Over the past decade, there has been rapid expansion of antiretroviral therapy (ART) in resource-limited settings. By 2011, an estimated 9.7 million people in low- and middle-income countries were receiving ART.1 During this growth, HIV programs have adapted services to cope with the increasing numbers of patients requiring ART. As programs mature and increase in size, the need to ensure long-term retention in care of patients receiving ART while continuing timely initiation of new patients onto treatment presents an ongoing challenge to policy makers and health-care providers alike.
Over the past decade, the conditions in which ART programs operate in low- and middle-income countries have changed considerably. The numbers of patients have increased rapidly and often disproportionately to the number of health-care workers providing care.2 Guidelines for the initiation of ART have been simplified, and context-specific recommendations have been adapted to facilitate improving and expanding access to treatment. Furthermore, eligibility criteria for ART initiation have evolved, and recommended CD4 count thresholds for starting treatment have recently been increased to improve patient outcomes.3
The effectiveness of ART in reducing morbidity and mortality depends on patient adherence to therapy and on the ability of HIV programs to retain patients in care. Previous analyses examining temporal trends in long-term program outcomes in resource-limited settings have reported conflicting results. Whereas data from South Africa showed increasing loss to follow-up (LTFU) by calendar year of enrollment,4–6 systematic reviews of sub-Saharan African cohorts and findings from a large Kenyan cohort reported improvements in patient retention in recent years.7–9 Evaluating program outcomes, assessing temporal trends in program retention, and investigating the factors associated with poor outcomes are essential to improve the long-term effectiveness of ART services in low- and middle-income countries. Understanding associations between program expansion and treatment outcomes is particularly relevant in the context of the “Treatment 2.0” initiative to scale-up HIV treatment through promoting innovation and efficiency gains and of the ambitious goal of expanding ART to 15 million people by 2015.1,10
The objective of this study was to describe temporal trends in patient characteristics at ART initiation and in ART outcomes using data from resource-limited countries where Médecins Sans Frontières (MSF) supports the provision of HIV treatment. We also examined associations between individual-level risk factors, absolute program size, and rate of ART program expansion and mortality and LTFU.
We analyzed patient electronic health records from 25 sites in 8 countries where MSF supports the provision of ART care. Cohorts were located in Democratic Republic of Congo, India, Kenya, Malawi, Mozambique, Myanmar, Uganda, and Zimbabwe. Details of these programs have been described previously.11 All programs provided care and treatment free of charge. Criteria for ART initiation followed World Health Organization (WHO) guidelines. The analysis included all adults 16 years or older who initiated ART between March 2001 and September 2011 in one of the programs.
Data Collection and Definitions
Characteristics at ART initiation, including sex, age, CD4 cell counts, WHO clinical stage, body mass index (BMI), and date of ART initiation were prospectively collected with the FUCHIA software (Epicentre, Paris, France). Throughout, reference to baseline is a reference to time of ART initiation.
Patient follow-up began at the date of ART initiation and was censored at the earliest of death, transfer out, last clinic visit, or analysis closure. Sites began initiating patients onto ART between 2001 and 2007. Only patients with a minimum of 6 months of follow-up were included with analysis closure preceding the database closure date by 6 months. The database closure ranged from September 30, 2011, to March 20, 2012. Deaths were events recorded before the analysis closure date. LTFU was defined as having no visit in the 6 months before analysis closure. Patients who initiated treatment but did not return were given 1 day of follow-up time so that they would contribute to survival analysis.12 Program retention was defined as being in care (ie, not dead or LTFU) at the time of analysis closure.
To quantify the size of the ART programs, 2 variables, program size and the rate of program expansion, were defined. For each patient, program size was calculated as the total number of both pre-ART and ART patients receiving care in the program at the end of the calendar year of patient ART initiation. To define the rate of program expansion, the rank of each patient enrolled by site was divided by the duration of ART provision in the site up to the date of enrollment, in months. For example, if the 100th patient at the site was enrolled 4 months after the program started, the rate of program expansion for that patient would be 25 (100/4).
