Access to this full-text is provided by PLOS.
Content available from PLOS One
This content is subject to copyright.
RESEARCH ARTICLE
Improved HIV case finding among key
populations after differentiated data driven
community testing approaches in Zambia
Joseph Kamanga
1
, Kayla StankevitzID
2
*, Andres Martinez
2
, Robert Chiegil
2
,
Lameck Nyirenda
1
, Florence Mulenga
1
, Mario Chen
2
, Mulamuli Mpofu
3
, Sam Lubasi
1
,
Moses Bateganya
2
1FHI 360, Lusaka, Zambia, 2FHI 360, Durham, North Carolina, United States of America, 3FHI 360,
Gaborone, Botswana
*kstankevitz@fhi360.org
Abstract
Introduction
Open Doors, an HIV prevention project targeting key populations in Zambia, recorded low
HIV positivity rates (9%) among HIV testing clients, compared to national adult prevalence
(12.3%), suggesting case finding efficiency could be improved. To close this gap, they
undertook a series of targeted programmatic and management interventions. We share the
outcomes of these interventions, specifically changes in testing volume, HIV positivity rate,
and total numbers of key populations living with HIV identified.
Methods
The project implemented a range of interventions to improve HIV case finding using a Total
Quality Leadership and Accountability (TQLA) approach. We analyzed program data for key
populations who received HIV testing six months before the interventions (October 2017–
March 2018) and 12 months after (April 2018–March 2019). Interrupted time series analysis
was used to evaluate the impact on HIV positivity and total case finding and trends in positiv-
ity and case finding over time, before and after the interventions.
Results
While the monthly average number of HIV tests performed increased by only 14% post-
intervention, the monthly average number of HIV positive individuals identified increased by
290%. The average HIV positivity rate rose from 9.7% to 32.4%. Positivity rates and case
finding remained significantly higher in all post-intervention months. Similar trends were
observed among FSW and MSM.
Conclusions
The Open Doors project was able to reach large numbers of previously undiagnosed key
populations by implementing a targeted managerial and technical intervention, resulting in a
significant increase in the HIV positivity rate sustained over 12 months. These results
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 1 / 13
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Kamanga J, Stankevitz K, Martinez A,
Chiegil R, Nyirenda L, Mulenga F, et al. (2021)
Improved HIV case finding among key populations
after differentiated data driven community testing
approaches in Zambia. PLoS ONE 16(12):
e0258573. https://doi.org/10.1371/journal.
pone.0258573
Editor: Catherine G. Sutcliffe, Johns Hopkins
University, UNITED STATES
Received: August 21, 2020
Accepted: October 1, 2021
Published: December 2, 2021
Copyright: ©2021 Kamanga et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: The data underlying
the results presented in the study are available
from the Harvard Dataverse: https://dataverse.
harvard.edu/dataset.xhtml?persistentId=doi:10.
7910/DVN/JAHLU5&faces-redirect=true.
Funding: This work is made possible by the
support of the American People through the United
States Agency for International Development
(USAID) and the U.S. President’s Emergency Plan
for AIDS Relief (PEPFAR). Financial assistance was
demonstrate that differentiated, data-driven approaches can help close the 95-95-95 gaps
among key populations.
Introduction
HIV testing services (HTS) are critical to achieve the Joint United Nations Programme on
HIV/AIDS (UNAIDS) 95-95-95 targets of 95% of people living with HIV (PLHIV) knowing
their status; 95% who know their status on treatment; and 95% on treatment virally suppressed
[1]. While substantial progress has been made, HIV case finding among key and other priority
populations remains challenging [2,3]. New, innovative approaches are required to reach
HIV-positive individuals, especially those who have challenges accessing facility-based testing
or have never previously been tested.
Reaching key populations (KPs)—including men who have sex with men (MSM), transgen-
der persons (TG), and female sex workers (FSWs)—with HIV prevention and testing services
is a crucial last mile to ending the HIV epidemic [2]. In 2018, more than half of new HIV infec-
tions among adults worldwide occurred among KPs and their partners [4]. Many factors,
often stemming from criminalization of certain behaviors, make it difficult for KPs to access
facility-based testing services [5]. Yet, until recently, most HTS were delivered through facil-
ity-based settings.
Several approaches have been developed to improve access to HIV testing for individuals
not reached through traditional facility-based services, including home-based, mobile, index
(testing of sexual partners and family members), social network, and self-testing [6–8]. Uptake
and positivity rates of these approaches vary among different populations [8]. A recent system-
atic review found that while community-based testing modalities resulted in lower HIV posi-
tivity rates in the general population, positivity rates among KPs were high, from 24% to 28%
[6]. Index testing and social network strategy (SNS) have also demonstrated high positivity
rates among KPs [9,10]. However, concerns have been raised regarding index testing, espe-
cially with respect to confidentiality and client rights [11]. Thus, a mix of interventions is
needed. The effectiveness of the strategies depends on the context of the local epidemic, exist-
ing gaps in testing coverage, and the needs of the target populations [12].
