Application of an extended information-motivation-behavioral skills model to predict HPV vaccine acceptability among Chinese gay, bisexual, and other men who have sex with men

ABSTRACT

Background: Human papillomavirus (HPV) causes substantial disease burden among gay, bisexual, and other men who have sex with men (GBMSM), yet HPV vaccination pathways for men remain evolving in mainland China. Purpose: To apply a culturally adapted, extended Information-Motivation-Behavioral Skills (IMB) model to identify psychosocial pathways associated with HPV vaccine acceptability among Chinese GBMSM. Methods: A cross-sectional online survey was conducted between December 2024 and January 2025 among 1,045 GBMSM living in mainland China. Measures included HPV-related information, health beliefs, anticipated regret, subjective norms, self-efficacy, and HPV vaccine acceptability. Measurement and structural equation models estimated direct and indirect pathways. Results: HPV vaccine acceptability was moderate (mean 3.40, SD 1.14). Subjective norms were the strongest predictor of acceptability (β = 0.68, p < .001), with additional positive effects from health beliefs (β = 0.17, p < .01), information (β = 0.14, p < .001), and self-efficacy (β = 0.13, p = .02). Anticipated regret showed a positive indirect effect through self-efficacy (unstandardized B = 0.056, 95% CI 0.01 to 0.12) but a negative direct association with acceptability (β = -0.11, p = .03). Conclusions: HPV vaccine acceptability among Chinese GBMSM was shaped predominantly by social norms, with self-efficacy acting as a key pathway through which psychosocial factors were associated with acceptability.

Introduction

Human papillomavirus (HPV) is one of the most common sexually transmitted infections worldwide and is causally linked to anal, penile, and oropharyngeal cancers. Gay, bisexual, and other men who have sex with men (GBMSM) experience a disproportionate burden of HPV infection and HPV-related disease because of high anal HPV prevalence, frequent multisite infection, and persistent exposure within sexual networks.¹ In mainland China, HPV vaccination policy and public messaging have historically focused on women and girls, while vaccination opportunities for men have been comparatively limited. During the survey period, a male HPV vaccination pathway was only beginning to emerge: the quadrivalent HPV vaccine was approved for males aged 9–26 y in January 2025, while male-facing services and public messaging remained limited and largely self-paid.^2,3 As a result, a substantial proportion of Chinese GBMSM remained unvaccinated or uncertain about where men could obtain vaccination during the period in which primary prevention could confer the greatest benefit.^4–9

Evidence from China suggests multilevel barriers to vaccination among sexual minority men. Structural barriers include high out-of-pocket costs, limited service visibility, uncertainty about where men can be vaccinated, and the lack of a stable male-facing vaccination pathway during the study period. At the individual level, previous studies and meta-analytic evidence have highlighted low HPV awareness, low perceived susceptibility, stigma related to sexual orientation, limited clinician recommendation, and inconsistent linkage between knowledge and action.^4–10 However, most existing studies have been descriptive or have examined isolated predictors, providing limited insight into how these factors interact to shape vaccine decision-making.

The Information-Motivation-Behavioral Skills (IMB) model offers a parsimonious account of health behavior change, positing that accurate information, sufficient motivation, and behavioral skills jointly support action.¹¹ HPV-related research has increasingly shown that IMB-informed approaches can be relevant to vaccine decision-making, including work among Malaysian men who have sex with men and mainland Chinese college women.^12,13 Within the model, behavioral skills are commonly represented through perceived capability or self-efficacy, which is central for translating intention into action.¹⁴

For HPV vaccination among Chinese GBMSM, the motivation component may require further refinement because cognitive, affective, and social motivations may operate through different pathways. In collectivist social contexts, perceived expectations from peers, partners, and healthcare professionals may act as a distinct social-motivational pathway rather than a residual component of personal attitudes.^15,16 Likewise, anticipated regret captures an affective forecast that is conceptually different from cognitive health beliefs such as perceived susceptibility, benefits, and barriers.^17–22 Prior studies have shown that anticipated regret is associated with vaccination intentions and uptake, and HPV vaccine acceptability has been examined among gay and bisexual men.^20,22,23

We therefore applied a culturally adapted, extended IMB model to examine how information, personal motivation, social motivation, and behavioral skills were associated with HPV vaccine acceptability. Using a large cross-sectional online survey of Chinese GBMSM, we aimed to (1) describe levels of information, motivation, behavioral skills, and HPV vaccine acceptability; (2) estimate direct and indirect pathways linking IMB constructs to acceptability; and (3) clarify the role of anticipated regret within the model.

