Sexual Behavior and Perceived Risk for Oropharyngeal Cancer Among Men Who Have Sex With Men: A Psychometric Scale Validation

The results from this cross-sectional study demonstrate that men who have sex with men assess their risk for human papillomavirus–associated oropharyngeal cancer based on lifetime numbers of cisgender male sex partners, rimming partners, and fellatio partners.

Abstract

Background

Men who have sex with men (MSM) are at increased risk for human papillomavirus–associated oropharyngeal cancer (HPV-OPC). The objective of this analysis was to create a psychometrically validated scale to measure perception of risk for HPV-OPC.

Methods

We conducted an exploratory and a confirmatory factor analysis to determine and confirm the latent factor structure. We used a path diagram to evaluate the relationship between the validated scale and perceived risk for HPV-OPC. The model was determined to be a good fit if it met all criteria: root mean square error of approximation ≤0.06, standardized root mean residual ≤0.08, Comparative Fit Index ≥0.90, and Tucker-Lewis Index ≥0.90. We report standardized estimates and 95% confidence intervals.

Results

This cross-sectional study recruited 1315 MSM. A majority (73.33%) of MSM had performed fellatio on ≥20 partners, 36.98% had rimmed ≥20 partners, and 5.31% had performed cunnilingus on ≥10 partners in their lifetime.

Six sexual history survey items loaded onto 2 latent factors: sexual risk behaviors: class 1 and sexual risk behaviors: class 2. The final model statistics indicated good fit: root mean square error of approximation = 0.064, standardized root mean residual = 0.059, Comparative Fit Index = 0.996, and Tucker-Lewis Index = 0.993. Sexual risk behaviors: class 1 was associated with greater perceived risk for HPV-OPC (0.217; 95% confidence interval, 0.138–0.295). Age, HIV status, HPV vaccination status, and sexual risk behaviors: class 2 were not associated with perceived risk for HPV-OPC.

Conclusion

Men who have sex with men assessed risk for HPV-OPC based on their lifetime number of cisgender male sexual partners, rimming partners, and fellatio partners but not other sexual behaviors. Men who have sex with men may be responsive to future HPV-OPC educational interventions and opportunities for screening.

Oropharyngeal cancer (OPC) is cancer of the posterior one-third of the tongue, soft palate, tonsils, or posterior wall of the oropharynx.¹ Oropharyngeal cancer is one of the few head and neck cancers that has steadily increased in incidence the past decade.² The increasing incidence of OPC is attributable to human papillomavirus (HPV), which is a sexually transmissible infection. Oropharyngeal cancer incidence differs by gender, with 4.68 cases per 100,000 (95% confidence interval [CI], 4.62–4.75) in men versus 0.91 cases per 100,000 in women (95% CI, 0.88–0.93) diagnosed each year.³

There are 2 main classifications of OPC that are defined by separate etiologies and risk factors: classic and HPV-associated oropharyngeal (HPV-OPC). Classic OPC is related to heavy tobacco and alcohol use, whereas HPV-OPC is caused by persistent infection with HPV.¹ Patients with HPV-OPC have improved responses to cancer treatments and greater 5-year survival compared with HPV-negative OPC patients.⁴ Still, advanced HPV-OPC imparts significant morbidity and not insignificant mortality. Thus, early detection of HPV-OPC would have a substantial impact on morbidity, mortality, and quality of life.

Oral HPV infection is associated with the development of HPV-OPC.⁵ Analyses of data from the National Health and Nutrition Examination Survey reported that older age, infection with human immunodeficiency virus (HIV), lack of HPV vaccinations, history of sexually transmitted infections (STIs), earlier age at sexual debut, number of sexual partners, and oral sex are associated with oral HPV infections in men.^6–8 Among men who reported having same gender sex partners, the prevalence of high-risk (HR) HPV infection was 12.7%, whereas among men who did not report this behavior, the prevalence of HR-HPV infection was 6.8%.⁷ In particular, the prevalence of HR-HPV infection was highest (22.2%) among men who reported having 2 or more same gender lifetime oral sex partners.⁷ Heck et al.⁹ reported that the strongest risk factor for HPV-OPC in men was sex with another man (odds ratio [OR], 8.89; 95% CI, 2.14–36.8).

