Wu et al.

BMC Public Health (2021) 21:689

https://doi.org/10.1186/s12889-021-10580-8

RESEARCH ARTICLE Open Access

Impact of immediate initiation of

antiretroviral therapy among men who
have sex with men infected with HIV in
Chengdu, southwest China: trends analysis,
2008–2018
Chenyao Wu1, Baiyang Zhang1, Zhen Dai2, Qianwen Zheng1, Zhenhua Duan2, Qinying He2* and Cairong Zhu1*

Abstract
Background: Given the rampant HIV epidemic among men who have sex with men (MSM) in Chengdu, southwest
China, Treat All policy, defined as immediate antiretroviral therapy (ART) initiation after HIV diagnosis, was
implemented since 2014. Real-world research evaluating impacts of immediate ART on HIV epidemics is needed to
optimize policy-making as national and international guidelines have been lowering ART eligibility threshold. The
purpose of this study is to: assess temporal trends of the HIV epidemic and impacts of Treat All policy among MSM;
and lay foundation for HIV-related policy evaluation using longitudinal routine data from health information
systems.
Methods: Data used in this study were HIV sentinel seroprevalence, annual reported HIV cases and ART coverage
rate among MSM in Chengdu from 2008 to 2018, derived from national HIV/AIDS information system. Temporal
trends of the HIV epidemic were described using Joinpoint Regression Program. Interrupted time-series method
was deployed to evaluate Treat All policy.
(Continued on next page)

* Correspondence: [email protected]; [email protected]

2
Department of AIDS&STD Control and Prevention, Chengdu Center for
Disease Control and Prevention, 610041 No. 4 Longxiang Road, Sichuan,
Chengdu, China
1
Department of Epidemiology and Health Statistics, West China School of
Public Health and West China Fourth Hospital, Sichuan University, 610041
No. 17 Section 3, Renmin South Road, Chengdu, Sichuan, China

© The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the
data made available in this article, unless otherwise stated in a credit line to the data.
Wu et al. BMC Public Health (2021) 21:689 Page 2 of 11

(Continued from previous page)

Results: HIV sentinel seroprevalence rose from 11.20% in 2008 to 17.67% in 2013 and Annual Percent Change (APC)
was 8.25% (95% CI − 2.40%, 20.07%), then decreased to 5.17% in 2018 (APC = − 19.63%, 95% CI − 27.54%, −
10.86%). Newly reported HIV cases increased from 168 cases in 2008 to 1232 cases in 2015 (APC = 26.99%, 95% CI
21.32%, 32.93%), and reduced to 1014 cases in 2018 (APC = − 8.80%, 95% CI − 18.45%, 2.01%). ART coverage rate
has been climbing from 11.11% in 2008 to 92.29% in 2018 and Average Annual Percent Change was 16.09% (95%
CI 11.76%, 20.59%). Results of interrupted time-series models showed that compared to an annual increase of 0.87%
during pre-policy period, there was a decline of 3.08% (95% CI − 0.0366%, − 0.0250%) per year of HIV sentinel
seroprevalence since 2014; and compared to an annual increase of 116 cases before 2014, there was an annual
drop of 158 newly reported HIV cases (95% CI − 194.87%, − 121.69%) during the post-policy period.
Conclusions: Immediate ART after HIV diagnosis could potentially curb HIV transmission at population level among
MSM, along with other strategies. Future assessment of HIV prevention and control policy can be carried out using
routinely collected longitudinal data from health information systems.
Keywords: HIV infections, Homosexuality, male, Antiretroviral therapy, highly active, HIV seroprevalence, Health
policy, Interrupted time series analysis

Background Despite the individual-level benefits, impacts of immedi-

The HIV pandemic continues to be a major global pub- ate ART on HIV infection at population level is un-
lic health issue. World Health Organization (WHO) esti- known. Mathematic models predicted that, among
mated there were approximately 37.9 million people MSM, early ART might reduce HIV incidence and
living with HIV (PLHIV) by the end of 2018 [1]. The dis- prevalence [28, 29], pending further empiric evidence
tribution of China’s HIV epidemics is unbalanced by (we used early ART, immediate ART and Treat All
geographic region and subpopulation group. Both preva- interchangeably). Real-world research has been listed as
lence and incidence of HIV are especially high in south- high priority area in MSM HIV prevention and control
west China, and men who have sex with men (MSM) [30], whereas there are limited researches evaluating the
bear a disproportionate burden of HIV [2], accounting effects of HIV policies targeted at MSM in a real-world
for 25.5% of national new HIV infections in 2018, rising setting. Real-world policy evaluation is necessary to pro-
from 2.5% in 2006 [3, 4]. vide empiric evidence for policy makers and health care
MSM, whose risk of acquiring HIV is 22 times higher workers. Hence real-world researches in MSM HIV epi-
than the general population, are listed as key population demics are urgently needed, given the especially high
in the HIV pandemic by the Joint United Nations Pro- HIV incidence among MSM, to mitigate the knowledge
gram on HIV and AIDS (UNAIDS) [5]. Preventive mea- gap and facilitate policy making in MSM HIV prevention
sures such as condom use help contain HIV infection and control [30].
[6], but they don’t reduce infectiousness (defined by Sichuan province, the most populous and developed
plasm viral load) or HIV-related morbidity for PLHIV. region in southwest China, has ranked first nationally
Whilst antiretroviral therapy (ART) can optimize health with annual reported HIV cases for years [31]. The pro-
outcomes for PLHIV and reduce onward HIV transmis- vincial capital of Sichuan province, Chengdu, is one of
sion by lowering plasm viral loads [7–9]. Specially, start- the five urban regions in China where HIV is spreading
ing ART at early stages of infection, compared with a at an alarming rate with substantial increases in MSM
deferred initiation, has been proved to have individual- population [2]. Given the raging HIV epidemic among
level benefits, including less morbidity and mortality and MSM in Chengdu [2], the local Center for Disease Con-
reduced rate of linked partner transmissions [10–17]. trol and Prevention (CDC) determined that Treat All
Based on clinical significance of immediate ART after policy be implemented among MSM since 2014; under
HIV diagnosis and development of new antiretroviral this policy, MSM can receive free ART immediately after
drugs with fewer treatment-related toxic effects [10, 12, HIV diagnosis regardless of CD4+ cell count. Treat All
17–22], CD4+ cell threshold eligible for ART in both na- policy began two years earlier prior to the 2016 national
tional and international guidelines for HIV treatment and WHO guidelines recommending ART be offered to
has been relaxed to shorten time from diagnosis to ART all HIV-positive adults and adolescents, irrespective of
in recent years [23, 24]. Nevertheless, several random- CD4+ cell count. With national and international guide-
ized controlled trials found that HIV incidence in the lines lowering ART threshold, assessment of Treat All
study population wasn’t substantially reduced in the policy should be performed since its implementation, to
context of immediate ART for all PLHIV [25–27]. shed a light on whether “treat all” helps to contain the
Wu et al. BMC Public Health (2021) 21:689 Page 3 of 11

