Page 138 - SAHCS HIVMed Journal Vol 20 No 1 2019

Page 5 of 7 Original Research

management intervention to target patients failing treatment through the ROTF and simplifying ongoing access to care
and in the ART delivery model provided immediately after and treatment, with peer support, through the AC model
suppression. To date there is limited evidence on the outcomes immediately after suppression could be responsible for the
of high-risk patients in ART delivery models differentiated for positive outcomes.
stable patients.
These results should be viewed in light of a number of
We observed retention and viral suppression outcomes limitations. Firstly, a control group was not obtained,
matching or exceeding those of retention benchmarks and making comparison of these results difficult. Because of the
meta-analyses from sub-Saharan Africa through 18 months retrospective nature of the study and the ability of patients to
of follow-up. Retention in sub-Saharan Africa was estimated self-select into AC care or routine clinic care after ROTF, any
10
to be 81% at 12 months, significantly less than the 94.8% control group chosen would be biased. We chose to compare
retention in this cohort of recently suppressed patients. the outcomes to broader benchmark goals for all ART
While we report 12-month retention from AC enrolment, programmes. Importantly, our analysis begins at AC
not ART initiation, this remains significant. Twelve-month enrolment and not ART start. Because the largest drop in
retention is only slightly lower than the 12-month 97.0% retention occurs immediately after ART start, care must be
retention observed in stable patients in ACs at the same taken when comparing these results to those of newly
15
clinic and the 99.0% retention observed in stable patients enrolled patients. Secondly, tracing of patients lost to follow-
in a similar community AC cohort and the 12-month up to identify undocumented transfers was not completed.
36
31
retention in ACs across the Cape Metro. In a 2015 However, any bias this limitation created would serve to
systematic review of VL suppression, 12-month suppression reduce observed retention. Thirdly, patients were given the
in sub-Saharan Africa was estimated to be 64.2%. We choice to join an AC or return to routine clinic care after
37
observed 83.9% viral suppression at 12 months after AC completing the ROTF intervention. It is therefore possible
enrolment. These outcomes also compare well to those of that only the motivated patients joined ACs, and our results
patients switched to second-line regimens. Analysis of a are not representative of all patients who have experienced
cohort in Durban, South Africa, found 25.0% virological an elevated VL. This scenario is unlikely because fewer
failure every 6 months after switching to second-line than 10.0% of patients chose to return to routine care after
regimens. In a European cohort of treatment-experienced achieving suppression. Regardless, by allowing patients to
38
patients who recently achieved viral suppression, 31.0% self-select into care models, the probability that they will find
of patients experienced viral rebound within 1 year. a model of care that suits their life, and thus maintain
35
Considering 100.0% of the study cohort experienced recent adherence, increases. Fourthly, it is possible that transmitted
elevated VLs and 79.0% had either recently been switched resistance to first-line regimens was responsible for the
to a second-line regimen within the ROTF intervention or positive response in patients switched to second line. This
entered ROTF on a second-line regimen, our results are would only account for a small proportion of patients given
promising. the relative infrequency of transmitted resistance in South
Africa and the extensive evidence indicating that non-
45
Ninety per cent of patients retained in care after 18 months adherence is the primary cause of an elevated VL. 32,46 It is also
were still receiving their care in ACs, suggesting a high level possible that patients switched to second-line therapy had
of satisfaction with the service delivery model. This result is positive viral suppression results despite continued poor
47
important because patient satisfaction is a strong predictor adherence because of a switching effect. However, this
of adherence to ART regimens. While patients on effect is thought to be minimal as most patients who fail
39
ART experience elevated VLs (including viral rebound) for second-line treatment after switching do so within the first
a variety of reasons, the single largest predictor is sub- 2 years and the outcome would be seen within our follow-up
optimal treatment adherence. 4,40,41,42,43 Therefore, it follows period. Unfortunately, we are unable to differentiate the
48
that patients have better outcomes in models that better fit patients who entered the ROTF intervention on second-line
their lives. The AC model may support patients struggling regimens from those who were switched to second line
with adherence in routine care by reducing or removing during the intervention and are susceptible to this switching
barriers to adherence. effect. In addition, because of the limited number of patients,
there was insufficient power to analyse associations between
In addition to the differentiated nature of the ACs, the outcomes and patient demographics. An attempt was
model was also differentiated from routine care, providing made to include all patients who participated in both ROTF
more intensified integrated adherence and clinical care and ACs by cross-referencing both databases; however, it is
through the ROTF intervention to meet the needs of possible that patients were missed in our sampling approach.
patients as they attempted to achieve viral suppression In addition, data were collected from routine clinical
after elevated VLs and remain in care. This model of VL- databases and thus may be subject to data quality error.
informed differentiated care has been shown to be effective Finally, it is possible that the exclusion criteria that were
and cost-efficient in supporting patients who experience applied biased the results towards increased retention and
elevated VLs in routine care. In other words, it is possible viral suppression. This possibility was minimised by cross-
44
that both the intense support in achieving suppression checking multiple data sources and excluding the entire

http://www.sajhivmed.org.za 131 Open Access