Inequities in willingness to disclose contact information to access free mpox self-testing kits among MSM in China: a COM-B–informed multi-center study

Abstract

Background

Free mpox self-testing often requires disclosure of personal contact information (e.g., phone numbers or social media IDs). Among men who have sex with men (MSM), this requirement can function as an access gate with unequal impacts across socioeconomic and regional contexts. Using the Capability, Opportunity, Motivation–Behavior (COM-B) framework, we examined psychosocial and socioeconomic correlates of willingness to disclose among MSM in China.

Methods

We conducted an anonymous online cross-sectional survey (November 2023–March 2024) in six provincial regions representing three development tiers. MSM (n = 2,403) were recruited via community-based organizations using snowball sampling. Guided by the COM-B framework, we fitted hierarchical logistic regression models with sequential adjustment for sociodemographic, health, and behavioral covariates to estimate associations with disclosure willingness. Prespecified stratified analyses by education, income, and regional development assessed heterogeneity.

Results

Overall, 57.9% reported willingness to disclose contact information to obtain free mpox self-testing kits. COM-B factors were consistently associated with willingness across models: higher prevention self-efficacy (OR = 1.028, 95% CI: 1.009–1.046) and prior HIV voluntary counseling and testing (VCT) (OR = 1.535, 95% CI: 1.240–1.899) were positively associated with willingness, whereas greater healthcare-related concerns were negatively associated with it (OR = 0.983, 95% CI: 0.975–0.990). College education (OR = 1.542, 95% CI: 1.233–1.928) and higher income (≥ 6,001 CNY: OR = 1.349, 95% CI: 1.003–1.814) were positively associated. Compared with highly developed regions, MSM in moderately developed regions showed lower odds of willingness to disclose (OR = 0.607, 95% CI: 0.491–0.750). Subgroup analyses indicated that social support was more influential in less developed regions, while privacy- and discrimination-related concerns were more salient in moderately developed regions.

Conclusions

Requiring contact disclosure to access free self-testing can create inequitable barriers. Equity-oriented implementation should: (1) strengthen self-efficacy through peer-led practice and prompts; (2) reduce healthcare-related concerns via privacy-by-design (data minimization, transparent data-use, anonymous pickup); and (3) leverage HIV testing infrastructure. Moderately developed regions warrant priority attention given systematically lower willingness to disclose.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12939-026-02761-x.

Background

The 2022–2024 mpox outbreak revealed entrenched health inequities, with men who have sex with men (MSM) bearing a disproportionate burden globally [1, 2]. By August 2024, over 100,000 confirmed cases across more than 120 countries had been reported by the World Health Organization (WHO), underscoring the scale and persistence of this zoonotic threat [3]. Inequities were particularly acute in resource-limited settings, where stigma, discrimination, and insufficient diagnostic capacity constrained timely access to prevention and care [4, 5]. Pronounced regional economic heterogeneity in China heightens concerns about unequal access to essential services [6, 7].

In China, mpox control has relied on clinical testing, contact tracing, and isolation in designated public healthcare facilities since the first imported case in September 2022, with a surge in domestic cases from June 2023 predominantly among MSM [6, 8, 9]. No formal mpox self-testing program has been implemented nationwide, and access to diagnosis and treatment is free through public health systems without health insurance requirements, prioritizing high-risk groups for surveillance and vaccination [9]. While self-testing is not yet part of the strategy, our study examines potential barriers to access in a hypothetical free self-testing program, drawing on existing prevention efforts to inform future implementations.

Early and accessible testing is central to interrupting transmission, especially where stigma and logistical barriers impede clinic-based testing [10]. Mpox self-testing offers a pragmatic complement to facility-based services by reducing informational, geographic, and stigma-related barriers [5, 11]. Evidence from human immunodeficiency virus (HIV) programs shows that self-testing increases uptake among underserved MSM, particularly where health system access is constrained [12, 13]. By lowering user-side costs, self-testing can mitigate socioeconomic gradients in access; however, implementation often differs across development tiers, with the potential to reproduce or widen existing disparities [12, 14].

In many programs, obtaining free self-testing kits requires disclosure of personal contact information (e.g., phone numbers or social media IDs). This requirement functions as a prerequisite access behavior—a step that must be completed before obtaining diagnostic resources [15, 16]. Disclosure is conceptually distinct from testing itself: testing refers to performing a self-administered diagnostic, whereas disclosure is an upstream behavior determining whether individuals can access testing [17]. In practice, disclosure typically involves registering contact details via online platforms or community organizations to enable kit distribution, result notification, or linkage to care [16, 18]. For MSM living amid structural stigma or weak privacy protections, such disclosure can be a decisive deterrent rather than a trivial administrative step [19, 20]. In less developed areas, limited anti-discrimination safeguards and constrained service capacity may further suppress uptake [13]. More broadly, socioeconomic disadvantage compounds barriers to care, such that even no-cost services can remain inequitably accessed [21, 22].

The Capability, Opportunity, Motivation–Behavior (COM-B) framework provides a focused lens to examine disclosure as a prevention-related behavior shaped by knowledge and self-efficacy (capability), social and structural conditions (opportunity), and readiness to act (motivation) [23]. Prior work on HIV testing among MSM demonstrates that capability and opportunity jointly influence testing behaviors under structural constraints, supporting the application of COM-B to mpox self-testing contexts [24].

