HHS Public Access

Author manuscript
Lancet HIV. Author manuscript; available in PMC 2021 April 01.
Author Manuscript

Published in final edited form as:

Lancet HIV. 2020 April ; 7(4): e229–e237. doi:10.1016/S2352-3018(19)30402-3.

Point-of-care HIV viral load testing combined with task shifting

to improve treatment outcomes (STREAM): findings from an
open-label, non-inferiority, randomised controlled trial
Paul K Drain,
Department of Global Health, School of Public Health and School of Medicine, Department of
Medicine, School of Medicine, Department of Epidemiology, School of Public Health, University of
Author Manuscript

Washington, Seattle, WA, USA

Jienchi Dorward,
Nuffield Department of Primary Care Health Sciences, University of Oxford, Oxford, UK, Centre
for the AIDS Programme of Research in South Africa, University of KwaZulu-Natal, Durban,
South Africa

Lauren R Violette,
Department of Medicine, School of Medicine, University of Washington, Seattle, WA, USA

Justice Quame-Amaglo,
Department of Global Health, School of Public Health and School of Medicine, University of
Washington, Seattle, WA, USA
Author Manuscript

Katherine K Thomas,
Department of Global Health, School of Public Health and School of Medicine, University of
Washington, Seattle, WA, USA

Natasha Samsunder,
Centre for the AIDS Programme of Research in South Africa, University of KwaZulu-Natal,
Durban,South Africa

Hope Ngobese,

Correspondence to Dr Paul K Drain, Department of Global Health, University of Washington, Seattle 98104–2420 WA, USA
[email protected].
Author Manuscript

Contributors
PKD and NG conceived the trial and are the co-principal investigators. PKD, NG, JD, JQ-A, NS, HN, KM, PM, RVB, KN, and CC
designed the study. NG, JD, PKD, NS, J-QA, HN, KM, PM and KN implemented the study and acquired the data. LRV, and DD did
the statistical analysis. All authors critically revised the manuscript and consented to final publication.
Declaration of interests
All other authors declare no competing interests.
Data sharing
We will share individual participant data that underlie the results reported in this article, after de-identification (text, tables, figures,
and appendices). The study protocol will be available. Patient data will be available beginning 9 months and ending 36 months after
article publication. The data will be made available by the investigators whose proposed use of the data has been approved by an
independent review committee (“learned intermediary”) identified for this purpose. Proposals can be submitted up to 36 months
following article publication. After 36 months the data will be available in our university’s data warehouse but without investigator
support other than deposited metadata.
See Online for appendix
 Drain et al. Page 2

Prince Cyril Zulu Communicable Disease Clinic, Durban Municipality, Durban, South Africa,
Author Manuscript

National Health Laboratory Service, Durban, South Africa

Koleka Mlisana,
School of Laboratory Medicine and Medical Sciences, University of KwaZulu-Natal, Durban,South
Africa

Pravikrishnen Moodley,
School of Laboratory Medicine and Medical Sciences, University of KwaZulu-Natal, Durban,South
Africa, Department of Virology, Inkosi Albert Luthuli Central Hospital, Cato Manor, South Africa

Deborah Donnell,
Department of Global Health, School of Public Health and School of Medicine, University of
Washington, Seattle, WA, USA

Ruanne V Barnabas,
Author Manuscript

Department of Global Health, School of Public Health and School of Medicine, Department of
Medicine, School of Medicine, University of Washington, Seattle, WA, USA

Kogieleum Naidoo,
Centre for the AIDS Programme of Research in South Africa, School of Nursing and Public
Health, CAPRISA-MRC HIV-TB Pathogenesis and Treatment Research Unit, and Doris Duke
Medical Research Institute, University of KwaZulu-Natal, Durban,South Africa

Salim S Abdool Karim,

Centre for the AIDS Programme of Research in South Africa, School of Nursing and Public
Health, University of KwaZulu-Natal, Durban,South Africa, Department of Epidemiology,
Columbia University, New York, NY, USA
Author Manuscript

Connie Celum,
Department of Global Health, School of Public Health and School of Medicine, Department of
Medicine, School of Medicine, University of Washington, Seattle, WA, USA

Nigel Garrett
Centre for the AIDS Programme of Research in South Africa, School of Nursing and Public
Health, University of KwaZulu-Natal, Durban,South Africa

Summary
Background—Monitoring HIV treatment with laboratory testing introduces delays for providing
appropriate care in resource-limited settings. The aim of our study was to determine whether
point-of-care HIV viral load testing with task shifting changed treatment and care outcomes for
adults on antiretroviral therapy (ART) when compared with standard laboratory viral load testing.
Author Manuscript

Methods—We did an open-label, non-inferiority, randomised controlled trial in a public clinic in

Durban, South Africa. We enrolled HIV-positive adults (aged ≥18 years) who presented for their
first routine HIV viral load test 6 months after ART initiation. Individuals were randomly assigned
by a random number allocation sequence to receive either point-of-care viral load testing at
enrolment and after 6 months with task shifting to enrolled nurses (intervention group), or
laboratory viral load testing (standard-of-care group). The primary outcome was combined viral
suppression (<200 copies per mL) and retention at 12 months after enrolment. A non-inferiority

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 3

margin of 10% was used. Analysis was done by intention to treat. This study was registered with
Author Manuscript

ClinicalTrials.gov, NCT03066128.