HIV epidemic and set an example for studies of this information from the data can be regularly applied by
kind. policy makers to optimizing HIV prevention efforts. HIV
The purpose of this study, therefore, is to assess tem- sentinel surveillance and routine HIV case reporting
poral trends of the HIV epidemic and ART coverage provide valuable and perhaps the only reliable informa-
among MSM in Chengdu before and after Treat All pol- tion about HIV epidemic situations as well as evidence
icy; to evaluate impacts of Treat All policy on MSM HIV for HIV policy evaluation among MSM [39, 40], under
infection at population level; and to lay foundation for the circumstance that MSM come from a hidden and
HIV policy evaluation using longitudinal administrative stigmatized population whose size is hard to estimate
data, as well as provide evidence for policy makers to [41].
further improve early HIV treatment outcomes.
HIV sentinel seroprevalence
Methods HIV sentinel seroprevalence, defined as the proportion
Study design of the sampled MSM who are tested and confirmed as
This is a retrospective study using longitudinal and rou- HIV infected in the annual sentinel survey which is con-
tinely collected data from national health information ducted at fixed locations and lasts from April to June
system between 2008 and 2018. every year since 2005. It was calculated with number of
sampled MSM in the annual sentinel survey conducted
Data source and indicators in Chengdu as denominator and confirmed HIV cases
Indicators analyzed in this study were: HIV sentinel from those MSM as numerator.
seroprevalence, newly reported HIV cases and ART HIV sentinel surveillance is in the charge of local
coverage rate among MSM whose address was Chengdu CDCs designated by national health authority and exe-
city. The indicators were aggregated or calculated for cuted in accordance with Operational Manual of The
each year between 2008 and 2018, using data derived National AIDS Sentinel Surveillance Program. According
from National HIV/AIDS Comprehensive Response In- to the Operational Manual, MSM found to be previously
formation Management System (CRIMS). After years’ HIV infected should be included, and sample size de-
development since 1990’s, China’s web-based HIV infor- pends on HIV prevalence (250 if prevalence more than
mation system was established in 2005 to promote 10%, 400 otherwise for MSM). The MSM population in
timely reporting and data quality. With improvements HIV sentinel surveillance is defined as men who had oral
and expansion of HIV/AIDS data collection, a multifa- or anal intercourse with other males within the past
ceted web-based HIV/AIDS information system called year. Sampling methods are as below: (1) snowball sam-
CRIMS was put into operation since 1st January 2008. pling; (2) recruitment from venues frequented by MSM
To ensure data quality and completeness, there are laws like gay bars; (3) recruitment via networks. After in-
regulating data collecting process, staff training, oper- formed consent is signed, 3 to 5 mL venous blood is
ational manual for CRIMS and quality assessments car- sampled for HIV serological test and HIV serostatus is
ried out annually, thereby data from CRIMS can be determined by corresponding technical guidance.
deemed as reliable for our analyses.
As part of CRIMS, China’s HIV surveillance comprises Newly reported HIV cases
two main databases: (1) routine online HIV case report- Newly reported HIV cases used in this study were newly
ing by health care providers at all levels, legally man- confirmed cases registered in CRIMS whose route of
dated by national regulations; (2) HIV sentinel HIV infection was male homosexual transmission and
surveillance annually collecting cross-sectional data in- address was Chengdu city. Annual reported HIV cases
cluding HIV serostatus in sentinel sites across China, to were aggregated from 2008 to 2018.
consecutively monitor HIV epidemic situations among Given the hidden and stigmatized nature of MSM
key affected populations [32–35]. Chengdu is a south- [39], it’s deemed that newly reported HIV cases is a
west sentinel site of National HIV Sentinel Surveillance marker for HIV incidence, and trends in new HIV cases
System, and MSM have been its key monitoring group have been found to be consistent with trends in esti-
since 2004. Besides, databases of HIV counseling and mates of HIV incidence [42–44].
testing, ART management and the like are parts of
CRIMS as well. More detailed information about data ART coverage rate
collection, management and quality control can be ART coverage rate is defined as the percentage of living
found in previous study [36]. HIV-positive MSM registered in CRIMS who have been
Data from HIV surveillance (HIV seroprevalence and receiving ART. MSM in the treatment database (after re-
newly reported HIV cases) reflect trends of the HIV epi- ceiving treatment) were linked to their own records in
demic and have been used in many studies [2, 4, 37, 38], the epidemiology database (after confirmed as HIV
Wu et al. BMC Public Health (2021) 21:689 Page 4 of 11