Although most oral HPV infections become undetectable for unknown reasons, a small proportion of people infected with oncogenic HPV types will have persistent infections, potentially causing OPC.¹⁰ It is estimated that up to 73% of OPCs are caused by HPV.^1,11 Oncogenic HPV-16 is the most prevalent (>80%) type detected in OPC tumors and, to a lesser extent, HPV-18 (3%).¹² Several small case-control studies have reported associations between greater lifetime numbers of sex or oral sex partners with HPV-positive OPC compared with controls.^5,13,14 Men who have sex with men (MSM) in particular are at increased risk for HPV-OPC, given high oral HPV prevalence, high levels of sexual activity, and low HPV vaccine uptake.^9,15

Currently, there are no approved screening methods for the early detection of HPV-OPC, but understanding what behaviors MSM use to conceptualize their risk for this cancer may inform future preventative strategies.¹⁶ Specifically, little is known about MSM's perception of risk for HPV-OPC. Many theories of behavior, including the Health Belief Model and protection motivation theory, posit that an individual's belief about their risk of developing a disease (perceived risk) is a motivator for behavior change.¹⁷ Although overemphasis of a health threat may lead to inaction, understanding which factors MSM use to assess risk for HPV-OPC is useful to facilitate future health behavior change, including HPV vaccination and cancer screening uptake.¹⁸

The objective of this analysis is to measure MSM's perception of HPV-OPC risk by creating a psychometrically validated sexual risk scale. We conducted an exploratory factor analysis (EFA), confirmatory factory analysis (CFA), and structural equation model (SEM). We hypothesized that several sexual history questions, HPV vaccination status, HIV status, and age would predict perceived risk for HPV-OPC.

MATERIALS AND METHODS

Survey Design

This cross-sectional study recruited MSM from across the United States to complete a single online survey. Individuals were eligible to participate if they met all criteria: (1) 18 years or older; (2) reported having any type of sex with a man in the past 5 years; (3) identified their sexual orientation as gay, bisexual, or another sexual minority; and (4) self-identified as a man (including transgender men). Individuals who did not live in the United States were not eligible to participate.

The present study recruited MSM from 2 online dating sites (Scruff and Jack'd; Perry Street Software Inc., New York, NY) for 5 days between February and March 2022. During the phased, preprogrammed advertising period, all active users in a geo-targeted area were shown the recruitment ad with an embedded link to the survey screener (Qualtrics, Provo, UT). The survey was only available to individuals with a profile on one of the dating sites. If an individual met all eligibility criteria, he or she was directed to the consent sections and subsequently the main survey. Participants who completed the survey were compensated with a $50 Amazon e-gift card. The University of Minnesota Institutional Review Board approved all study materials.

Measures

The research team designed all sexual history survey questions. It was hypothesized that 6 sexual history survey items would measure “sexual risk.” For each item, participants were asked to report their number of lifetime sexual partners and lifetime oral sex partners. Because of low response rates for some categories (i.e., <5%), some response categories were collapsed. These items were all treated as ordinal for the latent variable analysis. Each item is described in more detail in the next sections.

Lifetime Sexual Partners

Participants were asked to self-report their lifetime number of different cisgender male sex partners. Cisgender male sex partners were defined by any sexual contact such as masturbating each other, oral, or anal sex. Ordinal responses included “1–10 men,” “11–20 men,” “21–100 men,” “101–1000 men,” and “>1000 men.”

In 2 additional questions, participants reported their lifetime number of different cisgender female sex partners and different transgender sex partners, defined as any sexual contact such as masturbating each other, oral, anal, or vaginal sex. Response options for both questions were categorized as “none” or “one or more partners” for this analysis.

Lifetime Oral Sex Partners

First, participants were asked to report the number of different partners with a penis on whom they had performed oral sex or fellatio (with or without ejaculation). For this analysis, responses were treated as ordinal and included “no fellatio partners,” “1–10 partners,” “11–20 partners,” “21–100 partners,” “101–1000 partners,” and “>1000 partners.”

Next, participants reported the number of different partners of any gender on whom they had performed oral-anal sex (rimming). The responses were treated as ordinal and categorized as “no partners rimmed,” “1–10 partners,” “11–20 partners,” “21–100 partners,” “101–1000 partners,” and “>1000 partners.”

Lastly, participants reported the number of different partners with a vagina on whom they had performed oral sex (cunnilingus). This item was analyzed categorically and included response options “No cunnilingus partners,” “1–10 partners,” and “>10 partners.”

Covariates

For this analysis, participants self-reported their HIV status as positive, negative, or unknown and their age in years. Human papillomavirus vaccination status was dichotomized as vaccinated (“at least one dose”) or other (“no doses” or “unsure of status”).