Methods

Study design and participants

We conducted a cross-sectional online survey between December 2024 and January 2025. A total of 1,398 adult GBMSM were screened; 42 declined to participate and 311 did not meet eligibility criteria. The final baseline analytical sample comprised 1,045 respondents from four macro-regions of mainland China: East China (n = 540, 52%), Central China (n = 260, 25%), West China (n = 181, 17%), and Northeast China (n = 64, 6%). The study focused on vaccine acceptability rather than verified uptake because routine HPV vaccination access for men was limited and only beginning to change during the survey period.

Eligible participants were Chinese-speaking men aged 18–45 y who self-identified as gay or bisexual, or who reported oral or anal sex with another man, were currently residing in mainland China, used WeChat, and were able to provide electronic informed consent. Exclusion criteria were inability to complete the web questionnaire, duplicate or invalid submission, and participation in prior HPV vaccination pilot activities. This inclusive definition extended beyond a behavior-only MSM classification to better capture the social and identity dimensions of Chinese GBMSM.

Recruitment and data collection

Recruitment was coordinated through the Suzhou Center for Disease Control and Prevention and the Red Ribbon Social Service Center, an NGO serving sexual minority men in Jiangsu Province. Digital study invitations were disseminated through WeChat official accounts, community group chats, and NGO partner networks. Participants were encouraged to share the survey link with eligible peers, resulting in a chain-referral sampling approach that extended recruitment beyond Jiangsu Province. The questionnaire was hosted on Wenjuanxing, an encrypted web-survey platform commonly used in China, and took approximately 10 minutes to complete. Electronic informed consent was obtained before participants entered the questionnaire. Because survey links were redistributed through anonymous peer networks and group chats, platform-level view counts could not be reliably retained; however, screening, eligibility, consent, and completion counts were recorded.

Conceptual framework and hypotheses

The extended IMB framework specified HPV-related information as the information component, health beliefs and anticipated regret as personal motivation, subjective norms as social motivation, and HPV vaccination self-efficacy as the behavioral-skills component. Health beliefs represented cognitive appraisals of susceptibility, effectiveness, and barriers, whereas anticipated regret represented an affective forecast about forgoing vaccination. Subjective norms were modeled separately because perceived peer and clinician recommendations may constitute a distinct social-motivational pathway in a highly networked GBMSM context. We hypothesized that information, health beliefs, anticipated regret, and subjective norms would be associated with HPV vaccine acceptability directly and indirectly through self-efficacy. The conceptual framework is shown in Figure 1.

Figure 1.

Conceptual framework of the culturally adapted, extended information-motivation-behavioral skills model for HPV vaccine acceptability among Chinese GBMSM. The model separates personal motivation (health beliefs and anticipated regret) from social motivation (subjective norms), specifies HPV-related information as an observed composite, and positions self-efficacy as the behavioral skills component linking psychosocial inputs to HPV vaccine acceptability.

Measures

Background variables included sexual orientation, age, household registration, ethnicity, education, marital status, income, relationship status, frequency of gay venue attendance, frequency of dating app use, number of sexual partners, and perceived awareness of HPV.

Information was measured using a 13-item knowledge about HPV and HPV vaccine scale adapted from prior Chinese MSM intervention work.²⁴ Each item was coded as correct (1) or incorrect/don’t know (0), and item scores were summed to create a total score ranging from 0 to 13, with higher scores indicating greater HPV-related information.

Motivation was divided into personal and social motivation. Personal motivation comprised health beliefs and anticipated regret. Health beliefs were assessed through perceived susceptibility, perceived effectiveness, and perceived barriers using items adapted from prior HPV vaccine studies.^25–28 Items used five-point Likert-type response options from 1 (strongly disagree) to 5 (strongly agree). Perceived barriers were coded so that higher values represented stronger barriers; when modeled as part of the higher-order health belief construct, the barriers dimension loaded in the opposite direction. Anticipated regret was measured using two five-point items assessing regret about forgoing vaccination.^26–28 Social motivation was measured through subjective norms using three five-point items on perceived recommendations from clinicians and peers.^26–28

Behavioral skills were operationalized as HPV vaccination self-efficacy and measured using a four-item five-point scale.^24,28 This operationalization focused on perceived capability to decide about vaccination, obtain vaccination when wanted, complete the full schedule, and maintain control over the vaccination decision.