positive). It was calculated with number of registered policy intervention. β2 estimates the change in level of yt
alive MSM HIV cases whose address was Chengdu city due to policy intervention (policy = 0 and policy = 1 indi-
as denominator and linked cases receiving ART as cate before and after the policy intervention respect-
numerator. ively), reflecting immediate impact of the policy. β3
Changes of ART coverage can reflect changes in CD4+ estimates the change in trend of yt after the policy inter-
cell count threshold for ART eligibility. Increasing vention, reflecting the long-term effect of the policy. β2
coverage was found to be related to a decreasing thresh- and β3 represent absolute annual change of outcome
old [13], thereby ART coverage can be used to directly variable. β2 answers whether there’s an immediate im-
reflect effects of the treatment eligibility threshold pact of, while β3 answers whether there’s a long-term or
changes. sustained impact of the policy intervention. εt is the ran-
dom error at time points [46].
Statistical methods Our outcomes of interest were HIV sentinel sero-
Temporal trends analysis was conducted with Joinpoint prevalence and annual reported HIV cases, a visual
Regression Program (Version 4.7.0.0 - February 2019, examination was used to assess the trend or non-
National Cancer Institute) to depict HIV seroprevalence, stationary characteristics of the data. Then possible
newly reported HIV cases and ART coverage among autocorrelation between values at serial time points was
MSM in Chengdu from 2008 to 2018. In this approach, assessed using the Durbin-Watson test, where both out-
Joinpoint models that best fitted the data were built and come variables showed autocorrelation. Hence general-
corresponding temporal trend curves were drawn, as- ized least squares estimator was applied, using Prais-
suming the change in outcome variable is constant over Winston method to correct data autocorrelation [47].
each time partition defined by the transition points The overall significance level was ɑ = 0.05.
(called joinpoints), but varies among different time parti-
tions. Annual Percent Change (APC) and Average An-
Results
nual Percent Change (AAPC) reflect relative change of
HIV sentinel seroprevalence and temporal trend
outcome variable. APC is an estimated annual percent-
HIV sentinel seroprevalence among MSM in Chengdu
age change where the outcomes of interest are assumed
showed a rising trend from 11.20% in 2008 to 17.67% in
to change at a constant percentage of that of the previ-
2013, then dropped all the way to 5.17% in 2018
ous year. AAPC is a summary measure of the trend to
(Table 1). Annual Percent Change (APC) between 2008
describe the average APCs over a period of multiple
and 2013 was 8.25% (95% CI − 2.40%, 20.07%), which
years.
wasn’t statistically significant, entailing uncertainty about
Interrupted time-series (ITS) method was applied to
the change direction of the trend before 2014. The sig-
evaluate impacts of Treat All policy on HIV transmis-
nificantly negative APC of 2013–2018 (− 19.63%, 95% CI
sion at population level among MSM [45]. By using lon-
− 27.54%, − 10.86%) reflected a downward trend of HIV
gitudinal and routinely collected data from health
seroprevalence after Treat All policy intervention (Fig. 1).
systems before and after a policy intervention, ITS ana-
Besides, Average Annual Percent Change (AAPC) of
lysis can evaluate impacts of health policies on outcome
2008–2018 (− 6.72%, 95% CI − 12.04%, − 1.09%) was sig-
of interest without requiring a control site. This method
nificantly below zero, indicating an overall declining
was based on fitting segmented linear regression model,
trend over the observation period.
which divided the time series into pre-policy (including
6 data points from 2008 to 2013) and post-policy (in-
cluding 5 data points from 2014 to 2018) segments. Newly reported HIV cases and temporal trend
Analysis was performed using Stata 14.0 (StataCorp. Newly reported HIV cases among MSM in Chengdu
2014). rapidly increased from 168 cases in 2008 to 1232 cases
The following model was established for ITS analysis in 2015, then decreased to 1014 cases in 2018, remaining
[46]: at a relatively high level during post-policy period com-
yt = β0 + β1*time + β2*policy + β3*postslope + εt, where pared with that in earlier years (Table 1). APC was
yt was the outcome variable at different time points, 26.99% (95% CI 21.32%, 32.93%) between 2008 and
variable time indicates time points and was coded num- 2015, suggesting a significant upward trend during this
ber 1–11 corresponding to 2008–2018 in year. Variable period. APC of 2015 to 2018 was − 8.80% (95% CI −
postslope was coded 0 up to the last point before the 18.45%, 2.01%), nonsignificant and entailing uncertainty
policy intervention and coded sequentially from 1 there- about the change direction of the trend after 2014
after. β0 estimates the baseline level of yt at the begin- (Fig. 2). While AAPC over 2008 to 2018 was 14.99%
ning of the time series. β1 estimates the structural trend (95% CI 10.79%, 19.34%), with an upper 95% CI limit
or natural growth rate of yt, independently from the smaller than the lower 95% CI limit of the APC of 2008
Wu et al. BMC Public Health (2021) 21:689 Page 5 of 11

Table 1 HIV sentinel seroprevalence, newly reported HIV cases and ART coverage rate among MSM in Chengdu between 2008 and
2018
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
a
HIV sentinel seroprevalence (%) 11.20 13.33 14.15 15.50 13.90 17.67 13.83 12.90 7.90 8.78 5.17
Newly reported HIV cases b 168 340 367 528 639 824 1061 1232 1124 1021 1014
c
ART coverage rate (%) 11.11 14.56 25.71 40.24 47.16 57.51 65.57 78.40 85.21 90.21 92.29
a
calculated with number of sampled MSM in the annual sentinel survey conducted in Chengdu as denominator and confirmed HIV cases from those MSM
as numerator
b
newly confirmed HIV cases in CRIMS whose route of infection was male homosexual transmission and address was Chengdu city
c
calculated with number of alive MSM HIV cases registered in CRIMS whose address was Chengdu city as denominator and linked cases receiving ART
as numerator

to 2015, indicating a slower growing rate on average Results of Model 1(Table 2) demonstrated that no sig-
over 2008 to 2018 than that during 2008 to 2015. nificant change in the level of HIV sentinel seropreva-
lence was observed immediately after the policy roll-out.
The statistically significant natural growth rate showed
ART coverage rate and temporal trend
an absolute year-to-year increase of 0.87%, indicating
ART coverage rate among MSM in Chengdu has been
there would have been an annual rise of 0.87% on aver-
climbing over the span of 2008 to 2018: from 11.11% in
age of HIV sentinel seroprevalence had the treatment
2008 to 92.29% in 2018 (Table 1). APC between 2008
policy change not been implemented. The change in
and 2015 (21.30%, 95% CI 13.45%, 29.68%) was higher
trend showed a significant absolute decrease of 3.08% of
than that of 2015 to 2018 (APC = 4.79%, 95% CI 2.03%,
HIV sentinel seroprevalence on average each year since
7.62%), implying ART coverage rate grew quickly prior
the policy implementation. Model 1 predicted that HIV
to 2015 and then slightly increased, gradually stabilizing
sentinel seroprevalence among MSM in Chengdu would
at a high level ever since Treat All policy was imple-
have reached 20.80% in 2018 without the intervention,
mented (Fig. 3).
while the actual figure was 5.17% for that year. There-
fore, the absolute effect of Treat All policy was a signifi-
Interrupted time-series analysis cant drop of 15.63% in HIV seroprevalence, equating to
Table 2 displays the results of segmented linear regres- a relative 75.14% decrease, four years after the treatment
sion model 1 and 2 built to evaluate the impacts of Treat policy intervention.
All policy on HIV sentinel seroprevalence and annual re- According to Model 2 (Table 2), annual reported HIV
ported HIV cases respectively among MSM in Chengdu. cases had an average increase of about 116 cases per

Fig. 1 Temporal trend of HIV sentinel seroprevalence among MSM in Chengdu between 2008 and 2018. *Indicates that the Annual Percent
Change (APC) is significantly different from zero at the alpha = 0.05 level. Joinpoint: year 2013
Wu et al. BMC Public Health (2021) 21:689 Page 6 of 11

Fig. 2 Temporal trend of newly reported HIV cases among MSM in Chengdu between 2008 and 2018. *Indicates that the Annual Percent Change
(APC) is significantly different from zero at the alpha = 0.05 level. Joinpoint: year 2015

year during the pre-policy period, indicating 116 more indicating on average some 158 fewer HIV cases were
cases would have been reported annually than the previ- reported annually since the policy implementation.
ous year had the treatment policy intervention not been Model 2 estimated that annual reported HIV cases
implemented. A significant increase immediately after among MSM in Chengdu would have been 1344 cases in
the policy implementation was observed, and we found a 2018 without the intervention, while the real number
downward trend where there was a decrease of about was 1014 cases that year. Therefore, the absolute effect
158 cases each year during the post-policy period, of Treat All policy was a significant reduction by 330