Despite rising attention to equitable mpox testing, little is known about how socioeconomic conditions shape willingness to disclose contact information, and how privacy concerns intersect with regional development disparities [18]. These dynamics align with the inverse care law and structural stigma, whereby those with greatest need face the greatest barriers. Requiring contact disclosure also raises issues of data justice—privacy functioning as a social determinant of health—potentially amplifying inequities when protections are weak [25].

Using the COM-B framework, this study examines psychosocial and socioeconomic correlates of willingness to disclose personal contact information to obtain free mpox self-testing kits among MSM in China, with explicit attention to regional development tiers. The findings aim to inform equity-oriented, context-sensitive implementation strategies that expand access while safeguarding privacy.

Methods

Study design and setting

This multicenter cross-sectional study was conducted from November 2023 to March 2024 using an anonymous online survey hosted on Wenjuanxing, a widely used and secure platform for survey administration in China. The study was conducted across six provincial regions representing three development tiers: Shanghai and Guangdong (eastern highly developed regions), Liaoning and Shaanxi (central moderately developed regions), and Yunnan and Xinjiang (western less developed regions) [26].

The selection of these study sites was grounded in empirical evidence regarding mpox epidemiology in China, regional case distributions, surveillance capacities, and the feasibility of conducting investigations among MSM. First, mpox cases in China are predominantly concentrated in eastern coastal provinces (e.g., Guangdong, Shanghai), which possess stronger public health surveillance infrastructure and higher reporting rates compared to central and western regions, where cases are fewer but surveillance capacity is relatively lower [8]. This epidemiological disparity, with approximately 95% of 2023 cases linked to MSM in eastern hotspots, ensures our selections capture diverse dynamics for representative data and targeted interventions [8]. Second, established networks of MSM-focused non-governmental organizations (NGOs) in these regions facilitated ethical recruitment, informed consent, and privacy protection, building upon our prior successful collaborations that enabled high participation in this study [27]. Third, the inclusion of eastern, central, and western regions ensures national representativeness, allowing for a comparative analysis of prevention strategies across different socioeconomic contexts [28]. Finally, the study’s feasibility was underpinned by proven partnerships with local NGOs in executing large-scale surveys and vaccination campaigns, rather than theoretical assumptions [29].

Local NGOs with established MSM networks distributed survey QR codes during online and offline outreach. NGOs were engaged solely for outreach and participant recruitment activities. Their responsibilities were limited to disseminating study information, facilitating access to potential participants through trusted networks, and assisting with data collection via administered surveys. NGOs did not participate in any aspect of data analysis, statistical interpretation, or inferential processes. All data processing, statistical analyses, and interpretation of results were conducted exclusively by the academic research team, comprising epidemiologists and public health experts, to ensure objectivity and scientific rigor. This delineation of roles was established at the study’s inception to maintain independence and minimize potential biases arising from non-academic involvement.

Participants completed the self-administered questionnaire on personal devices to maximize privacy and convenience. Given NGO-driven snowball recruitment, younger and more digitally connected MSM may be overrepresented. Probability weights were not feasible due to the absence of population benchmarks; therefore, findings should be interpreted as analytic generalizations to similar urban MSM networks.

Eligibility criteria

Eligible participants were biologically male, aged ≥ 18 years, reported sex with men in the past six months, and had resided primarily in one of the six regions for at least six months. We excluded submissions if the completion time was < 300 s, embedded attention/quality-control checks were failed, or the IP address geolocation was outside target regions. Electronic informed consent was obtained prior to survey initiation; the consent page detailed anonymity, data protection, and voluntary participation.

Survey instrument development

The questionnaire was developed with input from experts in public health and behavioral science and from MSM community representatives to ensure cultural relevance and clarity [26, 30]. A pilot test with 50 MSM assessed item comprehension, response burden, and psychometrics; minor revisions were made accordingly [26]. The final instrument included items on sociodemographic characteristics, health status, behaviors, COM-B constructs, and willingness to disclose contact information for free mpox self-testing.

Sample size and response validation

Sample size was informed by prior work in similar populations by the research team, targeting adequate power for multivariable models and subgroup analyses [30, 31]. Of 2,481 submissions, 2,403 valid responses remained after applying exclusion criteria (validity rate 96.86%). Region-specific counts were Shanghai (n = 569), Guangdong (n = 500), Liaoning (n = 502), Shaanxi (n = 199), Yunnan (n = 313), and Xinjiang (n = 320) (Supplementary Fig. 1) [32].

Primary outcome

The primary outcome was willingness to disclose personal contact information to obtain free mpox self-testing kits. This was conceptualized as an access-gate behavior—a prerequisite for obtaining self-testing resources—rather than testing behavior per se [33, 34]. Within COM-B, disclosure was treated as the target behavior (B). Participants selecting willingness to provide a phone number or social media ID were categorized as “disclosed”; those unwilling to provide such information were categorized as “non-disclosed.” This operationalization reflects programmatic practice in which contact information is required for kit distribution.