Perceived Risk

Perceived risk for HPV-OPC was obtained through a single survey question asking, “Compared to the average person, I believe my risk of getting oropharyngeal cancer is:” which was rated on a 5-point Likert scale (“strongly agree” to “strongly disagree”). The responses were approximately normally distributed (mean [SD], 3.283 [0.915]; skew, −0.381) and treated ordinally for the analysis.

Statistical Analysis

We report the frequency of self-reported demographics as well as sexual and medical characteristics for all study participants.

We conducted a missingness analysis to assess whether the missingness in observed analytic variables and the latent variables were associated with demographics of interest using the missing_compare function in the finalfit R package. Continuous data were compared with a Kruskal-Wallis test. Discrete data were compared with a χ² test. Results showed that the missingness in analytic variables was associated with age. In addition, the missingness in HIV, rimming, transgender sexual risk, and cunnilingus were associated with sexual orientation (Supplemental Digital Content Table 1, http://links.lww.com/OLQ/B39). Tests with reliable results are reported in Supplemental Digital Content Table 1.

The associations between individual sexual history items and perceived risk were stated a priori; however, we did not specify a latent factor structure for the sexual history items. Participants with missing data on any analytic variable were excluded. The factor structure for the 6 sexual history items was evaluated using an EFA with a maximum likelihood estimation approach. A scree plot was used to determine the number of factors included in the EFA. We allowed the factors to be correlated using an oblimin rotation. The factor structure was considered valid if the variance explained by the factor(s) was greater than 60%.¹⁹

After the EFA, a CFA was conducted to confirm the factor structure and improve overall model fit. The model was considered adequate if all the model fit index cutoff criteria were met: standardized factor loadings >0.3, root mean square error of approximation (RMSEA) ≤0.06, standardized root mean square residual (SRMR) ≤0.08, Comparative Fit Index (CFI) ≥0.90, and Tucker-Lewis Index (TLI) ≥0.90.^20,21 The only modifications implemented were incorporating covariances between item residuals (based on changes in χ² values).

Finally, using the results of the EFA and CFA, a full-hypothesized path diagram was created for use in a SEM (see the Results section). The SEM model was estimated with the “lavaan” package in R using the default estimation method settings.²² Because the sexual history items were treated as ordinal, “lavaan” defaulted to the diagonally weighted least squares estimator. All the modifications from the CFA were incorporated into the SEM. Model fit criteria for the SEM model were based on the same cutoff criteria for RMSEA, SRMR, CFI, and TLI.

Statistical analyses were conducted in STATA I/C 16.1 and R Studio. All statistical hypothesis tests were 2-sided and considered significant if P < 0.05. We reported the standardized regression estimates and 95% CIs.

RESULTS

The eligibility screener was initiated by 4464 individuals, 1836 completed consent, and 1722 surveys were deemed complete and valid. After removing individuals who did not complete the first question of the main survey on age (n = 23), self-reported a diagnosis of oral or OPC (n = 3), or did not provide complete data for all variables of interest (n = 381), the final sample size was 1315.

Demographic data are reported in Table 1. The average age of the sample was 41.45 years. Most participants identified as cisgender men (95.30%) and as gay (79.79%) or bisexual/pansexual (18.55%). The sample included 58.68% White non-Hispanic, 16.75% Black non-Hispanic, and 15.29% Hispanic participants.

TABLE 1.

Self-Reported Demographic Characteristics of MSM in the Study (n = 1648)

	n	%
Demographics
Age, mean (SD), y	41.5 (12.7)
Race/ethnicity
White non-Hispanic	967	58.7
African American/Black non-Hispanic	276	16.8
Asian non-Hispanic	53	3.2
Hispanic	252	15.3
American Indian or Alaska Native non-Hispanic	7	0.4
Native Hawaiian or other Pacific Islander non-Hispanic	4	0.2
Other non-Hispanic	14	0.9
≥2 races/other non-Hispanic	75	4.6
Current gender
Cisgender man	1602	95.3
Transgender man	15	0.9
Nonbinary	55	3.3
Self-described	9	0.5
Sex assigned at birth
Male	1679	99.0
Female	17	1.0
Sexuality
Gay	1346	79.8
Bisexual or pansexual	313	18.5
Queer or other self-described	28	1.7
Education
Less than high school	11	0.7
High school or GED	135	8.0
Some college	399	23.7
College degree or higher	1138	67.6
Relationship status
Not partnered	1083	64.3
Partnered	601	35.7
Annual individual income, median (IQR)‡	$50,000 ($29,000–$80,000)

GED indicates general education development or general education diploma; IQR, interquartile range.