Baseline characteristics were described by year of ART initiation using medians and interquartile ranges for continuous variables and proportions for categorical variables. Kaplan–Meier methods were used to describe cumulative probabilities of death, LTFU, and program retention after ART initiation and were analyzed overall, by calendar year of ART start, program size, and rate of expansion.
Cox proportional hazards models were used to assess associations between baseline patient characteristics and outcomes. Heterogeneity across sites was accounted for using random effects. To examine differences in risk factors over time, models were stratified by early (0–12 months) and late (12–72 months) follow-up periods. Adjusted models were built first by including all baseline characteristics (model 1) and then adding the program size variable (model 2), the rate of expansion variable (model 3), or both (model 4). The primary analysis only included patients with complete data on baseline characteristics (complete case analysis).
In sensitivity analyses, models including patients with missing baseline CD4 cell counts, clinical stage, and/or BMI as separate categories were used. We also assessed different measures of program size, including the number of patients who had previously initiated ART by site and the number of ART patients in care at each site at the end of the calendar year. Kaplan–Meier survival proportions and hazard ratios (HRs) stratified by site were assessed and compared with the primary results. Finally, to account for nondifferential censoring (eg, higher risk of death among patients LTFU early after ART start), models of time to death and LTFU were calculated using competing risk methods.13
Data were analyzed with STATA 12.0 (STATA Corporation, College Station, TX). All FUCHIA sites obtained agreement from health ministries for the prospective collection of data. No patient identifiers were included in datasets. The International Ethics Review Board of MSF reviewed the study and determined it did not require formal approval. The Human Subjects Research Ethics Committee of the Faculty of Health Sciences from the University of Cape Town reviewed and approved the data usage and analysis plan.
Characteristics at ART Initiation
Between 2001 and 2011, 132,334 individuals contributed 299,658 person-years of follow-up (median per patient, 1.75 years; interquartile range, 0.57–3.56). At ART initiation, the median age was 35 years, 61% of patients were females, 69% had CD4 cell counts <200 cells per microliter, and less than 5% had CD4 counts >350 cells per microliter (Table 1). Twenty-five percent of patients had clinical stage 4 disease, and one-third had a BMI below 18.5 kg/m2. More than half of the patients received treatment in Malawi (n = 37,657) or Zimbabwe (n = 30,783); Asian cohorts contributed 14% of patients.
Temporal Changes in Patient Baseline Characteristics and Outcomes
The number of patients initiating ART increased substantially each calendar year (Table 2), from 4427 in 2001–2003 to 22,863 in 2010. Median age was 35 years and remained constant over time. Median CD4 cell count increased over time from 97 cells per microliter in 2003 or earlier to 184 cells per microliter in 2011 (see Web Appendix 1, Supplemental Digital Content,http://links.lww.com/QAI/A547). However, every year, approximately 30% of patients initiated ART with a CD4 count <100 cells per microliter (see Web Appendix 2, Supplemental Digital Content,http://links.lww.com/QAI/A547); the range across countries being 22%–48%. The proportion of patients with clinical stage 4 disease at the time of ART initiation decreased from 44% to 14%.
Mortality decreased over the study period (Fig. 1A), from 17% to 5% at 12 months and from 22% to 9% at 36 months (Table 3). Larger programs (Fig. 1C) and those with greater rate of expansion (Fig. 1D) had lower rates of mortality. In contrast, LTFU increased substantially over time, from 6% to 15% at 12 months and from 11% to 21% at 36 months (Fig. 1B). LTFU increased with program size up to a number of 7500 patients (Fig. 1D). Program retention was highest between 2006 and 2009. Trends in outcomes were homogeneous across sites (see Web Appendix 3, Supplemental Digital Content,http://links.lww.com/QAI/A547).
