Scaling-up access to antiretroviral therapy for children: A cohort study evaluating care and treatment at mobile and hospital-affiliated HIV clinics in rural Zambia

Background: Travel time and distance are barriers to care for HIV-infected children in rural sub-Saharan Africa. Decentralization of care is one strategy to scale-up access to antiretroviral therapy (ART), but few programs have been evaluated. We compared outcomes for children receiving care in mobile and hospital-affiliated HIV clinics in rural Zambia. Methods: Outcomes were measured within an ongoing cohort study of HIV-infected children seeking care at Macha Hospital, Zambia from 2007 to 2012. Children in the outreach clinic group received care from the Macha HIV clinic and transferred to one of three outreach clinics. Children in the hospital-affiliated clinic group received care at Macha HIV clinic and reported Macha Hospital as the nearest healthcare facility. Results: Seventy-seven children transferred to the outreach clinics and were included in the analysis. Travel time to the outreach clinics was significantly shorter and fewer caretakers used public transportation, resulting in lower transportation costs and fewer obstacles accessing the clinic. Some caretakers and health care providers reported inferior quality of service provision at the outreach clinics. Sixty-eight children received ART at the outreach clinics and were compared to 41 children in the hospital-affiliated clinic group. At ART initiation, median age, weight-for-age z-scores (WAZ) and CD4+ T-cell percentages were similar for children in the hospital-affiliated and outreach clinic groups. Children in both groups experienced similar increases in WAZ and CD4+ T-cell percentages. Conclusions: HIV care and treatment can be effectively delivered to HIV-infected children at rural health centers through mobile ART teams, removing potential barriers to uptake and retention. Outreach teams should be supported to increase access to HIV care and treatment in rural areas. © 2014 van Dijk et al. The study was approved by the Ministry of Health of the Government of Zambia, the University of Zambia Biomedical Research Ethics Committee and the Institutional Review Board of the Johns Hopkins Bloomberg School of Public Health. Caretakers provided written informed consent and children 8–16 years of age provided assent to participate in the study. In 2004, the Government of Zambia initiated public sector ART programs, with roll-out beginning in primary care clinics in the Lusaka Urban District [26] and subsequently implemented throughout the country. The program provides ART and basic laboratory tests, including CD4+ T-cell counts, free of charge. The number of ART sites increased from four in 2004 to 454 in 2010 [27]. While 61% of the Zambian population resides in rural areas [28], ART services are primarily offered in urban areas and in district level hospitals where the infrastructure and human and technical resources are greatest. To reach those affected by HIV in rural areas and increase access to HIV services, the Zambian Ministry of Health introduced the national Mobile ART Services program in 2007 [29]. Under this program, mobile ART teams of medical professionals were created at district hospitals. Rural health centers (RHC) in the catchment area were selected as designated ART outreach sites to be visited every two weeks by the mobile ART teams. Rural health center staff provide reproductive, maternal and child health care, treatment for tuberculosis, HIV testing, and other basic services, but generally do not have the training or capacity to provide ART. The mobile ART team assists staff at the ART outreach site in providing ART services, builds capacity and coordinates laboratory services. This program has contributed to the decentralization of ART services to the primary health care level to maximize limited resources and reach the greatest number of people in need. This study was conducted at the rural HIV clinic at Macha Hospital in Southern Province, Zambia. The study setting and population have been described in detail elsewhere [30], [31]. In brief, Macha Hospital serves as a referral hospital for at least 13 rural health centers, providing services for patients within an 80 km radius. The catchment area of Macha Hospital is populated by traditional villagers living in small, scattered homesteads, characteristic of much of rural sub-Saharan Africa, with an estimated population size of over 150,000 persons. The HIV clinic has provided care to more than 8500 HIV-infected adults and children since 2005. HIV care services, including antiretroviral treatment, are provided through the Government of Zambia’s antiretroviral treatment program, with support from the President’s Emergency Plan for AIDS Relief. The clinic provides care and treatment free of charge by physicians, clinical officers and nurses. Children diagnosed with HIV infection are determined to be eligible for ART according to guidelines established by the Ministry of Health [32]. Children eligible