Estimating potential silent transfer using baseline viral load measures among people presenting as new to HIV care in Lusaka, Zambia: a cross-sectional study

Abstract

Objectives

To estimate potential silent transfer using baseline viral load measures among individuals presenting as new to HIV care in routine HIV clinical settings in Lusaka, Zambia.

Design

Cross-sectional study.

Setting

Two large, urban government-operated health facilities supported by the Centre for Infectious Disease Research in Zambia.

Participants

A total of 248 participants with an incident positive HIV rapid test.

Outcome measures

The primary outcome measure was HIV viral suppression at baseline (i.e., potential silent transfer), defined as having a viral load ≤1000 RNA copies(c)/mL at the time of initiating HIV care. We also examined viral suppression at ≤60 c/mL.

Methods

We surveyed and measured baseline HIV viral load as part of the national recent infection testing algorithm among people living with HIV (PLWH) presenting as new to care. Using mixed effects Poisson regression, we identified characteristics among PLWH associated with potential silent transfer.

Results

Among the 248 PLWH included, 63% were women with median age of 30, and 66 (27% (66/248)) had viral suppression at ≤1000 c/mL and 53 (21% (53/248)) at ≤60 c/mL thresholds, respectively. Participants aged 40+ years had a significantly higher adjusted prevalence of potential silent transfer (adjusted prevalence ratio (aPR): 2.10; 95% CI: 2.08, 2.13) compared with participants aged 18–24 years. Participants reporting no formal education had a significantly higher adjusted prevalence of potential silent transfer (aPR: 1.63; 95% CI: 1.52, 1.75) compared with those completing primary education. Among 57 potential silent transfers who completed a survey, 44 (77%) indicated having tested positive previously at ≥1 of 38 clinics in Zambia.

Conclusions

The high proportion of PLWH with potential silent transfer points to clinic shopping and/or co-enrolment at multiple care sites simultaneously, suggesting an opportunity to improve care continuity at the time of HIV care entry.

Correspondence to Dr Jake M Pry; [email protected]

Strengths and limitations of this study

• We present a method to identify and characterise potential silent transfer at initial presentation to care using baseline viral load measures, which could be used to improve HIV care continuity.

• We used a pragmatic operational definition of silent transfer that likely underestimated the true prevalence of silent transfer in the study population.

• We were unable to definitively establish silent transfer through participant linked medical record information or analysis of antiretroviral metabolites.

Introduction

While HIV treatment is now widely available in sub-Saharan Africa (SSA), challenges remain in monitoring people living with HIV (PLWH) across the HIV care continuum. As HIV testing services and antiretroviral therapy (ART) have scaled up via test and treat programmes in many countries, PLWH are also now healthier and more mobile, making movement between health facilities easier. In an HIV care landscape with multiple providers and service access points, and few national unique identifier systems able to reliably monitor PLWH across clinics, it has become increasingly difficult to follow individuals receiving HIV care as they change or access different treatment sites, particularly those who do so without officially notifying the health system. These so-called ‘silent transfers’, ‘self-transfers’ or ‘unofficial transfers’ have been occurring with increasing frequency in SSA, may have unique characteristics and health service needs and may obscure estimates of the number of PLWH ‘in care’ at a site at any given time.^1–6

Previous research has identified predictors of HIV care facility transfer, such as lack of transportation to the clinic and other structural barriers,⁷ and has shown PLWH experiencing silent transfer to be less likely to be in care than those who transfer officially at 1-year follow-up.⁸ In Zambia, we have observed increased time to re-engagement in care among silent transfers compared with official transfers.⁹ Although poorly described, there is also likely heterogeneity in the silent transfer population, which may include PLWH re-engaging in official HIV care through the ‘side door’ to the cascade, PLWH ‘shopping’ for a clinic to first establish care and PLWH already established in care looking to change their treatment site.^10–13