Findings—Between Feb 24, 2017, and Aug 23, 2017, we screened 657 participants, and 390
were enrolled and randomly assigned to either the intervention group (n=195) or standard-of-care
group (n=195). 175 (90%) individuals in the intervention group and 148 (76%) individuals in the
standard-of-care group had the primary outcome of retention with viral suppression, a difference
of 13·9% (95% CI 6·4–21·2; p<0·00040). 182 participants (93%) in the intervention group had
viral suppression compared with 162 (83%) in the standard-of-care group (difference 10·3%, 3·9–
16·8; p=0·0025); 180 (92%) and 162 (85%) were retained in care (7·7%, 1·3–14·2; p=0·026).
There were no adverse events related to point-of-care HIV viral load testing or task shifting.

Interpretation—Point-of-care viral load testing combined with task shifting significantly

improved viral suppression and retention in HIV care. Point-of-care testing can simplify treatment
and improve outcomes for HIV-positive adults receiving ART in resource-limited settings.
Author Manuscript

Funding—National Institute of Allergy and Infectious Diseases.

Introduction
Over half of the 37 million people living with HIV are accessing antiretroviral therapy
(ART), which has substantially reduced AIDS-related deaths worldwide.1 The goal of ART
is to maintain viral suppression, which restores immunological function, reduces
transmission, and improves patient outcomes. UNAIDS has set an ambitious target for 90%
of people receiving ART- to have viral suppression by 2020.2 Globally, less than half of all
people living with HIV have viral suppression, and ART coverage and HIV treatment
outcomes have large regional disparities, particularly in low-income and middle-income
countries.1
Author Manuscript

WHO recommends monitoring ART effectiveness with routine HIV viral load testing.3
Patients with HIV viraemia should receive prompt adherence counselling and potential,
whereas those with viral suppression might be candidates for differentiated care through task
shifting to other health-care workers or community-based ART delivery programmes. These
programmes aim to increase health-system efficiencies and improve patient care, and are
facilitated by availability of viral load monitoring.4 WHO estimates the annual demand for
viral load tests will increase from 10·7 million in 2016, to 28·5 million in 2021,5 but scaling
up viral load testing has been limited by infrastructure and human resources.6–8 Even where
laboratory viral load testing is available, such as in South Africa, some patients do not
receive their results in a timely manner, which can lead to emergence of drug resistance, risk
of ongoing transmission, and excess morbidity and mortality.9
Author Manuscript

Diagnostic point-of-care tests have rapidly emerged worldwide.10 The Xpert HIV-1 viral
load assay (Cepheid, Sunnyvale, CA, US) requires 1 mL of plasma to provide a quantitative
result (40–10 million copies per mL) in less than 2 h and has received European and WHO
regulatory approvals.11,12 The assay is done on an automated molecular GeneXpert
(Cepheid, Sunnyvale, CA USA) platform, which is widely used for the diagnosis of
tuberculosis.12 We showed good accuracy of Xpert HIV-1 viral load when done in a South
African clinic,13 but there have been no studies showing the effectiveness of point-of-care

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 4

viral load testing to improve treatment outcomes. Furthermore, although ART is mainly
Author Manuscript

provided by professional nurses in primary care clinics, there could be a role for task shifting
of ART provision to health-care workers with less training, such as enrolled nurses

Therefore, we compared point-of-care Xpert HIV-1 viral load testing with same-day
counselling and task shifting to enrolled nurses against standard-of-care laboratory viral load
testing for achieving viral suppression and retention in care for HIV-positive adults receiving
ART.

Methods
Study design and participants
We undertook the STREAM (Simplifying HIV TREAtment and Monitoring) study, a single
site, open-label, two-arm, non-inferiority, randomised controlled trial. The study site was at
Author Manuscript

the CAPRISA Ethekwini Clinical Research Site and the neighbouring Prince Cyril Zulu
Clinic, which is a large urban public clinic that provides HIV and primary care services for a
diverse and mobile population of more than 10 000 patients in Durban, South Africa. Details
of the research protocol have been published previously.14

Eligible study participants were HIV-positive adults, aged 18 years or older, who presented
for their first routine HIV viral load test 6 months after ART initiation. We enrolled
participants best suited to receive differentiated HIV care by excluding those who were
pregnant, had active tuberculosis, or required acute medical care by a physician. However, if
participants became pregnant or were diagnosed with tuberculosis during follow-up, they
were included in the final analysis.

All participants gave written informed consent. The study received ethical approval from the
Author Manuscript

University of KwaZulu-Natal, Durban, South Africa (BFC296/16) and University of

Washington, Seattle, WA, USA (STUDY00001466). This study is registered with
ClinicalTrials.gov, NCT03066128.

Randomisation and masking

We randomly assigned participants (1:1) to receive either point-of-care viral load testing
with same-day counselling and task shifted care by an enrolled nurse (registered nurse in the
USA), or standard-of-care laboratory-based viral load testing and care by a professional
nurse (nurse practitioner in the USA). The study statistician generated an allocation
sequence with random numbers using SAS (version 9.4). Those who enrolled and allocated
the participant were masked to the assignment.
Author Manuscript

Sequentially numbered, sealed, opaque envelopes containing a study group allocation and
participant identification number were opened once an eligible, consenting participant was
enrolled.

