Background: A significant proportion of those initiating antiretroviral treatment (ART) for HIV infection are lost to follow-up. Causes for discontinuing ART follow-up in resource-limited settings are not well understood.
Methods: A retrospective analysis was conducted of all adult patients receiving ART at an urban public clinic in Johannesburg, South Africa between April 2004 and June 2005. Patients discontinuing follow-up for at least 6 weeks were identified and further studied, and causes for treatment default were tabulated.
Results: Of 1631 adult patients studied, 267 (16.4%) discontinued follow-up during the study period. Gender, ethnicity, and age were not predictive of loss to follow-up. Of those discontinuing follow-up, 173 (64.8%) were successfully traced. Death accounted for 48% (n = 83) of those traced. Characteristics associated with death were older age at ART initiation (P = 0.022), lower baseline CD4 cell count (P = 0.0073), higher initial HIV RNA load (P = 0.024), and loss of weight on ART (P = 0.033). Date of death was known for 71% (n = 59) of patients traced deceased, of whom 83% (n = 49) had died within 30 days of active ART. Common nonmortality losses included relocation or clinic transfer (25.4%) and hospitalization or illness not resulting in death (10.4%). Few cited financial difficulty or medication toxicity as reasons for discontinuing follow-up.
Conclusions: Nearly 1 in 6 patients receiving ART in a resource-constrained setting had discontinued follow-up over a 15-month period. Early mortality was high, especially in those with profound immunosuppression. Improving access to care and streamlining patient tracking may improve ART outcomes.
From the *Washington University School of Medicine, St. Louis, MO; †Reproductive Health and HIV Research Unit (RHRU), University of the Witwatersrand, Johannesburg, South Africa; ‡Department of Medicine, Johannesburg Hospital, University of the Witwatersrand, Johannesburg, South Africa; and §International Centre for Reproductive Health, Mombasa, Kenya.
Received for publication June 3, 2007; accepted September 18, 2007.
Supported by grants from the President's Emergency Plan for AIDS Relief and the US Agency for International Development.
Data in this article were presented, in part, at the 13th Conference on Retroviruses and Opportunistic Infections, Denver, CO, February 5-8, 2006, and at the Second Annual International HIV Treatment Adherence Conference, Jersey City, NJ, March 29-31, 2007.
Correspondence to: Rishikesh P. Dalal, 7 Maryland Plaza, Apartment 601, St. Louis, MO 63108 (e-mail: firstname.lastname@example.org).
Antiretroviral therapy (ART) has transformed the management of HIV infection in sub-Saharan Africa. ART requires a large commitment from patients, however; adherence is demanding, and resistance develops rapidly after missed doses.1,2 Initiation of ART entails frequent clinic visits, extensive counseling, clinical and immunologic staging, and ongoing monitoring. This is resource-intensive for health care services and for those receiving ART. For patients, indirect costs of ART can include missed days of employment, transport costs, and other out-of-pocket expenses. Discontinuation of treatment and loss to follow-up may signify wastage of these resources, along with poorer health outcomes.1,3 In addition to reinitiation procedures and increased complexity of drug selection, further resources are consumed if the patient becomes ill after treatment default.
Although substantial rates of loss to follow-up have been noted in several studies of ART programs in resource-limited settings,4-8 the causes for loss to follow-up in these settings have received little attention.9 Before the advent of affordable ART medications, poverty was often cited10-12 as contributing to poor clinic attendance, given the substantial cost of the medications and attendance fees. The importance of poverty in loss to follow-up in the era of low-cost ART remains unclear, however. Additionally, more must be done to evaluate the role of drug toxicity and death among those discontinuing ART.13
Our objective was to characterize the population discontinuing follow-up at an urban ART clinic in a resource-constrained setting and determine causes of loss to follow-up, with the goal of informing future service delivery to improve outcomes.