Using plasma viral load measures collected on individuals with an incident HIV positive test, we estimate potential silent transfer at two health facilities in Lusaka, Zambia. We hypothesised that individuals self-reporting unknown HIV status upon presentation for HIV testing and found to have a suppressed viral load at that same visit would have a high probability of being a PLWH silently transferring care to that facility. In this paper, we estimate the proportion of PLWH with viral suppression at baseline and identify predictors of potential silent transfer in a routine HIV care setting.^{14 15}

Methods

Design

We conducted a cross-sectional substudy of a parent implementation research study called REACHZ active from 20 May 2021 to 10 March 2022. REACHZ focused on integrating testing with a rapid test for recent infection (RTRI) into index testing services at the point-of-care for persons with an incident positive HIV rapid antibody test. To this end, REACHZ sought to improve the efficiency and precision of partner elicitation and contact tracing by prioritising index testing clients with laboratory-confirmed recent HIV infection for more thorough assisted partner notification counselling. As part of REACHZ enrolment procedures, HIV-1/2 viral load was measured at baseline to enable assessment of recent infection. These procedures followed the national recent infection testing algorithm developed to support HIV surveillance in Zambia through the Tracking with Recency Assays to Control the Epidemic (TRACE) initiative.¹⁶ REACHZ participants were recruited from two Ministry of Health (MOH) health facilities in Lusaka District, Zambia that formed our study sites and which were supported by the Centre for Infectious Disease Research in Zambia (CIDRZ) with funding from the US President’s Emergency Plan for AIDS Relief (PEPFAR) through the United States Centers for Disease Control and Prevention.

Population

Eligibility criteria for the parent REACHZ study included: (1) ≥18 years old; (2) diagnosed with HIV infection using the national HIV testing algorithm; (3) consented to receive routine MOH index testing services; (4) willing and able to provide study written informed consent; and (5) willing to provide study locator information (e.g., working phone number).¹⁷ For this analysis, we included all participants from the REACHZ study who underwent baseline viral load testing, including all participants with RTRI recent infection from study inception and all study participants regardless of RTRI status from 1 September 2021 when universal baseline viral load testing became available through the study. Leveraging REACHZ study socio-demographic, clinical and laboratory data, for this analysis, we included REACHZ participants with a documented baseline viral load result to form the study population.

Study recruitment

After completing routine HIV testing at a study site, individuals with a positive rapid antibody test result and who consented to undergo routine index testing services offered as a part of HIV care enrolment procedures were referred on the same day to on-site REACHZ study staff for study screening. After verifying eligibility, these individuals were invited to enrol in the study. Following study informed consent, participants completed enrolment procedures, including RTRI testing, specimen collection for confirmatory RTRI and viral load testing at the CIDRZ central laboratory and an enrolment survey. A subgroup of participants found to have a suppressed viral load (i.e., ≤1000 copies(c)/mL) completed a brief survey at their first study follow-up visit to ascertain reasons for potential silent transfer.

Routine HIV testing procedures

The Zambian national HIV diagnosis algorithm requires two positive rapid antibody tests performed serially on blood drawn by fingerstick. The first test, used for screening purposes, is the Alere Determine (Abbott, Waltham, Massachusetts, USA) followed by a second, confirmatory, SD Bioline (Standard Diagnostics, Suwon, South Korea) test. Both tests are third generation rapid tests that detect the presence of anti-HIV-1/2 antibodies. Per national guidelines, routine baseline HIV-1/2 viral load testing was made available to all newly HIV-diagnosed individuals at the time of their initial HIV care evaluation.

Study HIV testing procedures

For consenting study participants, testing for recent HIV infection was done using the Asanté (Sedia Biosciences, Beaverton, Oregon, USA) RTRI, performed on whole blood from a fingerstick by trained healthcare workers at the site and repeated by trained laboratory technicians at a central laboratory. Recent infection was confirmed at the CIDRZ central laboratory with the addition of a plasma RNA HIV-1/2 viral load performed with an assay detection limit of 60 copies of viral RNA/mL (Aptima HIV-1 Quant Assay, Hologic Panther, Santa Fe Springs, California, USA).