The contribution of LTFU to overall program attrition increased with duration of ART (see Web Appendix 4, Supplemental Digital Content,http://links.lww.com/QAI/A547). During the first year of treatment, approximately half of the program losses were LTFU patients (6% of patients had died compared with 8% who were LTFU). However, after 5 years, two-thirds of the losses were LTFU patients (24% compared with 13% of deaths). Program retention decreased from 82% at 12 months to 73% at 36 months and 66% at 60 months. The smallest programs had the largest estimates of 12-month mortality: 12%, 11%, and 9% in sites with less than 500, 500–999, and 1000–2499 people, respectively. However, 12-month LTFU was largest in medium size programs. Similarly, programs with a slow rate of expansion had double the risk of 12-month mortality compared to those with fastest expansion (10.0% vs. 5.3%). LTFU was lowest in programs with a slow rate of expansion compared those with medium or fast expansion (Fig. 1F).
Risk Factors for Mortality
The risk of death decreased with each successive calendar year of enrollment (Table 3). After adjusting for program size and rate of expansion, this association was only seen for the 0-to 12-month period and up to 2007 [adjusted HR (aHR) = 0.76, 95% confidence interval (CI): 0.61 to 0.95 for 2007 and aHR = 1.00, 95% CI: 0.76 to 1.33, for 0–12 months; aHR = 1.21, 95% CI: 0.73 to 2.00, for 12–72 months, vs. ≤2003].
Larger program size was associated with decreased early mortality (aHR = 0.49, 95% CI: 0.31 to 0.77, for ≥20,000 vs. <500 patients). This association was not significant for late mortality (aHR = 0.34, 95% CI: 0.09 to 1.27).
Fully adjusted models did not show evidence of association between rate of expansion and early or late mortality (aHR = 1.13, 95% CI: 0.87 to 1.48, and aHR = 0.85, 95% CI: 0.55 to 1.31, respectively, for ≥125 vs. <25 patients per month).
Increased early and late mortalities were associated with male gender (aHR = 1.30 and 1.57, respectively, for male vs. female), older age (aHR = 1.46 and 1.48, respectively, for ≥45 vs. 16–25 years), and low BMI (aHR = 2.59 and 1.56, respectively, for ≤18.5 vs. 18.6–25.0 kg/m2). Death was also strongly associated with advanced clinical stage and lower CD4 count level, with the strongest associations observed with early mortality (aHR = 2.65 for stage 4 vs. stage 1 or 2 and aHR = 0.29 and 0.26–0.32 for 200–349 vs. <25 CD4 cells per microliter).
Risk Factors for LTFU
Increased risk of early and late LTFU was observed with each successive calendar year of ART initiation (Table 4). aHRs for early LTFU increased from 1.09 (95% CI: 0.83 to 1.43) in 2004 to 3.29 (95% CI: 2.42 to 4.46) in 2011, compared with the 2001–2003 period, and aHRs for late LTFU from 1.04 (95% CI: 0.84 to 1.28) to 6.86 (95% CI: 4.94 to 9.53), respectively. Larger program size was associated with a 7-fold increase in the risk of early (HR = 7.35, 95% CI: 5.55 to 9.73) and late LTFU (HR = 7.03, 95% CI: 4.30 to 11.48), but the association in final models attenuated for early LTFU (aHR = 1.77, 95% CI: 1.04 to 3.04) and was not observed for late LTFU (aHR = 0.53, 95% CI: 0.27 to 1.04). Rate of program expansion was strongly associated with an increased risk of early and late LTFU (early aHR increased from 1.26 in programs with rates of 25–59 patients per month to 2.31 in those with ≥125 patients per month and late aHR from 1.22 to 2.29 compared with programs with rates of 0–24 patients per month, respectively). Late LTFU was only higher among patients who received treatment in programs with 500–4999 HIV patients in care (aHR = 1.34, 95% CI: 1.00 to 1.80 for 2500–4999 vs. <500 program size).