Fig. 3 Temporal trend of ART coverage rate among MSM in Chengdu between 2008 and 2018. *Indicates that the Annual Percent Change (APC)
is significantly different from zero at the alpha = 0.05 level. Joinpoint: year 2015
Wu et al. BMC Public Health (2021) 21:689 Page 7 of 11

Table 2 Results of segmented regression models for HIV sentinel seroprevalence and annual reported HIV cases among MSM before
and after Treat All policy
Coefficient Standard error P value 95% Confidence interval (CI)
a
HIV sentinel seroprevalence (Model 1)
β0 (Baseline) 0.1123 0.0059 < 0.001*** (0.0982, 0.1264)
β1 (Structural trend) 0.0087 0.0015 0.001*** (0.0051, 0.0124)
β2 (Change in level) − 0.0005 0.0087 0.959 (− 0.0210, 0.0201)
β3 (Change in trend) − 0.0308 0.0025 < 0.001*** (− 0.0366, − 0.0250)
Newly reported HIV cases b (Model 2)
β0 (Baseline) 63.35 37.33 0.133 (−24.92, 151.62)
β1 (Structural trend) 116.38 9.68 < 0.001*** (93.48, 139.28)
β2 (Change in level) 463.23 54.65 < 0.001*** (334.01, 592.45)
β3 (Change in trend) −158.28 15.47 < 0.001*** (−194.87, −121.69)
a
calculated with number of sampled MSM in the annual sentinel survey conducted in Chengdu as denominator and confirmed HIV cases from those MSM
as numerator
b
newly confirmed HIV cases in CRIMS whose route of infection was male homosexual transmission and address was Chengdu city

newly reported HIV cases, a relative 24.55% decrease in to that in 2008). After CD4+ cell threshold for ART eli-
other words, four years after the treatment policy gibility was abolished, great progress has been made in
intervention. terms of the proportion of HIV-positive MSM who’re re-
ceiving ART, which was also observed in previous study
Discussion [13]. Scaling up ART coverage has been important com-
Our study showed HIV sentinel seroprevalence and an- ponent in actions against the growing HIV epidemic
nual reported HIV cases declined, and ART coverage among MSM [30]. Increased ART coverage reflects dir-
rate increased after Treat All policy was implemented in ect effect of lowering the treatment eligibility threshold
MSM population in Chengdu, illuminating the potential (CD4+ cell count), serving as the intermediate path from
of immediate ART initiation to contain the HIV epi- decreased threshold for ART eligibility to new HIV in-
demic. It also suggested that in real-world conditions, fection reduction. While real-world studies evaluating
early ART could be potentially efficacious in curbing impacts of lowering treatment threshold on ART cover-
HIV transmission at population level, which were con- age and HIV epidemics were scarce, series of clinical tri-
sistent with previous studies [48, 49]. als concerning this aspect were carried out. Both the
Treat All policy contributed to reducing new HIV in- TasP (ANRS 12249) and the HPTN 071 (PopART) trials
fections at population level among MSM and had a sus- conducted in South Africa didn’t see increased ART
tained impact. According to ITS analysis, compared to coverage or reduced incidence in the immediate ART
upward trends during the pre-policy period, there were arm compared to the control group [27, 51]. The ab-
statistically significant downward trends in HIV sentinel sence of decreased incidence can be explained by inci-
seroprevalence and annual reported HIV cases since the dent HIV infections from outside of the study
policy intervention, indicating the growing trends were population in light of multiple sex partners behavior
reversed after Treat All policy implementation. In- among MSM [52], suggesting ART accessibility for “all”
creased ART coverage is associated with decreased HIV is critical in fighting against HIV. Besides, low treatment
incidence [13, 49, 50], and the huge progress made in coverage in the trials added to the evidence of challen-
ART coverage reflected Treat All policy helped ensure ging situations where lack of funding, manpower and so
more efficient ART delivery and prompted more asymp- on may emerge in trial settings, not to mention in real-
tomatic MSM living with HIV to undergo treatment, world settings with more restricted budgets. Neverthe-
thus potentially turning down population-level HIV less, Treat All policy bore inspiring outcomes, which can
transmission, which were in line with similar real-world be due to great work and robust funding from local and
studies to show expanded access to ART helps curb the national authorities and organizations, plus budgets
growth of HIV epidemics [48, 49]. guarantee from China’s National Free Antiretroviral
The second “90” of UNAIDS’s “90–90–90” goals, Treatment Program [53].
which is 90% of those diagnosed with HIV receive ART, Lagged effect, referring to impacts taking time to
was achieved among MSM in this region. As is the case manifest after intervention [54], wasn’t observed in our
in this study, ART coverage rate reached 92.29% among study, which can be due to that the HIV epidemic was
MSM in 2018 (an absolute increase of 81.18% compared in dire situation among MSM in Chengdu and MSM
Wu et al. BMC Public Health (2021) 21:689 Page 8 of 11

