Our study reports pre-treatment mortality and losses to follow-up among ART-eligible children from a cohort who were yet to initiate treatment at a West-African clinic. The majority of deaths in our cohort occurred in the first three months after eligibility and during the period of preparation for ART. The pre-treatment mortality rate observed in our setting is relatively higher than reports from other sites; a study in rural Zambia reported 10% mortality among ART-eligible children with a mortality rate of 2.73 (95% CI: 1.7 – 4.18) per 100 child-years while another study reported that 1% of treatment-eligible children receiving care at an urban ART program in Zambia died prior to ART initiation. In a large cohort of 1766 children in Cote d’Ivoire, the reported loss-to-programme rate of 50.3/100 child-years of follow up was also much lower than that reported in this study though, it was not clear if the children in that cohort were eligible for ART. One possible reason for the relatively higher pre-treatment mortality and loss-to-programme rates among ART-eligible children in our setting is the advanced stage of clinical disease and immunosuppression at diagnosis. Many children in resource limited settings are only tested for HIV as part of workup for recurrent or severe ill health; we reported that 80% of the treatment-eligible children attending our clinic were assessed as eligible for ART either at or within 6 months of their being diagnosed with HIV infection. The inclusion of home visits as part of our patient follow-up protocol provides a more accurate and complete ascertainment of survival status that could have contributed to the higher mortality rates observed in our cohort.
The proportion of children who initiated ART in our clinic is comparable to that reported in a study in South Africa where 39.5% (96/243) of ART-eligible children started treatment but considerably lower than figures reported from Cote d’Ivoire and Zambia where 86% and 70% of ART-eligible children respectively started therapy. The median time of 5.1 months from eligibility to ART initiation in our cohort was comparable to that reported from Cambodia where ART was initiated after a median of 4.7 months, but more than twice as long as the 2.1 months reported from rural Zambia. Another study from Zambia reported that children eligible for ART at presentation in rural areas took longer time to initiate treatment than children in urban areas (3.6 vs. 0.9 months).
Delays in treatment initiation among eligible children are reported to be due to several factors[12–14, 19, 20] some of which were observed in our cohort. Firstly, the long process of pre-treatment counselling required repeated clinic visits to ensure that care-givers are both willing and capable of taking the responsibility of supporting the child to take life-long medications with a high level of adherence. Lack of caregiver readiness as assessed during counselling sessions could prolong the process further. This often proved difficult for patients from rural areas and those with working parents though parents/caregivers who were in HIV care, specifically those already ART and with good adherence, underwent ‘accelerated’ counselling with emphasis on the challenges of administering drugs to children, drug-storage and side-effects. We however did not collect information on the number of parents/caregivers who completed the 4 pre-ART counselling sessions among children who died or were LTFU.
A second factor for treatment delay was the time required to obtain approval to start ART from the national ART eligibility committee. The fortnightly occurrence of the meeting frequently resulted in eligible children, whose caregivers had completed pre-ART counselling, waiting for days (sometimes as long as one month if the meeting date fell on a public holiday) to receive approval to start treatment. The concept of the national eligibility committee arose in the early days of ART availability in The Gambia when ART supplies were limited, providers inexperienced and practical application of eligibility criteria was still being debated. Though the committee now meets weekly, the inconvenience of travel to attend the meeting, increased uptake of HCT services nationwide and the widespread availability of ART favour decentralization of the committee to centres involved in provision of ART and may help shorten the time patients have to wait in order to initiate life-saving treatment.