for ART must undergo counseling to ensure the family is prepared to initiate ART. Upon initiation, children are seen every two weeks for the first month and every month for the following two months. Thereafter, the child is seen at three-monthly intervals if adherence and clinical response are good. Standard ART regimens consist of zidovudine, stavudine or abacavir plus lamivudine, and nevirapine or efavirenz. Due to increasing patient numbers and considerable travel distance to access HIV services [30], the HIV clinic began a mobile ART program in 2007. Three rural health centers were selected as outreach clinics based on distance and size of the catchment populations, and were located 13 km (Mapanza RHC in Choma District), 21 km (Chilala RHC in Kalomo District) and 46 km (Moobola RHC in Namwala District) from the hospital. In 2010, Chilala RHC began providing ART services independently and children who were seen in the outreach clinic were officially transferred from the Macha HIV clinic. The selected outreach clinics are staffed with an average of one clinical officer and a minimum of two nurses. The outreach team from the Macha Hospital clinic consists of at least one clinical officer or licentiate, nurse, pharmacy dispenser, laboratory assistant, counselor and data entry clerk. Medications, medical consumables and transportation are provided by the hospital. As laboratory testing cannot be performed at the outreach clinics, blood samples are collected and transported from the outreach clinic to the Macha Hospital laboratory, and results are returned during the next outreach visit. Clinically stable patients with good adherence are provided with a 3-month supply of medication. Children are eligible for referral to the outreach clinic for care and treatment if they have been stable on ART for at least three months, demonstrated good adherence, have no opportunistic infections, and their caregiver requested to receive care closer to home. Children not yet eligible for ART and in stable condition can also be referred to the outreach clinic with the understanding that referral back to the hospital clinic might be needed once the child becomes eligible for ART or their condition worsens. Few children start ART at an outreach clinic. Beginning in September 2007, HIV-infected children younger than 16 years of age and registered at the HIV clinic at Macha Hospital were eligible for enrollment into an observational cohort study. This report describes a subset of these children receiving care between September 2007 and March 2012. Children were evaluated at study visits approximately every three months. At each visit, a structured questionnaire was administered to the caregiver to collect information on socio-demographics, household characteristics, and medical and treatment history. The child was examined to measure height and weight, and a blood specimen was obtained to measure CD4+ T-cell counts and percentages (Guava Easy CD4 system; Guava Technologics, Inc., Hayward, CA) as part of clinical care. Plasma levels of HIV RNA were quantified by reverse transcriptase polymerase chain reaction assay (Amplicor HIV-1 Monitor v. 1.5, Roche Molecular Systems; lower limit of detection of 400 copies/mL, upper limit of detection 750,000 copies/mL) as part of the study. Due to financial constraints, viral load testing was not performed at all study visits; samples for viral load testing were selected every three months during the first year of ART and every six months thereafter. Adherence was assessed at every visit by pill counts and syrup volume measurements. For children who missed study visits, home visits were attempted to ascertain their status. Upon implementation of the mobile ART program, some study children were transferred to outreach clinics and a study assistant was added to the outreach team to continue study procedures at outreach visits. A questionnaire was administered to the caretaker at one outreach clinic visit to obtain information on health care delivery system related factors, including access to care and perceived quality of care at the outreach clinic. Two analyses were conducted. The first analysis compared modes of transportation and travel time before and after transfer to the outreach clinic and assessed the perceived quality of care at the outreach clinics. HIV-infected children who transferred to the outreach clinic and whose caregiver completed a questionnaire both at study entry and upon transfer to the outreach clinic were eligible to be included in this analysis. The second analysis compared treatment outcomes between children receiving ART who transferred to the outreach clinics and children receiving ART at the Macha HIV clinic who reported Macha Hospital as their hospital-affiliated rural health center. This comparison group was selected as these children live in the vicinity of the Macha HIV clinic and would not have been transferred to an outreach clinic, thereby removing potential confounding by distance to the clinic which is known to impact clinical outcomes. All