Explanatory variables: COM-B framework

Explanatory variables were grounded in the COM-B framework, which conceptualizes health behavior as the product of three interrelated domains: capability (knowledge and skills, including self-efficacy), opportunity (social and structural conditions that enable or constrain action), and motivation (reflective and automatic processes that drive willingness to act) [23, 24]. In this study, disclosure of personal contact information to obtain free mpox self-testing kits was explicitly operationalized as a prevention behavior required to access diagnostic resources and enable early detection. We therefore mapped study measures to COM-B domains a priori and treated disclosure willingness as the target behavior, aligning measurement and analysis with the framework’s emphasis on how individual capacities, contextual opportunities, and motivational states jointly shape preventive action.

Capability

We assessed capability to engage in mpox prevention, including willingness to disclose contact information to obtain a test kit, using three measures. Mpox knowledge was measured with a 12-item scale awarding one point per correct answer; higher scores reflected greater understanding of transmission, prevention, and symptoms (Cronbach’s α = 0.684) (Table S1 in Supplementary File 1) [35]. Prevention judgment was captured with four 5-point Likert items assessing perceived infection likelihood under different partner characteristics or behaviors, indicating the strictness of individual risk appraisal (α = 0.906) (Table S2 in Supplementary File 1) [35]. Prevention self-efficacy was measured with six 5-point Likert items reflecting confidence in adopting mpox preventive actions (α = 0.859) (Table S3 in Supplementary File 1) [31]. Higher scores on all scales indicated greater capability.

Opportunity

Opportunity-related factors, conceptualized as external conditions shaping prevention and disclosure behaviors, were measured across four constructs. Perceived social norms regarding mpox were assessed with four 5-point Likert items, with higher scores indicating more supportive public attitudes (α = 0.946) (Table S4 in Supplementary File 1) [35, 36]. Perceived social support was measured with the 12-item Perceived Social Support Scale (range 12–84), capturing emotional and practical support from family, friends, and other sources; higher scores reflected stronger support (α = 0.958) (Table S5 in Supplementary File 1) [37]. Perceived discrimination related to sexual orientation was assessed with five items spanning daily life, healthcare, and employment contexts (α = 0.877), with higher scores indicating greater discrimination experiences (Table S6 in Supplementary File 1) [38]. Healthcare-related concerns were measured with 14 items on a 5-point Likert scale covering perceived barriers to mpox-related services, including cost, privacy, and mistreatment (α = 0.949); higher scores denoted greater concerns (Table S7 in Supplementary File 1) [30].

Motivation

Motivational factors captured behavioral history and psychological readiness to engage in mpox prevention and to disclose contact information when required for access. Prior HIV voluntary counseling and testing (VCT) was recorded as a binary variable (yes/no) to reflect previous engagement with health services that may normalize disclosure. Self-perception of mpox infection risk was measured with four 5-point Likert items capturing perceived susceptibility and readiness to take preventive action (α = 0.886), with higher scores indicating stronger perceived risk/readiness (Table S8 in Supplementary File 1) [39]. Global self-esteem was assessed using the 10-item Rosenberg Self-Esteem Scale (α = 0.872), where higher scores indicated higher self-esteem and lower scores reflected reduced self-worth that may influence disclosure confidence (Table S9 in Supplementary File 1) [40].

Covariates

To address potential confounding, models adjusted for sociodemographic, health status, and behavioral characteristics. Sociodemographic variables included age; education (high school or below vs. college or above); monthly income in Chinese Yuan (CNY) (≤ 3,000; 3,001–6,000; ≥6,001); employment (mental labor, physical labor, student, freelance/unemployed); marital status (single, married, divorced/widowed); region of residence defined by region development tier (Shanghai/Guangdong as highly developed; Liaoning/Shaanxi as moderately developed; Yunnan/Xinjiang as less developed); length of local residence (local, < 5 years, ≥ 5 years); and sexual orientation (homosexual vs. others). Health status covariates comprised self-reported HIV test results (positive vs. negative/unknown), diagnosis of other sexually transmitted diseases (STDs) in the past six months (e.g., syphilis, gonorrhea, genital warts), and common chronic conditions (hypertension, diabetes, hyperlipidemia). Behavioral covariates included past six-month smoking, alcohol use, and drug use, as well as sexual behaviors condomless anal intercourse (CAI) and participation in group sex.

Statistical analysis

Statistical analysis was conducted in R (version 4.5.1). Continuous variables were summarized as medians with interquartile ranges (IQRs) given non-normal distributions, and categorical variables as counts and percentages. Between-group differences by disclosure status were evaluated using independent t tests or Mann-Whitney U tests for continuous variables, and chi-square tests for categorical variables. Variables associated with the outcome at p < 0.05 in univariate analyses were considered for multivariable modeling.

Guided by COM-B, we specified a hierarchical logistic regression to assess the incremental contribution of factor domains. The base model included COM-B variables only and treated disclosure willingness as the target behavior (B), with predictors a priori mapped to Capability, Opportunity, and Motivation domains. Model 1 additionally adjusted for sociodemographic characteristics; Model 2 further included health status variables; and Model 3 incorporated behavioral covariates. Results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). Prespecified subgroup analyses stratified models by education, income, and regional development tiers to assess heterogeneity.

