Substantial progress has been made over the past decade in the scale-up of pediatric antiretroviral therapy (ART) in resource-limited settings and about 630,000 children in low- and middle-income countries are estimated to be receiving ART as of the end of 2012.1 However, based on data from 22 priority countries, only one-third of those in need are receiving treatment compared with 65% ART coverage for adults, and the increase in the proportion of children receiving ART in the last year has been half that of adults (11% vs 21%).1 Improving access to pediatric ART in resource-limited settings faces continuous challenges that range from difficulties in diagnosis, complex ART formulations, perceived complexity of pediatric treatment, and limited capacity of health care providers to initiate ART in children and provide follow-up for those already on ART.2
For adults, a critical enabling strategy that has supported scale-up of ART over the last decade has been the delegation of responsibility for ART initiation and maintenance from highly qualified health workers to health workers with fewer qualifications and shorter trainings. Such task shifting has been promoted by the World Health Organization (WHO) for adults since 20083 and has promoted ART scale-up even in some of the most poorly resourced regions of sub-Saharan Africa, and this approach is strongly recommended by the most recent WHO consolidated guidelines for the use of Antiretrovirals (ARVs).4 Evidence from randomized trials and program cohorts assessing outcomes for adults have found no differences in safety and efficacy of whether ART is delivered by doctors or nurses.5–9
Although for adults task shifting is now a broadly practiced approach to improve access to ART, the application of this approach to support ART scale-up in children has been less well accepted, likely due to a combination of less clear policies and evidence to support this approach, and less prioritization given to pediatric ART scale-up. Some high HIV burden countries, including South Africa and Zimbabwe, have adopted task shifting as policy for both adults and children, but implementation is limited by lack of confidence and skills among nonspecialist staff. A recent review of pediatric ART care in Zimbabwe, where coverage for children is half that of adults (42% vs 86%), recommended task shifting of pediatric ART services as a priority short-term action to improve treatment access.10
We undertook this review to systematically assess the evidence on the use of task shifting for treatment and care of HIV-infected children.
We developed a sensitive search strategy that combined key terms for task shifting, antiretroviral therapy, and children (defined as aged ≤14 years or as defined by the studies). Using these terms, we searched independently and in duplicate (M.P., N.F.) for published studies in the Cochrane HIV/AIDS Review Group Trials Register, The Cochrane Library, MEDLINE via Pubmed, EMBASE, and CENTRAL from inception till August 01, 2013, according to a review protocol (see Appendix, Supplemental Digital Content,http://links.lww.com/QAI/A477). We further screened abstracts of all conferences of the International AIDS Society (up to Kuala Lumpur, June 2013) and all Conference on Retroviruses and Opportunistic Infections (up to Atlanta, March 2013). We also searched for unpublished and ongoing studies by considering prospective clinical trial registries [Current Controlled Trials (www.controlled-trials.com)] and by contacting research organizations and experts in the field. Bibliographies of all included articles were screened for additional studies. No language, date, or geographical restrictions were applied.
Randomized and quasi-randomized trials, comparative and noncomparative prospective or retrospective cohorts, qualitative studies and cost-effectiveness studies were eligible for inclusion. Case reports and case series <10 patients were excluded.
We sought studies that reported outcomes among children in whom ART was initiated or maintained by a nonphysician (eg, nurse, midwife, or medical assistant). When possible, these interventions were compared with initiation or maintenance of ART by a physician or specialist (eg, pediatrician or obstetrician).
Information on study characteristics, methodological quality, and study outcomes was extracted by 1 reviewer (M.P.) and independently verified by a second reviewer (N.F.). Any discrepancies were resolved by consensus with a third reviewer (M.D.). If insufficient data were available in the study publication, further information was sought from the publication authors. There is no single agreed tool for assessing the quality of observational studies.11 For this review, we assessed study quality independently and in duplicated (M.P., N.F.) using commonly used indicators of methodological quality of observational studies as follows: study design, availability of patient level data, assessment of objective outcomes, comparison with standard of care, conduction of an adjusted analysis, and publication in peer-review journal. Primary outcomes of interest that contributed to the quantitative synthesis were mortality and loss to follow-up; secondary outcomes were virological suppression/failure, CD4 gain, new AIDS-defining illness, development of resistance mutations, time to initiation, cost, and patient/provider satisfaction.