Male gender (aHR = 1.25 and 1.21, respectively, for early and late periods), younger age (aHR = 0.58 and 0.48, respectively, for ≥45 vs. 16–25 years), and advanced clinical stage (aHR = 1.56 and 1.28, respectively, for stage 4 vs. stage 1 or 2) were also associated with an increased risk of LTFU. Patients with a CD4 cell count of less than 25 cells per microliter had a higher risk of early LTFU (aHR decreased from 0.89 among patients with 25–49 cells per microliter to 0.60 among those with 200–349 cells per microliter). No association between CD4 cell count and late LTFU was observed.
Sensitivity analyses including patients with missing data (see Web Appendices 5 and 6, Supplemental Digital Content,http://links.lww.com/QAI/A547) and stratification by site provided similar results. Estimates from competing risk models were slightly reduced but did not change results (see Web Appendices 7 and 8, Supplemental Digital Content,http://links.lww.com/QAI/A547).
This multicenter cohort study included over 130,000 HIV-infected patients initiated on ART in 8 low- and middle-income countries where the provision of ART has rapidly expanded over the last 10 years. In this challenging context where programs needed to adapt to the growing numbers of patients in need for care, we observed a gradual improvement in measures of disease severity at ART initiation and a decrease in mortality over time. Despite these findings, every year, 30% of the total number of patients who initiated ART had CD4 cell counts less than 100 cells per microliter. Furthermore, 12-, 24-, and 36-month LTFU rates among patients initiating ART in successive calendar years doubled between the periods 2001–2003 and 2008.
Over the 10-year study period, ART provision was rapidly implemented and programs expanded to reach a growing number of people infected with HIV. Expansion rates ranged from 0–25 to 125–192 new patients per month and program size from <500 to 20,000–23,995 patients. Mortality gradually improved with 6-month estimates decreasing from 14% in the years before 2004 to less than 4% in 2011, which is consistent with reports from South Africa.4 The observed decreased estimates of mortality may result from improved access to ART, as suggested by the increased number of patients initiating ART with less severe disease. The rapid growth of programs might have also led to poorer outcome ascertainment, with greater number of deaths occurring in recent years misclassified as LTFU.14 Linkage to national death registries is not available in these cohorts, highlighting the need to improve outcome ascertainment for program evaluation in resource-limited countries.
The rapid growth of ART programs is related to a combination of several factors, including the long-term availability of antiretroviral drugs, the implementation of new simplified guidelines for ART, and the widespread availability of CD4 enumeration to identify ART-eligible individuals. One-third of patients initiated ART with a CD4 cell count below 100, placing a substantial strain on the health-care system requiring intensive support from clinically trained staff.15,16 Our findings do not suggest that increases in treatment guideline thresholds prevent patients with more advanced HIV disease from accessing care but rather that all programs still face challenges in timeously seeking, testing, and linking patients to ART care.
Even with continued challenges to improving access to ART, LTFU seems to present a ubiquitous challenge to the long-term effectiveness of ART programs. After 2 years of ART, 15% of the cohort was LTFU and retention was 77%, and after 5 years of treatment, two-thirds of patients remained in care. This is slightly less than the estimate from 23 countries with cohorts of more than 2000 people, which reported 72% retention after 5 years.1 A temporal trend of increasing LTFU was observed with LTFU contributing an increasing proportion of overall program attrition.4,17 These findings confirm the urgent need to refocus efforts to improve long-term retention and contradict systematic reviews from sub-Saharan Africa, suggesting that program outcomes are improving.7,17 A novel finding of our study is that the rate of program expansion, more than the size of the HIV program, was associated with the high levels of LTFU. This is likely to relate to the need for timely adjustments in programs to cope with the increase in activity, independently of financial and human resource allocation. Associations observed between LTFU and male gender, younger age, and advanced clinical stage are consistent with previous reports.4,18–20