Given the cross-sectional design, all associations should be interpreted as correlational rather than causal. For continuous predictors on multi-item Likert scales, per-unit ORs reflect one-point changes; cumulative differences across plausible score ranges (e.g., 10 points) may represent meaningful variation in odds [41, 42].

Results

Participant characteristics by disclosure status

Of 2,403 respondents, 57.9% reported willingness to disclose personal contact information to obtain free mpox self-testing kits (Table 1). Compared with non-disclosers, disclosers were more likely to have college education or above (82.8% vs. 72.7%) and higher monthly income (≥ 6,001 CNY: 47.7% vs. 38.1%; both p < 0.001). Regional composition differed: participants from highly developed regions were overrepresented among disclosers (49.7% vs. 37.4%), whereas those from moderately developed regions were more prevalent among non-disclosers (p < 0.001). Age, marital status, and length of residence did not differ significantly between groups.

Table 1.

Sociodemographic and behavioral characteristics of participants by personal contact information disclosure status for accessing free mpox self-testing kits (N = 2,403)

Participants Characteristics	Overall (N = 2,403)	Non-disclosed (n = 1,012)	Disclosed (n = 1,391)	p-Value
Sociodemographic factors
Age [MD (IQR)]	29 (25–35)	29 (25–34)	29 (25–35)	0.710
Education level [n (%)]				< 0.001
High school or below	515 (21.4)	276 (27.3)	239 (17.2)
College or above	1888 (78.6)	736 (72.7)	1152 (82.8)
Monthly income, CNY [n (%)]				< 0.001
3000	502 (20.9)	249 (24.6)	253 (18.2)
3001–6000	851 (35.4)	377 (37.3)	474 (34.1)
6001	1050 (43.7)	386 (38.1)	664 (47.7)
Employment status [n (%)]				0.019
Mental labor	386 (16.1)	139 (13.7)	247 (17.8)
Physical labor	774 (32.2)	319 (31.5)	455 (32.7)
Student	320 (13.3)	136 (13.4)	184 (13.2)
Freelance/unemployed	923 (38.4)	418 (41.3)	505 (36.3)
Marital status [n (%)]				0.134
Single	2035 (84.7)	841 (83.1)	1194 (85.8)
Married	255 (10.6)	122 (12.1)	133 (9.6)
Divorced/widowed	113 (4.7)	49 (4.8)	64 (4.6)
Regions [n (%)]				< 0.001
Highly developed regions	1069 (44.5)	378 (37.4)	691 (49.7)
Moderately developed regions	701 (29.2)	357 (35.3)	344 (24.7)
Less developed regions	633 (26.3)	277 (27.4)	356 (25.6)
Length of residence [n (%)]				0.177
Local	936 (39.0)	403 (39.8)	533 (38.3)
< 5 years	870 (36.2)	377 (37.3)	493 (35.4)
5 years	597 (24.8)	232 (22.9)	365 (26.2)
Sexual orientation [n (%)]				0.037
Homosexual	1887 (78.5)	774 (76.5)	1113 (80.0)
Others	516 (21.5)	238 (23.5)	278 (20.0)
Health status factors
Self-reported HIV test results [n (%)]				0.142
Negative/unknown	2204 (91.7)	938 (92.7)	1266 (91.0)
Positive	199 (8.3)	74 (7.3)	125 (9.0)
Diagnosis of other STDs [n (%)]				0.003
No	2182 (90.8)	940 (92.9)	1242 (89.3)
Yes	221 (9.2)	72 (7.1)	149 (10.7)
Common chronic diseases [n (%)]				0.042
No	2078 (86.5)	892 (88.1)	1186 (85.3)
Yes	325 (13.5)	120 (11.9)	205 (14.7)
Behavioral factors
Smoking [n (%)]				0.188
No	1464 (60.9)	601 (59.4)	863 (62.0)
Yes	939 (39.1)	411 (40.6)	528 (38.0)
Alcohol consumption [n (%)]				0.773
No	1122 (46.7)	476 (47.0)	646 (46.4)
Yes	1281 (53.3)	536 (53.0)	745 (53.6)
Drug use [n (%)]				0.631
No	2334 (97.1)	981 (96.9)	1353 (97.3)
Yes	69 (2.9)	31 (3.1)	38 (2.7)
CAI [n (%)]				0.013
No	1395 (58.1)	617 (61.0)	778 (55.9)
Yes	1008 (41.9)	395 (39.0)	613 (44.1)
Multiple sexual partners [n (%)]				0.748
No	1206 (50.2)	504 (49.8)	702 (50.5)
Yes	1197 (49.8)	508 (50.2)	689 (49.5)
Capability-related factors
Mpox knowledge [MD (IQR)]	8 (7–10)	8 (6–10)	8 (7–10)	0.011
Prevention judgment [MD (IQR)]	17 (16–20)	16 (15–20)	17 (16–20)	0.063
Self-efficacy [MD (IQR)]	23 (20–25)	22 (19–25)	23 (20–26)	< 0.001
Opportunity-related factors
Social norms [MD (IQR)]	15 (11–16)	16 (11–17)	15 (11–16)	0.094
Social support [MD (IQR)]	59 (48–70)	56 (48–69)	60 (48–70)	< 0.001
Perceived discrimination [MD (IQR)]	1 (0–3)	1 (0–4)	1 (0–3)	0.001
Healthcare-related concerns [MD (IQR)]	51 (42–57)	53 (44–60)	50 (42–56)	< 0.001
Motivation-related factors
VCT [n (%)]				< 0.001
No	457 (19.0)	230 (22.7)	227 (16.3)
Yes	1946 (81.0)	782 (77.3)	1164 (83.7)
Self-perception of infection [MD (IQR)]	15 (12–16)	15 (12–17)	14 (12–16)	0.135
Self-esteem [MD (IQR)]	30 (26–34)	30 (27–34)	29 (26–33)	0.054