For the primary outcome of mortality, the proportion of patients who had died was estimated for each time point together with corresponding 95% confidence intervals (CIs). These data were then pooled using the DerSimonian-Laird random-effects method.12 Proportions were transformed before pooling using the Freeman–Tukey double arcsine transformation for variance stabilization13 and then back-transformed to the original scale.14 We calculated the τ2 statistic using DerSimonian and Laird method of moments estimator12 to assess between-study heterogeneity.15 Sensitivity analyses were conducted to assess the potential influence of studies that only reported site-level outcomes and studies in which maintenance of ART was task shifted to nurses, but not initiation.
Anticipating a limited number of comparative studies, we undertook 2 preplanned analyses comparing outcomes reported by the task-shifting programs included in this review with outcomes reported by traditional models in which children were managed by doctors or specialists.
In the first comparison, we analyzed outcome data from the International epidemiologic Databases to Evaluate AIDS-southern Africa (IeDEA-SA) collaboration, a large multi-site pediatric cohort from sub-Saharan Africa (www.iedea-sa.org). We used IeDEA-SA data only from sites in which children are managed by physicians and specialists, but to improve comparability, we excluded tertiary care sites where very sick children were managed as none of the task-shifting studies included children managed in tertiary centers. We compared the mortality and loss to follow-up at different time points against those reported by studies in which task shifting occurred.16 In the second comparison, we compared the pooled proportion of mortality and the pooled proportion of loss to follow-up at 1 year with mortality and retention outcomes reported by a recent systematic review of pediatric ART outcomes17; after removing 1 duplicate study (ie, a task-shifting study that was also identified by our review18), we calculated a pooled estimates for mortality and for loss to follow-up data from these studies using random-effects models as described.
All analyses were conducted using Stata version 12 (StataCorp LP, College Station, TX).
Of 1773 records retrieved (668 identified through database searching and 1105 through conference proceedings), 14 full-text articles were assessed for eligibility, and 8 studies, including 7 published studies,18–24 and 1 unpublished cohort (MSF, 2013) were included in the review (Fig. 1). Together, these studies reported outcomes on 11,828 children across 10 countries (Democratic Republic of Congo, Kenya, Lesotho, Malawi, Mozambique, Rwanda, South Africa, Tanzania, Zambia, and Zimbabwe). All studies contributed to the meta-analysis summarizing mortality and loss to follow-up, but only 2 studies21,23 provided sufficient information to allow for within-study comparisons between task shifting and traditional models of care.
All studies were carried out in southern or eastern Africa from 2003 onwards. Five studies,18–21,24 contributing outcome data for 5899 patients, investigated delegation to nonphysician staff of both ART initiation and maintenance in governmental health facilities; a number of these programs reported receiving technical support from nongovernmental partners including Médecins Sans Frontières (MSF),19,20 the International Center for AIDS Care and Treatment Programs,21 and the Lighthouse Trust.23 The number of children included in the studies ranged from 31 to 8153. Study characteristics are summarized in Table 1.