Thirdly, caregivers failed to attend follow-up appointments; almost 20% of treatment eligible children attended only one clinic visit before being lost to follow-up. Poor uptake of paediatric HIV services is multifactorial and reasons include facility-related factors such as long queues, overcrowding, negative staff attitudes; unemployment, lack of money for food and transport, difficulty getting time off work, absenteeism from school and fear of social rejection from disclosure to family and neighbours. While the cost of transport was not a hindrance to clinic attendance in our setting due to the practice of reimbursement of transport costs for all patients, disclosure of HIV-status to a family member or friend was a major factor for loss to follow-up among our adult cohort and suggests that unwillingness to disclose the child’s status to another caregiver may have contributed significantly to failure to keep clinic appointments as well as delays in initiation of ART among eligible children. Disclosure of paediatric infection is multifaceted being dependent on caregiver and family characteristics such asbiologic relationship, caregiver permanence, caregiver beliefs and psychosocial function, as well as child-specific factors such as age, developmental stage, cognitive abilities and psychosocial function. A child’s positive status usually indicates maternal infection and the resultant anxiety, fear of blame, social and healthcare discrimination as well as marital abandonment may negatively influence the mother’s acceptance of the HIV status of the child[21, 23]. Denial of results, despair or depression hinder health-care seeking and social support, and may lead to difficulties in reacting to the options and advice given by health workers[23–25]. Maternal depression (in cases of vertical transmission) and concurrent HIV or other comorbidities may also affect the mother’s ability to care for the child; the belief that the child might die any moment may cause her not to take proper care of the child anymore. In circumstances where the biologic parents may have died or may be too ill to care for the child, responsibility for care is shifted to one or more relatives or family friends; poor coordination amongst multiple caregivers may also compromise the quality of care especially where secondary caregivers have not been disclosed to and therefore do not understand the importance of regular clinic visits or adherence to prescribed medications. Studies from sub-Saharan Africa suggest that whilst many caregivers appreciate shared childcare, sustained adherence and community support as benefits of disclosing a child’s HIV status to others, many still believe that disclosure to others could have negative effects[21, 23, 26]. Community-based support groups have been shown to play a major role in providing continuous support to caregivers of HIV-infected children and though several support groups were linked to our clinic, we did not routinely collect data of patient/caregiver membership with such groups or assess what impact this had on their quality of life. Disclosing a diagnosis of HIV infection to a child has been suggested to have direct benefits for adherence to ART and among adolescents is associated with higher retention in care[26, 28, 29]. The average age of disclosure reported among children in sub-Saharan Africa is 8 years[30, 31]; 75% of those LTFU in our cohort (data not shown) were less than 5 years of age at eligibility and therefore too young for a disclosure their HIV status.
Lastly, we report that tuberculosis was the most common cause of death among treatment-eligible children and may have also been an underlying cause of death among those who reportedly died at home. This is consistent with data from paediatric cohorts in resource limited settings in which tuberculosis has been cited as a major reason for delayed treatment initiation[19, 20] as well as the most common co-morbidity associated with early deaths among both ART naïve and experienced children[13, 14, 32]. As such, treatment for concurrent tuberculosis could also have contributed to the delay in initiation of ART.
Overall, 60% of ART-eligible children in our programme died or were lost to follow-up without starting treatment giving a loss-to-care incidence rate of 128 per 100 child-years of follow-up. We observed that the children who were lost to care were significantly more likely to be less than 2 years of age at ART eligibility or to have advanced clinical disease (predominantly tuberculosis and severe malnutrition - data not shown). Children aged less than 2 years or in WHO stage 3 or 4 had almost twice the risk of being lost to the programme before starting treatment respectively compared with those 5 years of age or older and those in WHO stage 1 or 2. The observation that just over half of children LTFU were in WHO stage 3 or 4 coupled with the relatively shorter median time to LTFU from eligibility compared with the time to initiation of ART (4.2 vs 5.1 months), suggests that many of those LTFU may have died as a result of rapid disease progression and advanced HIV disease during the time of preparation for ART and may have not even completed the process. This observation supports the 2010 WHO recommendation that ART be initiated for all HIV-infected infants and children between 12 and 24 months of age irrespective of their immunological threshold or clinical stage. In our study the vital status of the parent was not related to death or loss to care, which was surprising as HIV-infected orphans in sub-Saharan have been shown to have delayed access to HIV care and treatment as well as reduced clinic attendance[33, 34] .
Children assessed as eligible for ART between January 2006 and January 2010 using the revised 2006 guidelines were observed to have had a significantly higher risk of being lost to programme compared with those assessed as eligible between June 2004 and December 2005 based on the 2003 WHO treatment guidelines. As the paediatric cohort was a sub-cohort of a much larger adult cohort, this may be attributable to the increase in clinic enrolment over time and the reduced ability of clinic staff to effectively manage the patient load. This risk remained albeit insignificantly after adjusting for the effect of age, clinical stage and other variables.
As the only centre in the country performing virological tests for the diagnosis of HIV infection in children <18 months of age and the major paediatric HIV referral clinic in The Gambia, we experienced a good uptake of paediatric HIV services at our HIV clinic, which however, is not representative of paediatric HIV care in West Africa as patients had the advantage of reimbursement of all transport costs incurred in clinic visits, free treatment, nutritional support and provision and school fees for all children enrolled in school up till the age of 18 years. This coupled with the small size of our paediatric cohort, limit the generalizability of findings. There were however several limitations to this study; data were incomplete for many key variables such as WHO stage or CD4 T-cell percentage at eligibility and these children were classified based on immunologic or clinical criteria alone. Secondly, children with moderate or severe malnutrition in our setting were initially classified as WHO stage 3 or 4 respectively without assessing their response to nutritional support and treatment and thus may have been misclassified as eligible for treatment. Unreported deaths among ART-eligible children LTFU in our cohort suggest that the reported mortality rate may be underestimated; in addition, a good proportion of the children ineligible for ART and also LTFU may have progressed to an advanced stage of HIV-disease and died as they became eligible.