since has been the key intervention group, so that quick to be taken from a systematic perspective. First of all, ex-
and great efforts were taken to roll out Treat All policy pand HIV testing in that it underlies implementation of
citywide. Besides, concurrent endeavors like expanded nearly all other interventions. As supplement to volun-
HIV testing and HIV/AIDS supportive environment tary counseling and testing, home-based “HIV self-
construction may also play a part. testing” was suggested by WHO to improve HIV testing
It should be noted that those recruited in the annual [59]. There are self-testing kits available now in
sentinel surveillance at venues frequented by MSM (like Chengdu. Moreover, to encourage engagement, HIV
gay bars, bath room, etc.) are more likely to be sexually testing services can be applied on CBO’s websites and
active with more risky sexual behaviors, thereby HIV provided in places frequented by MSM, which have
sentinel seroprevalence may be somewhat overestimated helped improve testing volumes in Chengdu. Second,
[37]. But this influence isn’t prone to vary over time. strengthen collaboration with international, governmen-
Moreover, the proportion of previously HIV infected tal and non-governmental organizations to guarantee ro-
MSM in the surveyed sample has remained stable since bust funding. Chengdu CDC worked with UNAIDS to
the national HIV surveillance system scale-up in 2008, successfully execute strategic planning of AIDS preven-
so HIV seroprevalence since 2008 can reasonably reflect tion and treatment for MSM (2011–2015) to facilitate
changes in the trend of new HIV infections among field work and community empowering. Besides, during
MSM [37]. Hence HIV seroprevalence obtained from China’s “12th five-year plan on AIDS prevention and
serial cross-sectional surveys could still reveal informa- treatment”, Chengdu CDC collaborated with China
tion about changing trends of the HIV epidemic and be CDC, the central institution in Beijing, to conduct ques-
used to evaluate effects of HIV prevention and control tionnaire interviewing and serological testing among
efforts among MSM. MSM. Finally, a Continuum of HIV Services should be
Additionally, despite a descending trend since 2014, established with CBOs, CDCs and hospitals coordinated
annual reported HIV cases among MSM remained at a as a whole, to improve health care efficiency, expand
relatively high level, compared with that in the pre- coverage and facilitate sustained accessibility. One-stop
policy period. This may be attributed to the following as- HIV testing, confirmation and treatment services are
pects: optimization of national HIV information system now available at the same hospital in Chengdu. The
making sure more complete and timely case reporting; “Tongle” (“happy together”) organization, a Chengdu-
increased number of MSM community-based organiza- based CBO which serves as a hinge to link MSM, local
tions (CBOs) and their frequent activities prompting CDCs and hospitals, sets as a good example that support
more hidden HIV-positive MSM to be detected; and ris- from peers and CBO workers trusted by MSM are ne-
ing number of HIV-positive MSM with an ongoing HIV cessary while delivering HIV services.
epidemic in MSM population whose size is growing as
well. Limitations
Our study reveals practical value of Treat All policy to There are some limitations in this study. Firstly, HIV
potentially reduce HIV infection at population level, sentinel seroprevalence is obtained using non-probability
while challenges remain at the frontline. Treat All policy sampling methods, which might lead to selection bias
demands enormous investments to achieve effective and undermine representativeness [39]. However, prob-
ART expansion and sustained viral suppression, the ul- ability sampling is not practical considering the hidden
timate goal of HIV control, which requires sustained nature of MSM population. The same sampling design
ART accessibility. However, this goal is yet a challenge used in each survey round guarantees any bias caused by
to be overcome financially and administratively, espe- nonrandom sampling can be assumed to be relatively
cially for low- and middle-income countries where re- stable [40], so samples of sub-groups of MSM still pro-
sources of all sorts are limited [55]. Furthermore, due to vide important information of the HIV epidemic as well
social and structural factors like stigmatization, key pop- as effects of policy intervention. Secondly, the number of
ulations affected by HIV including MSM have poor ac- time points used in ITS models was relatively small. Al-
cess to HIV services. But they are groups who benefit though there were only 6 and 5 time points before and
most from early ART [56]. Core-group theory posits that after the policy intervention respectively, the analysis
prevention among a relatively few at the highest risk of outcomes presented statistically significant trends, indi-
acquisition and transmission, can protect many along a cating robust statistical power to detect changes in the
potential transmission chain [57, 58]. Therefore, MSM trends. Thirdly, viral suppression rate, the third UNAIDS
have been and will still be a priority group in HIV pre- “90” goal, is supposed to be taken into consideration in
vention and control strategies globally. treatment policy evaluation. It wasn’t included due to
To maximize the benefits of Treat All policy, experi- data incompletion in the early period when the informa-
ence can be learned from Chengdu and measures need tion system was being built up. Further research should
Wu et al. BMC Public Health (2021) 21:689 Page 9 of 11

work on evaluation of viral suppression to fully interpret Availability of data and materials
impacts of HIV policy interventions. Finally, an observa- The data that support the findings of this study are not publicly available
due to data protection and confidentiality and restrictions apply to the
tional study, our study was based on an ecological per- availability of these data, which were used under license for the current
spective, potential confounding might not be controlled. study. Data are however available from the authors upon reasonable request
Any causal relationships can’t be assumed between the and with permission of Chengdu CDC.