Self-reported medical characteristics are presented in Table 2. Only 32.68% of participants self-reported at least 1 dose of the HPV vaccine. A large proportion (66.62) of participants reported being diagnosed with at least 1 STI in their lifetime, and 16.85% were living with HIV.

TABLE 2.

Self-Reported Sexual and Medical Characteristics of MSM (n = 1396)

	n	%
Sexual characteristics
Lifetime no. cisgender male sex partners*
1–10 partners	167	12.0
11–20 partners	137	9.8
21–100 partners	461	33.0
101–1000 partners	511	36.6
>1000 partners	120	8.6
Lifetime no. cisgender female sex partners*
None	750	53.9
≥1 partner	642	46.1
Lifetime no. transgender sex partners*
None	1098	79.2
≥1 partner	288	20.8
Lifetime no. fellatio partners†
None	9	0.7
1–10	198	14.2
11–20	165	11.9
21–100	492	35.4
101–1000	414	29.8
>1000	114	8.2
Lifetime no. rimming partners‡
None	141	10.2
1–10	524	38.0
11–20	204	14.8
21–100	315	22.8
>100	195	14.1
Lifetime no. cunnilingus partners§
None	851	61.9
1–10	452	32.9
>10	73	5.3
Medical characteristics
HPV vaccination status
Vaccinated (≥1 dose)	530	32.7
Other (0 doses or unsure)	1092	67.3
HIV status
HIV-positive	231	16.9
HIV-negative or unknown	1140	83.1
History of STIs¶
≥1 infection	868	66.6
Never	435	33.4
Outcome
Perceived risk for HPV-OPC
Much lower	70	5.0
Somewhat lower	127	9.1
About the same	645	46.1
Somewhat higher	450	32.2
Much higher	106	7.6

Survey question asked “Compared to the average person, I believe my risk of getting oropharyngeal cancer is:”

*Sex partners: number of different partners with whom participant engaged in mutual masturbation, oral, anal, or vaginal sex.

^†Fellatio partners: number of different partners with a penis on whom participant performed oral sex.

^‡Rimming partners: number of different partners anuses on whom participant performed oral sex.

^§Cunnilingus partners: number of different partners with a vagina on whom participant performed oral sex.

^¶STIs options: syphilis, gonorrhea, chlamydia, HPV, herpes simplex virus (genital or anal), hepatitis A, hepatitis B, hepatitis C, nongonococcal urethritis/nonspecific urethritis, and/or trichomoniasis.

The sample reported high lifetime numbers of sexual partners (Table 2). More than 78% of participants reported sexual contact with more than 20 cisgender male partners in their lifetime. A smaller proportion reported at least 1 cisgender female sexual partner (46.12%) or 1 transgender sexual partner (20.78%) in their lifetime. More than 65% of participants reported performing fellatio on 21 to 1000 different partners, and 8.20% reported more than 1000 lifetime fellatio partners. Although less prevalent, 36.98% had rimmed 21 or more partners and 5.31% had performed cunnilingus on more than 10 partners.

Based on the results of the EFA and CFA, the 6 sexual history survey items loaded onto 2 correlated latent factors: sexual risk behaviors: class 1 (ordinal α = 0.889) and sexual risk behaviors: class 2 (ordinal α = 0.745). Factor loadings were about 0.30 or larger, indicating that each factor explains the variance among its respective items well.

The CFA structure was included in the final SEM path diagram (Fig. 1). The results from the SEM analysis are presented in Table 3. The model fit indices indicated good model fit: RMSEA = 0.064, SRMR = 0.059, CFI = 0.996, and TLI = 0.993. The reference items were lifetime male sexual partners and lifetime female sexual partners for the sexual risk behaviors: class 1 and sexual risk behaviors: class 2 factors, respectively. From the regression model, a 1-SD increase in sexual risk behaviors: class 1 was associated with a 0.217-SD increase (on average) in perceived risk for HPV-OPC among men in the sample, holding all other covariates constant. This standardized estimate is relatively precise (95% CI, 0.138–0.295). Standardized estimates for other predictors were not statistically significant: sexual risk behaviors: class 2 (0.080; 95% CI, −0.073 to 0.233), HIV status (0.018; 95% CI, −0.042 to 0.078), HPV vaccination status (0.033; 95% CI, −0.033 to 0.098), and age (−0.060; 95% CI, −0.141 to 0.021).