Models of Care
In 6 studies, task shifting was implemented along side a decentralized model of care delivery, with initiation and provision of ART at primary health level. In these settings, the full range of HIV services was integrated into the package of primary health care offered at the health facility.18,20,22 Nurses followed up children at monthly visits ordering laboratory tests, interpreting results, initiating ART and following up non-complex cases. For settings with full integration of tuberculosis (TB) and HIV activities nurses also administered TB medications.24 In 2 studies, a home-based care program using nurses and community volunteers was set up to provide additional care at home for children requiring extra educational, nutritional, or adherence support.18,20 Three studies described the role of physicians in the model of care18,19,22; in these studies, physicians were involved in the management of complex clinical cases, training/mentoring, and supervision. One study described further task shifting in which tasks traditionally performed by nurses were in some cases transferred to receptionists (for administrative work and data collection/monitoring), laboratory staff for blood collection, and counselors and community support groups for counseling.19
The training provided to nurses was described by 4 studies. In one study, bedside training was provided to nurses who received training on pediatric care including ART as part of their national ART provider certification.23 A second study reported that nurses received in-service theoretical and practical training on HIV-related conditions management, including clinical training adapted from IMAI20; this study also reported that targeted training was provided to address specific gaps and supervision visits by a trained physician, initially on a weekly or biweekly basis but with decreasing frequency as the program matured. Two studies reported the use of pediatric-specific clinical care protocols and standardized guidance to recognize side-effects of ARVs and indications for referral to physicians.18,19
Risk of Bias
Among the included studies, all but 1 (MSF) were published in peer-reviewed journals, 4 were retrospective analyses, and 3 described prospective collection of data (see Table S1, Supplemental Digital Content,http://links.lww.com/QAI/A477). Patient-level data were reported by all but 2 studies21,23; these 2 studies reported site-level outcomes, comparing sites where nurses and doctors managed ART. Adjusted analysis was not performed in any of the studies included and the influence of known and unknown bias cannot be excluded.
Mortality was reported by 6 published studies18–22,24 and 1 unpublished cohort at varying time points; and for the pooled analysis, different studies contributed data to different time points. These results are summarized in Figure 2. Overall, mortality was <10% for all studies across all time points with the exception of 1 study that reported 10.4% at 24 months.20 The cumulative pooled proportion of deaths was estimated to be 3.2% (95% CI: 2.0 to 4.5) at 6 months 4.6% (95% CI: 2.1 to 7.1) at 12 months, 6.2% (95% CI: 3.7 to 8.8) at 24 months, and 5.9% (95% CI: 3.5 to 8.3) at 36 months. Heterogeneity was high (τ2 0.01). In sensitivity analysis, excluding studies that reported only site level outcomes, or studies in which nurses maintained but did not initiate ART, did not influence these results (data not shown).
Two studies described higher mortality in younger children.18,22 In the first study, from Zambia, 20.2% of children younger than 18 months receiving ART died over 732 child-years of follow-up (mortality rate: 21.1 deaths per 100 child-years, 95% CI: 15.4 to 28.2); mortality was particularly high within 90 days of starting ART.18 Mortality decreased to 9.5% (mortality rate: 7.6 deaths per 100 child-years, 95% CI: 5.9 to 9.8) and to 5.9% (mortality rate: 4.5 deaths per 100 child-years, 95% CI: 3.6 to 5.6) when considering children on ART aged 18–59 or >60 months, respectively.18 A similar trend was observed in the second study, from KwaZulu Natal, where the risk of dying was significantly higher in children aged <18 months (mortality rate: 4.4 per 100 child-years, 95% CI: 3.3 to 5.8) compared with children aged 18–59 months (mortality rate: 2.6 per 100 child-years, 95% CI: 1.8 to 3.8), and children aged ≥60 months (mortality rate 2.0 per 100 child-years, 95% CI: 1.3 to 3.0).22 One study compared outcomes from sites where nurses and doctors managed ART care. Mortality at 2 years was 8.7% (95% CI: 7.7 to 9.7) in nurse-led sites and 11.0% (95% CI: 10.1 to 11.9) in doctor-led sites, corresponding to a relative risk of 0.78 (95% CI: 0.69 to 0.91).
Loss to follow-up was reported by 6 published studies,18–22,24 and 1 unpublished cohort with different studies contributing data to different time-points and definition of loss to follow-up was not standardized across studies. Overall, the pooled proportion for loss to follow-up was 3.15% (95% CI: 0 to 7.2) at 6 months, 4.7% (95% CI: 0.9 to 8.4) at 12 months, 10.6% (95% CI: 1.5 to 19.8) at 24 months, 11.9% (95% CI: 4.6 to 19.1) at 36 months, and 3.9% (95% CI: 2.1 to 5.6) at 48 months (Fig. 3).
Comparisons With Results From Non–Task-Shifted Models of Care
No important difference in mortality and loss to follow-up comparing task-shifted programs with the IeDEA-SA sites where children were managed by doctors and specialists (Fig. 3). Similarly, the pooled proportion of mortality and loss to follow-up at 1 year were comparable to those reported by a systematic review of pediatric outcomes at 1 year in which children were managed by doctors or specialists as follows: mortality was 7.2% vs 6.2% for task-shifted care (P = 0.6) and loss to follow-up was 3.6% vs 4.7% for task-shifted care (P = 0.6). These results are summarized in Figure 3.