In Zambia, key populations have a high burden of HIV, yet there is limited data available
due to the criminalization of sex work and homosexuality in the country [13]. In 2018 PEP-
FAR estimated that prevalence of HIV was 41.6% among FSW and 17.1% among MSM [14].
From July to September 2017, the Open Doors Project (ODP) in Zambia reported a posi-
tivity rate of 9% among KPs reached by the project, substantially lower than the estimated
prevalence in these populations, as well as the Zambia national adult prevalence of 12.3%
[15]. The low positivity rate was an indication that testing approaches implemented prior to
September 2017 were not effective or were not deployed with fidelity. In response, the proj-
ect undertook a review of strategies and technical approaches used and planned remedial
actions.
We describe the design and implementation of an improved testing and linkage approach
and the associated changes in number of KP individuals tested (volume tested), proportion
who tested positive (positivity rate), and number who test positive (case finding rate).
Results from this real-world approach to improving HIV testing efficacy will help other
HIV testing providers adapt the large number of testing modalities and approaches to their
own contexts.
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 2 / 13
provided by USAID to FHI 360. The contents of this
report are the sole responsibility of FHI 360 and do
not necessarily reflect the views of USAID,
PEPFAR, or the United States Government. The
funder provided support in the form of some
portion of salaries for authors JK, KS, AM, RC, LN,
FM, MC, MM, SL, MB; but did not have any
additional role in the study design, data collection
and analysis, decision to publish, or preparation of
the manuscript. The specific roles of these authors
are articulated in the ‘author contributions’ section.
Competing interests: The authors have declared
that no competing interests exist.
Methods
Setting and population
The USAID Open Doors project is funded by the U.S. President’s Emergency Plan for AIDS
Relief through the United Stated Agency for International Development (USAID). The ODP is
the largest KP project in Zambia and is implemented by FHI 360 and the Zambia Health Edu-
cation and Communications Trust (ZHECT), a local nongovernmental organization, in col-
laboration with several KP-led civil society organizations (CSOs). The project works in 8 high
HIV burden districts (prevalence range 6.9%-16.1%), with major centers of economic activity
and transit routes within five provinces, as shown in Table 1.
The ODP has operated since May 2016. With an implementation design based on Zambia
National HIV and AIDS Strategic Framework for HIV testing services (HTS) [16], the pro-
gram aims to successfully identify KP, offer HTS, and link those who test negative to preven-
tion interventions and those who test positive to ART. The project has supported testing at
project facilities (e.g., wellness centers) and in the community through outreach.
Intervention
To improve fidelity in case identification, ODP applied the Total Quality Leadership and
Accountability (TQLA) intervention, an FHI 360 adaptive management approach [17]. TQLA
is a strategy for improving leadership, strengthening data collection systems, and improving
utilization of data to improve outcomes.
The core principles of the TQLA approach are adaptive leadership and regular review of
data. Project leaders and staff convene meetings with site-level health care workers to discuss
and prioritize project challenges and co-create solutions. After key interventions are agreed
upon, program staff put in place a plan to meet regularly to review data, identify gaps, and con-
tinue to implement solutions. These regular data reviews, also called Situation Room Meetings
(SRM), entail review of granular, site-level data on a daily or weekly basis to inform where
technical assistance (TA) should be targeted. TA is provided over the phone or in person.
To address the gaps identified in case identification in the OPD project, outreach workers
from the eight project districts and representatives from the KP community were engaged in
March 2018. Interventions to improve case identification were adopted, as summarized in
Table 2. KPs were engaged as peer promoters and were trained and directly supported imple-
mentation of SNS and index testing approaches at community level.
Consequently, the ODP introduced daily (Monday-Friday) SRMs project wide, and used
the framework to evaluate district performance, pivot resources and cascade accountability
throughout the team. District and community teams reported data daily, on key indicators
(# tested, # declined test, # known HIV-positive status, # positive, # linked to treatment) which
were aggregated, visualized, and the trends analyzed, and explanatory factors discussed. Photos
from SRMs and examples of data reviewed are available in S1 Fig. At each SRM, staff discussed
performance against targets and made decisions regarding resource allocation, site-level target
Table 1. Provinces and districts supported by the ODP.