Procedures
At enrolment, we collected sociodemographic, medical information, and other potential risk
factors for viral failure or loss to follow-up, which included the WHO Alcohol Use Disorder

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 5

Identification Tool,15 WHO Violence Against Women instrument (female participants only),
16 and Patient Health Questionnaire-2 to screen for depression.17 All participants- were
Author Manuscript

followed up for 12 months and received care according to the South African HIV treatment
guidelines, including testing for viral load and creatinine at enrolment, and viral load,
creatinine, and CD4 cell count after 6 months.18 Any participant with a viral load of more
than 1000 copies per mL received enhanced adherence counselling and repeat viral load
testing after 2 months. If the repeat viral load was more than 1000 copies per mL, the
participant was offered to switch to a standard second-line ART regimen. We did
retrospective HIV genomic resistance testing using standard Sanger consensus sequencing at
enrolment and at 12 months. Participants in both groups were reimbursed 100 South African
rand per clinic visit, in accordance with South African research guidelines.19

In the intervention group, participants with viral suppression at enrolment and no

comorbidities were seen by an enrolled nurse, rather than a professional nurse. In South
Author Manuscript

Africa, enrolled nurses receive 2 years of training and are less highly qualified than
professional nurses, who have 4 years of training and are the standard providers of HIV care.
For this trial, enrolled nurses had received the standard South African Nurse Initiated and
Managed ART training, which covers basic HIV management, ART side-effects, adherence
counselling, and appropriate referral systems.20 At routine clinical visits, enrolled nurses
noted vital signs, screened for tuberculosis, assessed adherence by asking individuals about
missing one or more ART doses in past 4 days, and issued predispensed ART packages. If
enrolled nurses detected clinical symptoms or abnormal vital signs, then the participant was
referred up to a professional nurse for more comprehensive clinical management. All routine
testing in the intervention group was done onsite using point-of-care Xpert HIV-1 viral load,
Statensor Xpress-i (Nova Biomedical, Waltham, MA, USA) for creatinine, and Pima (Alere,
Waltham, MA, USA) for CD4 cell count. Point-of-care viral load testing and associated
Author Manuscript

clinical care was done by the research team.

Participants were asked to wait in the clinic for the point-of-care test results, and waiting
times were minimised by initiating point-of-care testing at the start of the clinical visit.
Therefore, the intention was to deliver results to the participants in the intervention group on
the same day and during the same clinical visit.

In the standard-of-care group, participants were seen by a professional nurse, who had also
received Nurse Initiated and Managed ART training. The nurse did the same procedures but
sent all blood specimens to the routine National Health Laboratory Service for viral load
testing using m2000 RealTime (Abbott, Chicago, IL, USA). In accordance with standard
practice in the South African health-care system, participants were expected to receive their
Author Manuscript

viral loads at their next clinical visit, which is scheduled at the clinician’s discretion and
typically occurs after 4 weeks (28 days) or 8 weeks (56 days).

Based on the South African policy, the study clinic participated in a programme for
clinically stable participants to collect ART at community pick-up points, such as private
pharmacies and community-based organisations.21 Participants in both groups were eligible
for referral into the programme if they were not pregnant, were clinically stable, both their
enrolment and month 6 viral loads were less than 40 copies per mL, and CD4 count was

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 6

more than 200 cells per µL. Once referred, participants collected ART every 2 months at
Author Manuscript

community pick-up points and were reviewed at the clinic by a professional nurse after 6
months. All blood testing occurred at the clinic. Participants not eligible for community-
based ART delivery continued with clinic visits every 2 months. Any participant more than 2
weeks late for a scheduled clinic or community pick-up point visit received one phone call
from the clinical team and and was asked to resume ART collection at the clinic, per South
African guidelines.14

Outcomes
The primary outcome was a composite measure of retention in care with viral suppression
(<200 copies per mL) after 12 months. We used this composite because both components are
necessary to have positive treatment outcomes for HIV-positive adults. Retained in care was
defined as collecting ART at the study clinic or a community pick-up point between 44
Author Manuscript

weeks and 56 weeks after enrolment. Participants not retained in care at 56 weeks were
tracked by the study team up to 60 weeks for viral load testing. All viral load testing for the
primary outcome was done using a laboratory-based Cobas 6800/8800 machine (Roche,
Basel, Switzerland).

A priori secondary outcomes detailed in the study protocol14 were viral suppression using
thresholds of 1000 and 50 copies per mL, viral suppression and retention at any clinic,
retention in care alone at any clinic, time to detection of virological failure and switch to
second-line ART, and presence of HIV drug resistance mutations among participants with
viraemia at study exit. We used laboratory and clinical records to assess the timing of viral
load test results available in health-information systems, when test results were
communicated to participants, and when participants were referred into the community-
based ART delivery programme.
Author Manuscript

Statistical analysis
The study sample size was determined using a non-inferiority hypothesis for the intervention
group compared with the control group, assuming that 80% of adults in the control group
would be retained with viral suppression at 12 months. Using a non-inferiority margin of
10%, 390 participants were sufficient to provide 80% power to rule out a decrease of 10%,
based on a one-sided 95% CI. If non-inferiority was successfully shown, we planned to
assess superiority, as stated in the study protocol.14 We recorded all clinical data using
standardised electronic case report forms in the iDataFax system (version 2014.1.1). All data
entry underwent three stages of quality control, including immediate source document
review, internal quality audits, and weekly quality reports.
Author Manuscript

The primary analysis was done in the intention-to-treat population enrolled in the study and
compared the proportions who had the composite primary outcome by estimating the
absolute risk difference between the intervention and control groups. The same population
was used for all secondary outcomes, unless noted. For the outcomes assessing viral
suppression among those who had a viral load done, we restricted the population to those
with a viral load at study exit. Non-inferiority of the intervention group was defined as the
one-sided 95% Newcombe-Wilson score CI for the difference in proportions to exclude a

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 7

non-inferiority margin of 10%.22,23 Superiority was assessed with Fisher’s exact test, and
Author Manuscript

two-sided 95% CIs were generated for the difference in proportions. We separately
compared each component of the primary outcome by group, as well as secondary binary
outcomes, using the same methods. For time-to-event outcomes, we compared groups using
hazard ratios (HRs) generated from Cox proportional models, except when differences were
so extreme that all events in the intervention group occurred before all events in the control
group, resulting in perfect separation of events. In such cases, we could not compute a HR
and report a log-rank test p value instead. We used Kaplan-Meier curves to estimate time to
referral into the community-based ART delivery programme for each study group. We
compared health-care utilisation variables using Fisher’s exact test and t tests when
appropriate. We did all analyses using SAS (version 9.4). We did not have a data monitoring
committee because no investigational product was used.