This analysis of long-term outcomes was based on substantial follow-up of a large number of patients treated in several resource-limited ART programs. All sites offered care and treatment free of charge and followed WHO recommendations for ART initiation, and site heterogeneity was accounted for using random effects in the Cox models. These data are not likely to be representative of all ART programs because MSF provides additional resources and technical support. For example, quality and completeness of routine electronic data are challenging, but MSF provides considerable means to address these concerns.2 Without linkage to death registries, estimates of program attrition may misclassify some proportion of deaths as LTFU.5,14,18,21,22 LTFU may be overestimated further because of administrative errors, incomplete records of patient decentralization and unrecorded transfers, and the contribution of treatment interrupters.2,20,23–25 Furthermore, we were unable to minimize this bias as we did not have data on tracing for all or a sample of those who were LTFU.26 Our focus was limited to program outcomes after ART initiation acknowledging that a substantial proportion of patients were lost during pre-ART care.27–30 Although program size and the rate of expansion were adjusted for, residual confounding may be present if important internal organizational aspects were not sufficiently captured and from unmeasured factors. Sensitivity analyses confirmed our results investigating differences by program, including patients with missing covariate data, and adjusting for site heterogeneity. A competing risk approach led to very similar findings corroborating that the traditional Cox proportional hazards models are appropriate.31,32
Our findings explore the tension and challenges involved in pursuing the ambitious goals of expanding ART to 15 million people by 2015 and implementing Treatment 2.0 strategies in high-prevalence resource-limited settings.1,10 Recent Treatment as Prevention models assumed a long-term drop-out rate of 1.5% annually, which seems optimistic considering our estimates of long-term retention.33 With a high burden of acutely ill patients, ART programs will struggle to expand access to patients in earlier stages of HIV disease without additional resources and a change to the models of ART delivery. With new 2013 WHO guidelines to expand ART access to an additional 9.2 million people in low- and middle-income countries, we need to fully understand the individual and programmatic implications of earlier initiation.1
Over a decade, ART programs have expanded in high-prevalence resource-limited settings with a focus to increase access to care and thus patient numbers. Today, many sites hold high numbers of HIV-infected patients, above 20,000 at some sites. The quality of ART services and psychosocial counseling at these large sites may be taking a back seat as the drive for expansion continues. The conscious trade-off between numbers and quality deserves more discussion and close monitoring as the targets for expansion of treatment continue to increase. The findings of our study suggest that, potentially, ART sites should be capped at a maximum number of patients and the rate of enrollment restricted in favor of balancing growth with quality care. Human resource capacity of the site and program organization characteristics are likely to be important determinants to consider in achieving this balance and deserve further investigation to provide effective recommendations regarding maximum patient volume and expansion rate at program level.
For the first time, the WHO guidelines acknowledge the challenge of long-term retention in ART programs with explicit guidance on operations and service delivery, including adherence, retention, decentralization, and task shifting.3 Additional resources are needed to strengthen monitoring systems to ascertain true outcomes of children and adults lost to care pre- and post-ART initiation.14,34 Identification of ART patients disengaged from care is critical as they are at an increased risk of developing and transmitting drug-resistant strains of HIV.17 There is an urgent need to determine sustainable and optimal models of care for stable patients on lifelong ART, especially in large programs in high-prevalence resource-limited settings. Decreasing visit frequency by expanding intervals between prescription refills, decentralizing ART delivery into community-based patient-led groups, and introducing flexible systems to support mobile populations are all interventions that could be considered and assessed on a large scale.35–38
In summary, ART programs in resource-limited settings have grown rapidly over the last decade. However, significant work remains to continue expanding access while addressing the growing challenge of program attrition. Sustainable models of care for long-term retention of patients in large, high-prevalence resource-limited settings are urgently needed.
S.B., E.C., and J.L. established/maintained the cohorts and provided data. A.G. was responsible for writing the article and undertook the statistical analyses. A.G., G.v.C., L.M., and M.P.R. contributed to prepare the plan of analysis and to write the report and interpreted the data. All authors comment on the draft manuscripts and approved the final version.
1. World Health Organization. Global Update on HIV Treatment 2013: Results, Impact and Opportunities. Geneva, Switzerland: World Health Organization; 2013.