Province HIV Prevalence (2018) [14] Districts supported by ODP
Central 12.6% Kabwe and Kapiri Mposhi
Copperbelt 14.7% Kitwe and Chililabombwe
Lusaka 16.6% Lusaka Urban and Chirundu
North Western 6.3% Livingstone
Southern 12.5% Solwezi
https://doi.org/10.1371/journal.pone.0258573.t001
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 3 / 13
revisions, course correction, and technical strategies required to achieve project targets. Feed-
back was provided to sites through field visits, email, WhatsApp groups, and phone calls. Field
visits were informed by data and were often to sites that were not meeting targets. During each
visit, project staff helped frontline workers to review their performance against targets, identify
root causes for underperformance, adopt and scale up best practices. SRM meetings continued
from April 1 onwards, and have continued to date.
Statistical analysis
We analyzed routinely collected program data from before and after the implementation of
the improvement strategies and SRM meetings. The data are of program participants 16 years
and older who sought HTS during the six months before the intervention (October 1, 2017–
Table 2. Strategies used to improve HIV positivity rate.
Strategy Description
Hot spot mapping and targeting • Analyzed existing “hot spots” (i.e., bars/taverns, brothels,
shebeens, streets, nightclubs) and places where KPs congregate
by KP type to determine HIV test positivity during the three
months prior to intervention (January–March 2018)
• Applied the 80/20 rule to identify the top 20% of hot spots
responsible for 80% of all identified HIV-positive KPs during
the prior three months. These were largely hot spots with a
combination of greater KP population and higher HIV positivity
[18].
• Revised microplans and site maps to highlight priority hot spots
and characteristic features of KPs (size, type, place, and time of
congregation), and determined appropriate HTS approaches
and scheduling activities to reach KPs (daytime, nights, etc.)
• Conducted daily outreach to hot spots generating the greatest
number of positives
Moonlight testing • Adopted moonlight HIV testing to reach individuals who were
hard to reach during daytime testing hours
Risk assessments prior to testing • Trained and deployed peer promoters and lay counsellors to use
risk assessment tools to identify and prioritize HIV testing for
KP individuals likely to test positive
• Trained and deployed peer promoters and counsellors to classify
KPs as either MSM, FSW, or transgender using the PEPFAR
standard KP classification tool [11]
Enhanced peer outreach approach (EPOA)
training
• Peer educators trained on demand creation and reaching hard-
to-reach KP members [19]
Index case testing (ICT)/partner notification
services (PNS), bridge index testing
• Trained and deployed peer promoters and counsellors to adopt
ICT/PNS to target HTS to partners of index cases
• Offered HTS to sexual partners of each index client
• Offered HTS to male partners of index FSWs (referred to as
"bridge" clients) as well other FSWs they identified through PNS
Social network strategy (SNS) • Integrated HTS into existing social network activities, where the
index client introduces friends within her/his social network to
access HTS. “Friends” were not necessarily sexual partners of the
index client, but merely belonged to the same social network.
HIV self-test • Distributed self-test kits through index clients to their partners
and within social networks to reach hidden and hard-to-reach
KPs
• Those who tested HIV positive using self-test kits received an
HIV confirmatory test prior to linkage to ART
https://doi.org/10.1371/journal.pone.0258573.t002
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 4 / 13
March 31, 2018) and 12 months after the intervention (April 1, 2018–March 31, 2019).
Although some components of the intervention were rolled out over a few months, to simplify
the analysis we treated April 1 as the beginning of all intervention components. To receive
HTS, an individual must have self-identified as either an FSW, MSM or TG. Since the ODP
specifically supports key populations, individuals who do not identify as such (for instance, cli-
ents of sex workers) are referred elsewhere; therefore, they are not included in this analysis.
An M&E officer at the ODP central office generated a dataset with the HTS data between
the specified dates. HTS data were de-identified by removing participant names and unique
IDs. The M&E officer then transferred the de-identified datafiles to the HQ-based analysis
team for final cleaning and analyses.
To describe the overall sample characteristics, we calculated n’s, positivity rates, and case
finding rates stratified by intervention period and KP type, age group, testing modality, and
district. We compared the overall positivity rates for the 6 pre-intervention months to those of
the 12 post-intervention months, explored the trends in the monthly positivity rates and case
finding, and used a time series regression analysis to study the impact of the intervention.
More specifically, we fitted a segmented linear regression model with a Newey-West variance
estimator to account for the serial correlation between months [20]. We also fit this time series
model by population type and testing modality. We excluded transgender population and
facility testing modality from our models by population type and modality due to low sample
sizes in some months. In fitting these models, we generally used a 1-month lag for the serial
correlation structure; the serial correlation tests didn’t allow us to reject the null hypothesis the
serial correlation in the time series died out after 1 month. We used Stata version 15 for all
data management and analysis, and Linden’s user-contributed commands for conducting the
interrupted time series analysis [21–23].