Role of the funding source

Author Manuscript

Cepheid loaned the GeneXpert instruments for this study at no cost. Cepheid and the funder
of the study had no role in the study design, data collection, data interpretation, or writing of
the report. The corresponding author had full access to all the data in the study and had final
responsibility for the decision to submit for publication.

Results
Between Feb 24, 2017, and Aug 23, 2017, we screened 657 patients, of whom 267 were
ineligible or chose not to enroll, and 390 were enrolled, randomly assigned, and evaluated
for the primary outcome (figure 1, appendix p 1). The median patient age was 32 years (IQR
27–38) and 235 (60%) were female (table 1). Most participants earned less than 4000 South
African rand (approximately USD$280) per month, took public transportation to the clinic,
Author Manuscript

and estimated their travel distance to be more than 5000 m and travel time less than 60 min.

128 (33%) of 390 participants had been diagnosed HIV positive at least 12 months earlier,
and 18 (5%) reported receiving ART before initiation of the current regimen. The median
CD4 cell count was 366 cells per µL (IQR 204–546) at ART initiation, and 468 cells per µL
(309–666) at study enrolment. Of the 18 (5%) participants with unsuppressed viral load
(>1000 copies per mL) at enrolment, 14 (78%) had evidence of any HIV drug resistance
mutations. Sociodemographic, psychosocial, and clinical measures were similar between the
two study groups (table 1).

After 12 months of clinical follow-up, 175 (90%) of 195 participants in the intervention
group and 148 (76%) of 195 participants in the standard-of-care group were retained in care
Author Manuscript

with viral suppression, a difference of 13·9% (95% CI 6·4–21·2; p<0·00040; table 2). Each
component of the primary outcome was higher in the intervention group than the standard-
of-care group: 10·3% higher rate of viral suppression and a 7·7% higher rate of retention
(table 2). There were no adverse events related to point-of-care HIV viral load testing or task
shifting care.

We evaluated several secondary study endpoints. All secondary outcomes, using higher and
lower viral load thresholds, and including retention in any clinic, showed improvements for

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 8

point-of-care viral load testing with task shifting versus standard-of-care, although not all
Author Manuscript

were significant (table 2). When the composite outcome included retention in care at any
clinic (ie, not just the Prince Cyril Zulu Clinic), which had less certainty, point-of-care viral
load testing increased retention and viral suppression by 12·3% (p=0·0011; table 2). When
evaluating the composite outcome with the WHO viral load threshold of less than 1000
copies per mL, or a more strict threshold of less than 50 copies per mL, the intervention
group had 12% and 14% improvements in retention with viral suppression, respectively
(table 2). When comparing each individual measure of HIV viral load of less than 1000
copies per mL or viral load of less than 50 copies per mL, the intervention point-of-care
viral load testing group had significantly higher rates of viral suppression than did the
standard-of-care group (table 2). When restricted to those with a viral load result at study
exit, the proportion of patients with viral suppression (<200 copies per mL) was 5·3% higher
in the intervention group than in the standard-of-care group (p=0·051). Reasons for not
Author Manuscript

being retained in care are presented in the appendix (p 2).

All 398 point-of-care viral load results except for one (no reason noted) were entered into
the health information system on the same day, whereas laboratory-based results were
entered into the record and available to clinicians a median of 2 days (IQR 1–4) after blood
draw (table 3). Overall, 397 of 398 point-of-care viral load results were communicated to
participants by their provider, compared to 304 of 373 (82%) in the standard-of-care group,
an improvement of 18·3% (p<0·0001; table 2). Almost all (99%) point-of-care viral load
results were communicated to participants on the same day; three participants received their
result the next day and one received a result 7 days later. Laboratory-based viral load results
were communicated at participants’ next clinical visits, which were a median of 28 days
(IQR 28–54) after the blood draw.
Author Manuscript

During the study, seven (4%) participants in the intervention group and nine (5%) in the
standard-of-care group developed viral failure, defined as two consecutive viral loads of
more than 1000 copies per mL. Among these, the intervention improved the median time to
detection of viral failure from 123 days (IQR 98–162) to 55 days (55–57). Of those with
viral failure, six (100%) participants in the intervention group were appropriately switched
to second-line ART; three participants were switched on the same day, and two participants
were switched within 1 week. Among nine participants with viral failure in the standard-of-
care group, four (44%) were appropriately switched to second-line ART, a median of 76
days (IQR 20–134) after the second confirmatory viral load blood draw.

At study end, five (2·6%) participants in the intervention group and seven (3·6%)
participants in the standard-of-care group had presence of any HIV drug resistance
Author Manuscript

mutations (data not shown). Overall, during the course of the study participants in the
intervention group had fewer total clinic visits and more visits with an enrolled nurse
(p<0·0001; appendix p 3).

Patients were eligible for referral into the community ART delivery programme if they met
the criteria, which included a suppressed viral load after 6 months of study enrolment, which
was 12 months after ART initiation. Overall, 116 (56%) participants in the intervention
group and 52 (27%) in the standard-of-care group were referred into the community-based

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 9

ART delivery programme, an increase of 32·8% (95% CI 23·2–41·6; table 2). Among those
Author Manuscript

referred into the community-based ART delivery programme, the time to referral for the
intervention group was 168 days (IQR 168–175) and for the standard-of-care group was 261
days (231–28; table 3). Overall, participants in the intervention group had a 3·5-fold (95%
CI 2·5–4·8) higher probability of referral into the community-based ART delivery
programme than did participants in the standard-of-care group (figure 2).