2. McGuire M, Pinoges L, Kanapathipillai R, et al.. Treatment initiation, program attrition and patient treatment outcomes associated with scale-up and decentralization of HIV care in rural Malawi. PLoS One. 2012;7:e38044.
3. World Health Organization. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for a Public Health Approach. Geneva, Switzerland: World Health Organization; 2013.
4. Cornell M, Grimsrud A, Fairall L, et al.. Temporal changes in programme outcomes among adult patients initiating antiretroviral therapy across South Africa, 2002-2007. AIDS. 2010;24:2263–2270.
5. Kranzer K, Govindasamy D, Ford N, et al.. Quantifying and addressing losses along the continuum of care for people living with HIV infection in sub-Saharan Africa: a systematic review. J Int AIDS Soc. 2012;15:17383.
6. Nglazi MD, Lawn SD, Kaplan R, et al.. Changes in programmatic outcomes during 7 years of scale-up at a community-based antiretroviral treatment service in South Africa. J Acquired Immune Deficiency Syndromes. 2011;56:e1–e8.
7. Fox MP, Rosen S. Patient retention
in antiretroviral therapy programs up to three years on treatment in sub-Saharan Africa, 2007-2009: systematic review. Trop Med Int Health. 2010;15:1–15.
8. Rosen S, Fox MP. Retention
in HIV care between testing and treatment in sub-Saharan Africa: a systematic review. PLoS Med. 2011;8:e1001056.
9. Ochieng-Ooko V, Ochieng D, Sidle JE, et al.. Influence of gender on loss to follow-up
in a large HIV treatment programme in western Kenya. Bull World Health Organ. 2010;88:681–688.
10. Duncombe C, Ball A, Passarelli C, et al.. Treatment 2.0: catalyzing the next phase of treatment, care and support. Curr Opin HIV AIDS. 2013;8:4–11.
11. Pujades-Rodriguez M, O'Brien D, Humblet P, et al.. Second-line antiretroviral therapy in resource-limited settings: the experience of Medecins Sans Frontieres. AIDS. 2008;22:1305–1312.
12. Grimsrud AT, Cornell M, Egger M, et al.. Impact of definitions of loss to follow-up
(LTFU) in antiretroviral therapy program evaluation: variation in the definition can have an appreciable impact on estimated proportions of LTFU. J Clin Epidemiol. 2013;66:1006–1013.
13. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509.
14. Boulle A, Van Cutsem G, Hilderbrand K, et al.. Seven-year experience of a primary care antiretroviral treatment programme in Khayelitsha, South Africa. AIDS. 2010;24:563–572.
15. Geng EH, Hunt PW, Diero LO, et al.. Trends in the clinical characteristics of HIV-infected patients initiating antiretroviral therapy in Kenya, Uganda and Tanzania between 2002 and 2009. J Int AIDS Soc. 2011;14:46.
16. Kigozi IM, Dobkin LM, Martin JN, et al.. Late-disease stage at presentation to an HIV clinic in the era of free antiretroviral therapy in Sub-Saharan Africa. J Acquir Immune Defic Syndr. 2009;52:280–289.
17. Rosen S, Fox MP, Gill CJ. Patient retention
in antiretroviral therapy programs in sub-Saharan Africa: a systematic review. PLoS Med. 2007;4:e298.
18. Van Cutsem G, Ford N, Hildebrand K, et al.. Correcting for mortality
among patients lost to follow up on antiretroviral therapy in South Africa: a cohort analysis. PLoS One. 2011;6:e14684.
19. Boyles TH, Wilkinson LS, Leisegang R, et al.. Factors influencing retention
in care after starting antiretroviral therapy in a rural South African programme. PLoS One. 2011;6:e19201.
20. Kranzer K, Lewis JJ, Ford N, et al.. Treatment interruption in a primary care antiretroviral therapy program in South Africa: cohort analysis of trends and risk factors. J Acquir Immune Defic Syndr. 2010;00:1–7.