Figure 1.

Fitted path diagram with standardized estimates for the SEM. ***P < 0.0001.

TABLE 3.

Standardized Estimates, 95% CIs, and Fit Indices From SEM Predicting Perceived Risk for HPV-OPC

	Standardized Estimate	95% CI
Regression
HPV vaccination status	0.033	−0.033 to 0.098
Age	−0.060	−0.141 to 0.021
HIV status	0.018	−0.042 to 0.078
Sexual risk behaviors: class 1	0.217	0.138 to 0.295*
Sexual risk behaviors: class 2	0.080	−0.073 to 0.233
Loadings
Sexual risk behaviors: class 1—lifetime cisgender male sexual partners	0.817	0.754 to 0.880*
Sexual risk behaviors: class 1—lifetime fellatio partners	0.842	0.781 to 0.904*
Sexual risk behaviors: class 1—lifetime rimming partners of any gender	0.785	0.727 to 0.842*
Sexual risk behaviors: class 2—lifetime cisgender female sexual partners	0.469	0.333 to 0.606*
Sexual risk behaviors: class 2—lifetime transgender sexual partners	0.505	0.355 to 0.655*
Sexual risk behaviors: class 2—lifetime cunnilingus partners	0.331	0.191 to 0.470*
Covariances
Sexual risk behaviors: class 1—perceived lower risk behaviors	0.381	0.248 to 0.514*
Sexual risk behaviors: class 1—HIV status	0.246	0.195 to 0.298*
Sexual risk behaviors: class 2—HIV status	0.111	−0.008 to 0.229
HPV vaccination status—age	−0.344	−0.391 to −0.297*
Sexual risk behaviors: class 1—age	0.253	0.195 to 0.312*
Sexual risk behaviors: class 2—age	0.352	0.226 to 0.478*
Lifetime cisgender female sexual partners—lifetime cunnilingus partners	0.757	0.710 to 0.803*
Lifetime cisgender male sexual partners—lifetime fellatio partners	0.648	0.537 to 0.760*
Lifetime transgender sexual partners—lifetime cunnilingus partners	0.374	0.282 to 0.465*
Variances
Lifetime cisgender male sexual partners—lifetime cisgender male sexual partners	0.333	0.230 to 0.436*
Lifetime fellatio partners—lifetime fellatio partners	0.291	0.187 to 0.394*
Lifetime rimming partners of any gender—lifetime rimming partners of any gender	0.384	0.294 to 0.475*
Lifetime cisgender female sexual partners—lifetime cisgender female sexual partners	0.780	0.652 to 0.908*
Lifetime transgender sexual partners—lifetime transgender sexual partners	0.745	0.594 to 0.897*
Lifetime cunnilingus partners—lifetime cunnilingus partners	0.891	0.798 to 0.983*
Perceived HPV-OPC risk—perceived HPV-OPC risk	0.935	0.899 to 0.970*
Model fit
RMSEA (good fit criteria: ≤0.06)	0.064
SRMR (good fit criteria: ≤0.08)	0.059
CFI (good fit criteria ≥0.90)	0.996
TLI (good fit criteria ≥0.90)	0.993

*P < 0.0001.

DISCUSSION

To our knowledge, this is the first study to measure the impact of sexual risk on the perception of HPV-OPC in MSM. Men in the study accurately perceived themselves to be at greater risk for HPV-OPC compared with the average person based on their lifetime number of cisgender male sexual partners, rimming partners, and fellatio partners. However, MSM did not perceive themselves to be at risk for this cancer based on their lifetime number of cisgender female sexual partners, transgender sexual partners, cunnilingus partners, age, HIV status, or HPV vaccination status. Given the results of this study, MSM may be amenable to future novel screening methods for this cancer and may benefit from targeted interventions to increase knowledge of HPV-OPC and related sexual risk factors.

Assessment of MSM's risk was consistent with previous studies demonstrating that oral HPV and HPV-OPC risk increases with the lifetime number of sex partners and oral sex partners.^5,10,14,23 Several studies have established that greater lifetime numbers of cunnilingus partners, rimming partners, and oral sex partners in general are associated with increased risk for oral HPV infection.^6,10,23 However, one study reported that recent performance of fellatio was not associated with oral HPV infection (hazard ratio, 0.85; 95% CI, 0.29–2.54).¹⁰ Although performing fellatio has not been widely reported in previous literature, MSM accurately perceived themselves to be at risk based on increasing lifetime numbers of oral sex partners.