policy intervention and the analyzed indicators. Though Declarations

ITS analysis didn’t control for other events that may
have influenced the outcome, single group time series Ethics approval and consent to participate
This study was considered to be a policy evaluation research and didn’t
still address threats to internal validity and provide a require research ethical approval by the Ethics Committee. We inquired IRB
methodologically acceptable design for studying inter- of Sichuan University and the board suggested ethics approval be waived
vention effects [54]. Besides, our study is limited in given that all data have been completely deidentified and aggregated
before access and analysis. All the data used in the study were collected
Chengdu, hence careful consideration and comparisons with permission and in accordance with national regulation by Chengdu
of geographical and subpopulation patterns of HIV epi- Center for Disease Control and Prevention.
demics among populations at provincial and national
level should be taken before cautious generalization to Consent for publication
Not applicable.
broader populations.
Competing interests
Conclusions The authors declare that they have no competing interests.
The policy of immediate ART after diagnosis in MSM Received: 27 July 2020 Accepted: 7 March 2021
population in Chengdu improved ART coverage, enab-
ling more MSM living with HIV to receive treatment.
And the downward trends of HIV sentinel seropreva- References
1. World Health Organization (WHO). HIV/AIDS Key facts (15 November 2019)
lence and newly reported HIV cases demonstrated the [Available from: https://www.who.int/en/news-room/fact-sheets/detail/hiv-a
potential of Treat All policy to turn the tide of the ram- ids.
pant HIV epidemic in MSM population. A real-world 2. Zhang L, Chow EP, Jing J, Zhuang X, Li X, He M, et al. HIV prevalence in
China: integration of surveillance data and a systematic review. Lancet
evaluation of treatment policy intervention among Infect Dis. 2013;13(11):955–63. https://doi.org/10.1016/S1473-3099(13)70245-
MSM, our study laid foundation for future assessment of 7.
HIV prevention and control policy using longitudinal 3. National Health Commission of the People's Republic of China. Transcript of
Regular Press Conference National Health Commission of the People's
routine data from health information systems, particu- Republic of China. 2019 [Available from: http://www.nhc.gov.cn/wjw/xwdt/2
larly in resource-limited low- and middle-income coun- 01811/b0ca3817224e44899a3248a0f6e48948.shtml.
tries. In the meanwhile, further coordinated strategies 4. Wang L, Wang L, Norris JL, Li DM, Guo W, Ding ZW, et al. HIV prevalence
and influencing factors analysis of sentinel surveillance among men who
are needed to maximize the benefits of early ART to have sex with men in China, 2003-2011. Chin Med J. 2012;125(11):1857–61.
contain HIV epidemics. 5. UNAIDS. Global HIV & AIDS statistics — 2019 fact sheet. UNAIDS. Available
from: https://www.unaids.org/en/resources/fact-sheet. Accessed 31 Dec
Abbreviations 2019.
ART: Antiretroviral therapy/treatment; AIDS: Acquired immunodeficiency 6. Riddell J, Amico KR, Mayer KH. HIV Preexposure prophylaxis: a review. JAMA.
syndrome; APC: Annual Percent Change; AAPC: Average Annual Percent 2018;319(12):1261–8. https://doi.org/10.1001/jama.2018.1917.
Change; CDC: Center for Disease Control and Prevention; CRIMS: National 7. Johnson LF, Mossong J, Dorrington RE, Schomaker M, Hoffmann CJ, Keiser
HIV/AIDS Comprehensive Response Information Management System; O, et al. Life expectancies of South African adults starting antiretroviral
CI: Confidence interval; HIV: Human immunodeficiency virus; ITS: Interrupted treatment: collaborative analysis of cohort studies. PLoS Med. 2013;10(4):
time series; MSM: Men who have sex with men; PLHIV: People living with e1001418.
HIV; UNAIDS: The Joint United Nations Program on HIV and AIDS; 8. Mills EJ, Bakanda C, Birungi J, Chan K, Ford N, Cooper CL, et al. Life
WHO: World Health Organization Expectancy of Persons Receiving Combination Antiretroviral Therapy in
Low-Income Countries: A Cohort Analysis From Uganda. Ann Internal Med.
Acknowledgements 2011;155(4):209.
We thank all the staff at Chengdu CDC for their field work of conducting the 9. O'Connor J, Vjecha MJ, Phillips AN, Angus B, Cooper D, Grinsztejn B,
annual sentinel surveillance survey and their routine work of collecting and Lopardo G, Das S, Wood R, Wilkin A, Klinker H, Kantipong P, Klingman KL,
archiving loads of data every day. We also thank health care practitioners at Jilich D, Herieka E, Denning E, Abubakar I, Gordin F, Lundgren JD, INSIGHT
every level for HIV detecting and timely registering in CRIMS, as well as START study group. Effect of immediate initiation of antiretroviral therapy
community workers who help convene MSM to participate in sentinel on risk of severe bacterial infections in HIV-positive people with CD4 cell
surveillance survey and facilitate all the field work. counts of more than 500 cells per μL: secondary outcome results from a
randomised controlled trial. Lancet HIV. 2017;4(3):e105–e12. https://doi.org/1
Authors’ contributions 0.1016/S2352-3018(16)30216-8.
CZ and QH conceived of the study. CZ, QH and CW made the statistical 10. Lodi S, Sharma S, Lundgren JD, Phillips AN, Cole SR, Logan R, Agan BK,
plan. QH, ZD1, ZD2 and QZ collected and aggregated the research data. CW Babiker A, Klinker H, Chu H, Law M, Neaton JD, Hernán MA, INSIGHT
and BZ performed the analyses. CW wrote the manuscript. All authors Strategic Timing of AntiRetroviral Treatment (START) study group. The per-
advised on the results interpretation and helped improve the manuscript. All protocol effect of immediate versus deferred antiretroviral therapy initiation.
authors have read and approved the final manuscript. AIDS. 2016;30(17):2659–63. https://doi.org/10.1097/QAD.0000000000001243.
11. Lifson AR, Grund B, Gardner EM, Kaplan R, Denning E, Engen N, Carey CL,
Funding Chen F, Dao S, Florence E, Sanz J, Emery S, INSIGHT START Study Group.
No funding. Improved quality of life with immediate versus deferred initiation of
Wu et al. BMC Public Health (2021) 21:689 Page 10 of 11