Ethical review
Approval to analyse these data was obtained from FHI 360 Office of International Research
Ethics (OIRE) and the ERES Converge IRB in Lusaka, Zambia. The study received a waiver of
informed consent as the research involved no more than minimal risk to participants and
could not be practicably carried out if informed consent was not waived. Approval to publish
the manuscript was provided by the National Health Research Authority in Zambia (Ref No:
NHRA00001/16/04/2021).
Results
Between October 1, 2017 and March 31, 2019, 30,911 KP individuals (75.2% FSWs, 22.4%
MSM, and 2.4% TG) 16 years and older received HTS—9,408 in the six months pre-interven-
tion and 21,503 during the 12 months intervention period (Table 3). The mean age was 27
years. Table 3 shows the number of tests and HIV positivity rates for different KP groups, age
categories, testing modalities, and by district before and after the intervention. Average
monthly number of HIV tests increased 14% after the intervention, from an average of 1,568
individuals tested per month during pre-intervention to an average of 1,792 individuals tested
per month during post-intervention. While it had the lowest positivity rate, the Community/
Outreach testing modality reached the majority of patients (92.3% pre-intervention and 83.4%
post-intervention) and found the majority of positive cases (86.7% pre-intervention, and
78.4% post-intervention).
The positivity rate in the six pre-intervention months was 9.7% and increased to 32.4% dur-
ing the 12 post-intervention months. Fig 1 shows the absolute number of cases identified per
month. The monthly average number of positive cases identified went from 149 pre-
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 5 / 13
intervention to 581 post-intervention, a 290% increase. Increases were also seen across all KP
groups: from 10.7% to 36% among FSWs; 5.6% to 22.5% among MSM; and 5.0% to 25.5%
among. Likewise, positivity rate increases were observed in all testing modalities and in each
region, even in Lusaka where the pre-intervention positivity was already high at 25.3%.
Table 3. Participant characteristics and positivity by intervention period.
Characteristics Pre-intervention (6 months) Post-intervention (12 months) Total
n positivity Case finding n positivity Case finding n
KP group
FSW 7,533 10.7% 806 15,724 36.0% 5,661 23,257
MSM 1,656 5.6% 93 5,266 22.5% 1,185 6,922
TG 219 5.0% 11 513 25.5% 131 732
Age group
<18 771 2.2% 17 224 8.0% 18 995
18–24 4,021 5.6% 225 6028 18.2% 1,097 10049
24–29 2,352 11.2% 263 6166 29.9% 1,844 8518
29–34 1,274 15.9% 203 4381 38.7% 1,695 5655
34–39 586 19.5% 114 2614 48.5% 1,268 3200
39 or older 404 21.8% 88 2090 50.0% 1,045 2494
Testing modality
Community/Outreach 8,684 8.9% 773 17,923 30.5% 5,467 26,607
Facility 724 19.3% 140 985 31.2% 307 1,709
Index/PNS 0 - 771 70.0% 540 771
SNS 0 - 1,824 35.9% 655 1,824
District
Chililabombwe 1,224 7.4% 91 2,132 33.7% 718 3,356
Chirundu 890 4.6% 41 1,543 25.3% 390 2,434
Kabwe 963 5.8% 56 2,213 28.5% 631 3,177
Kapiri Mposhi 2,238 4.4% 98 2,051 27.3% 560 4,463
Kitwe 1,044 6.0% 63 3,833 25.6% 981 4,879
Livingstone 1,105 10.7% 118 2,557 39.9% 1,020 3,664
Lusaka 833 25.3% 211 4,457 35.2% 1,569 5,293
Solwezi 1,111 20.3% 226 2,717 40.3% 1,095 3,830
Total 9,408 9.5% 894 21,503 32.4% 6,967 30,911
https://doi.org/10.1371/journal.pone.0258573.t003
Fig 1. HIV case finding by month.
https://doi.org/10.1371/journal.pone.0258573.g001
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 6 / 13
Based on the interrupted time series analysis (Table 4,Fig 2), the post-intervention monthly
positivity increased by 20.6% (95% CI: 5.1%-36.1%, p = 0.01) and the monthly number of cases
identified increased by 327 (95% CI:118.6–536.0, p = 0.005). Pre-intervention, case finding
was increasing by 24.8 cases per month (95% CI: 6.6, 43.0; p = 0.01). Further, the slope of the
post-intervention trend line did not differ statistically from the slope of the pre-intervention
trend line.