Secondary outcomes for studies in which care was provided through a task shifting approach were variously and inconsistently reported. Two studies reported virological suppression (both defined as <400 copies/mL). In the first study, from Rwanda, the proportion of children achieving virological suppression was 73% at 12 months and 82.8% at 24 months.19 In the second study, from South Africa, 73.5% of children achieved virological suppression within 12 months after treatment initiation.22 Two published studies18,19 and 1 unpublished cohort reported immunological recovery at 12 months, either as CD4% or CD4 absolute count. Similarly, overall increase in CD4 count was reported for TB-coinfected children, despite being severely immunosuppressed at treatment initiation.24 Improvement was also observed in hemoglobin levels and growth parameters suggesting a general improvement of clinical conditions. Finally, 1 study assessed the extent of agreement in the ART formulation prescribed by nurses and clinical officers, either using split adult tablets or pediatric fixed-dose combinations. Agreement was observed to be very good (κ = 0.93), and calculations of adherence and next appointments were reported to be similar to clinical officers although clinical officers used an electronic system.23
Task shifting has been suggested as a way to provide quality, cost-effective care to more patients compared with physician-dependent ART delivery,5 and many countries have adopted this model for adult care: a survey completed in 2012 found that nurse initiation of ART had been adopted by 10 of 16 surveyed countries in sub-Saharan Africa.25 However, there has been reluctance to apply this approach in providing care to HIV-infected children.
The findings of this review suggests that delegation of initiation and maintenance of pediatric ART by nonphysician staff can result in outcomes comparable with traditional doctor or specialist managed care, and we found no evidence to suggest that task shifting resulted in an increased risk of mortality or is associated with poorer clinical and program outcomes. Although these findings are based on low quality study designs that prevent definitive conclusions from being drawn, results are consistent across the majority studies and outcomes, including settings where integrated management of TB and HIV is provided.24 These findings are also consistent with reports from randomized trials6,7 and cohort studies8,9 assessing ART outcomes in adults which found no difference in mortality outcomes between nurse-led and physician-led care.
Earlier guidance on task shifting was unclear about its applicability to children, and nurse training to support task shifting has tended to give greater focus to adult ART management.26 The broader spectrum of clinical presentation of opportunistic infections and other HIV-related diseases might make differential diagnosis challenging for nonphysicians.27 Treating HIV-infected children requires consideration of a number of factors such as an initial establishing of HIV infection status, age, weight, breastfeeding status, exposure to drugs for the prevention of mother-to-child transmission of HIV, CD4 count and/or percentage, and need for TB cotreatment. Of note, in the study reporting data on a cohort of TB-coinfected children, only half of the patients initiated ART within 4 weeks, reflecting the complexity of TB cotreatment management in this age group.24 These challenges can be overcome by simplifying testing and treatment algorithms and providing targeted training and frequent supervision. This implies that the establishment of task-shifting approaches may require initial investments,28 but such costs can be offset by the long-term savings that can be gained through decentralization of ART services.29
Limitations of This Review
This review was unable to identify randomized comparative evidence supporting task shifting for children, reflecting a general insufficiency of the evidence base in contrast with adults. Most of the evidence in this review was noncomparative and observational, and therefore subject to biases inherent in such study designs. Two studies only provided aggregated program data that do not allow consideration of patient-level outcomes: however, these studies were included as they both provide within-study comparisons of ART provided by nurse or doctors; only 1 of these studies contributed to the meta-analysis, and dropping it from the analysis did not change the overall result.