Patient and public involvement

For the analysis presented here, we aimed to describe factors associated with baseline viral suppression to gain insight into potential silent transfer and opportunities to improve patient experiences with continuity of care at HIV care entry. While public feedback on the specific data collection tools was not obtained, public input on study design and implementation was obtained from the National Health Research Authority of the Zambian MOH and various constituencies at the study sites, including MOH health workers and community representatives. Final dissemination to participants and stakeholders of parent study results, including for the analysis presented here, are being planned and will be done at the conclusion of the primary analysis.

Data collection

Study data were collected via hard copy forms by trained study research assistants and subsequently entered into District Health Information Software 2 (DHIS2) database on a secure CIDRZ server. Facility level data on the number of incident HIV positive rapid tests and PLWH consenting to routine index testing services during the study period were obtained from the PEPFAR Data for Accountability Transparency and Impact (DATIM) platform. Survey tool development was informed by previous silent transfer research.^{9 10 13 18 19}

Outcomes

We defined our primary outcome, baseline HIV viral suppression, as ≤1000 copies/mL of HIV RNA in plasma at the time of enrolment in the parent study, regardless of the recency status of their HIV infection. We considered any participant in the parent study with suppressed baseline viral load to be a ‘potential silent transfer’.²⁰ As part of a sensitivity analysis, we used an alternative definition for viral suppression of ≤60 RNA copies/mL based on the lower limit of detection for the testing platform.

Analysis

We describe the study population by viral suppression status (i.e., suppressed, ≤1000 copies/mL vs unsuppressed, >1000 copies/mL), comparing patient characteristics by suppression status using the t-test for continuous variables and χ² test for binary and categorical variables (significance attributed at alpha <0.05). Confounding variables considered were based on the extant literature and the minimal adjustment set was identified by directed acyclic graph. Next, we used mixed effects Poisson regression to estimate adjusted prevalence ratios for potential silent transfer at two viral suppression thresholds (≤1000 copies/mL and ≤60 copies/mL). Finally, we obtained marginal probability estimates from our adjusted Poisson regression model by age category and sex to illustrate the association between viral suppression (≤1000 copies/mL) and key demographic characteristics. All analyses were conducted using Stata V.17 (StataCorp, College Station, Texas, USA).

Results

During the parent study period (ie, 20 May 2021 to 10 March 2022), approximately 1224 persons at the study sites had an incident positive HIV rapid test and consented to routine index testing services. Of these, 393 (32%) were screened for study participation, and of these 344 (88%) met study eligibility criteria. Of those eligible, 321 (93%) provided informed consent and completed REACHZ enrolment procedures.

A total of 248 (77%) REACHZ participants had a documented baseline viral load result and were thus eligible for this analysis. Of these, 66 (27%) were noted to have a suppressed viral load. Most of these participants were women (63%) and had a median age of 30 years (IQR: 25–37) (table 1).

Table 1

Population characteristics by viral suppression status (unsuppressed, >1000 copies/mL and suppressed, ≤1000 copies/mL) (N=248)