Study Site and Population
The Johannesburg Hospital adult HIV clinic began providing ART in April 2004 as part of the South African plan for Comprehensive Care for HIV/AIDS in the Public Health Sector.14 The clinic serves a diverse patient population from the Johannesburg area, including patients from the inner city center and from surrounding townships. Adults were considered eligible for ART based on the World Health Organization's (WHO's) 2002 recommendations,15 which include a prior AIDS diagnosis (WHO stage 4 disease) or a blood CD4 count <200 cells/μL. In addition, patients attend a series of counseling visits during the pretreatment phase stressing the importance of adherence to ART medications. The choice of antiretrovirals is consistent with WHO guidelines,16 and first-line therapy for treatment-naive patients consists of 2 nucleoside reverse transcriptase inhibitors (NRTIs), stavudine (d4T) and lamivudine (3TC), along with a nonnucleoside reverse transcriptase inhibitor (NNRTI), efavirenz or nevirapine. Provision is made for switching to a second-line protease inhibitor-based regimen in the event of virologic failure. Individual drugs are substituted in the event of severe toxicity.
During the period of study, an access fee based on documented income levels was charged per visit, regardless of medication, clinical, or laboratory resource use. Unemployed patients were eligible to apply for a fee waiver, although the process involved significant administrative steps. Other potential sources of support included a state “disability grant” (approximately ZAR 720 per month; at the time of the study, approximately $114 US) and pensions for those older than 60 years of age. Patients with prior ART exposure with good response were continued on prior regimens when possible.
Patients were generally scheduled for a follow-up visit at the clinic every 30 days and received a 30-day supply of medications at each follow-up visit.
Study Design and Data Collection
By mid-2005, it became apparent that a significant proportion of patients receiving ART were not returning for treatment, as measured by missed follow-up appointments. A study was designed to describe the characteristics of this group and to investigate reasons for loss to follow-up. Ethical clearance for the study was obtained from the Human Research Ethics Committee, University of the Witwatersrand (protocol number M050815).
A retrospective analysis was conducted of all adult patients (age >18 years) at the Johannesburg Hospital HIV clinic receiving ART medications at any time between April 2, 2004 and June 23, 2005. Baseline demographic characteristics of age, gender, and ethnic background of all patients included in the study were analyzed.
Patients discontinuing follow-up, arbitrarily defined as missing a scheduled follow-up appointment for at least 6 weeks, were identified and further studied. Additional demographic and clinical characteristics, including ART regimens, baseline blood CD4 cell count, and HIV RNA load, were collected. Clinical data recorded during ART were extracted from patient charts and electronic health records by skilled data collectors and were verified for accuracy by the authors to maximize the internal validity of the study. Tracing was conducted by telephone contact with patients or their families using contact information collected at clinic enrollment. Home visits were attempted if telephone contact was unsuccessful. On contact, health status (eg, living, deceased) and reasons for loss to follow-up were determined. Some patients were traced as deceased on the basis of review of electronic health records.
Data Management and Analysis
Data were analyzed using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). Patients returning to the clinic after missing an appointment for longer than 6 weeks and who subsequently defaulted again were counted only once in the analysis. For patients traced living, the length of time on ART medications was determined by adding 30 days from the last date of medication collection. For patients traced deceased and the date of death was within 30 days of last date of medication collection, the last date of ART was assumed to be the date of death. For patients traced deceased and the date of death was after 30 days of last medication collection or if the date of death was unknown, the length of time on ART medications was determined by adding 30 days from the last date of medication collection. Univariate comparisons for categoric variables were tested using the χ2 test. For continuous variables, the Student t test and Wilcoxon rank sum test were used for comparing data with normal and nonnormal distributions, respectively. Normality was assessed using the D'Agostino and Pearson omnibus normality test.17 For all statistical tests, statistical significance was defined as less than a 2-sided α-level of 0.05.
During the period from April 2, 2004 to June 23, 2005, 1631 adult patients received ART medications through the clinic. Baseline demographic data for this group were collected. The study population was found to be 66% female (1077 of 1631 patients) and 93.1% black African (1518 of 1631 patients) and had a mean age at enrollment of 36.4 years (SD = 9.1 years). Men were significantly older at clinic enrollment than women (mean: 38.9 vs. 35.2 years; P < 0.0001). Among the study population, 267 (16.4%) were lost to follow-up, with the remainder (1364 patients) found to be continuing active follow-up (Fig. 1).
Characteristics of Patients Lost to Follow-Up
Baseline and clinical characteristics of the 267 patients discontinuing active follow-up are described in Table 1. They were found to be 65.9% female and 92.1% black African and had a mean age at program enrollment of 36.6 years (SD = 9.7 years). These characteristics were not significantly different compared with those continuing active follow-up (P < 0.76, P < 0.93, and P < 0.97, respectively). A similar age differential between men and women discontinuing follow-up was noted (mean: 38.4 vs. 35.7 years; P = 0.028). Men discontinuing follow-up were significantly more likely to be married or cohabitating than women (33.3% vs. 16.0%; P = 0.0012). Although nearly two thirds of patients not returning to the clinic reported being unemployed, only 1 in 5 had received a fee waiver to collect ART medications free of charge.