21. Fox MP, Brennan A, Maskew M, et al.. Using vital registration data to update mortality
among patients lost to follow-up from ART programmes: evidence from the Themba Lethu Clinic, South Africa. Trop Med Int Health. 2010;15:405–413.
22. Brinkhof MW, Pujades-Rodriguez M, Egger M. Mortality
of patients lost to follow-up in antiretroviral treatment programmes in resource-limited settings: systematic review and meta-analysis. PLoS One. 2009;4:e5790.
23. Kranzer K, Ford N. Unstructured treatment interruption of antiretroviral therapy in clinical practice: a systematic review. Trop Med Int Health. 2011;16:1297–1313.
24. Geng EH, Nash D, Kambugu A, et al.. Retention
in care among HIV-infected patients in resource-limited settings: emerging insights and new directions. Curr HIV/AIDS Rep. 2010;7:234–244.
25. McGuire M, Munyenyembe T, Szumilin E, et al.. Vital status of pre-ART and ART patients defaulting from care in rural Malawi. Trop Med Int Health. 2010;15(suppl 1):55–62.
26. Henriques J, Pujades-Rodriguez M, McGuire M, et al.. Comparison of methods to correct survival estimates and survival regression analysis on a large HIV African cohort. PLoS One. 2012;7:e31706.
27. Ingle SM, May M, Uebel K, et al.. Outcomes in patients waiting for antiretroviral treatment in the Free State Province, South Africa: prospective linkage study. AIDS. 2010;24:2717–2725.
28. Harries AD, Schouten EJ, Libamba E. Scaling up antiretroviral treatment in resource-poor settings. Lancet. 2006;367:1870–1872.
29. Lahuerta M, Ue F, Hoffman S, et al.. The problem of late ART initiation in Sub-Saharan Africa: a transient aspect of scale-up or a long-term phenomenon? J Health Care Poor Underserved. 2013;24:359–383.
30. Bastard M, Nicolay N, Szumilin E, et al.. Adults receiving HIV care before the start of antiretroviral therapy in sub-Saharan Africa: patient outcomes and associated risk factors. J Acquir Immune Defic Syndr. 2013;64:455–463.
31. Schöni-Affolter F, Keiser O, Mwango A, et al.. Estimating loss to follow-up
in HIV-infected patients on antiretroviral therapy: the effect of the competing risk of death in Zambia and Switzerland. PLoS One. 2011;6:e27919.
32. Clouse K, Pettifor A, Maskew M, et al.. Initiating antiretroviral therapy when presenting with higher CD4 cell counts results in reduced loss to follow-up
in a resource-limited setting. AIDS. 2013;27:645–650.
33. Granich R, Gilks C, Dye C, et al.. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet. 2009;373:48–57.
34. Fenner L, Brinkhof MW, Keiser O, et al.. Early mortality
and loss to follow-up
in HIV-infected children starting antiretroviral therapy in Southern Africa. J Acquir Immune Defic Syndr. 2010;54:524–532.
35. Decroo T, Telfer B, Biot M, et al.. Distribution of antiretroviral treatment through self-forming groups of patients in Tete Province, Mozambique. J Acquir Immune Defic Syndr. 2011;56:e39–e44.
36. Luque-Fernandez MA, Van Cutsem G, Goemaere E, et al.. Effectiveness of patient adherence groups as a model of care for stable patients on antiretroviral therapy in Khayelitsha, Cape Town, South Africa. PLoS One. 2013;8:e56088.
37. Ford N, Mills EJ. Simplified ART delivery models are needed for the next phase of scale up. PLoS Med. 2011;8:e1001060.
38. Wilkinson LS. ART adherence clubs: a long-term retention
strategy for clinically stable patients receiving antiretroviral therapy. South Afr J HIV Med. 2013;14:48–50.
antiretroviral therapy (highly active); loss to follow-up; mortality; program outcomes; retention; survival analysis
Supplemental Digital Content
© 2014 by Lippincott Williams & Wilkins