Characteristic	Total n (%)	Unsuppressed n (%)	Suppressed n (%)	P value
N	N=248	N=182	N=66
Sex of the participant				0.21
Female	157 (63.3)	111 (61.0)	46 (69.7)
Male	91 (36.7)	71 (39.0)	20 (30.3)
Age, median (IQR)	30 (25–37)	29.5 (24–36)	32 (27–38)	0.24
Age category				0.47
18–24 years	57 (23.0)	46 (25.3)	11 (16.7)
25–34 years	107 (43.1)	78 (42.9)	29 (43.9)
35–44 years	69 (27.8)	47 (25.8)	22 (33.3)
45+ years	15 (6.0)	11 (6.0)	4 (6.1)
Marital status				0.67
Single	85 (34.3)	64 (35.2)	21 (31.8)
Married	116 (46.8)	83 (45.6)	33 (50.0)
Divorced/separated	38 (15.3)	27 (14.8)	11 (16.7)
Widowed	9 (3.6)	8 (4.4)	1 (1.5)
Educational attainment				0.42
No formal education	4 (1.6)	2 (1.1)	2 (3.0)
Primary	59 (23.8)	40 (22.0)	19 (28.8)
Secondary	169 (68.1)	127 (69.8)	42 (63.6)
College/university	16 (6.5)	13 (7.1)	3 (4.5)
Employment status				0.93
Not working/employed	82 (33.1)	60 (33.0)	22 (33.3)
Employed/working	52 (21.0)	37 (20.3)	15 (22.7)
Self-employed	100 (40.3)	74 (40.7)	26 (39.4)
Student	8 (3.2)	6 (3.3)	2 (3.0)
Retired	3 (1.2)	3 (1.6)	0 (0.0)
Other	3 (1.2)	2 (1.1)	1 (1.5)
Have mobile phone				0.52
No	39 (15.7)	27 (14.8)	12 (18.2)
Yes	209 (84.3)	155 (85.2)	54 (81.8)
Tested and received previous HIV test results				0.01
No	36 (14.5)	29 (15.9)	7 (10.6)
Yes	209 (84.3)	153 (84.1)	56 (84.8)
Do not know	3 (1.2)	0 (0.0)	3 (4.5)
Consume alcohol				0.99
No	128 (51.6)	94 (51.6)	34 (51.5)
Yes	120 (48.4)	88 (48.4)	32 (48.5)
Used condom at last sex				0.49
No	179 (72.2)	128 (70.3)	51 (77.3)
Yes	68 (27.4)	53 (29.1)	15 (22.7)
Do not know	1 (0.4)	1 (0.5)	0 (0.0)
Health facility				0.001
Clinic 1	81 (32.7)	49 (26.9)	32 (48.5)
Clinic 2	167 (67.3)	133 (73.1)	34 (51.5)
Month of study enrolment				<0.001
June 2021	8 (3.2)	2 (1.1)	6 (9.1)
July 2021	0 (0.0)	0 (0.0)	0 (0.0)
August 2021	11 (4.4)	3 (1.6)	8 (12.1)
September 2021	66 (26.6)	47 (25.8)	19 (28.8)
October 2021	35 (14.1)	32 (17.6)	3 (4.5)
November 2021	21 (8.5)	18 (9.9)	3 (4.5)
December 2021	12 (4.8)	10 (5.5)	2 (3.0)
January 2022	48 (19.4)	36 (19.8)	12 (18.2)
February 2022	47 (19.0)	34 (18.7)	13 (19.7)

Note: Suppressed ≤1000 copies RNA/mL; no July 2021 enrolments due to COVID-19 lockdown in Zambia.

ART, antiretroviral therapy; IQR, Interquartile range.

We observed a significant difference in adjusted prevalence ratios (aPR) for baseline viral load suppression by age, sex, marital status, study facility and month of study enrolment (table 2). Older age groups, specifically PLWH aged 30–34 years (aPR: 1.84; 95% CI: 1.48, 2.29) and PLWH aged 40+ years (aPR: 2.10; 95% CI: 2.08, 2.13), had significantly higher adjusted prevalence of suppressed baseline viral load compared with clients aged 18–24 years. Though not statistically significant, women had a higher adjusted prevalence of being suppressed (≤1000 copies/mL) at presentation to care compared with men (aPR: 1.41; 95% CI: 0.99, 2.02). Clinic 1, the slightly smaller of the two study clinics, and one at a lower level of the health system, had a significantly higher adjusted prevalence of participants with baseline viral suppression compared with Clinic 2 (aPR: 1.90; 95% CI: 1.79, 2.02) (table 2).