Approximately 2 in 5 patients had a history of prior or current tuberculosis (TB) infection, and 18% reported prior exposure to ART medications (defined as any ART other than single-dose nevirapine for prevention of mother-to-child transmission [PMTCT]). The median time on ART before discontinuation of follow-up was 84 days (interquartile range [IQR]: 43 to 168 days, range: 13 to 392 days), compared with 205 days (IQR: 108 to 332 days, range: 13 to 448 days) for those in active care. The median initial CD4 count was 76 cells/μL (IQR: 20 to 146 cells/μL), and the median plasma viral load was 5.14 log10 copies/mL (IQR: 4.45 to 5.65 log10 copies/mL). Patients with prior ART exposure had higher median (IQR) baseline blood CD4 counts than those without prior exposure (89 [IQR: 29 to 219] vs. 70 (IQR: 18 to 141) cells/μL; P = 0.046). Most (92.5%) were on either of 2 first-line NRTI/NNRTI regimens at the time of treatment default, with many (89 [38.7%] of 230 patients) reporting at least 1 side effect during treatment. The most common side effects were rash, diarrhea, and vomiting (data not shown). Relatively few (16 [6%] of 267 patients) changed regimens during treatment; when the reason was known, changes were attributable to desired pregnancy, side effects, or presumed resistance as the result of persistently elevated plasma viral loads.
Tracing of Patients Discontinuing Follow-Up
Results of telephone and/or physical tracing of patients lost to follow-up are shown in Table 2. Of the 267 patients discontinuing follow-up, 94 (35.2%) were unable to be traced because of incorrect contact information and 173 (64.8%) were successfully traced. Among these patients, death was the single most common outcome noted, accounting for nearly half (83 [48%] of 173) of those traced. Overall, the known mortality rate for the cohort was 5.1% (83 of 1631 patients). Date of death was known for 59 (71%) of the 83 patients traced dead, of whom 83% (49 of 59 patients) had died within 30 days of active ART. The remaining (10 of 58) patients had discontinued ART for a median of 163 days (IQR: 108 to 244, range: 68 to 336 days) before death. Causes of death, when known, are shown in Table 3 and included AIDS-defining and non-AIDS-defining events.
Physical relocation and clinic transfers accounted for 25.4% of those discontinuing ART (44 of 173 patients), with more than two thirds of these patients reporting continuation of ART at a different clinic, of whom 7 (23.3%) of 30 were obtaining medications from private physicians. Seven patients (4.0%) cited cost as a reason for discontinuing follow-up at the clinic, of whom 5 transferred to other clinics to obtain medications for a lower fee or more conveniently. Hospitalization or illness not resulting in death accounted for another 10.8% of those traced, with a third of these lost to follow-up being attributable to TB infection. Eight patients (4.6%) voluntarily discontinued follow-up because of feeling well. Toxicity or side effects of ART were cited by 5 (2.95%) of 173 patients traced, including 3 cases of lactic acidosis (all on d4T-containing regimens) and 2 cases of immune reconstitution inflammatory syndrome (IRIS). Other reasons reported for discontinuing follow-up are noted in Table 2.
Comparison of Patients Traced Living, Deceased, and Untraceable
For further analysis, the group traced living (n = 90) was compared with the group traced deceased (n = 83) and those untraceable (n = 94) (see Table 1, columns A to C). Compared with those traced living, those who died were found to be significantly older at clinic enrollment (mean: 39.4 vs. 36.0 years; P = 0.022), had lower initial blood CD4 counts (median: 33 vs. 98 cells/μL; P = 0.0073), had higher initial HIV RNA loads (median: 5.28 vs. 4.96 log10 copies/mL; P = 0.024), and had loss of weight since commencing ART (median: −1.0 vs. +0.5 kg; P = 0.033). In addition, significantly fewer patients traced living were found to have current TB than those traced deceased or untraceable, whereas similar rates of prior TB infection were noted. In all respects, except for the fraction with current TB, the untraceable group was not significantly different than the group traced living. Assessment of fees was not associated with increased mortality.