antiretroviral therapy in early asymptomatic HIV infection. AIDS. 2017;31(7): What’s New 2015 [Available from: http://apps.who.int/iris/bitstream/10665/1
953–63. https://doi.org/10.1097/QAD.0000000000001417. 98064/1/9789241509893_eng.pdf?ua=1.
12. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, 24. Mayer KH, Krakower DS. Antiretrovirals for HIV treatment and prevention the
Kumarasamy N, Hakim JG, Kumwenda J, Grinsztejn B, Pilotto JH, Godbole challenges of success. Jama-Journal of the American Medical Association.
SV, Mehendale S, Chariyalertsak S, Santos BR, Mayer KH, Hoffman IF, 2016;316(2):151–3. https://doi.org/10.1001/jama.2016.8902.
Eshleman SH, Piwowar-Manning E, Wang L, Makhema J, Mills LA, de Bruyn 25. Havlir D, Charlebois E, Balzer L, Clark T, Kwarisiima D, Ayieko J, et al. SEARCH
G, Sanne I, Eron J, Gallant J, Havlir D, Swindells S, Ribaudo H, Elharrar V, community cluster randomized study of HIV "test and treat" using multi-
Burns D, Taha TE, Nielsen-Saines K, Celentano D, Essex M, Fleming TR, HPTN disease approach and streamlined care in rural Uganda and Kenya. J Int
052 Study Team. Prevention of HIV-1 infection with early antiretroviral AIDS Soc. 2018;21:164–5.
therapy. N Engl J Med. 2011;365(6):493–505. https://doi.org/10.1056/ 26. Iwuji CC, Orne-Gliemann J, Larmarange J, Balestre E, Thiebaut R, Tanser F,
NEJMoa1105243. Okesola N, Makowa T, Dreyer J, Herbst K, McGrath N, Bärnighausen T, Boyer
13. McManus H, Callander D, Donovan B, Russell DB, O'Connor CC, Davies SC, S, de Oliveira T, Rekacewicz C, Bazin B, Newell ML, Pillay D, Dabis F,
Lewis DA, Hellard ME, Chen MY, Petoumenos K, Varma R, Cogle A, Boyd Bärnighausen T, Herbst K, Iwuji C, Makowa T, Naidu K, Newell ML, Okesola
MA, Grulich A, Pollard J, Medland N, Fairley CK, Guy RJ. Early initiation of N, de Oliveira T, Pillay D, Rochat T, Tanser F, Viljoen J, Zuma T, McGrath N,
antiretroviral therapy for people newly diagnosed with HIV infection in Balestre E, Dabis F, Karcher S, Orne-Gliemann J, Plazy M, Prague M, Thiébaut
Australia: trends and predictors, 2004-2015. Med J Aust. 2019;210(6):269–75. R, Tiendrebeogo T, Boyer S, Donfouet H, Gosset A, March L, Protopopescu
https://doi.org/10.5694/mja2.50006. C, Spire B, Calmy A, Larmarange J, Inghels M, Diallo H, Calvez V, Derache A,
14. Caby F, French Hosp Database HF-A. CD4(+)/CD8(+) ratio restoration in Marcelin AG, Dray-Spira R, Lert F, el Farouki K, Lessells R, Freedberg K, Imrie
long-term treated HIV-1-infected individuals. Aids. 2017;31(12):1685–95. J, Chaix ML, Newell C, Hontelez J, Bazin B, Rekacewicz C. Universal test and
15. May MT, Gompels M, Delpech V, Porter K, Orkin C, Kegg S, Hay P, Johnson treat and the HIV epidemic in rural South Africa: a phase 4, open-label,
M, Palfreeman A, Gilson R, Chadwick D, Martin F, Hill T, Walsh J, Post F, community cluster randomised trial. Lancet HIV. 2018;5(3):E116–E25. https://
Fisher M, Ainsworth J, Jose S, Leen C, Nelson M, Anderson J, Sabin C, UK doi.org/10.1016/S2352-3018(17)30205-9.
Collaborative HIV Cohort (UK CHIC) Study. Impact on life expectancy of HIV- 27. Hayes RJ, Donnell D, Floyd S, Mandla N, Bwalya J, Sabapathy K, et al. Effect
1 positive individuals of CD4(+) cell count and viral load response to of Universal Testing and Treatment on HIV Incidence - HPTN 071 (PopART).
antiretroviral therapy. Aids. 2014;28(8):1193–202. https://doi.org/10.1097/QA New England J Med. 381(3):207–18.
D.0000000000000243. 28. Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary
16. Zhang F, Dou Z, Ma Y, Zhang Y, Zhao Y, Zhao D, Zhou S, Bulterys M, Zhu H, HIV testing with immediate antiretroviral therapy as a strategy for
Chen RY. Effect of earlier initiation of antiretroviral treatment and increased elimination of HIV transmission: a mathematical model. Lancet. 2009;
treatment coverage on HIV-related mortality in China: a national 373(9657):48–57. https://doi.org/10.1016/S0140-6736(08)61697-9.
observational cohort study. Lancet Infect Dis. 2011;11(7):516–24. https://doi. 29. Sorensen SW, Sansom SL, Brooks JT, Marks G, Begier EM, Buchacz K, et al. A
org/10.1016/S1473-3099(11)70097-4. mathematical model of comprehensive test-and-treat services and HIV
17. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy incidence among men who have sex with men in the United States. PLoS
N, Hakim JG, Kumwenda J, Grinsztejn B, Pilotto JH, Godbole SV, Chariyalertsak ONE [Electronic Resource]. 2012;7(2):e29098.
S, Santos BR, Mayer KH, Hoffman IF, Eshleman SH, Piwowar-Manning E, Cottle 30. Beyrer C, Baral SD, Collins C, Richardson ET, Sullivan PS, Sanchez J, Trapence
L, Zhang XC, Makhema J, Mills LA, Panchia R, Faesen S, Eron J, Gallant J, Havlir G, Katabira E, Kazatchkine M, Ryan O, Wirtz AL, Mayer KH. The global
D, Swindells S, Elharrar V, Burns D, Taha TE, Nielsen-Saines K, Celentano DD, response to HIV in men who have sex with men. Lancet. 2016;388(10040):
Essex M, Hudelson SE, Redd AD, Fleming TR, HPTN 052 Study Team. 198–206. https://doi.org/10.1016/S0140-6736(16)30781-4.
Antiretroviral therapy for the prevention of HIV-1 transmission. N Engl J Med. 31. Chinese Center for Disease Control. Public Health Science Data Center.
2016;375(9):830–9. https://doi.org/10.1056/NEJMoa1600693. Available from: http://www.phsciencedata.cn/Share/edtShareNew.jsp?id=394
18. Petersen M, Balzer L, Kwarsiima D, Sang N, Chamie G, Ayieko J, Kabami J, 03. Accessed 6 Feb 2020.
Owaraganise A, Liegler T, Mwangwa F, Kadede K, Jain V, Plenty A, Brown L, 32. Mao Y, Wu Z, Poundstone K, Wang C, Qin Q, Ma Y, et al. Development of a
Lavoy G, Schwab J, Black D, van der Laan M, Bukusi EA, Cohen CR, Clark TD, unified web-based national HIV/AIDS information system in China. Int J
Charlebois E, Kamya M, Havlir D. Association of Implementation of a Epidemiol. 2010;39:II79–89.
universal testing and treatment intervention with HIV diagnosis, receipt of 33. Sun J, Liu H, Li H, Wang L, Guo H, Shan D, et al. Contributions of
antiretroviral therapy, and viral suppression in East Africa. Jama-J Am Med international cooperation projects to the HIV/AIDS response in China. Int J
Assoc. 2017;317(21):2196–206. https://doi.org/10.1001/jama.2017.5705. Epidemiol. 2010;39:II14–20.
19. Makhema J, Wirth KE, Pretorius Holme M, Gaolathe T, Mmalane M, Kadima 34. Wang L, Wang N. HIV/AIDS epidemic and the development of
E, Chakalisa U, Bennett K, Leidner J, Manyake K, Mbikiwa AM, Simon SV, comprehensive surveillance system in China with challenges. Chin Med J.
Letlhogile R, Mukokomani K, van Widenfelt E, Moyo S, Lebelonyane R, 2010;123(23):3495–500.
Alwano MG, Powis KM, Dryden-Peterson SL, Kgathi C, Novitsky V, Moore J, 35. Zhang L, Chow EPF, Zhang J, Jing J, Wilson DP. Describing the Chinese HIV
Bachanas P, Abrams W, Block L, el-Halabi S, Marukutira T, Mills LA, Sexton C, surveillance system and the influences of political structures and social
Raizes E, Gaseitsiwe S, Bussmann H, Okui L, John O, Shapiro RL, Pals S, stigma. Open AIDS J. 2012;6(1):163–8. https://doi.org/10.2174/18746136012
Michael H, Roland M, DeGruttola V, Lei Q, Wang R, Tchetgen Tchetgen E, 06010163.
Essex M, Lockman S. Universal testing, expanded treatment, and incidence 36. Xu J, Mao Y. Current status and prospect of China's national HIV/AIDS
of HIV infection in Botswana. N Engl J Med. 2019;381(3):230–42. https://doi. comprehensive response information management system (CRIMS). Chinese
org/10.1056/NEJMoa1812281. JAIDS STD. 2016;22(9):756–8.
20. Saag MS, Benson CA, Gandhi RT, Hoy JF, Landovitz RJ, Mugavero MJ, Sax PE, 37. Ge L, Li D, Li P, Guo W, Cui Y. Population specific sentinel surveillance for
Smith DM, Thompson MA, Buchbinder SP, del Rio C, Eron JJ Jr, Fätkenheuer HIV infection, syphilis and HCV infection in China during 2010-2015. DIS
G, Günthard HF, Molina JM, Jacobsen DM, Volberding PA. Antiretroviral SURVEILLANCE. 2017;32(2):111–7.
drugs for treatment and prevention of HIV infection in adults 2018 38. Qiao D, Liu Z, Dong Z, Li W, Yang H. Sentinel surveillance and factors
recommendations of the international antiviral society-USA panel. Jama-J affecting HIV infection among MSM in Zhengzhou, 2016-2018. Modern
Am Med Assoc. 2018;320(4):379–96. https://doi.org/10.1001/jama.2018.8431. Preventive Medicine. 2019;46(19):3603–7.
21. The INSIGHT START Study Group. Initiation of antiretroviral therapy in early 39. Strickler H, Hoover DR, Dersimonian R. Problems in interpreting HIV sentinel
asymptomatic HIV infection. N Engl J Med. 2015;373(9):795–807. https://doi. seroprevalence studies. Ann Epidemiol. 1995;5(6):447–54. https://doi.org/10.1
org/10.1056/NEJMoa1506816. 016/1047-2797(95)00070-4.
22. Danel C, Moh R, Gabillard D, Badje A, Le Carrou J, Ouassa T, et al. A trial of 40. Zhang D, Bi P, Hiller JE, Lv F. Web-based HIV/AIDS behavioral surveillance
early Antiretrovirals and isoniazid preventive therapy in Africa. N Engl J Med. among men who have sex with men: potential and challenges. Int J Infect
2015;373(9):808–22. https://doi.org/10.1056/NEJMoa1507198. Dis. 2008;12(2):126–31. https://doi.org/10.1016/j.ijid.2007.06.007.
23. World Health Organization (WHO). Policy Brief: Consolidated Guidelines on 41. Beyrer C, Baral SD, van Griensven F, Goodreau SM, Chariyalertsak S, Wirtz AL,
the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Brookmeyer R. Global epidemiology of HIV infection in men who have sex
Wu et al. BMC Public Health (2021) 21:689 Page 11 of 11