Table 2

Adjusted prevalence ratio results for viral suppression by threshold (N=248)

Covariate	≤1000 copies/mL (n=66)		≤60 copies/mL (n=53)
	aPR	95% CI	aPR	95% CI
Sex
Male	1.00 (ref)	Ref	1.00 (ref)	Ref
Female	1.41	(0.99, 2.02)	1.55	(0.97, 2.47)
Age (category)
18–24 years	1.00 (ref)	Ref	1.00 (ref)	Ref
25–29 years	1.25	(0.51, 3.08)	1.03	(0.52, 2.04)
30–34 years	1.84	(1.48, 2.29)	1.99	(1.91, 2.07)
35–38 years	1.77	(0.86, 3.63)	1.93	(0.64, 5.85)
40+ years	2.10	(2.08, 2.13)	2.41	(1.49, 3.89)
Marital status
Single	1.00 (ref)	Ref	1.00 (ref)	Ref
Married	0.96	(0.72, 1.27)	1.22	(0.64, 2.31)
Divorced/separated	0.89	(0.64, 1.24)	1.07	(0.77, 1.47)
Widowed	0.28	(0.10, 0.83)	0.40	(0.13, 1.24)
Educational attainment
No formal education	1.63	(1.52, 1.75)	0.78	(0.12, 4.99)
Primary	1.00 (ref)	Ref	1.00 (ref)	Ref
Secondary	0.88	(0.62, 1.24)	0.73	(0.71, 0.76)
College/university	0.68	(0.21, 2.16)	0.57	(0.31, 1.04)
Facility
Clinic 1	1.90	(1.79, 2.02)	1.95	(1.83, 2.07)
Clinic 2	1.00 (ref)	Ref	1.00 (ref)	Ref

Notes: Model exposure of interest was age adjusting for sex, facility and month of enrolment (random effect only). CI: confidence interval

aPR, adjusted prevalence ratio.

The marginal probability of being virally suppressed at baseline was highest among women aged 40+ years at 42% (95% CI: 39.3%, 44.3%) (figure 1, table 3). Men aged 18–24 years had the lowest probability of baseline viral suppression at 12% (95% CI: 4.6%, 19.0%).

Table 3

Marginal probability results of interaction between age and sex

Interaction term	Marginal probability %	95% CI %
18–24 years and female	17.1	(13.3, 20.8)
18–24 years and male	11.8	(4.6, 19.0)
25–29 years and female	25.0	(0.0, 50.8)
25–29 years and male	17.3	(6.2, 28.3)
30–34 years and female	37.2	(23.8, 50.7)
30–34 years and male	25.8	(24.9, 26.7)
35–39 years and female	34.5	(7.9, 61.1)
35–39 years and male	23.9	(0.0, 51.7)
40+ years and female	41.8	(39.3, 44.3)
40+ years and male	28.9	(19.3, 38.6)

Note: Estimates based on model adjusted for marital status, educational attainment, month of testing and facility. CI: confidence interval

Figure 1. Marginal probability of new client having baseline viral suppression at presentation by age and sex (N=248).

Fifty-seven of 66 (86%) participants with a suppressed baseline viral load (≤1000 copies/mL) completed the silent transfer survey. Of these, 44 (77%) indicated that they had tested positive for HIV previously across one or more of 38 different clinics in Zambia. Eight (14% (8/57)) indicated that they had previously received HIV treatment at a different clinic. Of those reporting having received HIV treatment previously, 7 (88% (7/8)) cited travel and convenience-related concerns among the reasons for attending the study clinic, including lower transportation cost (n=5) and closer proximity of the new clinic to home (n=2).

Discussion

We describe characteristics associated with viral suppression for ‘new’ patients at the time of presentation to HIV care, which may be indicative of silent transfer in Lusaka, Zambia. Overall, we found that 27% of the study population fit our silent transfer definition. We noted that the strongest predictor of silent transfer was marital status, followed by age, study clinic, educational attainment and sex. Older women had a higher adjusted prevalence of silent transfer compared with younger people of either sex, whereas those who had been widowed had the lowest adjusted prevalence of silent transfer.