In this study, we characterized the patients lost to follow-up in an urban ART clinic in a resource-limited setting. Over a 15-month period, nearly 1 in 6 patients receiving ART had discontinued follow-up. Commonly collected demographics (gender, age, and ethnicity) did not identify those at high risk for loss to follow-up. Tracing of patients discontinuing follow-up revealed that nearly half of those traced had died, with most deaths occurring within 30 days of active ART. Factors associated with death included older age at clinic enrollment, low baseline CD4 cell count, high initial HIV RNA viral load, and loss of weight with ART. Addressing factors promoting late commencement of ART may improve ART outcomes.
The high proportion of women initiated on ART in this study is mirrored in several other large ART sites in South Africa4,18 and elsewhere in sub-Saharan Africa.6,19 This may reflect differing attitudes and greater use of voluntary counseling and testing (VCT) services by women in these areas20 as well as the rise of PMTCT programs,21,22 resulting in more women knowing their HIV status. Interestingly, this study suggests that once men access ART programs, their rate of loss to follow-up is similar to that of women. Men on ART were more than twice as likely to report being married than women, which may reflect greater disclosure of HIV status by women to their partners.23
Although only 4% of traced patients directly cited cost as a reason for discontinuing follow-up at the clinic, care must be taken before inferring that cost is not a barrier to treatment in this group. The high level of unemployment reported in this study (>60%) has been noted in another study24 in this population, and for the unemployed, paying for ART medications without a fee waiver would be expected to represent a significant financial burden. Additional data, such as spousal income and level of family support, are needed to evaluate the role of cost of treatment fully in loss to follow-up in the resource-constrained setting.
Side effects of ART medications were frequent among those discontinuing ART. Although other studies have noted that side effects of ART can be a barrier to daily medication adherence,11,25,26 we found that these side effects infrequently resulted in actual loss to follow-up. The side effects reported were predictable for the ART regimens prescribed and were generally transient and mild, consistent with an earlier study among this population.27
A high prevalence of previous and current TB infection was noted among those lost to follow-up. It is estimated that more than 60% of patients who have TB in South Africa have HIV coinfection,28,29 and more than 12% of deaths in patients who have TB are attributable to HIV infection.28 In our study, we found that TB resulted in significant morbidity and mortality, accounting for a third of hospitalizations leading to loss to follow-up, and was the most common AIDS-defining illness resulting in death. Patients traced living were less than half as likely to have current TB than those traced dead or untraceable. Although some patients with TB coinfection received medications for TB and HIV through the ART clinic, most received TB treatment at other clinics. It has been suggested that integration of TB and HIV treatment may maximize the effectiveness of interventions aimed at reducing the morbidity and mortality of both diseases.30-32
Although virologic and immunologic responses to ART in resource-limited settings are comparable to those observed in Western settings,7,33,34 it has been noted that mortality is higher in low-income countries.33 In the present study, death accounted for nearly half of patients lost to follow-up who were traced. Those who died had a median baseline CD4 count of 33 cells/μL, indicating significant immunodeficiency at program enrollment. Patients traced deceased lost more weight on treatment, which may reflect a greater burden of TB, IRIS, and other opportunistic infections in this population. Most patients died within a month of active ART rather than after a period of time off of treatment, suggesting that tracing methods are unlikely to have an impact on mortality in this group. Coupled with indicators of advanced immunodeficiency, this suggests that many of the patients who died may have benefited from earlier initiation of ART. The median time on ART of 80 days among those who died is consistent with other studies showing that mortality is greatest in the first few months of ART in resource-constrained settings because of advanced immunodeficiency.33,35
Potential causes for late initiation of ART include poor access to VCT, infrequent measurement of blood CD4 cell counts in patients with HIV, delays in referral to ART clinics, waiting times to enter programs, and stringent criteria for initiation of ART. It is estimated that only 10% to 15% of those in urgent need of ART in developing countries are receiving treatment.36 Earlier initiation of ART may be associated with lower mortality, better immune improvement, and less drug-related toxicity and may also reduce HIV transmission.37 At the time of this writing, eligibility for ART in the South African national program was based on the WHO's 2002 guidelines, which include stage 4 disease or a blood CD4 count <200 cells/μL.15 The updated WHO 2003 recommendation16 relaxes the requirement to include stage 3 disease with a blood CD4 count <350 cells/μL. Less stringent guidelines should clearly result in yet more patients eligible for ART, but whether inefficiencies in health care systems can be overcome to meet the greater demand for ART is unclear.