In studies with HPV-OPC cases, associations between lifetime number of any sexual partners and HPV-OPC are consistently reported; however, the strength of the association varies.^5,6,13 D'Souza et al.⁵ reported that 26 or more lifetime vaginal sex partners and 6 or more oral sex partners were both associated with OPC (adjusted ORs of 3.1 [95% CI, 1.5–6.5] and 3.4 [95% CI, 1.3–8.8], respectively). Drake et al.¹⁴ reported that more than 10 lifetime oral sex partners were associated with 4.3 times greater odds of HPV-OPC compared with control. Dahlstrom et al.¹³ reported that patients with HPV-OPC had greater mean lifetime numbers of sex partners (oral, anal, or vaginal; P = 0.009) and more oral sex partners than HPV-negative OPC patients (P = 0.049). Notably, most studies with HPV-OPC patients report that lifetime vaginal sex partners, lifetime cunnilingus partners, and recent cunnilingus partners are associated with OPC risk.^5,10,14 This may be due to higher viral loads of HPV in the cervical mucosa than on the penis, leading to greater viral dose exposure when performing oral-vaginal sex.²⁴ However, MSM in our study did not perceive themselves to be at increased risk for HPV-OPC based on their lifetime number of cisgender female or transgender sexual partners, or lifetime number of cunnilingus partners. This may be due to low overall prevalence of these sexual behaviors relative to sex with men or partners with a penis, or may be attributable to low knowledge of these risk factors.

We report that neither age, nor HPV vaccination status, nor HIV status was associated with perceived risk for HPV-OPC. In this study, approximately one-third of MSM self-reported receiving at least 1 dose of the HPV vaccine, which is consistent with the prevalence of HPV vaccination in other studies with populations of MSM.^15,25 Up to 92% of HPV-associated cancers, including head and neck cancers, could be prevented by timely vaccination with HPV vaccine.²⁶ Because of the low prevalence of HPV vaccination and greater numbers of lifetime sexual partners in this study, MSM may be susceptible to carcinogenic oral HPV infections and potentially HPV-OPC. In addition, infection with HIV is associated with increased risk for oral HPV infections, persistence of HPV infections, and potentially increased risk for HPV-OPC compared with HIV-negative individuals.²³ It is possible that MSM did not view these factors as impactful on the development of OPC because of overall low knowledge of HPV transmission and OPC development.

Establishing HPV awareness and cancer knowledge is an important step for MSM to more accurately assess their potential risk for OPC. A previous systematic review reported that MSM have low knowledge of how HPV is transmitted with less than 75% reporting awareness that HPV could be transmitted through sexual activity.²⁷ In another study of MSM, men were more likely to attribute other HPV-associated cancers (like anal cancer) to sexual behaviors but not OPC.²⁸ Besides low knowledge of HPV transmission, studies have consistently reported low awareness of the association between HPV and OPC in MSM. In several studies with MSM, fewer than 50% of participants were aware that HPV could cause OPC.^27,28 Recently, it has been reported that sexual minority men were more likely to be aware of the association between HPV and OPC compared with heterosexual men, perhaps reflecting an increase in accessible information as OPC incidence increases.²⁹ Future research should evaluate educational interventions for MSM on the importance of HPV vaccinations, risk for oral HPV infections, and HPV-OPC, especially if screening becomes available.

This study has limitations. Human papillomavirus vaccination, HIV status, and sexual history were all self-reported. Like other populations, MSM may be subject to recall or response bias. Men who have sex with men did not report information regarding sexual debut or deep kissing, which have been reported as risk factors for oral HPV infections and HPV-OPC.¹⁴ This sample is not nationally representative and participants may be more sexually active than other populations of MSM, so the results may not be generalizable. Lastly, a proportion of participants did not complete the entire survey, so consideration of missing data and bias is important.

This study has several strengths including a large sample of age and racially diverse MSM. Sexual behaviors were ascertained by gender and the body part on which participants had performed sexual acts. Ascertaining history of sexual behaviors through anatomically accurate questions is critical in a sample of gender and sexually diverse participants where binary labels such “man” or “woman” may otherwise lead to misclassification. In addition, we designed and validated a sexual risk scale to comprehensively measure MSM's perception of risk based on lifetime sexual behaviors. The psychometrically validated scale and results from this study could be used by intervention researchers to promote HPV vaccine uptake, HPV-OPC education, and self-screenings in MSM.