For this analysis, we used an operational definition of silent transfer that may not fully capture all potential types of silent transfer, which we limited to those with viral suppression as a marker of previous HIV care and ART receipt. Specifically, we do not account for PLWH who silently transferred after a gap in HIV care sufficient to allow viral load rebound and/or PLWH in HIV care with an inadequate ART regimen or HIV resistance leading to unsuppressed viral load. As such, our estimate of a little over one-quarter of patients being potential silent transfers is likely conservative. It is important to note that some participants may have had a viral load ≤1000 copies/mL at baseline due to natural viral suppression, however, given the relative rarity of this phenomenon, we believe HIV controllers play a negligible role in explaining our findings.^{21 22}

Though typically thought of as an option reserved for well-resourced settings, our findings suggest that ‘clinic shopping’—or experiencing multiple HIV care clinics to identify the most favourable one—may be increasingly possible in Lusaka, Zambia, and similar settings where HIV treatment scale up has reached maturity. Previous research in Zambia has outlined drivers of HIV care disengagement, such as rude interactions with healthcare providers.^{9 23 24} As HIV care coverage expands, additional attention must be paid to how PLWH may move between different clinics and providers in search of more convenient and patient-centred services, and how these movements may differ by age, gender and other socio-demographic and clinical characteristics. In response, the health system can make further investments in differentiated care models, as well as in systems for unique identifiers, patient monitoring across different HIV care sites along the care continuum and quality improvement for HIV service delivery. Given the observed association between low educational attainment and potential silent transfer, patient education may have a role in improving transfer processes and options for PLWH in Zambia.

There may be several explanations for the apparent silent transfers we observed. It is possible that people who fear or lack confidence to request an official transfer, or who do not have information about the processes involved, opt for repeating the ‘new patient’ process with which they are familiar (despite the substantial time involved). It is also possible that some individuals are not transfers at all, but rather are PLWH receiving care from multiple clinics simultaneously. Motivations for attending multiple clinics concurrently may be to collect ART for partners, friends or family members, or to stock up on ART to avoid the need to visit the clinic for refills. With the scale up of 6-month medication dispensation, these reasons may feature less prominently now than in the past. Nevertheless, prior research from Zambia has emphasised the importance of advertising processes for ‘buddy’ ART pick-up and options to receive extended or expedited ART refills from other clinics in case of planned travel.²⁵

Recognising that new patient enrolment is a time and resource intensive process for the health system, programmatic efficiencies can be gained by identifying individuals already in HIV care elsewhere at the time of presentation to care, which, in turn, can improve programme and individual patient monitoring of the HIV care cascade. Improving HIV care enrolment procedures in this way could be achieved by systematic screening for silent transfer at the time of HIV care initiation, and offering those who disclose silent transfer tangible benefits, such as immediate or ‘fast track’ eligibility for available differentiated service delivery model options.

Limitations

This study had several limitations. First, only two urban HIV care facilities were included in this study, and only about one-third of the target population were screened for study eligibility, which may limit the generalisability of these findings in Zambia and SSA. Second, as noted above, we used a pragmatic definition of silent transfer that may not have captured all types of silent transfer, thereby underestimating silent transfer in our analysis. To enable a more definitive assessment of silent transfer, the addition of ART metabolite analysis in samples such as hair, would help identify individuals with prior ART exposure. Third, there may have been participants in our parent study who silently transferred care but who were not virally suppressed at baseline and thus were excluded from our definition of silent transfer. Unsuppressed PLWH may have different motivations for silent transfer than those included in our study and may be driven more by factors associated with treatment interruption or HIV care re-engagement. Finally, although over three-quarters of participants disclosed having had a previous HIV positive test result and ART identification number, we could not confirm this in our search of the national electronic medical record, which limits our understanding of the clinical histories of potential silent transfers in our study. The addition of unique identifying methods such as biometrics may enable more efficient patient tracking in the medical record across health facilities.

Conclusions

We observed a relatively high proportion of PLWH with suppressed viral load at baseline suggesting that silent transfer is a major issue in routine HIV care settings. These persons tended to be older, women and without prior formal education with heterogeneity between clinics. These observations suggest the possibility of patients who clinical shop and/or are enrolled in care at multiple clinics simultaneously and warrant further study and attention in routine HIV programmes.

We would like to thank the study participants who volunteered for the study. We also thank the Zambian Ministry of Health for their leadership on all facets of the national HIV programme. We would also like to thank the healthcare workers and CIDRZ research staff who made this work possible.