The results of the study revealed inefficient patient tracking in this clinic. More than a quarter of traced patients lost to follow-up had actually relocated or transferred clinics, which was largely undocumented in patient records. A full third of patients lost to follow-up were untraceable because of incorrect contact information. Clinical parameters suggest that many of the untraceable patients are likely to be alive and may or may not be receiving ART at other clinics. A number of patients discontinued ART for periods of up to 11 months before death, and these patients may have benefited from efforts at program retention. As more patients begin long-term ART, there must be efficient patient tracking to allow for continuity of care for an increasingly mobile population of young adults.
This analysis has several limitations. First, the data were retrieved from a clinic with limited staff and a high workload, limiting full clinical and demographic data collection to the group lost to follow-up. Additionally, the retrospective study design examining the first 15 months of clinic activity resulted in a variable duration of follow-up of patients because patients enrolled at the clinic at different times. Finally, this study was limited to an adult population, and the results may not generalize to pediatric populations.
In summary, this analysis of patients receiving ART in an urban resource-limited setting showed that a significant proportion of patients were lost to follow-up during the period of study. Early death was the most common outcome, suggesting that timely ART initiation rather than aggressive patient tracking would decrease mortality in this population. Additional research is needed to investigate strategies to reduce mortality among those with advanced immunosuppression at the time of ART initiation. Streamlined patient tracking is important for accurate evaluation of ART program outcomes and remains essential for continuity of care for populations in resource-constrained settings.
The authors thank the Johannesburg Hospital HIV clinic's patients and staff.
1. Oyugi JH, Byakika-Tusiime J, Ragland K, et al. Treatment interruptions predict resistance in HIV-positive individuals purchasing fixed-dose combination antiretroviral therapy in Kampala, Uganda. AIDS
2. Adje C, Cheingsong R, Roels TH, et al. High prevalence of genotypic and phenotypic HIV-1 drug-resistant strains among patients receiving antiretroviral therapy in Abidjan, Cote d'Ivoire. J Acquir Immune Defic Syndr
3. Nachega JB, Hislop M, Dowdy DW, et al. Adherence to highly active antiretroviral therapy assessed by pharmacy claims predicts survival in HIV-infected South African adults. J Acquir Immune Defic Syndr
4. Bekker LG, Myer L, Orrell C, et al. Rapid scale-up of a community-based HIV treatment service: programme performance over 3 consecutive years in Guguletu, South Africa. S Afr Med J
5. Wools-Kaloustian K, Kimaiyo S, Diero L, et al. Viability and effectiveness of large-scale HIV treatment initiatives in sub-Saharan Africa: experience from western Kenya. AIDS
6. Karcher H, Omondi A, Odera J, et al. Risk factors for treatment denial and loss to follow-up in an antiretroviral treatment cohort in Kenya. Trop Med Int Health
7. Nacher M, El Guedj M, Vaz T, et al. Risk factors for follow-up interruption of HIV patients in French Guiana. Am J Trop Med Hyg
8. van Oosterhout JJ, Bodasing N, Kumwenda JJ, et al. Evaluation of antiretroviral therapy results in a resource-poor setting in Blantyre, Malawi. Trop Med Int Health
9. World Health Organization. Perspectives and Practice in Antiretroviral Treatment. The Lighthouse Clinic. A Centre for Comprehensive HIV/AIDS Treatment and Care in Malawi
. Geneva, Switzerland: World Health Organization; 2004.
10. Iliyasu Z, Kabir M, Abubakar IS, et al. Compliance to antiretroviral therapy among AIDS patients in Aminu Kano Teaching Hospital, Kano, Nigeria. Niger J Med
11. Weiser S, Wolfe W, Bangsberg D, et al. Barriers to antiretroviral adherence for patients living with HIV infection and AIDS in Botswana. J Acquir Immune Defic Syndr
12. Byakika-Tusiime J, Oyugi JH, Tumwikirize WA, et al. Adherence to HIV antiretroviral therapy in HIV+ Ugandan patients purchasing therapy. Int J STD AIDS
13. Ive P, Conradie F, Xaba S, et al. Causes of loss to follow-up in patients taking antiretroviral therapy in the national rollout program of South Africa. Presented at: Third International AIDS Society Conference on the HIV Pathogenesis and Treatment; 2005; Rio de Janeiro.