We also analyzed changes in positivity and case finding by population and modality (Fig 3).
Statistical results of these models are included in S1 Table. The immediate intervention effect
shows statistically significant increases in positivity and case finding among both MSM and
FSW, as well as among clients reached via community outreach. Positivity and case finding for
both populations stay higher than pre-intervention rates throughout the intervention period.
While improvements in community outreach positivity and case finding are maintained post-
intervention, the trend in case finding post-intervention is negative. This can mostly be attrib-
utable to a large initial intervention effect that recedes over time.
Discussion
To reach the 95-95-95 goals, programs must implement evidence-based interventions with
fidelity to identify populations at risk, provide HTS, and link HIV-positive individuals to ART
for sustained viral suppression. Results from this analysis of program data before and after
implementation of managerial and technical interventions show that implementation of evi-
dence-based interventions—implemented under an adaptive management approach focused
on data use for decision-making—can lead to program improvement.
While resource constraints have led to an increased focus on improving testing efficiency,
some stakeholders have cautioned that focusing on positivity rate as a primary criterion to
Table 4. Results of interrupted time series analysis on positivity and case finding (n = 30,911).
Coef. Newey-West Std. Err. p-value 95% CI
Positivity rate
Pre-intervention trend 0.005 0.003 0.129 (-0.002, 0.013)
Effect of intervention 0.206 0.072 0.012 (0.052, 0.360)
Change in trend during intervention -0.002 0.008 0.773 (-0.019, 0.014)
Case finding
Pre-intervention trend 24.8 8.484 0.011 (6.6, 43.0)
Effect of intervention 327.3 97.307 0.005 (118.6, 536.0)
Change in trend during intervention -22.0 14.336 0.147 (-52.8, 8.7)
https://doi.org/10.1371/journal.pone.0258573.t004
Fig 2. Percent of HIV tests that were positive and number of positive cases, by intervention period.
https://doi.org/10.1371/journal.pone.0258573.g002
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 7 / 13
Fig 3. Percent of HIV tests that were positive and number of positive cases by population and testing modality.
https://doi.org/10.1371/journal.pone.0258573.g003
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 8 / 13
measure success of HTS could lead to a decrease in the total number of positives identified
[24]. In this analysis we have shown that programs can improve the positivity rate and more
efficiently increase the total number of positive individuals identified simultaneously.
Other studies have shown that while SNS and index/PNS can lead to high positivity rates
compared to other modalities, the absolute number of HIV-positive individuals identified is
often low [6,25,26]. This is confirmed through this analysis, with only 17% of new positives
being identified by SNS and index testing/PNS. This is ultimately due to the volume of clients
reached through these approaches—only 12.1% of clients tested were reached through the new
high yield modalities. Thus, a combination of testing approaches is critical, balancing the HIV
positivity rate on one hand and reach/case finding on the other. In this program, community/
outreach and facility testing, though achieving lower positivity rates compared to index/PNS
and SNS, resulted in more HIV-positive KP individuals being identified. While the introduc-
tion of high yield testing modalities can explain some of the improvement in case finding,
community/outreach and facility testing case finding rates also increased. This suggests that
while the introduction of SNS and index/PNS were important components of success, other
interventions focused on improving community outreach such as hotspot mapping and
moonlighting also contributed to the impact of the intervention.
In ODP, we achieved higher positivity than other studies in Zambia and sub-Saharan
Africa. A recent study of targeted testing strategies among the general population in Zambia
showed positivity of index testing at 44.7% and other community testing (including mobile
and testing at standalone VCT centres) at 23.2% [27]. A similar study in Zimbabwe found that
index testing (32.6%) had much higher positivity than facility-based testing (4.1%) [28]. While
positivity in our program was higher than these studies, some of the difference is likely attrib-
uted to our focus on KPs. In fact, a recent systematic review of testing positivity found that
while there were only four studies specifically targeting KPs, those studies recorded higher pos-
itivity (24–55%), when compared to community-based general population (6–11%) or facility
VCT (18–20%) testing strategies [29]. The success of KP projects in identifying new positive
cases, and lack of published literature, suggests scale up of KP-focused strategies is urgently
needed [30].
Involvement of target populations in planning and implementation is key in reaching
unreached members of targeted populations [31–33]. The stigmatization and criminalization
of many KPs make it increasingly important to garner acceptance among these populations.
While KP partners were part of the project from the beginning, ODP directly involved KPs
in the co-design of the managerial and technical interventions. Involving KPs helped the
project identify gaps in the current approaches and decide to implement differentiated mod-
els of testing. Involvement of KPs in mobilizing peers for HIV services (EPOA), has been
shown to be effective in identify unreached networks with HIV services [19,34,35]. This
project successfully employed the approach to effectively reach KPs and achieved high
positivity.