Data availability statement

Data are available upon reasonable request. Data will be made available upon publication of the main effects paper for the parent (REACHZ) study. Additionally, appropriately de-identified data may be made available, consistent with the rules and regulations of the Zambian Ministry of Health, upon reasonable request to the study co-principal investigator ([email protected]).

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).

Ethics approval

The study protocol was approved by ethical review bodies of the University of Zambia (#1157-2020), the Zambian National Health Research Authority and the University of Alabama at Birmingham (#300006066). Participants gave informed consent to participate in the study before taking part.

Footnote

Twitter @jakepry_epi

MEH and SI contributed equally.

Contributors JMP: Guarantor, writing original draft, conceptualisation, data management, data analysis, critical review and incorporation of coauthor input, final draft preparation. CM: Data curation and data management. HK: Data curation and data management. MMu: Data curation and data management. CF: Data curation and data management. MMo: Data curation and data management. TS: Critical draft review. CB-M: Critical draft review. MEH: Writing original draft, conceptualisation, data management, secured funding, final draft preparation. SI: Conceptualisation, secured funding, final draft preparation.

Funding Funding for this work was provided through the Centers for Disease Control and Prevention (CDC) and the President’s Emergency Plan For AIDS Relief (PEPFAR) through grant number: NU2GGH001920.

Competing interests None declared.

Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

References

¹ Geng EH, Glidden DV, Bwana MB, et al. Retention in care and connection to care among HIV-infected patients on antiretroviral therapy in Africa: estimation via a sampling-based approach. PLoS One 2011; 6: e21797. doi:10.1371/journal.pone.0021797

² Kranzer K, Lewis JJ, Ford N, et al. Treatment interruption in a primary care antiretroviral therapy programme in South Africa: cohort analysis of trends and risk factors. J Acquir Immune Defic Syndr 2010; 55: e17–23. doi:10.1097/QAI.0b013e3181f275fd

³ Mutevedzi PC, Lessells RJ, Newell ML. Disengagement from care in a decentralised primary health care antiretroviral treatment programme: cohort study in rural South Africa. Trop Med Int Health 2013; 18: 934–41. doi:10.1111/tmi.12135

⁴ Nglazi MD, Lawn SD, Kaplan R, et al. Changes in programmatic outcomes during 7 years of scale-up at a community-based antiretroviral treatment service in South Africa. J Acquir Immune Defic Syndr 2011; 56: e1–8. doi:10.1097/QAI.0b013e3181ff0bdc

⁵ Wilkinson LS, Skordis-Worrall J, Ajose O, et al. Self‐transfer and mortality amongst adults lost to follow‐up in ART programmes in low‐and middle‐income countries: systematic review and meta‐analysis. Trop Med Int Health 2015; 20: 365–79. doi:10.1111/tmi.12434

⁶ Fox MP, Bor J, Brennan AT, et al. Estimating retention in HIV care accounting for patient transfers: a national laboratory cohort study in South Africa. PLoS Med 2018; 15: e1002589. doi:10.1371/journal.pmed.1002589

⁷ Geng EH, Odeny TA, Lyamuya R, et al. Retention in care and patient-reported reasons for undocumented transfer or stopping care among HIV-infected patients on antiretroviral therapy in Eastern Africa: application of a sampling-based approach. Clin Infect Dis 2016; 62: 935–44. doi:10.1093/cid/civ1004

⁸ Hickey MD, Omollo D, Salmen CR, et al. Movement between facilities for HIV care among a mobile population in Kenya: transfer, loss to follow-up, and Reengagement. AIDS Care 2016; 28: 1386–93. doi:10.1080/09540121.2016.1179253

⁹ Sikombe K, Mody A, Kadota J, et al. Understanding patient transfers across multiple clinics in Zambia among HIV infected adults. PLoS One 2020; 15: e0241477. doi:10.1371/journal.pone.0241477

¹⁰ Beres LK, Schwartz S, Simbeza S, et al. Patterns and predictors of incident return to HIV care among traced, disengaged patients in Zambia: analysis of a prospective cohort. J Acquir Immune Defic Syndr 2021; 86: 313–22. doi:10.1097/QAI.0000000000002554