Abbreviations: MD = Median; IQR = Interquartile Range; CNY = Chinese Yuan

Health and behavioral profiles were largely comparable, with two exceptions: disclosers had higher prevalence of other STDs in the past six months (10.7% vs. 7.1%) and of common chronic conditions (14.7% vs. 11.9%; both p < 0.05). Most behavioral indicators showed no between-group differences, although CAI was more frequent among disclosers (44.1% vs. 39.0%, p = 0.013).

COM-related factors and disclosure status

Disclosers scored higher on mpox knowledge (p = 0.011) and prevention self-efficacy (p < 0.001). They reported greater perceived social support and lower levels of perceived discrimination and healthcare-related concerns (all p < 0.001). Prior HIV VCT was more common among disclosers than non-disclosers (p < 0.001). No significant differences were observed for prevention judgment, perceived social norms, self-perception of mpox infection, or self-esteem.

Regression analysis

Hierarchical logistic models showed that COM-B constructs were consistently associated with willingness to disclose across all adjustment stages, indicating effects independent of sociodemographic, health, and behavioral covariates (Table 2). Higher prevention self-efficacy was linked to greater odds of disclosure (OR = 1.028, 95% CI: 1.009–1.046, p < 0.05), whereas greater healthcare-related concerns were associated with lower odds (OR = 0.983, 95% CI: 0.975–0.990, p < 0.001). Prior HIV VCT was positively associated with disclosure (OR = 1.535, 95% CI: 1.240–1.899, p < 0.001).

Table 2.

Hierarchical logistic regression of COM-B-related factors associated with personal contact information disclosure for accessing free mpox self-testing kits (N = 2,403)

	Base Model	Model 1	Model 2	Model 3
	OR (95%CI)	OR (95%CI)	OR (95%CI)	OR (95%CI)
COM-related factors
Mpox knowledge	1.030 (0.996, 1.066)	1.009 (0.974, 1.045)	1.008 (0.974, 1.044)	1.010 (0.975, 1.046)
Self-efficacy	1.028 (1.010, 1.047) *	1.027 (1.009, 1.046) *	1.027 (1.009, 1.046) *	1.028 (1.009, 1.046) *
Social support	1.006 (1.000, 1.012)	1.003 (0.997, 1.009)	1.003 (0.997, 1.009)	1.003 (0.997, 1.009)
Perceived discrimination	0.948 (0.907, 0.990) *	0.963 (0.920, 1.007)	0.955 (0.913, 1.000)	0.956 (0.913, 1.001)
Healthcare-related concerns	0.984 (0.977, 0.991) **	0.983 (0.976, 0.990) **	0.983 (0.976, 0.990) **	0.983 (0.975, 0.990) **
VCT
No	(Reference)	(Reference)	(Reference)	(Reference)
Yes	1.499 (1.217, 1.845) **	1.532 (1.240, 1.894) **	1.515 (1.225, 1.874) **	1.535 (1.240, 1.899) **
Sociodemographic factors
Education level
High school or below		(Reference)	(Reference)	(Reference)
College or above		1.535 (1.229, 1.916) **	1.544 (1.235, 1.930) **	1.542 (1.233, 1.928) **
Monthly income, CNY
≤3000		(Reference)	(Reference)	(Reference)
3001–6000		1.221 (0.932, 1.599)	1.212 (0.924, 1.588)	1.204 (0.918, 1.579)
≥6001		1.362 (1.014, 1.829) *	1.362 (1.013, 1.831) *	1.349 (1.003, 1.814) *
Employment status
Mental labor		(Reference)	(Reference)	(Reference)
Physical labor		0.935 (0.717, 1.220)	0.932 (0.713, 1.218)	0.936 (0.717, 1.223)
Student		0.979 (0.671, 1.429)	1.018 (0.696, 1.489)	1.011 (0.691, 1.480)
Freelance/unemployed		0.794 (0.613, 1.028)	0.793 (0.612, 1.028)	0.799 (0.616, 1.035)
Regions
Highly developed regions		(Reference)	(Reference)	(Reference)
Moderately developed regions		0.602 (0.489, 0.742) **	0.596 (0.483, 0.737) **	0.607 (0.491, 0.750) **
Less developed regions		0.843 (0.673, 1.056)	0.841 (0.670, 1.055)	0.845 (0.673, 1.061)
Sexual orientation
Others		(Reference)	(Reference)	(Reference)
Homosexual		1.091 (0.889, 1.339)	1.106 (0.900, 1.358)	1.103 (0.898, 1.355)
Health status factors
Diagnosis of other STDs
No		–	(Reference)	(Reference)
Yes		–	1.488 (1.095, 2.022) *	1.420 (1.039, 1.939) *
Common chronic diseases
No		–	(Reference)	(Reference)
Yes		–	1.436 (1.114, 1.849) *	1.431 (1.111, 1.843) *
Behavioral factor
CAI
No		–	–	(Reference)
Yes		–	–	1.158 (0.971, 1.380)