14. South African Department of Health. National Antiretroviral Treatment Guidelines
. Pretoria, South Africa: Jacana; 2004.
15. World Health Organization. Scaling up Antiretroviral Therapy in Resource-Limited Settings: Guidelines for a Public Health Approach; Executive Summary
. Geneva, Switzerland: World Health Organization; 2002.
16. World Health Organization. Scaling Up Antiretroviral Therapy in Resource-Limited Settings: Guidelines for a Public Health Approach; 2003 Revision
. Geneva, Switzerland: World Health Organization; 2002.
17. D'Agostino RB, Pearson ES. Tests for departures from normality. Empirical results for the distributions of b2 and p b1. Biometrika
18. Coetzee D, Hildebrand K, Boulle A, et al. Outcomes after two years of providing antiretroviral treatment in Khayelitsha, South Africa. AIDS
19. Sow PS, Otieno LF, Bissagnene E, et al. Implementation of an antiretroviral access program for HIV-1-infected individuals in resource-limited settings: clinical results from 4 African countries. J Acquir Immune Defic Syndr
20. Steen TW, Seipone K, Gomez Fde L, et al. Two and a half years of routine HIV testing in Botswana. J Acquir Immune Defic Syndr
21. Volmink J, Siegfried NL, van der Merwe L, et al. Antiretrovirals for reducing the risk of mother-to-child transmission of HIV infection. Cochrane Database Syst Rev
22. Creek TL, Ntumy R, Seipone K, et al. Successful introduction of routine opt-out HIV testing in antenatal care in Botswana. J Acquir Immune Defic Syndr
23. Miller AN, Rubin DL. Motivations and methods for self-disclosure of HIV seropositivity in Nairobi, Kenya. AIDS Behav
24. Skogmar S, Shakely D, Lans M, et al. Effect of antiretroviral treatment and counseling on disclosure of HIV-serostatus in Johannesburg, South Africa. AIDS Care
25. Golub SA, Indyk D, Wainberg ML. Reframing HIV adherence as part of the experience of illness. Soc Work Health Care
26. Carrieri MP, Leport C, Protopopescu C, et al. Factors associated with nonadherence to highly active antiretroviral therapy: a 5-year follow-up analysis with correction for the bias induced by missing data in the treatment maintenance phase. J Acquir Immune Defic Syndr
27. Hudspeth J, Venter W, Van Rie A, et al. Access to and early outcomes of a public South African adult antiretroviral clinic. S African J Epid Infection
28. Corbett EL, Watt CJ, Walker N, et al. The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med
29. World Health Organization. Global Tuberculosis Control: Surveillance, Planning, Financing
. Geneva, Switzerland: World Health Organization, 2007.
30. Micek M. Integrating TB and HIV Care in Mozambique: Lessons from an HIV Clinic in Beira. CORE TB/HIV Case Study
. Seattle, WA: Health Alliance International; 2005.
31. Zachariah R, Spielmann MP, Chinji C, et al. Voluntary counseling, HIV testing and adjunctive cotrimoxazole reduces mortality in tuberculosis patients in Thyolo, Malawi. AIDS
32. Martinson N, Hale M, Karstaedt A, et al. Causes of death in HIV-infected adults with TB admitted to 2 hospitals in Soweto, South Africa. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; 2007; Los Angeles.
33. Braitstein P, Brinkhof MW, Dabis F, et al. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet
34. Laurent C, Ngom Gueye NF, Ndour CT, et al. Long-term benefits of highly active antiretroviral therapy in Senegalese HIV-1-infected adults. J Acquir Immune Defic Syndr
35. Lawn SD, Myer L, Harling G, et al. Determinants of mortality and nondeath losses from an antiretroviral treatment service in South Africa: implications for program evaluation. Clin Infect Dis
36. WHO/UNAIDS. Progress on Global Access to HIV Antiretroviral Therapy: a Report on “3 by 5” and Beyond
. Geneva, Switzerland: WHO/UNAIDS; 2006.
37. Holmberg SD, Palella FJ Jr, Lichtenstein KA, et al. The case for earlier treatment of HIV infection. Clin Infect Dis
Keywords:© 2008 Lippincott Williams & Wilkins, Inc.
antiretroviral therapy; HIV; loss to follow-up; resource-limited settings; South Africa; tracing