The intervention focused heavily on increased frequency of reporting and use of data to
inform decisions, including at daily SRMs, an important aspect of TQLA. The use of data to
inform program decisions is widely accepted, yet in many HIV programs more emphasis is
placed on data collection than data use [36–38]. Recently, data use and increased monitoring
has been promoted, with PEPFAR even recommending that many testing and treatment indi-
cators be monitored on a weekly basis rather than the required reporting frequency of quar-
terly or semi-annually [11,39]. While it is not possible to determine the role of intervention
components singly, this analysis showed that implementing a combination of interventions,
including daily reporting and continuous review of data and performance trends, allowed
timely course correction and resulted in improvement over time.
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 9 / 13
This analysis, and the TQLA approach, have important limitations. Given the lack of an
experimental design, observed changes in positivity rate cannot be attributed to TQLA, or spe-
cific management approaches or technical activities. It is not possible to determine whether
the newly implemented technical approaches would have been successful without the new
management approaches and daily data review, or vice versa. Further, the intervention compo-
nents were introduced in a phased manner and often in combination making attribution of
changes over time hard to determine. TQLA approaches such as convening meetings, invest-
ing in collection for daily data review, and daily targeted TA can be resource intensive. Cost
data are not available to show if this approach or individual components are cost effective.
Lastly, lack of accurate size estimation or prevalence data on different KPs makes it impossible
to compare the case identification rate and positivity to the target populations as a whole.
Nonetheless, this analysis has many strengths. We used validated, individual level HIV test-
ing data that allowed us to look at changes by covariates such as population type and site. Fur-
ther, we analyzed data from six months before the intervention and 12 months after, allowing
us not only to present a before and after scenario, but also to examine changes monthly over
time. Often an intervention can provide an immediate improvement in the targeted outcome,
yet over time that outcome comes back to pre-intervention levels. By including 12 post-inter-
vention months, we were able to see that even after the initial improvement recedes, positivity
rates settled into a more stable pattern, that is, even 12 months after the intervention, still sig-
nificantly higher than positivity in all pre-intervention months.
Conclusion
Implementing a targeted managerial and technical intervention in combination resulted in
substantially greater HIV-testing positivity and case identification over the 12 months. These
findings have important implications for HIV testing programs, especially those that target
key populations, who often face access challenges because of criminalization and stigma. The
ODP was able to reach large numbers of previously unreached KP individuals and identify
those who were HIV positive, demonstrating that differentiated, data-driven, community-
focused approaches can help close the 95-95-95 gaps.
Supporting information
S1 Fig.
(PNG)
S1 Table.
(PDF)
Acknowledgments
We would like to acknowledge the Open Doors project staff whose hard work and dedication
were essential to the success of this study.
Author Contributions
Conceptualization: Joseph Kamanga, Kayla Stankevitz, Robert Chiegil, Lameck Nyirenda,
Florence Mulenga, Mulamuli Mpofu, Moses Bateganya.
Data curation: Joseph Kamanga, Lameck Nyirenda, Florence Mulenga, Sam Lubasi.
Formal analysis: Kayla Stankevitz, Andres Martinez, Lameck Nyirenda, Mario Chen.
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 10 / 13
Investigation: Joseph Kamanga, Lameck Nyirenda, Moses Bateganya.
Methodology: Joseph Kamanga, Kayla Stankevitz, Andres Martinez, Lameck Nyirenda, Flor-
ence Mulenga, Mario Chen, Moses Bateganya.
Supervision: Joseph Kamanga.
Validation: Andres Martinez.
Visualization: Kayla Stankevitz, Lameck Nyirenda.
Writing – original draft: Joseph Kamanga, Kayla Stankevitz.
Writing – review & editing: Joseph Kamanga, Kayla Stankevitz, Andres Martinez, Robert
Chiegil, Lameck Nyirenda, Florence Mulenga, Mario Chen, Mulamuli Mpofu, Sam Lubasi,
Moses Bateganya.
References
1. UNAIDS. UNAIDS. Fast-track commitments to end AIDS by 2030. Geneva, Switzerland; 2017.
2. Marsh K, Eaton JW, Mahy M, Sabin K, Autenrieth CS, Wanyeki I, et al. Global, regional and country-
level 90–90–90 estimates for 2018: assessing progress towards the 2020 target. LWW; 2019. https://
doi.org/10.1097/QAD.0000000000002355 PMID: 31490781
3. Doshi RH, Sande E, Ogwal M, Kiyingi H, McIntyre A, Kusiima J, et al. Progress toward UNAIDS 90-90-
90 targets: A respondent-driven survey among female sex workers in Kampala, Uganda. PLoS One.