HPV vaccine acceptability was defined as stated willingness or readiness to seek, initiate, and complete HPV vaccination and was measured using a five-item five-point scale.^27,28 For multi-item psychosocial scales, item scores were averaged, with higher scores indicating higher levels of the construct.

Scale adaptation and validation

All scales adapted from prior studies were translated and culturally tailored for Chinese GBMSM using forward translation, blinded back-translation, expert review, and pilot testing with 25 community-recruited GBMSM. Minor wording refinements were made to ensure clarity and cultural appropriateness before field deployment. Psychometric validation was then conducted in the full baseline sample. Internal consistency was assessed using Cronbach’s α. Convergent validity was evaluated using standardized factor loadings, composite reliability, and average variance extracted. Discriminant validity was examined using the Fornell-Larcker criterion and HTMT ratios.

Data analysis

Descriptive statistics were generated for all variables. Background characteristics were first examined in relation to HPV vaccine acceptability using bivariable analyses, including independent-samples t tests, one-way analysis of variance, or simple linear regression, as appropriate. Candidate variables for the multivariable linear regression model were selected based on theoretical relevance and previous literature, with preliminary analyses used to confirm model specification. The final model was interpreted as an exploratory analysis of background correlates rather than as a causal model.

Measurement and structural models were estimated in AMOS version 29 using maximum likelihood. Information was specified as an observed composite because modeling 13 dichotomous knowledge items directly within the SEM would have substantially increased parameterization and identification complexity. Latent constructs included health beliefs, anticipated regret, subjective norms, self-efficacy, and HPV vaccine acceptability. Model fit was evaluated separately for the measurement model and structural model using χ²/df, comparative fit index (CFI), Tucker-Lewis index (TLI), normed fit index (NFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR), with model adequacy assessed according to established cutoff criteria.²⁹ Indirect effects were tested using bias-corrected bootstrap procedures with 2,000 resamples. Standardized path coefficients (β) with corresponding p values were reported for structural paths, while mediation estimates were reported consistently as unstandardized coefficients.

The baseline dataset contained complete responses for all indicators and composite variables included in the SEM; therefore, no mean/median substitution, deletion for missingness, or imputation was required. This reflected the web-survey design, which required valid responses to core SEM items before questionnaire submission. A post hoc sensitivity SEM excluding participants who reported prior HPV vaccination was also conducted to examine whether the structural paths were driven by previously vaccinated respondents. Distributions of quantitative scale scores were assessed using skewness and kurtosis (Supplementary Table S5). As a sensitivity analysis for the regression model, a generalized linear model with a gamma distribution and log link was fitted for HPV vaccine acceptability; because the direction and substantive conclusions were similar, linear regression was retained for interpretability; full results are reported in Supplementary Table S6.

Results

Participant characteristics

Among the 1,045 participants, 66.0% were aged 25–40 y, 70.2% identified as gay, and 20.0% as bisexual. Overall, 56.2% had an urban household registration and 47.0% reported monthly disposable income of 5,000 RMB or more. Approximately one-fifth (20.5%) reported attending gay venues at least monthly, and 31.7% reported using dating apps at least monthly. Most participants (85.1%) had heard of HPV, but only 22.2% reported knowing a lot about it (Table 1). Fifty-one participants (4.9%) reported having received at least one HPV vaccine dose before the study, indicating that vaccine uptake was uncommon in the baseline sample.

Table 1.

Participant characteristics (N = 1,045).

