Introduction

Sexual health and well-being have been shown to be associated with overall health (Davison et al., 2009; Dogan et al., 2013; Holmberg et al., 2010; Sprecher & Cate, 2004), and sex remains important to some older adults (Lindau et al., 2007). Yet, research that explores the sexual health of older gay men is lacking compared to literature that expounds on the sexual health of heterosexual and younger populations. Additionally, research is limited that examines factors associated with erectile dysfunction (ED; a common health condition among aging men; Laumann et al., 2007), which is linked with negative health outcomes, such as an increased risk of depression (Liu et al., 2018).

Erectile function is a critical mechanism for many cisgender men to express their sexuality and engage in sexual activities. The “inability to achieve or maintain an erection,” known as ED (Morgentaler, 1999), often threatens the ability of men to engage in certain sexual activities or sexual behaviors, such as penetration (Smith et al., 2007). It has been estimated that over 20% of men aged 40 or older have experienced some form of ED, and this percentage increases with age (Laumann et al., 2007). Complications from ED are known to decrease overall sexual satisfaction (Smith et al., 2007). ED may also cause stress, anxiety, depression, or low self-confidence (Latini et al., 2006; Shabsigh et al., 1998; Tomlinson & Wright, 2004).

Health-Related Factors and Erectile Dysfunction

Prior studies have suggested that ED may be caused by several physical and psychological factors. For instance, literature suggests that one independent risk factor for ED diagnosis or treatment is age (Mulhall et al., 2016). Erectile function has been linked to cardiovascular health (Gandaglia et al., 2014), diabetes (Penson et al., 2009), and prostate cancer (and treatment; Nelson et al., 2011). Additionally, normal blood flow to the penis is essential for erectile function. Hypertension or high blood pressure may prevent dilation of the penis (i.e., erection) because blood may be trapped, preventing it from entering the penis (Burchardt et al., 2000; Feldman et al., 1994; Foy et al., 2019; Heikkilä et al., 2017). A meta-analysis found that ED occurs 1.39 times more in depressed patients than patients without depression (Liu et al., 2018), and another study found a bidirectional relationship between the two (Shiri et al., 2007). Behavioral and mutable factors such as cigarette smoking (Mannino et al., 1994; Tostes et al., 2008), lack of regular exercise (Silva et al., 2017), side effects from taking medication (Rosen & Marin, 2003), and cannabis use (Pizzol et al., 2019) are also associated with ED.

ED is more common among gay men than heterosexual men by 1.5 times (Bancroft et al., 2005; Barbonetti et al., 2019), yet little is known about how these factors play a role in ED among older gay men. Additionally, several of these physical and mental health factors that are associated with ED in the general population are more prevalent among gay men (Gonzales & Henning-Smith, 2017; Operario et al., 2015). For instance, hypertension (Jackson et al., 2016) and illicit drug use are more common (Rhodes, et al., 2007) in gay men compared to heterosexual men. Thus, gay men may be at increased risk of experiencing ED.

Social Factors and Erectile Dysfunction

There is evidence suggesting that factors associated with ED go beyond health factors (Aytaç et al., 2000). For instance, studies have shown that some racially and ethnically diverse men are at a heightened risk of ED (Saigal et al., 2006; Selvin et al., 2007; Smith et al., 2009). Thus, it is essential to examine the potential influence of social factors on ED among older gay men. For example, internalized gay ageism, a term coined by Wight et al. (2015), is the internalization of ageist messages in the context of aging as a gay man, and may contribute to ED. Understanding social factors is particularly important given the vast differences between gay and straight cultures. Gay culture reinforces ageist stereotypes (Slevin & Linneman, 2010), and said stereotypes may lead to feelings of increased self-consciousness about their bodies and reduced sexual self-esteem among gay men in general (Filice et al., 2019; Slevin & Linneman, 2010). These permissible acts of ageism in the gay community may stem from the internalized homophobia endured by gay men and the effects of the HIV/AIDs crisis (Meyer, 1995). Therefore, such factors could potentially play a role in ED among older gay men, especially as there is emerging evidence that suggests a negative relationship between internalized ageism and sexual health and well-being among older adults in the general population (Syme & Cohn, 2021). However, little is known about the influence of internalized gay ageism on ED among aging gay men.

To guide this study, we adapted the stereotype embodiment theory (Levy, 2009), which posits that individuals unknowingly internalize stereotypes, influencing health outcomes through pathways, such as a physiological pathway. We hypothesized that older gay men who report higher rates of internalized gay ageism will report higher rates of ED, and this relationship will be mediated by blood pressure, even when sociodemographic characteristics and other health-related factors known to influence ED are controlled.

Method

Participants

The Study on Aging and Sexual Satisfaction Among Gay Men was a cross-sectional online survey that assessed sexual health and well-being among gay men 50 years or older who resided in the Midwestern United States. Eligibility for participation in the survey included being 50 years or older (assessed by “What is your age range?”), (2) identifying as gay (assessed by “Do you identify as gay?”), (3) having been assigned male at birth (assessed by “What sex was originally listed on your birth certificate?”), (4) identifying as male (assessed by “What is your primary gender identity today?”), and (5) residing in a Midwestern state (Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin) at the time of the survey.

Data were collected from December 2021 to May 2022 using Qualtrics online survey software (a paper survey option was available upon request). The survey took participants approximately 15 to 20 min to complete. Recruitment of the sample was accomplished through several avenues, including word of mouth, email blasts, phone calls, and presentations. Places of recruitment included a combined total of over 300 contacts to lesbian, gay, bisexual, and transgender + organizations, word-of-mouth contacts, and several personal contacts among others. A total of 195 responses were collected and after data cleaning, 181 participants were analyzed. Informed consent language was displayed at the beginning and end of the survey. A $10 Amazon gift card incentive was offered to participants who completed the survey and provided their home mailing address on a separate form linked to the main survey.

