and HIV-2, which can be used at decentralized sites. Direct project RETRO-CI costs were US$13.36 for HIV rapid tests, US$37.23 for CD4 cell count, and US$65.95 for viral load. Hematological and biochemical test costs have not been included.
As these costs do not reflect full market prices, we also modeled two other hypothetical test cost estimates:
1. High test costs (double RETRO-CI estimates) as these are more accurate reflections of current prices which might be available from a commercial laboratory
2. Low-test costs (half RETRO-CI estimates) as these might model what could be achieved if there are significant technological breakthroughs to identify low-cost alternative tests and/or if industry makes significant price reductions, as already demonstrated by many pharmaceutical companies for antiretroviral drugs.
Between 1 August 1998 and 31 July 2000, 2625 persons presented to the DAI–CI, of these 1.8 % were HIV negative, 90% were HIV-1 infected, 2.8% HIV-2 infected, and 5.3% dually infected with HIV-1 and HIV-2. Of the 2576 persons found to be HIV positive, 131 (5%) were under 18 years of age and 307 (12%) had previously received antiretroviral therapy. Of the remaining 2138 adult antiretroviral-naive patients, 1070 (50%) were female, the median age was 36 years, and 74% reported an income < US$ 200/month and 36% were unemployed. Patients generally presented with very advanced HIV disease and median CD4 cell count was 172 × 106 cells/μl.
Results of the serial screening algorithm to identify patients eligible for the DAI–CI
Of the 2138 HIV-positive, antiretroviral-naive adults who came for a screening visit, 2057 persons (96%) were eligible for antiretroviral treatment by DAI–CI criteria, 35 (2%) were ineligible by clinical, CD4 cell count and viral load criteria, and a further 46 (2%) were ineligible by clinical and CD4 cell count criteria but did not return for follow-up and/or had no viral load results (Fig. 1). Most persons had advanced CDC stage B or C at screening (1610 persons or 75% of those screened), a further 392 (18%) were eligible with a CD4 cell count less than 500 × 106 cells/μl in the absence of advanced clinical disease [including 180 (8%) with a CD4 cell count less than 200 × 106 cells/μl] and only 55 (2.6%) were eligible due to a viral load greater than 10 000 copies/ml alone.
Results of the serial screening algorithm modeling current US recommendations
Had the current US recommendations been used to define eligibility for subsidized ART in the same patient population, similar results would have been obtained (Fig. 2). Of the 2138 HIV-positive, antiretroviral-naive adults screened, 1977 persons (92%) would have been identified as eligible for antiretroviral treatment, 80 (4%) would have been ineligible by clinical, CD4 cell count and viral load criteria and a further 81 (4%) ineligible by clinical and CD4 cell count criteria but who did not return and/or had no viral load result. The same number of persons (1610) would have been identified through advanced disease at screening, a further 319 (15%) would have been eligible with a CD4 cell count less than 350 × 106 cells/μl and only 48 (2.2%) would have been eligible due to a viral load greater than 55 000 copies/ml alone.
Since not all the patients attended for a second visit and had the opportunity to receive viral loads, the proportion of patients identified as eligible through viral load testing represents a minimum estimate. Even in the worst-case scenario, when we assign all patients missing viral loads as ‘eligible’, only 6% of patients screened could be eligible using either the DAI–CI (5%) or current US criteria (6%). In the more likely scenario in which we assume that clients missing viral loads had similar viral load patterns to those tested, we estimated viral load screening would have identified 83 (3.9%) eligible patients in the DAI–CI (viral load > 10 000 copies/ml) and 78 (3.6%) with current US criteria (viral load > 55 000 copies/ml) (Fig. 1 and Fig. 2 and Table 1).
Programme costs for eligibility screening
The direct costs and benefits of the laboratory screening are shown in Table 1. Performing both HIV tests and CD4 cell count tests on all 2138 patients screened (the strategy used in the DAI–CI) cost approximately US$28 000 for HIV tests and US$80 000 for CD4 cell counts. Had viral loads been included in a parallel screening strategy it would have cost US$249 269 and resulted in expenditure of 57% of the laboratory screening budget in order to identify less than 4% of eligible clients. A serial screening strategy (total cost of US$57 197) was much cheaper than a parallel strategy (total cost of US$249 269). The use of clinical criteria with HIV testing cost an average of US$18 per eligible person identified. The addition of CD4 cell count in a serial algorithm cost an average of US$50 per eligible person identified through meeting CD4 cell count criteria without clinical symptoms. The additional cost of a person identified through viral load criteria alone (as part of a serial algorithm) was US$108. Using the current US recommendations these additional costs increased to US$62 and 177 for CD4 cell count and viral load screening tests respectively. A WHO screening approach with sequential clinical evaluation and CD4 cell counts would have been the cheapest strategy (total cost US$48 221).
