healthmedicinet

This is a programmatic study of the impact of a government ART intervention program.
In this setting, ART enrollment fluctuated over the years but consistently declined
from 2009 to 2011. While Nigerian national treatment coverage for children has increased
during 2005–2011 as the number of facilities providing HIV/AIDS care has also increased,
our study demonstrates that individual facilities, especially those that have provided
services for over 5 years, may be reaching a “saturation point” leading to declines
in enrollment.

One of the findings in this study is the prediction of mortality by age and WHO stage
in pediatric HIV patients, as we expected. Probability of survival was greatest for
those commencing ART between age 2 and 9. Mortality was high in children under 2 years
and children above 10 years and for children at WHO stages 3 and 4. As expected, lower
hemoglobin values were associated with higher mortality although the association was
not statistically significant, possibly because of missing values for this study variable.
CD4 counts (log-transformed to allow sensitivity to extreme values for children under
5 years) and age (as continuous covariates) were significantly predictive of mortality
using a multivariable Cox regression analysis model. Baseline hemoglobin values (due
to missing values) and sex were not significantly predictive of mortality.

Exploring age and gender differences in baseline characteristics using ANOVA resulted
in statistically significant differences by age group for CD4, mortality, and year
of ART initiation. As we expected, CD4 values in the 0–2 and 3–4-year-olds were significantly
different from the other groups, but not from each other because of the rapid physiologic
changes in this age group. WHO stage was not significantly different by age as median
WHO stage at ART initiation was 3 regardless of age. Children in this population were
at an advanced stage of HIV before commencing ART. This might be because interim and
national guidelines had recommended initiating ART for all patients aged 12 months
with confirmed HIV infection regardless of WHO stage, CD4 count, or viral load since
2008. ART has shown tremendous improvements in short term outcomes in survival for
children with HIV in Africa 11], 12]. Overall, ART led to significant reduction in mortality which can reach 50 % in children
below 2 years without any intervention 13]. After ART initiation, mortality in our study was similar to other studies in both
similar and diverse settings: 2.67 per 100 child years in our study vs. 2.31 per 100
child years in China 14] and 10 % in our study vs. 7.7 % in South Africa 11]. However, mortality rates should be interpreted with caution when loss to follow-up
is excessive 15].

Massey Street Children Hospital offers the family-centered model of care in which
an ‘index’ patient, in this case a child, is the entry point for other family members
in need of care. Few studies have documented the impact of this model on child outcomes
16], probably because comprehensive HIV care is now widespread with index patients being
adults, pregnant women, or children. However, barriers to care still hinder many people
living with HIV and interfere with retention in care 17].

Median age at ART initiation in our study was 41 (inter-quartile range 18–77) months,
which is similar to 18] or younger 16], 19], 20] than the median age found in other pediatric HIV studies in Africa and elsewhere
21]. The hospital is predominantly a children’s center and receives early referrals from
other hospitals, which is most likely the cause of this age discrepancy. There is
an opportunity to examine the impact of early intervention in this cohort of children
because it is associated with better prognosis through immune and growth recovery.
Strengthening the prevention and/or elimination of mother to child transmission program
encourages early detection of infection and subsequent enrollment in HIV care 22]. Globally there are significant shortfalls and inequities in pediatric vs. adult
HIV care leading to treatment gaps 1]. These disparities need to be addressed, and more sites should be opened or services
should be effectively decentralized to other levels of care. However, decentralization
of pediatric HIV care is fraught with concerns about quality of care, loss to follow-up,
and drug resistance; studies reporting high success rates with decentralization 23], 24] have employed measures which are often not generalizable to other settings.

Overall participant retention rate in our study was 64 %, which is lower than but
comparable to the 76.8 % rate (25) in a recent study conducted across purposively
sampled sites with dedicated clinical staff for pediatric HIV. However, only about
50 % of the earliest cohort (2005) in our study were still on treatment at the end
of 2011.

This analysis is one of the first in Nigeria to assess long-term outcomes of pediatric
HIV care and uses routine program real-world data as opposed to trial or study data.
The children in our study center include a rich mix of both hospital and self-referred
patients which makes the interpretation of our study results appropriate for a wider
local and international audience.

Limitations of our study include loss of patients to follow-up, which is a recognized
potential confounder in chronic care. This factor was explored in the sensitivity
analysis before any deductions were made. We have shown that this was not a major
cause of bias. Selection bias for access to the ART program between clients delivering
at health facilities and those delivering at home has been explored as an outcome
of interest, i.e., association of age with survival. Missing data might also constitute
a selection bias. Efforts were made to fill missing gaps with other sources of data.
Because hemoglobin data were significantly missing, our analysis compared the characteristics
of participants with missing data vs. other participants. We found that a child who
died was less likely to have hemoglobin test results available. Thus, we propose that
anemia might have been found to be a risk factor for mortality in this study population
if the data had been available. Other biases such as subject error (e.g., recall of
date of birth), instrument error (malfunction or calibration errors of measurement),
or observer error (errors in recording or transferring data) are probably non-differential.
Routine efforts in the HIV program minimize these errors through internal and external
quality assurance, capacity building, and review meetings.

In Nigeria, the HIV epidemic is widespread and stable, and there are plans to scale
up decentralization of services to primary care level to reduce inequalities and sustain
access to care, especially in HIV infected children, who are especially vulnerable.
An organized health care system was identified as one of the reasons for success in
China’s pediatric HIV program, which has very low losses to follow-up 25]. High quality HIV care has also led to success of pediatric HIV programs in Nigeria
26]. However, the extent to which quality of HIV care in Nigeria is limited by political
instability, poverty, and overstretched health workers should be explored before a
reliable conclusion is made. It is essential that HIV programming in high prevalence
settings continue to take up holistic strategies to create an organized system of
care which ensures that children initiated in ART are retained. More importantly,
elimination of pediatric HIV is imperative, as the burden of HIV on children, their
families, the health system, and the nation as a whole is immense.