Domain	Category	n (%)
Age, years	18–24	228 (21.8)
	25–30	353 (33.8)
	31–40	337 (32.2)
	41–45	127 (12.2)
Sexual orientation	Gay	734 (70.2)
	Bisexual	209 (20.0)
	Prefer not to say	99 (9.5)
	Other	3 (0.3)
Household registration	Rural	458 (43.8)
	Urban	587 (56.2)
Ethnicity	Han nationality	1003 (96.0)
	Other nationalities	42 (4.0)
Highest education level	Junior high and below	131 (12.5)
	High/secondary school	195 (18.7)
	Junior college or technical secondary school	252 (24.1)
	College and above	467 (44.7)
Employment status	Employed	785 (75.1)
	Student	118 (11.3)
	Unemployed	142 (13.6)
Marital status	Married	211 (20.2)
	Unmarried	746 (71.4)
	Divorced	88 (8.4)
Monthly disposable income (RMB)	3000 or less	221 (21.1)
	3001–5000	333 (31.9)
	5000 or more	491 (47.0)
Relationship status	In relationship	522 (50.0)
	Single	523 (50.0)
Participation in gay venues	Never	324 (31.0)
	At least once a month	214 (20.5)
	Several times a year	253 (24.2)
	Once a year or less often	254 (24.3)
Frequency of dating app usage	Never	120 (11.5)
	At least once a month	331 (31.7)
	Several times a year	287 (27.5)
	Once a year or less often	307 (29.4)
Age at first same-sex intercourse	17 or below	278 (26.6)
	Above 17	632 (60.5)
	Never have	135 (12.9)
Number of sexual partners	None	122 (11.7)
	1–5	495 (47.4)
	6–10	202 (19.3)
	11+	226 (21.6)
Frequency of anal intercourse	Not at all	178 (17.0)
	1–2 times in last 12 months	242 (23.2)
	3–10 times	288 (27.6)
	More than 11 times	337 (32.2)
Last HIV test at clinic	Within 6 months	412 (39.4)
	More than 6 months ago	633 (60.6)
HIV status	Definitely negative	623 (59.6)
	Probably negative/positive	287 (27.5)
	Definitely positive	43 (4.1)
	Don’t know/Prefer not to say	92 (8.8)
Perceived awareness of HPV	Never heard of it	156 (14.9)
	Very little	332 (31.8)
	Some	325 (31.1)
	A lot	232 (22.2)

Percentages may not total 100 because of rounding.

In the final exploratory multivariable regression model, higher HPV vaccine acceptability was independently associated with urban household registration, monthly disposable income of 3,001–5,000 RMB, more frequent dating app use, having one to five recent sexual partners, and greater perceived awareness of HPV (Table 2). The highest VIF was 4.29, indicating no substantial multicollinearity.

Table 2.

Multivariable linear regression of background characteristics associated with HPV vaccine acceptability.

Domain	Comparison	B	SE	β	p value	VIF
Household registration	Urban vs. rural	0.24	0.06	0.10	<.001	1.23
Monthly disposable income (RMB)	3001–5000 vs. 3000 or less	0.27	0.10	0.11	.01	2.73
	5000 or more vs. 3000 or less	0.18	0.10	0.08	.08	3.16
Dating app usage	At least once a month vs. never	0.24	0.10	0.10	.02	3.21
	Several times a year or less often vs. never	0.32	0.11	0.13	<.001	3.09
Number of sexual partners	1–5 vs. none	0.27	0.12	0.12	.02	4.29
	6–10 vs. none	0.01	0.13	0.00	.96	3.08
	11+ vs. none	0.06	0.14	0.02	.66	3.10
Perceived awareness of HPV	Very little vs. never heard of it	0.37	0.10	0.15	<.001	2.95
	Some vs. never heard of it	1.07	0.10	0.46	<.001	3.15
	A lot vs. never heard of it	1.52	0.13	0.45	<.001	2.15

The highest VIF was 4.29, suggesting moderate but acceptable overlap among behavioral covariates.

Psychosocial measures and measurement model

HPV vaccine acceptability was moderate (mean 3.40, SD 1.14). The item reflecting consideration of vaccination had the highest mean score, whereas willingness to attend a sexual health clinic for vaccination was lowest. Perceived susceptibility was the weakest health-belief component, while anticipated regret, subjective norms, and self-efficacy all scored above the scale midpoint. Distributions of the quantitative psychosocial measures were not strongly skewed (skewness range, −0.77 to 0.23; kurtosis range, −0.79 to 0.33; Supplementary Table S5), and visual patterns are shown in Supplementary Figure S1.

All latent constructs demonstrated acceptable psychometric properties, with Cronbach’s α ranging from 0.72 to 0.94, composite reliability from 0.72 to 0.94, and AVE from 0.46 to 0.80 (Table 3). Discriminant validity was supported by Fornell-Larcker criteria, as the square root of AVE for each construct exceeded its correlations with other constructs. The correlation between subjective norms and HPV vaccine acceptability was high (r = 0.81) but remained below the square root of AVE for subjective norms (0.89). Additional HTMT analyses supported discriminant validity under the 0.90 criterion; however, the HTMT ratio between subjective norms and HPV vaccine acceptability was 0.876, indicating that this distinction should be interpreted cautiously (Supplementary Table S4).