Abbreviations: OR, odds ratio; CI, confidence interval.  * p < 0.05  ** p < 0.001

Model 1: Controlling for sociodemographic factors

Model 2: Controlling for sociodemographic factors and health status factors

Model 3: Controlling for sociodemographic factors, health status factors and behavioral factor

Socioeconomic gradients were evident. Compared with participants with high school education or below, those with college education or above had higher odds of disclosure (OR = 1.542, 95% CI: 1.233–1.928, p < 0.001). Higher income (≥ 6,001 CNY) was also positively associated (OR = 1.349, 95% CI: 1.003–1.814, p < 0.05). Two health indicators—recent diagnosis of other STDs (OR = 1.420, 95% CI: 1.039–1.939, p < 0.05) and the presence of common chronic conditions (OR = 1.431, 95% CI: 1.111–1.843, p < 0.05)—were associated with higher willingness.

Regional disparities were pronounced. Relative to highly developed regions, MSM in moderately developed regions had substantially lower odds of disclosure (OR = 0.607, 95% CI: 0.491–0.750, p < 0.001), while estimates for less developed regions did not differ significantly from the reference. In the fully adjusted model, employment status, CAI, and sexual orientation were not significantly associated with disclosure.

Subgroup analyses

Subgroup analyses indicated heterogeneity in the associations between COM-related factors and disclosure across education, income, and regional strata (Figs. 1, 2 and 3). Among participants with higher education (college or above), greater self-efficacy (OR = 1.026, 95% CI: 1.006–1.047, p < 0.05) and prior HIV VCT (OR = 1.497, 95% CI: 1.184–1.894, p < 0.05) were positively associated with disclosure, while perceived discrimination (OR = 0.946, 95% CI: 0.899–0.996, p < 0.05) and healthcare-related concerns (OR = 0.992, 95% CI: 0.984–0.999, p < 0.05) were negatively associated. Among participants with lower education (high school or below), self-efficacy (OR = 1.047, 95% CI: 1.003–1.093, p < 0.05) showed a positive association, whereas healthcare-related concerns (OR = 0.954, 95% CI: 0.938–0.971, p < 0.001) showed a negative association.

Fig. 1.

Adjusted ORs and 95% CIs for associations between COM-B-related factors and personal contact information disclosure for accessing free mpox self-testing kits, stratified by education level. Group A: high school or below; Group B: college or above. Estimates are derived from subgroup logistic regression models adjusted for sociodemographic covariates within each education stratum. ORs greater than 1.0 indicate higher odds of disclosure; ORs less than 1.0 indicate lower odds. The vertical dashed line represents OR = 1.0 (null effect)

Fig. 2.

Adjusted ORs and 95% CIs for associations between COM-B-related factors and personal contact information disclosure for accessing free mpox self-testing kits, stratified by monthly income. Group A: ≤3,000 CNY; Group B: 3,001–6,000 CNY; Group C: ≥6,001 CNY. Estimates are derived from subgroup logistic regression models adjusted for sociodemographic covariates within each income stratum. ORs greater than 1.0 indicate higher odds of disclosure; ORs less than 1.0 indicate lower odds. The vertical dashed line represents OR = 1.0 (null effect)

Fig. 3.

Adjusted ORs and 95% CIs for associations between COM-B-related factors and personal contact information disclosure for accessing free mpox self-testing kits, stratified by regional development level. Group A: highly developed regions; Group B: moderately developed regions; Group C: less developed regions. Estimates are derived from subgroup logistic regression models adjusted for sociodemographic covariates within each regional stratum. ORs greater than 1.0 indicate higher odds of disclosure; ORs less than 1.0 indicate lower odds. The vertical dashed line represents OR = 1.0 (null effect)

Income-stratified analyses showed distinct patterns. In the ≤ 3,000 CNY group, self-efficacy (OR = 1.047, 95% CI: 1.005–1.092, p < 0.05) and prior HIV VCT (OR = 1.621, 95% CI: 1.022–2.571, p < 0.05) were positively associated with disclosure, and healthcare-related concerns (OR = 0.961, 95% CI: 0.945–0.977, p < 0.001) were negatively associated. In the 3,001–6,000 CNY group, self-efficacy (OR = 1.039, 95% CI: 1.006–1.074, p < 0.05) and social support (OR = 1.020, 95% CI: 1.008–1.031, p < 0.05) were positively associated, while healthcare-related concerns (OR = 0.968, 95% CI: 0.956–0.980, p < 0.001) were negatively associated. In the ≥ 6,001 CNY group, prior HIV VCT (OR = 1.715, 95% CI: 1.255–2.343, p < 0.05) was positively associated with disclosure.