Studies included outcomes for children of all ages, and generally included older children, so these cannot be taken as indicative of outcomes for infants and young children. This is particularly important in the context of more complex tasks being required for the management of these children such as undertaking timely infant diagnosis, use of syrups, and frequent dose adjustment. One study did not include young children because the required formulations were not available at the primary care level, highlighting the need for better pediatric products that suit decentralization to remote heath facilities.24
Different studies contributed data to the mortality and loss to follow-up estimates reported at different time points. As such these findings are not cumulative and should rather be understood as indicative of possible outcomes that can be achieved by different programs. The substantial heterogeneity between studies was compensated by applying a random-effects analysis and presenting these outcomes together with uncertainty intervals. We restricted our meta-analysis to the primary outcomes of mortality and loss to follow-up due to limited reporting and varied definition and measurement of other outcomes. In particular, assessment of CD4 and growth parameters could not be compared across studies due to the limited comparability between mean change, age-adjusted Z score, and absolute values.
This review includes data from a range of countries and settings, which lends support to the generalizability of findings, at least within southern Africa, but also means that outcomes are combined from a range of different service models, resulting in substantial heterogeneity. Notably, outcomes are likely to be influenced by changes in differing patient characteristics and evolutions in guidelines and program approaches at different stages of ARV rollout, differences in outcome ascertainment between studies, notably definitions of loss to follow-up, and ascertainment of vital status among patients lost to care. These differences could not be formally explored due to inconsistent reporting. We note, however, that a number of programs adopting task shifting had financial and technical support from nongovernmental organizations or external partners who provided targeted support for training, mentoring, and supervision. As these programs mature and support is gradually withdrawn in a number of settings, it will be important to assess whether same outcomes can be achieved and sustained.
We used a broad search strategy that attempted to identify both published and unpublished evidence. Anticipating a paucity of high quality studies, we allowed for the inclusion of broader range of lesser quality study designs and as a result we are able to report on outcomes for almost 12,000 children across 10 countries. Nevertheless, the fact that we were only able to identify 8 studies stands in stark contrast to the large body of evidence supporting task shifting for ART care for adults5 and serves as a reminder for the need for research efforts to include children.
To overcome the lack of direct comparisons within studies, we developed indirect comparison with existing published data and updated data from a large cohort in South Africa. It should however be noted that reference studies were not systematically identified but extracted from a published meta-analysis17 that had the goal to assess immunological and virological outcomes. These studies were conducted in facilities providing routine laboratory monitoring which are more likely to be well functioning and may therefore bias the results toward better clinical and program outcomes; such bias would not however compromise the comparisons made by this review as they would be expected to favor the non–task-shifting studies. Nevertheless, we did not adjust comparisons for age or different disease severity characteristics of children at different facilities. Infants and young children and those with more severe disease are more likely to be managed at better-resourced facilities where task shifting has not been used, and this should be taken into account when interpreting these comparisons.
Implications for Policy
Our review provides a summary of evidence that can support country considerations to apply task shifting to adults and children alike, as recommended by the most recent revision of WHO guidelines.4 The delegation of pediatric ART provision to nurses and other nonphysician providers should be seen as a service delivery model that facilitates decentralization and integrated family-centered care,30,31 with the goal of improving access to ART for a neglected population.32 The most recent revision of WHO guidelines, which recommends starting ART in any child younger than 5 years irrespective of clinical or immunological status and the use of age-specific ART regimens,4 is expected to facilitate this shift and encourage nurses to uptake new responsibilities in the management of HIV-infected children. A recent review of Zimbabwe national ART program found that nurses were able and willing to take on extended responsibilities in managing ART services with appropriate training, adequate staff complement, supportive supervision, and mentoring. Training, mentoring, and supervision packages will require adaptation as clinical management is improved and new components are added to the standard package of care for HIV-infected children. Since the impact of task shifting, integrated family-centered and decentralized care is optimized through health workers motivation, strategies to provide regular updates and feedback about challenges and successes should also be developed.
Implications for Research
This review serves to highlight the paucity of published evidence on task shifting in pediatric ART. Although high quality evidence is always desirable, the high mortality among HIV-infected untreated children and the persistent low ART coverage and delayed ART initiation among this group calls for innovative approaches. Programs should be encouraged to report outcome data as task shifting is implemented to improve the evidence base on the impact of task shifting on program performance and pediatric treatment scale-up. Qualitative studies to assess caregivers and children perceptions and satisfaction with regard to this approach could be undertaken to inform long-term success of this approach in a population group for whom psychological aspects of care are particularly important. Further research will also be important in defining how task shifting should be best implemented and how continued professional development of health workers can ensure appropriate building of pediatric clinical and program capacity.