¹¹ Hallett TB, Eaton JW. A side door into care Cascade for HIV-infected patients. J Acquir Immune Defic Syndr 2013; 63 Suppl 2: S228–32. doi:10.1097/QAI.0b013e318298721b

¹² Ehrenkranz P, Rosen S, Boulle A, et al. The revolving door of HIV care: revising the service delivery Cascade to achieve the UNAIDS 95-95-95 goals. PLOS Med 2021; 18: e1003651. doi:10.1371/journal.pmed.1003651

¹³ Pry J, Chipungu J, Smith HJ, et al. Patient-reported reasons for declining same-day antiretroviral therapy initiation in routine HIV care settings in Lusaka, Zambia: results from a mixed-effects regression analysis. J Int AIDS Soc 2020; 23: e25560. doi:10.1002/jia2.25560

¹⁴ Bock P, Fatti G, Ford N, et al. Attrition when providing antiretroviral treatment at CD4 counts >500Cells/ΜL at three government clinics included in the HPTN 071 (PopART) trial in South Africa. PLOS ONE 2018; 13: e0195127. doi:10.1371/journal.pone.0195127

¹⁵ Gabagaya G, Rukundo G, Amone A, et al. Prevalence of undetectable and suppressed viral load in HIV-infected pregnant women initiating option B+ in Uganda: an observational study nested within a randomized controlled trial. BMC Infect Dis 2021; 21: 907. doi:10.1186/s12879-021-06608-4

¹⁶ TRACE Initiative. Tracking with Recency assays to control the epidemic. 2022. Available: https://trace-recency.org [Accessed 13 Sep 2022 ].

¹⁷ Zambia Ministry of Health. Zambia Consolidated guidelines for treatment and prevention of HIV infection; 2020.

¹⁸ Holmes CB, Sikazwe I, Sikombe K, et al. Estimated mortality on HIV treatment among active patients and patients lost to follow-up in 4 provinces of Zambia: findings from a multistage sampling-based survey. PLoS Med 2018; 15: e1002489. doi:10.1371/journal.pmed.1002489

¹⁹ DATIM. Data accountability, transparency, and impact [ DHIS2 ]. 2023. Available: https://www.datim.org/dhis-web-commons/security/login.action

²⁰ Ellman TM, Alemayehu B, Abrams EJ, et al. Selecting a viral load threshold for routine monitoring in resource-limited settings: optimizing individual health and population impact. J Int AIDS Soc 2017; 20 Suppl 7: e25007. doi:10.1002/jia2.25007

²¹ Dinoso JB, Kim SY, Siliciano RF, et al. A comparison of viral loads between HIV-1-infected elite suppressors and individuals who receive suppressive highly active antiretroviral therapy. Clin Infect Dis 2008; 47: 102–4. doi:10.1086/588791

²² Pereyra F, Palmer S, Miura T, et al. Persistent low-level Viremia in HIV-1 elite controllers and relationship to immunologic parameters. J Infect Dis 2009; 200: 984–90. doi:10.1086/605446

²³ Mwamba C, Sharma A, Mukamba N, et al. 'They care rudely!': resourcing and relational health system factors that influence retention in care for people living with HIV in Zambia. BMJ Glob Health 2018; 3: e001007. doi:10.1136/bmjgh-2018-001007

²⁴ Zanolini A, Sikombe K, Sikazwe I, et al. Understanding preferences for HIV care and treatment in Zambia: evidence from a discrete choice experiment among patients who have been lost to follow-up. PLoS Med 2018; 15: e1002636. doi:10.1371/journal.pmed.1002636

²⁵ Eshun-Wilson I, Mukumbwa-Mwenechanya M, Kim H-Y, et al. Differentiated care preferences of stable patients on antiretroviral therapy in Zambia: A discrete choice experiment. J Acquir Immune Defic Syndr 2019; 81: 540–6. doi:10.1097/QAI.0000000000002070