Regional analyses also identified differences. In highly developed regions, higher mpox knowledge (OR = 1.071, 95% CI: 1.012–1.134, p < 0.05) and prior HIV VCT (OR = 2.004, 95% CI: 1.475–2.723, p < 0.001) were positively associated with disclosure. In moderately developed regions, higher self-efficacy (OR = 1.040, 95% CI: 1.007–1.075, p < 0.05), lower perceived discrimination (OR = 0.837, 95% CI: 0.775–0.903, p < 0.001), and prior HIV VCT (OR = 1.529, 95% CI: 1.022–2.289, p < 0.05) were associated with disclosure. In less developed regions, social support (OR = 1.019, 95% CI: 1.008–1.030, p < 0.001) and perceived discrimination (OR = 1.148, 95% CI: 1.046–1.259, p < 0.05) were positively associated, while healthcare-related concerns (OR = 0.975, 95% CI: 0.961–0.989, p < 0.001) were negatively associated with disclosure.

Discussion

In this study, 57.9% of MSM indicated willingness to disclose personal contact information to obtain free mpox self-testing kits, reflecting moderate acceptance. Self-efficacy was a central correlate of disclosure, aligning with evidence that confidence in performing preventive behaviors supports engagement with health services [43]. Marked regional heterogeneity was observed: participants from moderately developed regions had the lowest likelihood of disclosure relative to those from highly developed regions (OR = 0.607, 95% CI: 0.491–0.750, p < 0.001). Mpox knowledge showed a stronger association with disclosure in highly developed regions, potentially reflecting greater resource availability and health education infrastructure [44]. Healthcare-related concerns and perceived discrimination were inversely associated with disclosure, consistent with literature implicating stigma and structural barriers in reduced service uptake [45].

The hierarchical logistic regression clarified how factor domains contribute to disclosure behavior. Starting from a COM-B base model (capability, opportunity, motivation), we added sociodemographic variables (Model 1), health status (Model 2), and behavioral factors (Model 3). Self-efficacy, healthcare-related concerns, and prior HIV VCT emerged as core correlates in the base model and retained similar magnitude and significance across all models—self-efficacy (OR = 1.028), healthcare-related concerns (OR = 0.983), and prior HIV VCT (OR = 1.535)—indicating robust associations net of confounding. Introducing sociodemographic factors revealed independent contributions of education, income, and regional development. Perceived discrimination was significant in the base model but attenuated after sociodemographic adjustment, suggesting that structural conditions may account for individual-level stigma effects [45]. Overall, the hierarchical approach highlights the need to address both individual psychological correlates and broader socioeconomic context to improve disclosure and access to self-testing.

Subgroup analyses by education, income, and region further delineated variation in correlates. Among higher-educated participants, self-efficacy and social support showed the strongest positive associations, whereas healthcare-related concerns and stigma were salient barriers in middle-income groups. In less developed regions, social support was a key correlate, consistent with settings where informal networks and community organizations play a larger role in navigating care [46–48]. In moderately developed regions, associations pointed to the importance of reducing stigma and enhancing self-efficacy. An unexpected positive association between perceived discrimination and disclosure in less developed regions (OR = 1.148, 95% CI: 1.046–1.259, p < 0.05) warrants further investigation. Two speculative hypotheses may guide future research. First, a trust substitution mechanism: MSM experiencing discrimination in mainstream institutions may rely on trusted NGOs as intermediaries, attenuating perceived disclosure risk despite higher ambient discrimination. Second, a selection effect: in regions with limited distribution infrastructure, only highly motivated individuals willing to disclose despite discrimination may navigate access barriers, creating an artifactual positive association. Qualitative inquiry is needed to elucidate mechanisms.

Healthcare-related concerns—encompassing worries about privacy, mistreatment, and financial burden—were consistently associated with lower disclosure across subgroups [49]. Addressing these concerns requires system-level measures, including clear data minimization policies, robust privacy protections, transparent communication on data handling, and anonymous collection options (e.g., pickup codes or lockers). Prior HIV VCT was a strong predictor of disclosure (OR = 1.535, 95% CI: 1.240–1.899, p < 0.001), supporting the feasibility of leveraging HIV testing platforms to expand mpox self-testing access, consistent with evidence on integrated testing models [50]. By contrast, motivation-adjacent traits such as self-esteem showed limited influence, underscoring the comparatively greater role of external supports—social networks, community outreach, and health system features—in shaping disclosure.

These findings inform several actionable strategies. First, strengthen self-efficacy through peer-led demonstrations, practice opportunities, and timely behavioral prompts, with emphasis on populations with lower education and income [51]. Second, implement privacy-by-design in program workflows, ensure strict data governance, and communicate these safeguards clearly; provide anonymous pickup options to reduce perceived risks [52]. Third, integrate mpox self-testing within existing HIV testing infrastructure and outreach, with regional tailoring to reflect differences in social acceptance, economic development, and health resources [53]. Such context-specific approaches that integrate social, economic, and cultural considerations can improve equitable access and uptake of mpox self-testing.