If we simulate more realistic commercial testing prices by doubling these test price estimates we double these additional costs. This suggests that with this patient population the use of viral load as a serial screening tool will identify persons eligible for ART at an average cost in excess of US$200. Even if test prices decreased dramatically, to less than half of our project RETRO-CI subsidized estimates, viral load testing would still cost more than US$50 for each additional patient identified to be eligible.
Clinical screening with HIV confirmation identified the majority of patients eligible for ART at a low cost in the DAI–CI. A further 15–18% could be identified with CD4 cell count screening at a moderate cost. Current WHO recommendations would have identified 8% of asymptomatic persons with CD4 cell counts below 200 × 106 cells/μl in need of HAART. In contrast, the inclusion of quantitative viral load as a serial screening tool identified few additional persons eligible for ART and was extremely expensive. This result was robust across both the DAI–CI eligibility criteria and the current US recommendations: viral load screening identified 4% of persons screened to be eligible for ART, at a minimum cost in excess of US$100/eligible person. This result is primarily driven by the fact that most people seeking antiretroviral care in sub-Saharan Africa have very advanced HIV disease . The profile of patients seeking care in Côte d'Ivoire DAI–CI is very similar to that described for other African countries including Uganda , Kenya  and South Africa .
While there is much discussion of expanding access to care [7–9,12], this is unlikely to change the profile of patients seeking antiretroviral care to any large extent in the near future for two inter-related reasons. First, there are vast numbers of untreated patients with advanced disease in Africa [1,2]. Second, there is limited access to counseling and voluntary HIV testing services in most African countries so that relatively few persons with asymptomatic disease know they are infected with HIV. Stigma and discrimination represent further substantial but not insurmountable barriers to HIV testing. In Côte d'Ivoire, there is a similarly elevated antenatal HIV prevalence at all major urban centers, yet, at the beginning of 2002, there were only four stand-alone Voluntary Counseling Testing (VCT) centers in a country of more than 16 million people. (There are plans to increase the number of VCT centers dramatically in the next 5 years).
In comparison with those with advanced disease, persons with elevated viral loads in the absence of symptomatic HIV disease or low CD4 cell counts have less to gain from commencing therapy . There is ongoing debate about the benefit and harms of delaying treatment with CD4 cell counts in the range of 200–350 × 106 cells/μl [17,18]. The benefit of being identified as eligible with elevated viral load alone is of debatable clinical merit in the presence of an assured long-term supply of affordable drugs and of dubious benefit in their absence (as is the case in Côte d'Ivoire at present). It may be of both individual and public health benefit to delay commencing therapy until clinical or CD4 cell count criteria are met.
CD4 cell count and viral load baseline values may also have utility as part of monitoring the immunologic and virologic response to treatment. We did not consider this when defining ‘benefit’ as ‘being assessed as eligible for subsidized ART’. However, the recent national recommendations in Côte d'Ivoire recommend viral load testing as part of monitoring only in the absence of a good clinical and/or CD4 cell count response. Baseline samples (including filter paper samples) could be stored prior to commencing therapy and tested if clinically indicated without the enormous expenditure required for systematic baseline viral loads.
We are also performing analyses to evaluate the utility of other parameters or combinations of parameters (such as total lymphocyte count, hemoglobin, p24 antigen levels, nutritional indices and client age) for client screening as well as therapeutic response. Ongoing evaluations to define the predictive values of clinical screening, CD4 cell count, viral load and other parameters in various patient populations and contexts are needed to inform optimal screening and monitoring practices in a dynamic environment. In the not too distant future, we hope that sustained international commitment will result in simpler less-expensive laboratory screening tools and widespread affordable VCT and care services in many heavily affected countries such as Côte d'Ivoire.
Our data suggest that viral load testing should not be included as a mandatory part of eligibility screening as it is not an effective use of scarce public health resources. Health personnel should be trained to identify and refer clinically symptomatic persons for HIV care with ART. Overall, in sub-Saharan Africa, there are an estimated 28 million persons infected with HIV, and the first public health and humanitarian imperative must be to expand access to life-saving treatments to the millions of people with the most advanced disease.
