We conducted a cross-sectional online survey of 595 MSM in Shenzhen, China, during August-September 2020. Shenzhen is a major metropolitan city located in Guangdong Province in southern China, with a population of 13 million in 2020.

Participants (1) were Chinese-speaking men living in Shenzhen, (2) were aged at least 18 years, and (3) had oral or anal intercourse with at least 1 man in the past year. Those self-reporting to be HIV positive were excluded. Participants were recruited through multiple sources. Trained and experienced fieldworkers approached prospective participants in venues frequently visited by MSM (ie, bars, parks, and bathhouses) at different time slots on weekdays and weekends. The research team also conducted online outreaching by periodically posting study information on Weibo and WeChat, 2 commonly used social media platforms in China. Recruitment was supplemented by peer referrals. Fieldworkers briefed prospective participants about the study details on-site or using telephone/live chat apps and invited them to create the project’s official WeChat account. Through WeChat, fieldworkers screened the eligibility of prospective participants. Participants were assured that their identifiable information would be kept confidential, they had the right to discontinue participation in the study at any time, and their refusal or withdrawal from the study would not have any consequences. Participants signed an electronic consent form sent by WeChat. The fieldworkers approached 756 prospective participants in gay venues, 720 (95.2%) added the project official WeChat account, 685 (95.1%) were screened to be eligible through WeChat, 245 (35.8%) refused to participate, and 440 (64.2%) completed the online survey. Regarding online recruitment, 150 prospective participants contacted the fieldworkers, 132 (88%) were screened to be eligible through WeChat, 45 (34.1%) refused to participate, and 87 (65.9%) completed the survey. Of 115 prospective participants referred by peers, 98 (85.2%) were screened to be eligible through WeChat, 30 (30.6%) refused to participate and 68 (69.4%) completed the survey. A total of 595 participants completed this study. The main reasons for exclusion were (1) not having oral or anal intercourse with men in the past year (41/985, 4.2%), (2) being aged under 18 years (19/985, 1.9%), and (3) being HIV positive (10/985, 1%). The main reasons for refusals were lack of time and other logistic reasons. A flowchart of recruitment is shown in Figure 1.

We developed an online self-administered questionnaire using Questionnaire Star, a commonly used online survey platform in China. Quick response (QR) codes were generated and sent to the 595 participants through WeChat. The participants were asked to scan the QR code to complete the survey. Each mobile device was only allowed to access the online questionnaire once to avoid duplicate responses. The survey had 105 items (about 20 items per page for 5 pages), which took about 20 minutes to complete. The Questionnaire Star tool performed completeness checks before the questionnaire was submitted. The participants were able to review and change their responses through a Back button. An e-coupon of CNY 20 (US $2.97) was sent to the participants upon survey completion. All data were stored on the online server of Questionnaire Star and protected by a password. Only the corresponding author had access to the database.

Ethics approval was obtained from the Longhua District Centers for Disease Control and Prevention (CDC; reference: 2021009).

A panel consisting of 3 CDC staff, 2 public health researchers, a health psychologist, and 2 MSM volunteers was formed to develop the questionnaire used in this study. The questionnaire was pilot-tested among 10 MSM to assess clarity and readability. These 10 MSM did not participate in the actual survey. Based on their comments, the panel revised and finalized the questionnaire. The Chinese and English versions of the questionnaire are provided in Multimedia Appendix 1.

Background characteristics were collected, including age, relationship status, highest educational level attained, current employment status, monthly personal income, sexual orientation, and source of recruitment.

The dependent variable for this study was HIV testing utilization. In addition, 3 independent questions were used to assess whether participants performed a specific type of HIV testing (ie, HIV testing at community-based organizations [CBOs], public hospitals/the CDC, private hospitals, and other organizations in Shenzhen; HIV testing in a place other than Shenzhen; and home-based HIVST) in 3 reference periods. The first period was between November 2019 and January 2020. Soon after the China central government imposed a lockdown in Wuhan on January 23, 2020, Shenzhen initiated a tier 1 response (the highest level) to a major public health event on January 24, 2020 [17]. Therefore, the first reference period represented the time prior to the COVID-19 outbreak in China. The second period was between February and April 2020, after the lockdown in Wuhan was lifted in April 2020 and Shenzhen lowered its response level to tier 3 (the lowest level) in early May 2020 [17]. The second reference period hence represented the time after the COVID-19 outbreak and before the pandemic was under initial control in China. The last period was from May to July 2020, which represented the time after the COVID-19 pandemic was under initial control in China [17].