Discussion
In this randomised controlled trial that followed the South African guidelines for HIV
treatment and monitoring, point-of-care HIV viral load testing combined with task shifting
to enrolled nurses improved viral suppression and retention in care for HIV-positive adults
receiving ART. These improvements were achieved in part by ensuring rapid turnaround of
viral load results, which facilitated adherence counselling the same day and led to a faster
Author Manuscript

identification of viral failure and a subsequent switch to second-line ART. Viral failure was
identified significantly earlier in the intervention group due to the cumulative delays for
acting upon both the original and confirmatory laboratory viral load tests in the standard-of-
care group. Point-of-care HIV viral load testing and task shifting allowed earlier
identification of participants with viral suppression on the same day that their viral load was
taken. This same-day testing meant that patients could be rapidly referred into differentiated
care through the community-based ART delivery programme. In interviews and focus
groups, participants and health-care workers from our study reported that point-of-care
testing and taskshifting was acceptable (data not shown). In settings where patient follow-up
and HIV treatment outcomes remain suboptimal, use of point-of-care viral load testing with
task shifting could accelerate reaching the global UNAIDS target of 90% of people on ART
achieving viral suppression by 2020.
Author Manuscript

Additional clinical trials are evaluating point-of-care HIV viral load testing, but have not yet
reported results. Studies in Haiti (NCT03288246) and Uganda (NCT03553693) are
evaluating HIV testing and outcomes in children, adolescents, and adults. Another South
African study (NCT03187964) is comparing point-of-care to laboratory-based testing for the
difference in the time to adherence counselling and resistance testing. A clinical trial
(NCT03533868) of HIV-positive adults in Nigeria is evaluating on-site Xpert HIV-1 viral
load testing for reaching viral suppression (<1000 copies per mL) at 1 year after ART
initiation. Results from most of these trials are expected later this year.

We used the Cepheid Xpert HIV-1 viral load test in our study, which has high diagnostic
accuracy,24 and there are other point-of-care HIV viral load platforms.25,26 The SAMBA I/II
semi-Q (Diagnostics for the Real World, Cambridge, UK) can be done in 90 min on plasma
Author Manuscript

or whole blood and has a lower limit of quantification of 1000 copies per mL.27 The m-
PIMA, formerly Alere Q NAT, (Abbott, Chicago, IL, USA) can complete testing in 70 min
on plasma, and is currently the only point-of-care viral load test for both HIV-1 and HIV-2.28

Our study had several strengths and limitations. Point-of-care viral load testing and
associated clinical care was done by the research team, which might not always reflect
implementation in routine clinical services. Study reimbursements could have increased

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 10

retention in care, but we provided the same incentives for participants in both study groups
Author Manuscript

to minimise differences. The study was designed as an open-label clinical trial, and research
implementers were blinded to data by study groups. Study enrolment occurred 6 months
after ART initiation to coincide with the first viral load test in the routine South African
ART monitoring schedule. Therefore, our results might not be generalisable to all patients
initiating ART because they do not account for people who default early in their HIV care.
These people will require additional interventions to ensure retention until they are due their
first viral load test. Because our intervention contained two components, we cannot be
certain to what extent point-of-care testing or task shifting to enrolled nurses was
responsible for the improved clinical outcomes. Finally, the study was done at a single
clinical site serving an urban population of HIV-positive adults in South Africa, and
additional studies might be necessary to determine achievable outcomes in other clinics or
countries. Further assessment of the effect on quality of life, wider health-systems costs, and
Author Manuscript

cost-effectiveness is also warranted.

In conclusion, point-of-care HIV viral load testing combined with task shifting improved
treatment outcomes in an urban clinic in South Africa. To accommodate the rapid expansion
of ART, efficient methods to monitor HIV-positive people on ART will be required in
resource-limited settings. These innovative models of HIV care might be sustainable only if
they do not increase the existing burden on HIV care providers, laboratories, and health
systems. By improving and expediting delivery of viral load results to patients and their care
providers, HIV treatment programmes could use point-of-care testing for reducing barriers
to achieve viral suppression.

Supplementary Material
Author Manuscript

Refer to Web version on PubMed Central for supplementary material.

Acknowledgments
We thank all the people who participated in this study. This research was supported by the National Institute of
Allergy and Infectious Diseases (grant R21AI124719), the Center for AIDS Research at the University of
Washington (grant P30 AI027757), and the South African National Health Laboratory Service Research Trust
(grant 2018–1DEV-PMO01).

PKD reports receiving consulting and speaking fees from Gilead Science; and research support from the National
Institute of Health, Centers for Disease Control and Prevention, Gilead Sciences, and the Bill & Melinda Gates
Foundation, during the conduct of the study. KN and SSAK report grants from the National Institute of Health
during the conduct of the study. DD reports grants from USAID, outside of the conduct of this study; and grants
from the National Institute of Health during the conduct of the study.

References
Author Manuscript

1. Joint United Nations Programme on HIV/AIDS. Miles to go: closing gaps, breaking barriers,
righting injustices. 2018 http://www.unaids.org/sites/default/files/media_asset/miles-to-go_en.pdf
(accessed May 30, 2019).
2. Joint United Nations Programme on HIV/AIDS. 90–90–90. An ambitious treatment target to help
end the AIDS epidemic. 2014 http://www.unaids.org/sites/default/files/media_asset/90–90–
90_en.pdf (accessed May 30, 2019).

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 11

3. WHO. Interim technical update: technical and operational considerations for implementing HIV
viral load testing. July 2014 https://www.who.int/hiv/pub/arv/viral-load-testing-technical-update/en/
Author Manuscript

(accessed Jan 29, 2020).