All those involved in the design, analysis and support of the Drug Access Initiative, Côte d'Ivoire (DAI–CI) including: all the patients attending the DAI–CI; clinicians and staff at the DAI–CI accredited centers; the RETRO-CI staff: Gaston Djomand, Thierry Roels, Terry Chorba, Madeleine Sassan-Morokro, Rene Ekpini, Marie-Yolande Borget, Babette Gbaie and the clinical, data and laboratory support staff; the CDC Global AIDS Program: Stephan Wiktor, Eve Lackritz and Tedd Ellerbrock, the UNAIDS: Badara Samb and Joseph Saba and the staff from the National AIDS Control Program including Makan Coulibaly (former DAI–CI coordinator). The authors are also grateful for insightful review and comments from: CDC DHAP, Epidemiology Branch: Stefanie Sansom and Paul Weidle.
Contributors: Fabien Diomandé and Monica Nolan contributed to all phases of the study including the study concept, design, analysis, writing the first draft and subsequent revisions. Emmanuel Bissagnéné, John Nkengasong, Chantal Maurice, Ben Monga and Marie Laga contributed to the analysis and revisions of the manuscript for this nested study.
2. Harries AD, Nyangulu DS, Hargreaves NJ, Kaluwa O, Salaniponi FM. Preventing antiretroviral anarchy in sub-Saharan Africa. Lancet 2001; 358:410–414.
3. Ministère de la santé publique, Ministère de la lutte contre le SIDA and Projet RETRO-CI. Surveillance du VIH/SIDA et de la Syphilis en Côte d'Ivoire. Rapport d'activités 2000–2001. Ministère de la santé publique, Abidjan Côte d'Ivoire.
4. Djomand G, Roels T, Ellerbrock T, Hanson D, Diomandé F, Monga B, et al. Virologic and immunologic outcomes and programmatic challenges of an antiretroviral treatment pilot project in Abidjan, Côte d'Ivoire. AIDS 2003; 17 (Suppl 3):S5–S15.
5. Nkengasong JN, Borget MY, Maurice C, Boateng E, Kalou M, Djomand G, et al. Distribution of HIV-1 plasma RNA viral load and CD4+ T-cell count among HIV-infected Africans evaluated for antiretroviral therapy [Letter]. J Acquir Immune Defic Syndr 2001; 28:99–101.
6. Bartlett JA, DeMasi R, Quinn J, Moxham C, Rousseau F. Overview of effectiveness of triple combination therapy in antiretroviralnaive HIV-1 infected adults. AIDS 2001; 15:1369–1377.
7. Taegtmeyer M, Chebet K. Overcoming challenges to the implementation of antiretroviral therapy in Kenya. Lancet Infect Dis 2002; 2:51–53.
9. Binswager HP. HIV/AIDS treatment for millions. Science 2001; 292:221–222.
10. Attaran A, Sachs J. Defining and refining international donor support for combating the AIDS pandemic. Lancet 2001; 357(9249):57–61.
11. Farmer P, Léandre F, Mukherjee JS, Claude MS, Nevil P, Smith-Fawzi MC, et al. Community-based approaches in resource-poor settings. Lancet 2001; 358:404–409.
12. Hosseinipour MC, Kazembe PN, Sanne IM, van der Horst CM. Challenges in delivering antiretroviral treatment in resource poor countries. AIDS 2002; 16 (Suppl 4):S177–S187.
13. CDC. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR 1992; 41:(RR-17).
15. Weidle P, Malamba S, Mwebaze R, Sozi C, Rukundo G, Dowing R, et al. Assessment of a pilot antiretroviral drug therapy program in Uganda: patients' response, survival, and drug resistance. Lancet 2002; 360:34–40.
16. Kasper T, Hilderbrand K, Tshabane N, Louis F, Reuter H, Labatala V, et al. Antiretroviral therapy in primary health care centers in a South Africa Township. XIVth International AIDS conference, Barcelona, July 2002 [abstract MoOrB1095].
17. Weidle PJ, Mastro TD, Grant AD, Nkengasong J, Macharia D. HIV/AIDS treatment and vaccines for Africa. Lancet 2002; 359:2261–2167.
18. Youle M. Strategies of HIV management – when to start. AIDS 2002, 16(Suppl 4): S145–S149.
© 2003 Lippincott Williams & Wilkins, Inc.