2018; 13(9):e0201352–e. https://doi.org/10.1371/journal.pone.0201352 PMID: 30231030
4. UNAIDS. UNAIDS special analysis. 2019.
5. Macdonald V, Verster A, Baggaley R. A call for differentiated approaches to delivering HIV services to
key populations. J Int AIDS Soc. 2017; 20(Suppl 4):21658. https://doi.org/10.7448/IAS.20.5.21658
PMID: 28770592
6. Sharma M, Ying R, Tarr G, Barnabas R. Systematic review and meta-analysis of community and facil-
ity-based HIV testing to address linkage to care gaps in sub-Saharan Africa. Nature. 2015; 528(7580):
S77. https://doi.org/10.1038/nature16044 PMID: 26633769
7. Hensen B, Taoka S, Lewis JJ, Weiss HA, Hargreaves J. Systematic review of strategies to increase
men’s HIV-testing in sub-Saharan Africa. AIDS (London, England). 2014; 28(14):2133. https://doi.org/
10.1097/QAD.0000000000000395 PMID: 25062091
8. Suthar AB, Ford N, Bachanas PJ, Wong VJ, Rajan JS, Saltzman AK, et al. Towards universal voluntary
HIV testing and counselling: a systematic review and meta-analysis of community-based approaches.
PLoS medicine. 2013; 10(8):e1001496. https://doi.org/10.1371/journal.pmed.1001496 PMID:
23966838
9. Girault P, Green K, Clement NF, Rahman YA, Adams B, Wambugu S. Piloting a Social Networks Strat-
egy to Increase HIV Testing and Counseling Among Men Who Have Sex with Men in Greater Accra and
Ashanti Region, Ghana. AIDS and behavior. 2015; 19(11):1990–2000. https://doi.org/10.1007/s10461-
015-1069-z PMID: 25903507
10. Mahachi N, Muchedzi A, Tafuma TA, Mawora P, Kariuki L, Semo B-W, et al. Sustained high HIV case-
finding through index testing and partner notification services: experiences from three provinces in Zim-
babwe. J Int AIDS Soc. 2019; 22 Suppl 3(Suppl Suppl 3):e25321–e. https://doi.org/10.1002/jia2.25321
PMID: 31321918
11. PEPFAR. PEPFAR monitoring, evaluation, and reporting indicator reference guide. PEPFAR Washing-
ton (DC); 2020.
12. ICAP Approach to Strategic HIV Testing. Colombia University; 2017.
13. UNDP. Zambia: Removing legal and policy barriers for young key populations in the Zambia HIV
response. 2019.
14. PEPFAR. PEPFAR Zambia Country Operational Plan (COP/ROP) 2019 Strategic Direction Summary.
2019.
15. Zambia Population-Based HIV Impact Assessment: ZAMPHIA 2016. 2017.
16. National AIDS Strategic Framework (NASF) 2017–2021. National HIV/AIDS/STI/TB Council; 2017.
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 11 / 13
17. FHI360. TOTAL QUALITY LEADERSHIP AND ACCOUNTABILITY 2019 [https://www.fhi360.org/sites/
default/files/media/documents/resource-tqla-fact-sheet.pdf.
18. Reh FJ. Understanding Pareto’s principle-the 80–20 rule. The Balance; 2017.
19. Lillie TA, Persaud NE, DiCarlo MC, Gashobotse D, Kamali DR, Cheron M, et al. Reaching the
unreached: Performance of an enhanced peer outreach approach to identify new HIV cases among
female sex workers and men who have sex with men in HIV programs in West and Central Africa. PLoS
One. 2019; 14(4). https://doi.org/10.1371/journal.pone.0213743 PMID: 30943205
20. Newey WK, West KD. A simple, positive semi-definite, heteroskedasticity and autocorrelation consis-
tent covariance matrix. Прикладная эконометрикаNo1 (33) 2014. 2017 May 20: 125.
21. Linden A. Conducting interrupted time-series analysis for single-and multiple-group comparisons. The
Stata Journal. 2015 Jun; 15(2):480–500.
22. Linden A. A comprehensive set of postestimation measures to enrich interrupted time-series analysis.
The Stata Journal. 2017 Mar; 17(1):73–88.
23. StataCorp. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC; 2017.
24. De Cock KM, Barker JL, Baggaley R, El Sadr WM. Where are the positives? HIV testing in sub-Saharan
Africa in the era of test and treat. Aids. 2019; 33(2):349–52. https://doi.org/10.1097/QAD.