This study has several notable strengths. Its large sample size and broad geographic coverage enhance the generalizability of findings, while the application of the COM-B framework provides a systematic approach to analyzing behavioral correlates. Stratified analyses further contribute to nuanced insights, enabling the development of targeted interventions tailored to specific subgroups. However, several limitations warrant consideration. Our non-probability, NGO-based snowball sampling—primarily via online channels—limits external validity within study regions and nationally, likely overrepresenting MSM with stronger community or digital ties and inflating related outcomes. Despite multi-center coverage and analytic adjustments, heterogeneity in NGO network density and testing/surveillance access across regions introduces residual confounding that cautions against broad generalization. The cross-sectional design precludes causal inferences and our initial hypothesis that all COM-B domains would equally predict disclosure was only partially supported, with motivation factors showing weaker associations than expected. While our anonymous online survey design helped minimize response bias, self-reported data may introduce some bias. The counterintuitive finding that perceived discrimination positively predicted disclosure in less developed regions contradicts existing stigma literature and requires qualitative validation. Additionally, the study’s timing during heightened mpox awareness may limit generalizability. Key confounders such as institutional trust and prior healthcare experiences were not captured, potentially explaining unexpected regional variations. Future research should employ mixed-methods longitudinal designs to validate these findings and explore the mechanisms underlying regional disparities in disclosure behavior.

Conclusion

Using the COM-B framework, we identified consistent, cross-context correlates of willingness to disclose contact information for accessing free mpox self-testing among MSM. Higher self-efficacy and social support were associated with a higher likelihood of disclosure, whereas healthcare-related concerns were associated with a lower likelihood; prior HIV VCT showed a strong positive association with disclosure. Subgroup analyses underscored substantial heterogeneity by education, income, and regional development, with notably lower disclosure in moderately developed regions. These results support three priorities: strengthen self-efficacy (e.g., peer-led practice and prompts), reduce healthcare-related concerns through robust privacy safeguards and transparent data governance, and leverage HIV testing infrastructure to expand access. Region-specific tailoring remains essential—integrating services in highly developed regions, addressing privacy and discrimination in moderately developed regions, and reinforcing community support in less developed regions—to promote equitable uptake and strengthen outbreak response in vulnerable populations.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MSM

Men who have sex with men

WHO

World Health Organization

HIV

Human immunodeficiency virus

COM-B

Capability, Opportunity, Motivation–Behavior

NGOs

Non-governmental organizations

VCT

Voluntary counseling and testing

CNY

Chinese Yuan

STDs

Sexually transmitted diseases

CAI

Condomless anal intercourse

IQRs

Interquartile ranges

ORs

Odds ratios

CIs

Confidence intervals

Author contributions

ST, FH and YC conceptualized the study. ST, BW and YW developed the methodology. JZ, YW, and YC were responsible for funding acquisition and resources. ST, BW, YN, JB and YW managed data curation and investigation. ST, BW, JC, SC, and SL conducted formal analysis and software. BW, WY and NZ were responsible for visualization. ST, BW, NY and JC drafted the original manuscript, while ST, BW and JZ reviewed and edited the manuscript. YC, YW, and FH supervised the study and oversaw project administration. All authors have read and approved the final version of the manuscript.

Funding

This study was supported by the Key Discipline Projects of the Shanghai Three-Year Action Plan for Public Health (Grant No. GWVI-11.1-29), the Public Health Research Project from Shanghai Disease Control and Prevention Administration (Grant No. 2024GKM30), the National Health Commission of the People’s Republic of China (Grant No. Z155080000004), the Science and Technology Commission of Shanghai Municipality (Grant No. 20JC1410204) for the Seroepidemiological Study of Novel Coronavirus Pneumonia in Key Populations, the Tongren Hospital Talent Launch Project (Grant Nos. TR2023rc11, TR2024RC10), the Changning District Medical and Health Research Special Project (Grant No. CNKW2022Y08) and the Research Fund of Shanghai Tongren Hospital, Shanghai Jiaotong University School of Medicine (Grant Nos. TRYJ2024LC13, TRYJ2022LC10).

Data availability

The datasets generated and analyzed during this study are not publicly available due to concerns regarding participant confidentiality and the sensitive nature of the data involving stigmatized populations. To protect participant privacy, all data were de-identified and stored on access-restricted institutional servers with encrypted transmission protocols. Reasonable requests for data access may be directed to the corresponding author and will be considered on a case-by-case basis, subject to approval by the institutional ethics committee and in compliance with applicable data protection regulations and ethical guidelines.

Declarations

Ethics approval and consent to participate

This study adhered to the Declaration of Helsinki and was approved by the Ethics Committee of Shanghai University of Medicine & Health Sciences (Approval No. 2023-MSMMPOX-22-310222197604080237). Electronic informed consent was obtained prior to survey initiation; the consent page detailed anonymity, data protection, and voluntary participation. Data were transmitted via encrypted connections and stored on access-restricted servers. Only the research team had analytic access, and data were used solely for the purposes described in the informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.