4. Grimsrud A, Barnabas RV, Ehrenkranz P, Ford N. Evidence for scale up: the differentiated care
research agenda. J Int AIDS Soc 2017; 20: 22024. [PubMed: 28770588]
5. Habiyambere V, Dongmo Nguimfack B, Vojnov L, et al. Forecasting the global demand for HIV
monitoring and diagnostic tests: a 2016–2021 analysis. PLoS One 2018; 13: e0201341. [PubMed:
30231022]
6. United States Department of Health and Human Services. Guidelines for the use of antiretroviral
agents in adults and adolescents living with HIV. Department of Health and Human Services, 2019.
2019 http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf (accessed Jan 29,
2020).
7. Peter T, Zeh C, Katz Z, et al. Scaling up HIV viral load—lessons from the large-scale
implementation of HIV early infant diagnosis and CD4 testing J Int AIDS Soc 2017; 20: 9–15.
8. Lecher S, Ellenberger D, Kim AA, et al. Scale-up of HIV viral load monitoring—seven sub-Saharan
African countries. MMWR Morb Mortal Wkly Rep 2015; 64: 1287–90. [PubMed: 26605986]
Author Manuscript

9. Murphy RA, Court R, Maartens G, Sunpath H. Second-line antiretroviral therapy in sub-Saharan

Africa: it’s time to mind the gaps. AIDS Res Hum Retroviruses 2017; 33: 1181–84. [PubMed:
28793781]
10. Drain P, Hyle E, Noubary F, et al. Evaluating diagnostic point-of-care tests in resource-limited
settings. Lancet Infect Dis 2014; 14: 239–49. [PubMed: 24332389]
11. Cepheid. Monitoring for quality of life, Xpert HIV-1 viral load. 2015 http://www.cepheid.com/
administrator/components/com_productcatalog/library-files/
fadd5b909331644ab1415fe39ec94dd5-Xpert-HIV-1-Viral-Load-Brochure-CEIVD-3039–03.pdf
(accessed May 30, 2019).
12. epheid. Cepheid and FIND announce European approval of Xpert HIV-1 Viral Load. 2014 http://
ir.cepheid.com/releasedetail.cfm?releaseid=888964 (accessed May 30, 2019).
13. Garrett NJ, Drain PK, Werner L, Samsunder N, Abdool Karim SS. Diagnostic accuracy of the
point-of-care Xpert HIV-1 viral load assay in a South African HIV clinic. J Acquir Immune Defic
Syndr 2016; 72: e45–48. [PubMed: 26959192]
Author Manuscript

14. Dorward J, Garrett N, Quame-Amaglo J, et al. Protocol for a randomised controlled

implementation trial of point-of-care viral load testing and task-shifting: the Simplifying HIV
TREAtment and Monitoring (STREAM) study. BMJ Open 2017; 7: e017507.
15. Peltzer K, Simbayi L, Kalichman S, et al. Alcohol use in three different inner cities in South
Africa: AUDIT-C and CAGE. J Psychol Africa 2007; 17: 99–104.
16. Bernstein M, Phillips T, Zerbe A, et al. Intimate partner violence experienced by HIV-infected
pregnant women in South Africa: a cross-sectional study. BMJ Open 2016; 6: e011999.
17. Bhana A, Rathod SD, Selohilwe O, et al. The validity of the patient health questionnaire for
screening depression in chronic care patients in primary health care in South Africa. BMC
Psychiatry 2015; 15: 118. [PubMed: 26001915]
18. South Africa Department of Health. National consolidated guidelines for the prevention of mother-
to-child transmission of HIV (PMTCT) and the management of HIV in children, adolescents, and
adults. 2015 https://sahivsoc.org/Files/HIV%20clinicians%20ADULT%20GUIDELINES
%202017%20-%20David%20Stead.pdf (accessed May 30, 2019).
19. South African National Health Research Ethics Council. Payment of trial participants in South
Author Manuscript

Africa: ethical considerations for Research Ethics Committees (RECs). 2012 www.nhrec.org.za/
index.php/grids-preview?download=11:guidelines-for-payment (accessed May 30, 2019).
20. Georgeu D, Colvin CJ, Lewin S, et al. Implementing nurse-initiated and managed antiretroviral
treatment (NIMART) in South Africa: a qualitative process evaluation of the STRETCH trial.
Implement Sci 2012; 7: 66. [PubMed: 22800379]
21. Roberts P. Centralised chronic medicine dispensing and distribution: a public/private partnership to
increase access to HIV/chronic medication. 2018 http://files.icap.columbia.edu/files/uploads/
ICAP_Grand_Rounds_Project_Last_Mile_Slides.pdf (accessed Jan 29, 2020).

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 12

22. Wilson EB. Probable Inference, the Law of succession, and statistical inference. JASA 1927; 22:
209–12.
Author Manuscript

23. Newcombe RG. Interval estimation for the difference between independent proportions:
comparison of eleven methods. Statistics in Medicine 1998; 17: 873–90. [PubMed: 9595617]
24. Nash M, Huddart S, Badar S, Baliga S, Saravu K, Pai M. Performance of the Xpert HIV-1 viral
load assay: a systematic review and meta-analysis. J Clin Microbiol 2018; 56: 1–8.
25. Joint United Nations Programme on HIV/AIDS. HIV/AIDS diagnostics technology landscape;
semi-annual update 2015. 2015 https://aidsfree.usaid.gov/sites/default/files/hivaids_diag_tech.pdf
(accessed Jan 29, 2020).
26. Drain PK, Dorward J, Garrett N, et al. Point-of-care HIV viral load testing: An essential tool for a
sustainable global HIV/AIDS response. Clin Microbiol Rev 2019; 32: e00097–18. [PubMed:
31092508]
27. Goel N, Ritchie AV, Mtapuri-Zinyowera S, et al. Performance of the SAMBA I and II HIV-1 Semi-
Q Tests for viral load monitoring at the point-of-care. J Virol Methods 2017; 244: 39–45.
[PubMed: 28274744]
28. Jani IV, Meggi B, Vubil A, et al. Evaluation of the whole-blood Alere Q-nat point-of-care RNA
Author Manuscript

assay for HIV-1 viral load monitoring in a primary health care setting in Mozambique. J Clin
Microbiol 2016; 54: 2104–08. [PubMed: 27252459]
Author Manuscript
Author Manuscript