0000000000002096 PMID: 30557162
25. Muhula S, Memiah P, Mbau L, Oruko H, Baker B, Ikiara G, et al. Uptake and linkage into care over one
year of providing HIV testing and counselling through community and health facility testing modalities in
urban informal settlement of Kibera, Nairobi Kenya. BMC Public Health. 2016; 16(1):373. https://doi.
org/10.1186/s12889-016-3033-x PMID: 27142727
26. Nash SG FM, Gill ON, Connor N and contributors. HIV testing in England: 2017 report. London: Public
Health England; 2017.
27. Mwango LK, Stafford KA, Blanco NC, Lavoie MC, Mujansi M, Nyirongo N, et al. Index and targeted com-
munity-based testing to optimize HIV case finding and ART linkage among men in Zambia. Journal of
the International AIDS Society. 2020 Jun; 23:e25520. https://doi.org/10.1002/jia2.25520 PMID:
32589360
28. Mahachi N, Muchedzi A, Tafuma TA, Mawora P, Kariuki L, Semo BW, et al. Sustained high HIV case-
finding through index testing and partner notification services: experiences from three provinces in Zim-
babwe. Journal of the International AIDS Society. 2019 Jul; 22:e25321. https://doi.org/10.1002/jia2.
25321 PMID: 31321918
29. Sharma M, Ying R, Tarr G, Barnabas R. Systematic review and meta-analysis of community and facil-
ity-based HIV testing to address linkage to care gaps in sub-Saharan Africa. Nature. 2015 Dec; 528
(7580):S77–85. https://doi.org/10.1038/nature16044 PMID: 26633769
30. Wheeler T, Wolf RC, Kapesa L, Surdo AC, Dallabetta G. Scaling-up HIV responses with key popula-
tions in West Africa. JAIDS Journal of Acquired Immune Deficiency Syndromes. 2015 Mar 1; 68:S69–
73. https://doi.org/10.1097/QAI.0000000000000534 PMID: 25723993
31. Bertozzi SM, Laga M, Bautista-Arredondo S, Coutinho A. Making HIV prevention programmes work.
The Lancet. 2008; 372(9641):831–44. https://doi.org/10.1016/S0140-6736(08)60889-2 PMID:
18687457
32. Kim YM, Kalibala S, Neema S, Lukwago J, Weiss DC. Meaningful involvement of people living with HIV/
AIDS in Uganda through linkages between network groups and health facilities: An evaluation study.
Psychology, Health & Medicine. 2012; 17(2):213–22.
33. Cowan FM, Reza-Paul S, Ramaiah M, Kerrigan DL. Strategies to promote the meaningful involvement
of sex workers in HIV prevention and care. Current Opinion in HIV and AIDS. 2019; 14(5):401–8.
https://doi.org/10.1097/COH.0000000000000562 PMID: 31219886
34. Pines HA, Goodman-Meza D, Pitpitan EV, Torres K, Semple SJ, Patterson TL. HIV testing among men
who have sex with men in Tijuana, Mexico: a cross-sectional study. BMJ open. 2016; 6(2):e010388.
https://doi.org/10.1136/bmjopen-2015-010388 PMID: 26846899
35. Kimbrough LW, Fisher HE, Jones KT, Johnson W, Thadiparthi S, Dooley S. Accessing social
networks with high rates of undiagnosed HIV infection: the social networks demonstration project.
American journal of public health. 2009; 99(6):1093–9. https://doi.org/10.2105/AJPH.2008.139329
PMID: 19372521
36. World Health Organization (WHO)/Health Metrics Network. Framework and standards for country
health information systems: World Health Organization; 2008.
37. Aqil A, Lippeveld T, Hozumi D. PRISM framework: a paradigm shift for designing, strengthening and
evaluating routine health information systems. Health policy and planning. 2009; 24(3):217–28. https://
doi.org/10.1093/heapol/czp010 PMID: 19304786
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 12 / 13
38. Nicol E, Bradshaw D, Uwimana-Nicol J, Dudley L. Perceptions about data-informed decisions: an
assessment of information-use in high HIV-prevalence settings in South Africa. BMC Health Services
Research. 2017; 17(2):765. https://doi.org/10.1186/s12913-017-2641-1 PMID: 29219085
39. World Health Organization (WHO). Using unique patient identifiers for person-centred HIV patient moni-
toring and case surveillance (technical brief: HIV patient monitoring and case surveillance). World
Health Organization; 2017.
PLOS ONE
Improved HIV case finding among key populations in Zambia
PLOS ONE | https://doi.org/10.1371/journal.pone.0258573 December 2, 2021 13 / 13
Available via license: CC BY 4.0
Content may be subject to copyright.