Table 3.

Reliability, convergent validity, and discriminant validity of the latent constructs.

Construct	Cronbach’s α	CR	AVE	√AVE	HVA	PS	PE	PB	AR	SN	SE
HPV vaccine acceptability	0.93	0.93	0.73	0.86	1.00
Perceived susceptibility	0.94	0.94	0.80	0.89	0.33	1.00
Perceived effectiveness	0.93	0.93	0.73	0.85	0.54	0.36	1.00
Perceived barriers	0.72	0.72	0.46	0.68	−0.53	−0.25	−0.50	1.00
Anticipated regret	0.83	0.83	0.71	0.84	0.57	0.28	0.57	−0.45	1.00
Subjective norms	0.92	0.92	0.79	0.89	0.81	0.35	0.57	−0.54	0.61	1.00
Self-efficacy	0.87	0.87	0.63	0.79	0.72	0.31	0.55	−0.49	0.66	0.76	1.00

Cronbach’s α = internal consistency reliability; CR = composite reliability; AVE = average variance extracted. Diagonal values (√AVE) are shown in bold and should exceed the inter-construct correlations (Fornell – Larcker criterion). One construct had an AVE slightly below the conventional threshold of 0.50; however, composite reliability exceeded 0.70, indicating acceptable convergent validity according to the recommendations of Fornell and Larcker.

HVA, HPV vaccine acceptability; PS, perceived susceptibility; PE, perceived effectiveness; PB, perceived barriers; AR, anticipated regret; SN, subjective norms; SE, self-efficacy.

The CFA measurement model indicated good fit (χ² = 561.62, df = 278, χ²/df = 2.02, CFI = 0.987, TLI = 0.985, NFI = 0.975, RMSEA = 0.031, SRMR = 0.025). Standardized factor loadings ranged from 0.65 to 0.92 (Supplementary Tables S1-S3).

Structural model and mediation

The structural model also showed acceptable fit (χ² = 891.76, df = 311, χ²/df = 2.87, CFI = 0.974, TLI = 0.971, NFI = 0.961, RMSEA = 0.042, SRMR = 0.085). In this model, information (β = 0.14, p < .001), health beliefs (β = 0.17, p < .01), subjective norms (β = 0.68, p < .001), and self-efficacy (β = 0.13, p = .02) were positively associated with HPV vaccine acceptability (Table 4). Subjective norms also positively predicted self-efficacy (β = 0.56, p < .001), and anticipated regret was positively associated with self-efficacy (β = 0.32, p < .001). In contrast, the direct path from anticipated regret to acceptability was negative (β = −0.11, p = .03). The overall SEM is shown in Figure 2; the model explained 57.5% of the variance in HPV vaccine acceptability.

Table 4.

Structural equation model estimates for HPV vaccine acceptability.

Hypothesized path	Std. estimate	SE	CR	p value	Supported
Health beliefs → HPV vaccine acceptability	0.17	0.15	2.58	<.01	Yes
Subjective norms → HPV vaccine acceptability	0.68	0.07	11.86	<.001	Yes
Information → HPV vaccine acceptability	0.14	0.006	7.06	<.001	Yes
Anticipated regret → HPV vaccine acceptability	−0.11	0.07	−2.13	.03	No
Self-efficacy → HPV vaccine acceptability	0.13	0.08	2.40	.02	Yes
Health beliefs → self-efficacy	0.06	0.11	0.83	.41	No
Subjective norms → self-efficacy	0.56	0.04	11.27	<.001	Yes
Anticipated regret → self-efficacy	0.32	0.05	6.45	<.001	Yes
Information → self-efficacy	0.09	0.005	4.42	<.001	Yes

Std. estimate denotes the standardized path coefficient.

Figure 2.

Final overall structural equation model for HPV vaccine acceptability among Chinese GBMSM. Standardized structural path coefficients are displayed, and the model explained 57.5% of the variance in acceptability. Dashed arrows indicate non-significant pathways.

Bootstrap mediation analyses using consistently unstandardized estimates showed significant indirect effects through self-efficacy for anticipated regret (B = 0.056, 95% CI 0.010 to 0.120), subjective norms (B = 0.082, 95% CI 0.020 to 0.160), and information (B = 0.004, 95% CI 0.001 to 0.010), while the indirect effect of health beliefs was not significant (B = 0.016, 95% CI −0.020 to 0.060) (Table 5). Taken together, anticipated regret showed a negative direct association and a positive indirect association via self-efficacy, a pattern that should be interpreted as associational and potentially sensitive to model specification. A sensitivity SEM excluding the 51 participants who reported prior HPV vaccination (N = 994) produced the same substantive pattern of associations, indicating that the overall findings were not driven by previously vaccinated participants (Supplementary Table S7).