Independent variables included sexual risk behaviors, other HIV/STI prevention services, perceptions related to HIV testing, and structural barriers to HIV testing. Similar to measuring HIV testing behaviors, 3 independent questions were used to measure sexual risk behaviors and other HIV/STI prevention services in the 3 reference periods. The 4 different types of sexual risk behaviors assessed by the questionnaire were (1) condomless anal intercourse (CAI) with regular male sex partners (RPs), (2) CAI with nonregular male sex partners (NRPs), (3) CAI with male sex workers, and (4) sexualized drug use (SDU). An RP is defined as a stable boyfriend, while an NRP is defined as a man who is neither an RP nor a male sex worker. SDU is defined as the use of any of the following psychoactive substances before or during sexual intercourse: ketamine, methamphetamine, cocaine, cannabis, ecstasy, Dormicum/Halcion/Erimin 5/nonprescription hypnotic drugs, heroin, cough suppressant (not for curing cough), gamma-hydroxybutyric acid (GHB)/gamma-butyrolactone (GBL), 5-methocy-N,N-diisopropyltryptamine (Foxy), and mephedrone [18,19]. We created 4 variables comparing the presence of sexual risk behaviors before (period 1) and after (combined periods 2 and 3) the COVID-19 outbreak. The response categories of these variables were as follows: 1=no such behavior before or after the COVID-19 outbreak, 2=with such behavior only before the COVID-19 outbreak, 3=with such behavior both before and after the COVID-19 outbreak, and 4=with such behavior only after the COVID-19 outbreak.

The online survey also documented the participants’ use of other STI testing, other HIV/STI prevention services (eg, receiving free condoms, receiving peer education or education pamphlets, and attending lectures or seminars), and preexposure prophylaxis (PrEP) after the COVID-19 outbreak.

We applied the theory of planned behavior (TPB) as the theoretical framework to select perceptions related to HIV testing after the COVID-19 outbreak [20]. The TPB postulates that willingness to adopt a health-related behavior is a strong predictor of actual behavior. To form such an intention, one would evaluate the pros and cons of the behavior (positive and negative attitudes), consider whether their significant others would support such behavior (perceived subjective norm), and appraise how much control one has over the behavior (perceived behavioral control) [20]. In this study, 1 item measured the participants’ positive attitude toward HIV testing services during COVID-19 (ie, COVID-19 preventive measures taken by HIV testing service providers are effective); 2 other items measured some negative attitudes toward HIV testing services during COVID-19, such as the participants’ concerns related to the risk of contracting COVID-19 during HIV testing and inconvenience of undergoing HIV testing during the pandemic; and 2 single items measured the perceived subjective norm (ie, people who are important to you would support you to undergo HIV testing after the COVID-19 outbreak) and perceived behavioral control (ie, whether to undergo HIV testing after the COVID-19 outbreak is completely under control). The response categories for the latter 5 items were 1=strongly disagree, 2=disagree, 3=neutral, 4=agree, and 5=strongly agree. In addition, 1 item measured the perceived risk of HIV infection comparing the participants’ present situation with the situation before the COVID-19 outbreak (“When comparing your current situation versus the time before COVID-19, do you think your overall risk of HIV infection is higher, lower, or the same?”); the response categories were 1=much lower, 2=somewhat lower, 3=same, 4=somewhat higher, and 5=much higher.

The participants were also asked whether they adopted physical distancing after the COVID-19 outbreak (February-July 2020), including avoiding unnecessary travel and crowded places. Other structural barriers to utilizing HIV testing after the COVID-19 outbreak included whether HIV testing service providers were closed or had reduced their working hours and whether they had difficulty in obtaining HIVST kits and a history of home/centralized quarantine between February and July 2020.