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 13

Research in context
Author Manuscript

Evidence before this study

On May 30, 2019, we searched PubMed for studies assessing the outcomes of point-of-
care HIV viral load testing in people living with HIV and receiving antiretroviral therapy
(ART). We used the search terms “point-of-care,” “viral load,” and “HIV,” without
language restrictions, and did not find any randomised controlled trials or implementation
science studies. Therefore, we are not aware of any published data to evaluate the clinical
effectiveness or effect of point-of-care HIV viral load testing, when compared with
standard laboratory viral load testing, for improving treatment monitoring for HIV-
positive adults on ART.

Added value of this study

Author Manuscript

To our knowledge, this is the first randomised controlled trial to evaluate the effect of
point-of-care HIV viral load monitoring for people living with HIV receiving ART.
Overall, implementation of point-of-care HIV viral load testing, following the
recommended schedule for testing in South Africa, significantly increased the outcome of
retained in care with viral suppression. Additionally, the intervention substantially
accelerated switching to second-line ART for those with treatment failure (sustained viral
load >1000 copies per mL), as well as referral into a community-based ART delivery
programme.

Implications of all the available evidence

Our results suggest that ensuring rapid receipt of viral load results to patients and their
providers can improve HIV viral suppression and retention in care, lead to a faster switch
Author Manuscript

to second-line ART for those with viral failure, and improve referral into community-
based ART delivery programmes in resource-limited settings.
Author Manuscript

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 14
Author Manuscript
Author Manuscript

Figure 1:
Trial profile
Author Manuscript
Author Manuscript

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 15
Author Manuscript
Author Manuscript

Figure 2:
Kaplan-Meier estimates of time from study enrolment to referral into a community-based
ART delivery programme by intervention group and standard-of-care group
ART=antiretroviral therapy.
Author Manuscript
Author Manuscript

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 16

Table 1:

Baseline characteristics
Author Manuscript

Intervention group (n=195) Standard-of-care group (n=195)

Median (IQR) age (years) 31 (27–38) 32 (27–38)
Sex
Female 119 (61%) 116 (60%)
Male 76 (39%) 79 (40%)
Ethnicity
Black 193 (99%) 193 (99%)
Other 2 (1%) 2 (1%)
Highest level of education
None or primary school 10 (5%) 18 (9%)
Did not pass secondary school 90 (46%) 75 (39%)
Author Manuscript

Passed secondary school 64 (33%) 69 (35%)

Tertiary education 31 (16%) 33 (17%)
Monthly income
R <1000 (~US$70) 83 (43%) 79 (41%)
R 1000–4000 (~$70–280) 80 (41%) 63 (32%)
R >4000 (~$280) 29 (15%) 46 (24%)
Did not disclose 3 (2%) 7 (4%)
Has a regular or stable partner 155 (80%) 156 (80%)
Children
None 41 (21%) 33 (17%)
≥1 child 154 (79%) 162 (83%)
Primary method of travel to clinic
Author Manuscript

Walking 15 (8%) 16 (8%)

Public transportation 173 (89%) 172 (88%)
Private transportation 7 (4%) 7 (4%)
Distance travelled to clinic
≤5000 m 40 (21%) 46 (24)
≤60 min 179 (92%) 180 (92%)
Psychosocial and health behaviours

Positive depression screen* 19 (10%) 18 (9%)

Intimate partner violence in past year (women only) 20/119 (17%) 20/116 (17%)
AUDIT C score ≥4 (men only) 33/76 (43%) 24/79 (30%)
AUDIT C score ≥3 (women only) 22/119 (19%) 17/116 (15%)
Current smoker 32 (16%) 32 (16%)
Author Manuscript

Recreational drug use in past 6 months 15 (8%) 13 (7%)

Medical history, HIV care, and self-reported ART adherence
>12 months since HIV diagnosis 62 (32%) 66 (34%)
ART exposure before initiation of current regimen 8 (4%) 10 (5%)
Mean (SD) time from ART initiation to study enrolment (days) 183·8 (17·3) 183·1 (17·1)

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Drain et al. Page 17

Intervention group (n=195) Standard-of-care group (n=195)

ART regimen with tenofovir disoproxil fumarate, emtricitabine, and 192 (99%) 195 (100%)
efavirenz
Author Manuscript

Received previous treatment for tuberculosis 26 (13%) 34 (17%)

Self-report of one or more doses missed in past 4 days 43 (22%) 31 (16%)
Clinical and laboratory parameters
WHO clinical stage
Stage I 188 (96·4) 193 (99·0)
Stage II 7 (4%) 2 (1%)
Mean (SD) BMI (kg/m²) 25·6 (6·0) 26·0 (6·0)
Mean (SD) haemoglobin (g/L) 128·0 (17·0) 130·0 (15·0)

Mean (SD) creatinine clearance (mL/min†) 99·3 (29·0) 100·0 (27·2)

Hepatitis B surface antigen positive 12 (6·2) 11 (5·6)

Median (IQR) CD4 count at ART initiation (cells per L) 361 (171–546) 371 (230–551)
Author Manuscript