In conclusion, the findings of this review suggest that task shifting of ART care can result in outcomes comparable to routine physician care and that this approach should be considered as part of an overall strategy to scale-up pediatric treatment. Building pediatric clinical capacity among nonphysician staff with training, skills building, and mentorship will be critical to implementing this strategy in a responsible and sustainable way. Particular attention is required to ensure nurses are comfortable in managing HIV-infected infants and young children for whom mortality is the highest. Ensuring caregiver's trust and confidence and increasing acceptability of nurse-led management by mothers and care takers will be important. Further qualitative and quantitative research is needed to inform optimal implementation of task shifting in child health care settings.
Finally, as for adult care, task shifting does not imply that specialist care is not needed, and capacity to manage very sick children and complicated cases will still need to be maintained and improved. Task shifting should be viewed in the context of a broader strategy to optimize human resources available for pediatric HIV care within the context of maternal and child health care, where appropriate task shifting of stable uncomplicated cases frees up specialist care resources for those who need it.
The authors would like to thank the IeDEA-Southern Africa Collaboration for providing outcome data on children treated at collaborating sites. Specifically, the following sites (site investigator in brackets) provided this data: Gugulethu ART Program (Robin Wood); Harriet Shezi Children's Clinic (Harry Moultrie); Khayelitsha ART Program (Kathryn Stinson); McCord Hospital (Janet Giddy); Red Cross Children's Hospital (Brian Eley); Rahima Moosa Mother and Child Hospital (Karl Technau); Tygerberg Academic Hospital (Helena Rabie). IeDEA Southern Africa is funded by an NIH Grant (U01AI069924-08). This work was also supported by funds from the Bill and Melinda Gates Foundation. The authors are also grateful to those authors and investigators who provided additional information on study characteristics and results, particularly to Ruby Fayorsey, Suzue Saito, and Amir Shroufi.
1. Anon. Global Update on HIV Treatment 2013: Results, Impact and Opportunities. WHO Report, in Partnership With UNICEF and UNAIDS. Geneva, Switzerland: WHO; 2013.
2. Hansudewechakul R, Naiwatanakul T, Katana A, et al.. Successful clinical outcomes following decentralization of tertiary paediatric HIV care to a community-based paediatric antiretroviral treatment network, Chiangrai, Thailand, 2002 to 2008. J Int AIDS Soc. 2012;15:17358.
3. Anon. Treat, Train, Retain: Task Shifting
Global Recommendations and Guidelines. Geneva, Switzerland: WHO; 2008.
4. Anon. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for a Public Health Approach. Geneva, Switzerland: WHO; 2013.
5. Callaghan M, Ford N, Schneider H. A systematic review of task-shifting for HIV treatment and care in Africa. Hum Resour Health. 2010;8:8.
6. Sanne I, Orrell C, Fox MP, et al.. Nurse versus doctor management of HIV-infected patients receiving antiretroviral therapy (CIPRA-SA): a randomised non-inferiority trial. Lancet. 2010;376:33–40.
7. Fairall L, Bachmann MO, Lombard C, et al.. Task shifting
of antiretroviral treatment from doctors to primary-care nurses in South Africa (STRETCH): a pragmatic, parallel, cluster-randomised trial. Lancet. 2012;380:889–898.
8. Kiweewa FM, Wabwire D, Nakibuuka J, et al.. Noninferiority of a task-shifting HIV care and treatment model using peer counselors and nurses among Ugandan women initiated on ART: evidence from a randomized trial. J Acquir Immune Defic Syndr. 2013;63:e125–e132.
9. Boulle C, Kouanfack C, Laborde-Balen G, et al.. Task shifting
HIV care in rural district hospitals in Cameroon: evidence of comparable antiretroviral treatment-related outcomes between nurses and physicians in the Stratall ANRS/ESTHER trial. J Acquir Immune Defic Syndr. 2013;62:569–576.