HIV testing utilization was compared using McNemar tests. Since 1 of our objectives was to investigate factors associated with HIV testing utilization after the COVID-19 outbreak, we combined the utilization of any type of HIV testing in period 2 (February-April 2020) and period 3 (May-July 2020) and used it as the dependent variable in the subsequent analysis. First, associations between background characteristics and the dependent variable were analyzed using logistic regression models, and crude odds ratios (ORs) were obtained. After adjustment for those variables with P<0.05 in the univariate analysis, associations between independent variables of interest (HIV testing prior to the COVID-19 outbreak, other HIV/STI prevention service utilization after the COVID-19 outbreak, variables comparing the presence of sexual risk behaviors before and after the COVID-19 outbreak, perceptions related to HIV testing, and structural barriers to HIV testing after the COVID-19 outbreak) and the dependent variable were assessed by adjusted odds ratios (AORs). Each AOR was obtained by fitting a single logistic regression model, which involved 1 of the independent variables of interest and significant background variables. SPSS Statistics version 21.0 (IBM) was used for data analysis, with P<.05 considered statistically significant.

The majority of the participants were 18-30 years old (n=452, 75.9%), single (n=481, 80.8%), and employed full-time (n=433, 72.8%); had attained at least tertiary education (n=394, 66.2%), with a monthly personal income of CNY >=5000 (>=US$741.46) (n=346, 58.1%); see Table 1.

Relatively few participants (n=6-44, 1.0%-7.4%) reported the presence of sexual risk behaviors only after the COVID-19 outbreak (Table 2). The prevalence of sexual risk behaviors in different reference periods are shown in Multimedia Appendix 2. After the COVID-19 outbreak, 37-199 (6.2%-33.4%) participants used HIV/STI prevention services other than HIV testing (Table 3). Regarding perceptions related to HIV testing (Table 4), over half of the participants perceived their current risk of HIV infection was much/somewhat lower than that before COVID-19 (n=387, 65%) and agreed/strongly agreed that COVID-19 preventive measures taken by HIV testing service providers were effective (n=320, 53.7%). More than one-third of them had concerns related to COVID-19 infection during HIV testing (n=225, 37.9%) and the inconvenience of using HIV testing services during the pandemic (n=243, 40.8%). Regarding structural barriers to HIV testing, 58 (9.7%) and 63 (10.6%) participants reported that HIV testing service providers suspended and reduced their services, respectively, and 42 (7.1%) had difficulty in obtaining HIVST kits between February and July 2020 (Table 5).

About half of the participants underwent any types of HIV testing between February and July 2020 (n=331, 55.6%). Compared to the time before the COVID-19 outbreak (period 1, November 2019-January 2020), a significantly lower proportion of the participants underwent any type of HIV testing between February and April 2020 (period 2 vs period 1: n=262 vs 363, 44.0% vs 61.0%, P<.001). The proportion of testers did not increase significantly after the pandemic was under initial control in China (period 3, May-July 2020; period 2 vs period 3: n=262 vs 277, 44.0% vs 46.6%, P=.21; period 3 vs period 1: n=277 vs 363, 46.6% vs 61.0%, P<.001). We observed similar changes in the utilization of HIV testing at CBOs in Shenzhen, at public hospitals/the CDC in Shenzhen, at other organizations in Shenzhen, and in places other than Shenzhen, as well as the utilization of HIVST. In addition, 331 (55.6%) participants had undergone any type of HIV testing after the COVID-19 outbreak (combined periods 2 and 3); see Table 6. Patterns of HIV testing utilization across the study period are also shown in Figure 2.

In univariate analysis, participants who cohabited with or were married to a woman and identified themselves as heterosexual were less likely to undergo any type of HIV testing between February and July 2020 (Table 7).

After adjusting for these significant background characteristics, CAI with RPs and SDU both before and after the COVID-19 outbreak were associated with higher utilization of HIV testing after the COVID-19 outbreak. CAI with RPs and NRPs only after the COVID-19 outbreak was also positively associated with the dependent variable. Regarding HIV/STI prevention service utilization, utilization of HIV testing services prior to the COVID-19 outbreak was associated with higher HIV testing utilization after the COVID-19 outbreak. Users of other STI testing, other HIV/STI prevention services, and PrEP after the COVID-19 outbreak were more likely to undergo any types of HIV testing in the same period.