Median (IQR) CD4 count at enrolment (cells per µL‡) 462 (275–658) 473 (340–680)

HIV viral load <200 copies per mL at enrolment§ 179 (92%) 184 (94%)

HIV viral load >1000 copies per mL at enrolment§ 10 (5%) 8 (4%)

Presence of any HIV drug resistance at enrolment 7 (4%) 7 (3%)

Data are n (%), unless otherwise stated. AUDIT-C=Alcohol Use Disorder Identification Tool. ART=antiretroviral therapy. BMI=body-mass index.
R=South African rand.
*
Defined as Patient Health Questionnaire-2 score of 2 or greater.
†
Missing five observations for creatinine clearance (intervention group).
‡
Missing one observation for CD4 cell count at enrolment (standard-of-care group).
§
Viral loads at enrolment measured per protocol by study group.
Author Manuscript
Author Manuscript

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 2:

Primary and secondary outcomes

Intervention group (n=195) Standard-of-care group (n=195) Absolute risk difference, % (95% CI) p value
Drain et al.

Composite primary outcome 175 (90%) 148 (76%) 13·9% (6·4 to 21·2) 0·00040
Components of primary outcome
Viral suppression (<200 copies per mL) 182 (93%) 162 (83%) 10·3% (3·9 to 16·8) 0·0025
Retention in care at study clinic 180 (92%) 165 (85%) 7·7% (1·3 to 14·2) 0·026
Other secondary outcomes
Viral suppression (<200 copies per mL) and retention at any clinic 177 (91%) 153 (79%) 12·3% (5·2 to 19·4) 0·0011
Viral load <1000 copies per mL and retention at study clinic 176 (90%) 152 (78%) 12·3% (5·1 to 19·5) 0·0013
Viral load <50 copies per mL and retention at study clinic 167 (86%) 139 (71%) 14·4% (6·2 to 22·3) 0·0080
Viral load <1000 copies per mL 184 (94%) 166 (85%) 9·2% (3·2 to 15·4) 0·0041
Viral load <50 copies per mL 172 (88%) 152 (78%) 10·3% (2·8 to 17·6) 0·0099
Retention in care at any clinic at 12 months 182 (93%) 170 (87%) 6·2% (0·2 to 12·2) 0·059
Viral suppression among those with a result at study exit
Viral load <1000 copies per ml 184/189 (97%) 166/178 (93%) 4·1% (–0·3 to 9·0) 0·082
Viral load <200 copies per ml 182/189 (96%) 162/178 (91%) 5·3% (0·2 to 10·7) 0·051
Viral load <50 copies per ml 172/189 (91%) 152/178 (85%) 5·6% (–1·0 to 12·4) 0·11
Communication of viral load result to participant
Enrolment viral load communicated to participant 195/195 (100%) 171/192 (89%) 10·9% (6·8 to 16·1) <0·0001
Mid-study viral load communicated to participant 191/192 (>99%) 127/170 (75%) 24·8% (18·4 to 31·8) <0·0001
All viral load results communicated to participant 397/398 (>99%) 304/373 (82%) 18·3% (14·5 to 22·5) <0·0001

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

Referral into community-based ART delivery programme 116 (60%) 52 (27%) 32·8% (23·2 to 41·6) <0·0001

Data are n (%) or n/N (%), unless otherwise stated. ART=antiretroviral therapy.
Page 18
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 3:

Time to event results for secondary outcomes

Intervention group Standard-of-care group Hazard ratio (95% CI) p value

Drain et al.

n/N (%) Median, days (IQR) n/N (%) Median, days (IQR)

Time to entry of viral load result into health-information system*

Enrolment blood draw 195/195 (100%) 0 (0–0) 192/192 (100%) 3 (2–4) 96·0 (23·8–386·4) <0·0001
Mid-study blood draw 192/192 (100%) 0 (0–0) 170/170 (100%) 2 (1–3) 11·6 (7·8–17·3) <0·0001
Any blood draw 398/398 (100%) 0 (0–0) 373/373 (100%) 2 (1–4) 11·7 (8·9–15·3) <0·0001
Time to communication of viral load result to participant
Enrolment blood draw 195/195 (100%) 0 (0–0) 171/192 (89%) 28 (28–30) ··† <0·0001†
Mid-study blood draw 191/192 (>99%) 0 (0–0) 127/170 (75%) 41 (28–69) 12·7 (8·5–19·1) <0·0001
Any blood draw 397/398 (>99%) 0 (0–0) 304/373 (34%) 28 (28–54) 17·7 (13·0–24·2) <0·0001
Follow-up HIV treatment and care
Enrolment to detection of viral failure 7/195 (4%) 55 (55–57) 9/195 (5%) 123 (98–162) 0·8 (0·3–2·2) 0·69

Appropriate switch to second-line ART after viral failure‡ 6/6 (100%) 0 (0–7) 4/9 (44%) 76 (20–134) 10·9 (2·1–57·5) 0·0050

Enrolment to referral into community-based ART delivery program 116/195 (59%) 168 (168–175) 52/195 (27%) 261 (231–281) 3·5 (2·5–4·8) <0·0001

ART=antiretroviral therapy.
*
During the study, 2·9% of point-of-care viral load tests had invalid results, and all were repeated on the same day, whereas 1·1% of laboratory-based viral load tests had invalid results and were repeated.
†
Data excludes people evaluated for viral load after the 12-month study exit visit.
‡
Hazard ratio could not be calculated for time to communication of viral load results to participant for the enrolment blood draw because the events are separated by group: in the intervention group, all

Lancet HIV. Author manuscript; available in PMC 2021 April 01.

events were on day 0 or within 1 day of the blood draw, whereas for control participant events the earliest event was on day 12. The p value from a log rank test is provided instead.
Page 19