Table 5.

Mediation analysis of HPV vaccine acceptability through self-efficacy (bias-corrected bootstrap, 2,000 resamples).

Independent variable	a path (B)	b path (B)	Total effect (B)	Direct effect (B)	Indirect effect (B)	95% CI
Information	0.020***	0.182*	0.047***	0.044***	0.004*	0.001 to 0.010
Health beliefs	0.090	0.182*	0.393**	0.376**	0.016	−0.020 to 0.060
Anticipated regret	0.307***	0.182*	−0.087	−0.143*	0.056*	0.010 to 0.120
Subjective norms	0.453***	0.182*	0.856***	0.773***	0.082*	0.020 to 0.160

All mediation estimates are unstandardized coefficients. The a path refers to the path from the independent variable to self-efficacy; the b path refers to self-efficacy to HPV vaccine acceptability. *p < .05, **p < .01, ***p < .001.

Discussion

This study applied a culturally adapted, extended IMB model to HPV vaccine acceptability among Chinese GBMSM and found that both the measurement and structural models demonstrated acceptable fit. Subjective norms emerged as the strongest predictor of acceptability in the SEM, operating both directly and indirectly through self-efficacy. This finding is consistent with previous studies among MSM populations showing that perceived peer and clinician recommendations are strongly associated with vaccine acceptability and intention.^26,28 Information and health beliefs were also positively associated with acceptability, although their effects were smaller than that of subjective norms. Similar patterns have been reported in China and other settings, where knowledge alone may be insufficient to translate into vaccination intention without supportive social context.^23,27 Self-efficacy appeared to function as a key behavioral pathway linking information and motivation to vaccine acceptability. This is consistent with IMB-based research demonstrating that behavioral skills, particularly self-efficacy, are critical for translating motivation into health behavior. Together, these findings highlight the prominent role of social influence alongside individual capability in shaping vaccination-related decision-making in this population.

The high correlation and HTMT ratio between subjective norms and acceptability should be interpreted cautiously. Although Fornell-Larcker and HTMT results supported discriminant validity under conventional criteria, the SN-HVA HTMT value approached the 0.90 threshold. This may indicate partial conceptual proximity between perceived social endorsement and personal vaccine acceptability. Some subjective-norm items, particularly descriptive norm items referring to whether peers are getting vaccinated, may partially overlap with personal willingness in a highly networked GBMSM context. Future studies should further distinguish perceived external expectations, perceived community approval, and autonomous willingness to vaccinate.

A notable finding was the dual pathway observed for anticipated regret. Individuals who anticipated regretting non-vaccination reported higher self-efficacy, suggesting that affective concern may co-occur with greater perceived capability or planning. However, anticipated regret was negatively associated with acceptability through a direct pathway after other correlated predictors were included. Given the cross-sectional design and the two-item measurement of anticipated regret, this pattern should not be interpreted as evidence of a verified negative psychological mechanism. Alternative explanations include statistical suppression, high intercorrelations among motivational constructs, measurement constraints, omitted access-related variables, or model specification issues. Future longitudinal or experimental studies are needed to determine whether regret-based messages have beneficial, null, or counterproductive effects in this population.

These findings refine the IMB model in three ways. First, they support partitioning motivation into personal and social components when examining vaccination in collectivist or highly networked contexts. The strong subjective-norms pathway suggests that vaccine decisions among Chinese GBMSM are closely linked to peer and clinician expectations rather than solely individual attitudes. Second, the findings show that anticipated regret should not be treated as interchangeable with perceived susceptibility or severity. Cognitive appraisals estimate the likelihood and consequences of harm, whereas anticipated regret represents an affective forecast of self-evaluation after inaction.^17–22 Third, the results highlight the central role of self-efficacy as the mechanism through which information, norms, and affective motivation are translated into acceptability.

Accordingly, the anticipated-regret result should be used primarily to refine hypotheses rather than to justify stand-alone regret-based or fear-based communication. In practical terms, affective motivation may be more useful when paired with concrete coping support, such as where to obtain vaccination, how to manage cost, how to complete the schedule, and how to ask clinicians for non-stigmatizing guidance. This interpretation is consistent with the observed positive indirect association through self-efficacy but remains tentative because causal ordering cannot be established from the present data.