10. Anon. Rapid Paediatric HIV Assessment in Zimbabwe. Harare, Zimbabwe: Ministry of Health and Child Welfare of Zimbabwe, UNICEF, WHO; 2012.
11. Sanderson S, Tatt ID, Higgins J. Tools for assessing quality and susceptibility to bias in observational studies in epidemiology: a systematic review and annotated bibliography. Int J Epidemiol. 2007;36:666–676.
12. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.
13. Freeman M, Tukey J. Transformations related to the angular and the square root. Ann Math Stat. 1950;21:607–611.
14. Miller J. The inverse of the Freeman-Tukey double arcsine transformation. Am Stat. 1978;32:138.
15. Higgins JP, Thompson SG, Deeks JJ, et al.. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–560.
16. 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.
17. Ciaranello AL, Chang Y, Margulis AV, et al.. Effectiveness of pediatric antiretroviral therapy in resource-limited settings: a systematic review and meta-analysis. Clin Infect Dis. 2009;49:1915–1927.
18. Bolton-Moore C, Mubiana-Mbewe M, Cantrell RA, et al.. Clinical outcomes and CD4 cell response in children receiving antiretroviral therapy at primary health care facilities in Zambia. JAMA. 2007;298:1888–1899.
19. van Griensven J, De Naeyer L, Uwera J, et al.. Success with antiretroviral treatment for children in Kigali, Rwanda: experience with health center/nurse-based care. BMC Pediatr. 2008;8:39.
20. Cohen R, Lynch S, Bygrave H, et al.. Antiretroviral treatment outcomes from a nurse-driven, community-supported HIV/AIDS treatment programme in rural Lesotho: observational cohort assessment at two years. J Int AIDS Soc. 2009;12:23.
21. Fayorsey RN, Saito S, Carter RJ, et al.. Decentralization of pediatric HIV care and treatment in five sub-Saharan African countries. J Acquir Immune Defic Syndr. 2013;62:e124–e130.
22. Janssen N, Ndirangu J, Newell ML, et al.. Successful paediatric HIV treatment in rural primary care in Africa. Arch Dis Child. 2010;95:414–421.
23. Weigel R, Feldacker C, Tweya H, et al.. Managing HIV-infected children in a low-resource, public clinic: a comparison of nurse vs. clinical officer practices in ART refill, calculation of adherence and subsequent appointments. J Int AIDS Soc. 2012;15:17432.
24. Patel MR, Yotebieng M, Behets F, et al.. Outcomes of integrated treatment for tuberculosis and HIV in children at the primary health care level. Int J Tuberc Lung Dis. 2013;17:1206–1211.
25. Lynch S, Ford N, van Cutsem G, et al.. Public health. Getting HIV treatment to the most people. Science. 2012;337:298–300.
26. Nyasulu JC, Muchiri E, Mazwi SL, et al.. NIMART rollout to primary healthcare facilities increases access to antiretrovirals in Johannesburg: an interrupted time series analysis. S Afr Med J. 2013;103:232–236.
27. Kenny J, Mulenga V, Hoskins S, et al.. The needs for HIV treatment and care of children, adolescents, pregnant women and older people in low-income and middle-income countries. AIDS. 2012;26(suppl 2):S105–S116.
28. Lehmann U, Van Damme W, Barten F, et al.. Task shifting
: the answer to the human resources crisis in Africa? Hum Resour Health. 2009;7:49.
29. Humphreys CP, Wright J, Walley J, et al.. Nurse led, primary care based antiretroviral treatment versus hospital care: a controlled prospective study in Swaziland. BMC Health Serv Res. 2010;10:229.
30. Luyirika E, Towle MS, Achan J, et al.. Scaling up paediatric HIV care with an integrated, family-centred approach: an observational case study from Uganda. PLoS One. 2013;8:e69548.
31. Leeper SC, Montague BT, Friedman JF, et al.. Lessons learned from family-centred models of treatment for children living with HIV: current approaches and future directions. J Int AIDS Soc. 2010;13(suppl 2):S3.
32. Kredo T, Ford N, Adeniyi FB, et al.. Decentralising HIV treatment in lower- and middle-income countries. Cochrane Database Syst Rev. 2013;6:CD009987.