The current perceived risk of HIV infection was higher than that before the COVID-19 outbreak (AOR 1.15, 95% CI 1.01-1.30, P=.03), and perceived COVID-19 preventive measures taken by HIV testing service providers to be effective (AOR 1.52, 95% CI 1.29-1.78, P<.001) and perceived higher behavioral control to undergo HIV testing after the COVID-19 outbreak (AOR 1.18, 95% CI 1.00-1.40, P=.048) were associated with higher HIV testing utilization between February and July 2020. COVID-19 infection during HIV testing (AOR 0.78, 95% CI 0.68-0.89, P<.001), avoiding crowded places (AOR 0.68, 95% CI 0.48-0.98, P=0.04), and HIV testing service providers reducing their working hours (AOR 0.59, 95% CI 0.48-0.98, P=0.046) were associated with lower HIV testing utilization during the same period (Table 8).

To the best of our knowledge, this is 1 of the first studies investigating the impacts of the COVID-19 pandemic on HIV testing among MSM in China. A significant decline was observed in the utilization of facility-based HIV testing and HIVST comparing to the prepandemic era. The findings were similar to studies across countries [8-11,13]. A significant decline in sexual risk behaviors (CAI with RPs and NRPs) was also observed after the COVID-19 outbreak. Changes in sexual risk behaviors among MSM after the COVID-19 outbreak were mixed in the previous literature [21-27]. The level of sexual risk behaviors among our participants quickly rebounded to the prepandemic level after the COVID-19 pandemic was under initial control. This situation raised concerns about potential HIV/STI outbreaks among MSM in China in the postpandemic era. Currently, given the scale-up of COVID-19 vaccination, more countries are attempting to return to normal life. Our findings share some reference values for these countries regarding HIV prevention in the postpandemic era. After the control of the COVID-19 pandemic, local governments and service providers should rehire their personnel and resume their working hours for HIV prevention services. Given the implementation of physical distancing and the concerns about COVID-19 infection when using facility-based HIV testing, more efforts should be given to promote home-based HIVST with essential supporting services (eg, online counseling support and referral services for HIVST users) to mitigate the potential negative impacts caused by the pandemic.

Similar to previous findings, COVID-19 caused some structural barriers to accessing HIV testing [8,13-16]. During the pandemic, the Chinese government advocated physical distancing and recommended that people avoid unnecessary travel and crowded places [28,29]. In our study, about 70% of the participants reported avoiding crowded places after the COVID-19 outbreak. Avoiding crowded places was negatively associated with HIV testing utilization. Since facility-based HIV testing is usually provided by public hospitals, the CDC, and CBOs, it was likely that MSM would avoid these crowded places during the pandemic. About 10% of the participants reported that their HIV testing service providers reduced working hours during the pandemic, which was also a barrier. In China, public hospitals and the CDC reallocate some of the HIV prevention staff in order to implement COVID-19 prevention.

Our findings provide some empirical insights into service planning and intervention development. More attention should be given to MSM who cohabit with or are married to a woman or identify themselves as heterosexual, as in this study they reported lower HIV testing after the COVID-19 outbreak. Due to discrimination, MSM in China are sexual minorities and hidden in the population [30]. Some Chinese MSM marry a woman to conceal their homosexuality/same-sex behaviors and to deal with their parents’ expectations [30]. Since HIV is a highly stigmatized disease in China, female sexual partners knowing about the MSM’s HIV testing utilization might lead to some undesired consequences (eg, conflicts, exposure of homosexuality).

Use of HIV testing prior to the COVID-19 outbreak was associated with higher HIV testing utilization after the outbreak. Different health promotion strategies tailored to the needs of frequent and infrequent testers should be considered. Use of STI testing and other HIV/STI prevention services after the COVID-19 outbreak was also associated with higher HIV testing utilization during the same period. One explanation is that these services are usually performed simultaneously during HIV testing. COVID-19 did not have a significant impact on PrEP users, who reported higher HIV testing utilization, as they are required to undergo such tests every 3 months [31].