The strong association of subjective norms with acceptability suggests several intervention priorities. Peer-led communication within digital social networks and community organizations may be especially valuable for legitimizing HPV vaccination among GBMSM. A growing sexual-health literature suggests that social media and network-based outreach can be effective channels for reaching key populations, while platform-based recruitment strategies may improve reach among men with higher risk profiles.^30,31 Healthcare professionals may represent a key normative source; therefore, training clinicians and community-based providers to deliver clear, non-stigmatizing, and explicit vaccination recommendations could strengthen vaccine confidence and visibility.

Programs may also benefit from directly supporting self-efficacy. Practical navigation support, appointment information, cost transparency, reminders, and clinic-facing guidance are likely to reduce friction between willingness and action. Consistent with the mediating role of self-efficacy observed in this study, such strategies may help translate motivation into actionable vaccination behavior. Future research should also examine whether disclosure of sexual orientation or same-sex behavior to healthcare providers facilitates clinician recommendation, service navigation, and HPV vaccination uptake. Structural measures are essential: affordability support, clearer male-facing service pathways, and stronger public-health endorsement would help convert motivation into uptake, particularly for lower-income participants whose acceptability was lower in the multivariable regression model. At the policy level, male-inclusive vaccination pathways should be strengthened through clearer eligibility communication, more visible service provision, and affordability support for adult men at elevated risk.

Limitations

Several limitations should be acknowledged. The cross-sectional design precludes causal inference. Recruitment used convenience and online chain-referral sampling and resulted in overrepresentation of participants from eastern and urban regions, limiting generalizability to less-connected GBMSM. Survey view counts were not reliably available because links were redistributed through anonymous peer networks. All measures were self-reported, and HPV vaccination status was not clinically verified; 4.9% reported prior HPV vaccination, so acceptability among a small subset may reflect post-vaccination attitudes rather than prospective intention. A sensitivity SEM excluding the 51 participants who reported prior HPV vaccination produced a similar pattern of estimates. Behavioral skills were represented only by self-efficacy, as actual vaccine uptake behaviors could not be measured robustly in a low-access setting. Finally, the high SN-HVA correlation and HTMT ratio, although within conventional discriminant-validity criteria, suggest that future work should further disentangle external expectations from autonomous willingness.

Conclusion

HPV vaccine acceptability among Chinese GBMSM was moderate and was most strongly associated with social motivation, particularly subjective norms. Information and subjective norms were associated with acceptability both directly and indirectly through self-efficacy, whereas health beliefs were directly associated with acceptability but did not show a significant indirect association through self-efficacy. Anticipated regret showed a dual pattern, with a positive indirect association via self-efficacy and a negative direct association with acceptability, but this finding should be interpreted cautiously because suppression or model specification issues may contribute to the pattern. These findings provide a more nuanced understanding of how cognitive, social, and affective factors relate to HPV vaccine acceptability in this population. Interventions should therefore combine normative endorsement, practical self-efficacy support, and structural facilitation to improve uptake among Chinese GBMSM.

Biographies

Li Ping Wong is Professor in the Department of Social and Preventive Medicine, Faculty of Medicine, Universiti Malaya, and Head of the Centre for Population Health (CePH). Her research focuses on social and behavioral medicine, including vaccination, emerging infectious diseases, sexual and reproductive health, and environment-health interfaces. She holds a PhD in Epidemiology from Universiti Malaya.

Sin How Lim is Senior Lecturer in the Department of Social and Preventive Medicine, Faculty of Medicine, Universiti Malaya, Kuala Lumpur. He holds a PhD in Epidemiology from the University of Pittsburgh. His research centers on HIV/STI prevention among key populations, tobacco control, and syndemics, using mixed-methods and quantitative approaches.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All requests for materials and raw data should be directed to the corresponding authors. Deidentified individual participant data and the questionnaire may be available upon reasonable request, subject to institutional approval and applicable data protection requirements because the dataset contains potentially sensitive information from a sexual minority population.

Ethics approval

The study was approved by the Jiangsu College of Nursing Institutional Review Board (protocol #JSCN-ME-2024032801). Electronic informed consent was obtained from all participants before participation. Data were anonymized, and confidentiality was strictly maintained.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2026.2708426