with men. Lancet. 2012;380(9839):367–77. https://doi.org/10.1016/S0140-673 Publisher’s Note

6(12)60821-6. Springer Nature remains neutral with regard to jurisdictional claims in
42. Chen KT, Chang HL, Chen CT, Chen YA. The changing face of the HIV published maps and institutional affiliations.
epidemic in Taiwan: a new challenge for public health policy strategies.
AIDS Patient Care STDs. 2009;23(3):195–201. https://doi.org/10.1089/apc.2
008.0047.
43. Das M, Chu PL, Santos GM, Scheer S, Vittinghoff E, McFarland W, et al.
Decreases in community viral load are accompanied by reductions in new
HIV infections in San Francisco. PLoS One. 2010;5(6):e11068.
44. Wand H, Wilson D, Yan P, Gonnermann A, McDonald A, Kaldor J, Law M.
Characterizing trends in HIV infection among men who have sex with men
in Australia by birth cohorts: results from a modified back-projection
method. J Int AIDS Soc. 2009;12(1):19. https://doi.org/10.1186/1758-2
652-12-19.
45. Bernal JL, Cummins S, Gasparrini A. Interrupted time series regression for
the evaluation of public health interventions: a tutorial. Int J Epidemiol.
2017;46(1):348–55. https://doi.org/10.1093/ije/dyw098.
46. Lagarde M. How to do (or not to do) ... Assessing the impact of a policy
change with routine longitudinal data. Health Policy Plan. 2011;27(1):76–83.
https://doi.org/10.1093/heapol/czr004.
47. Judge GG. The theory and practice of econometrics: New York: Wiley; 1985.
48. Frank Tanser TB. Erofili Grapsa, Jaffer Zaidi, Marie-Louise Newell. High
coverage of ART associated with decline in risk of HIV Acquisition in Rural
KwaZulu-Natal, South Africa. Science. 2013;339(6122):966–71. https://doi.
org/10.1126/science.1228160.
49. Montaner JS, Lima VD, Harrigan PR, Lourenco L, Yip B, Nosyk B, et al.
Expansion of HAART coverage is associated with sustained decreases in
HIV/AIDS morbidity, mortality and HIV transmission: the "HIV treatment as
prevention" experience in a Canadian setting. PLoS One. 2014;9(2):e87872.
https://doi.org/10.1371/journal.pone.0087872.
50. Lima VD, Johnston K, Hogg RS, Levy AR, Harrigan PR, Anema A, Montaner
JSG. Expanded access to highly active antiretroviral therapy: a potentially
powerful strategy to curb the growth of the HIV epidemic. J Infect Dis.
2008;198(1):59–67. https://doi.org/10.1086/588673.
51. Tanser F, Kim HY, Vandormael A, Iwuji C, Barnighausen T. Opportunities and
challenges in HIV treatment as prevention research: results from the ANRS
12249 cluster-randomized trial and associated population cohort. Current
Hiv/Aids Reports. 2020;17(2):97–108. https://doi.org/10.1007/s11904-020-004
87-1.
52. Koblin BA, Husnik MJ, Colfax G, et al. Risk factors for HIV infection among
men who have sex with men. AIDS. 2006;20(5):731–9. https://doi.org/10.1
097/01.aids.0000216374.61442.55.
53. Chinese Center for Disease Control and Prevention. National Handbook of
Free HIV Antiviral Treatment (4th edition): People's Medical Publishing
House; 2016.
54. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression
analysis of interrupted time series studies in medication use research. J Clin
Pharm Ther. 2002;27(4):299–309. https://doi.org/10.1046/j.1365-2710.2002.
00430.x.
55. Baral S, Rao A, Sullivan P, Phaswana-Mafuya N, Diouf D, Millett G, Musyoki H,
Geng E, Mishra S. The disconnect between individual-level and population-
level goatee. HIV prevention benefits of antiretroviral treatment. Lancet HIV.
2019;6(9):E632–E8. https://doi.org/10.1016/S2352-3018(19)30226-7.
56. Molina J-M, Grund B, Gordin F, Williams I, Schechter M, Losso M, Law M,
Ekong E, Mwelase N, Skoutelis A, Wiselka MJ, Vandekerckhove L, Benfield T,
Munroe D, Lundgren JD, Neaton JD, INSIGHT START study group. Which
HIV-infected adults with high CD4 T-cell counts benefit most from
immediate initiation of antiretroviral therapy? A post-hoc subgroup analysis
of the START trial. The Lancet HIV. 2018;5(4):e172–e80. https://doi.org/10.101
6/S2352-3018(18)30003-1.
57. Garnett GP. Role of herd immunity in determining the effect of vaccines
against sexually transmitted disease. J Infect Dis. 2005;191(s1):S97–S106.
https://doi.org/10.1086/425271.
58. Anderson RMR. Infectious diseases of humans: dynamics and control.
Oxford: Oxford University Press; 1991.
59. World Health Organization (WHO). WHO recommends HIV self-testing 2016
2016 [Available from: http://apps.who.int/iris/bitstream/10665/251549/1/
WHO-HIV-2016.21-eng.pdf?ua=1.