Maintaining or increasing sexual risk behaviors (CAI with RPs and NRPs, and SDU) after the COVID-19 outbreak was significantly associated with higher HIV testing utilization during the same period. Participants might have perceived a lower risk of HIV infection due to the decline in sexual risk behaviors after the COVID-19 outbreak and hence perceived a lower need to undergo HIV testing. The perceived higher risk of HIV infection comparing to the prepandemic era was another facilitator of HIV testing utilization. However, although their sexual risk behaviors rebounded to the prepandemic level, more than 60% of the participants perceived their risk to be lower than the prepandemic level. Facilitating MSM to have an accurate HIV risk perception may be a useful strategy. A personalized HIV risk self-assessment tool may be helpful for MSM during the pandemic, which can be adapted from the HIV risk calculator developed by Chen and Dowdy [32].

Modifying perceptions related to HIV testing after the COVID-19 outbreak may also be useful. About 40% of the participants were concerned about COVID-19 infection when undergoing HIV testing. Such concern was associated with lower HIV testing utilization. Over half of the participants perceived COVID-19 preventive measures taken by HIV testing service providers to be effective. Such perception was a facilitator of HIV testing utilization. HIV testing service providers should make their COVID-19 preventive measures transparent to potential clients to reduce their concerns. The role of HIVST became more important during the COVID-19 pandemic. Previous studies have shown that the majority of MSM were willing to utilize HIVST during the social distancing period and that they preferred home delivery of HIVST kits and support of teleconsultation [33]. Recently, a novel HIVST service was implemented among Chinese MSM. A CBO sent a free HIVST kit through mail to users and provided real-time instructions and counseling through live chat apps, making the experience of HIVST similar to facility-based HIV testing. Such a service was effective in increasing HIV testing coverage and ensuring linkage to care [34,35]. This service could also improve perceived behavioral control to undergo HIV testing after the COVID-19 outbreak, which was another facilitator. Government organizations and CBOs in China should consider allocating more resources to implement HIVST services for MSM in the postpandemic era.

This study had a few limitations. First, the cross-sectional study design could not adequately determine the magnitude of the impact of the COVID-19 pandemic and routine testing frequency on HIV testing utilization. However, we believe the impact of routine testing frequency would be limited. Sexually active MSM are recommended by the China CDC to undergo HIV testing every 3 months. In the presence of a window period, all people who receive a negative HIV testing result are also advised to test again 3 months afterward. In this study, the duration of each reference period was in line with the recommended interval of HIV testing for MSM.

Second, HIV testing, sexual risk behaviors, and other HIV/STI prevention service utilization in different reference periods were based on self-reported data, so recall bias existed. The participants were likely to overreport HIV testing or other HIV/STI prevention utilization and underreport sexual risk behaviors due to social desirability.

Third, participants were recruited by nonprobabilistic sampling in 1 Chinese city. Compared to other Chinese cities with a general or a lower economy, there are more organizations providing HIV testing services in Shenzhen. In addition, given the relatively high-income level of the people in Shenzhen, MSM living in the city would have lower financial barriers to using chargeable HIV testing services provided by private clinics or purchasing HIVST kits. Therefore, the findings of this study could not be applied to other Chinese cities with a general or a lower economy. The COVID-19 pandemic might have a greater impact on HIV testing services in other smaller or less developed Chinese cities.

Fourth, we were not able to obtain the characteristics of MSM who refused to participate in the study. The characteristics of those who refused to join the study might be different from the participants, so selection bias existed. The response rate was relatively high compared to online surveys on similar topics.

Fifth, the items were constructed for this study and were not validated by other studies. Moreover, we only obtained cross-sectional associations and could not establish causal relationships.

In sum, utilization of facility-based and home-based HIVST among Chinese MSM declined after the COVID-19 outbreak and did not increase after the pandemic was under initial control. Removing structural barriers to accessing HIV testing, caused by COVID-19; modifying perceptions related to HIV testing; and making use of HIVST might be useful strategies to improve HIV testing among MSM during the pandemic.