children were required to have initiated ART before September 1, 2011 and have at least one study visit after ART initiation. Children in the outreach group were required to have transferred to an outreach clinic before September 1, 2011 and have at least one study visit after transfer. Children in both groups remained in the analysis until the first of death, transfer, loss to follow-up, or administrative censoring on March 1, 2012. Transfer in this context was defined as transfer of care to a clinic other than one of the outreach clinics. Children attending Chilala Clinic when it became independent were censored on their last study visit. Loss to follow-up was defined as failure to attend a study visit for at least six months prior to March 1, 2012. Descriptive statistics were used to compare the hospital-affiliated clinic group and outreach groups on characteristics at study entry and at ART initiation. A measure of socio-economic status (SES) was calculated based on the Demographic and Health Survey SES scale used in Zambia [33], with scores ranging from 0 to 24. SES percentiles were based on the predetermined cutoffs (<25th = 0–6; 26–50th = 7–12; 51–75th = 13–18; >75th = 19–24). Weight-for-age z-scores (WAZ) among children younger than 10 years of age were calculated based on the WHO growth standards [34], and children with z-scores below −2 were defined as underweight. Severe immunodeficiency was defined by CD4+ T-cell percentage according to the WHO 2006 treatment guidelines [35]. If laboratory tests were not available from the visit at which ART was initiated, results within three months prior to the date of initiation were used. Immunologic, clinical and virologic treatment outcomes were assessed, including CD4+ T-cell percentage, WAZ and viral suppression. For CD4+ T-cell percentage and WAZ, children were included if they had at least one measure available after ART initiation. Both outcomes were evaluated using linear regression with generalized estimating equations (GEE) with robust variance estimation to account for repeated measures per child. Models for each outcome compared: 1) pre-transfer visits in the outreach group to all visits in the hospital-affiliated clinic group; 2) post-transfer visits in the outreach group to all visits in the hospital-affiliated clinic group; and 3) changes between last pre-transfer visit and the 6-month post-transfer visit in the outreach group to changes between visits at a similar interval (6 months) in the hospital-affiliated clinic group. Models included the duration of ART and other covariates found to differ (p<0.10) between children in the outreach and hospital-affiliated clinic groups and known to be causally associated with each outcome. Trajectories of each outcome were explored using linear mixed effects models with random intercept, exchangeable correlation structure and robust standard error estimation. As neither outcome was linear over time, a spline term was added at 7.5 months, the upper window around the 6-month measure. The primary exposure in the models was clinic (hospital-affiliated vs. outreach clinic), which was treated as a time-varying covariate. Interactions between clinic and time were included in the models to determine whether trajectories of the outcomes differed between children in the hospital-affiliated and outreach clinics. For virologic outcomes, the proportion of children with viral suppression, defined as a HIV viral load below the limit of detection (400 copies/mL), was calculated for each visit after ART initiation. The proportion of children with viral suppression at each visit was compared between children in the hospital-affiliated and outreach clinic groups using chi-square tests. Unadjusted and adjusted logistic regression models with GEE were used to compare the odds of detectable viral load between children in the outreach and hospital-affiliated clinics (treated as a time-varying covariate) during follow-up. Detectable viral load was assessed at three thresholds: >400, >1000 and >10,000 copies/mL. Adherence during follow-up was also assessed. Caregivers were instructed to bring all unused medications to each clinic visit and adherence was measured by pill count or measurement of liquids for each drug prescribed. Adherence measures were capped at 100%. For children taking individual drugs, the adherence percentage of the drug to which the patient was least adherent was used. Optimal adherence was defined as taking more than 95% of drugs prescribed. The proportions of children with optimal adherence at each visit and at all visits were compared between children in the hospital-affiliated and outreach clinics using chi-square tests. Unadjusted and adjusted logistic regression models with GEE were used to compare the odds of optimal adherence between children in the outreach and hospital-affiliated clinics (treated as a time-varying covariate) during follow-up. All analyses were conducted using SAS for Windows version 9.1 (SAS Institute Inc., Cary, NC) and Stata, version 9 (StataCorp LP, College Station, TX).