Measures

Dependent Variable

Erectile Dysfunction ED was measured by using investigator-adapted items from a subscale of the Gay Male Sexual Difficulties Scale (McDonagh et al., 2016). The subscale used four items on a 6-point scale: 0-not applicable to 5-all the time. The original question asked, “During the past 6 months…” and was adapted to ask “During the past 12 months….” The four items for the scale included (1) “When you engaged in sexual activity, were you able to get an erection?” (2) “When you wanked (i.e., jerked), were you able to get an erection?” (3) “When you engaged in sexual activity, were you able to maintain your erection (i.e., keep it up)?” and (4) “When you wanked, were you able to maintain your erection?” Two of the response options were adapted to say “jerked off” instead of “wanked” to reflect American slang for masturbation. The scale score for ED was computed based on the mean score of the four items (non-applicable responses did not count toward the score and were recoded as “missing”) from the ED subscale of the Gay Male Sexual Difficulties Scale (McDonagh et al., 2016). The theoretical and actual range of scores was 1–5, with higher scores indicating higher ED. The final score was set to “missing” if there were any missing values for the four items. Cronbach’s alpha was (α = 0.94) with a skewness of 0.10.

Independent Variables

Internalized gay ageism Internalized gay ageism was measured using Wight et al.’s (2015) Internalized Gay Ageism Scale, which consisted of six statements. These six statements were: “As I get older, I feel good about myself as a gay man”; “I feel that older gay men are respected in the gay community”; “Aging is especially hard for me because I am a gay man”; “I am not too worried about looking older”; “As I get older, I feel more invisible when I am with other gay men”; and “I feel pressured to try to look younger than my age.” The response options (utilizing Wight et al. (2015) Table 1 response categories) ranged from 1 (strongly disagree) to 4 (strongly agree). The questions “As I get older, I feel good about myself as a gay man,” “I am not too worried about looking older,” and “I feel pressured to try to look younger than my age” were reverse coded. The scale score for internalized gay ageism was computed by taking the mean of the six items from the Internalized Gay Ageism Scale (Wight et al., 2015). The final score was set to “missing” if there were any missing values for any of the six items. Cronbach’s alpha was (α = 0.77) with a skewness of − 0.23. Higher scores indicated higher internalized gay ageism.

Table 1 Descriptive characteristics of Midwestern gay men 50 years or older (N = 181)
Full size table

Sociodemographic variables Participant demographic items included age, education level, race/ethnicity, relationship type, income, and residence.

Age Age was ascertained with the open-text question, “What is your age in years (whole numbers)?”

Education Education level was assessed based on the 2019 Behavioral Risk Factor Surveillance System survey (Centers for Disease Control & Prevention, 2019). The question was, “What is the highest grade or year of school you completed?” Response options were adapted, and they included less than high school, some high school, some college or technical school, community college degree (e.g., A.A.), undergraduate degree (e.g., B.S., B.A. etc.), graduate degree (e.g., M.S.W., M.A., Ph.D., J.D., M.D. etc.). Responses were collapsed into three categories: community college or below, undergraduate degree, and graduate degree.

Race and ethnicity To ask about race and ethnicity, a question from Hughes et al. (2016a) “Rethinking and Updating Demographic Questions: Guidance to Improve Descriptions of Research Samples” was used. Participants were asked, “Which categories describe you? Select all that apply to you” with response options of “American Indian or Alaska Native”; “Asian”; “Black or African American”; “Hispanic, Latino or Spanish origin”; “Middle Eastern or North African”; “Native Hawaiian or other Pacific Islander”; “White”; and “Some other race, ethnicity, or origin (please specify).” The analytic variable used three categories: Black, White, and other/multiracial and multiethnic (e.g., Hispanic, Latino, or Spanish: Middle Eastern or North African, Native Hawaiian or other Pacific Islander, and Some other race, ethnicity, or origin [please specify].

Income Income was assessed using an adapted question from the 2019 Behavioral Risk Factor Surveillance System (Centers for Disease Control & Prevention, 2019). It asked, “What is your annual household income from all sources?” The response options were adapted to include “don’t know/not sure,” “less than $25,000,” “$25,000 to less than $35,000,” “$35,000 to less than $50,000,” “$50,000 to less than $75,000,” and “$75,000 or more.” Don’t know/not sure was considered missing.

Relationship type Relationship type was assessed using an investigator-adapted version from Parsons et al. (2013). It asked, “What best describes your current relationship type?” The response options were “single (e.g., do not have main partner),” “monogamous (e.g., have a partner and agreed to only have sex with each other and no sex with casual partners),” “monogamish (e.g., have partners and agreed to have sex with others but only when the other member of the relationship was present),” “open (e.g., have a partner and both the partner and I have casual partners without the other partner present),” and “other (please specify in the box).” Single” and “widowed” were combined into one category for analysis.

Residence This question was adapted from the Michigan Transgender Health Survey 2018 (Kattari et al., 2020) and asked: “Would you consider where you live to be…?” Our response options included: “urban (metropolitan areas; cities of over 100,000 people [e.g., Detroit, Cleveland, Chicago, Milwaukee]),” “suburban (neighborhoods on the outskirts of near larger cities,” “small city (cities of 10,000 to 100,000 people [e.g., Jackson, Port Huron, Saginaw]), and “rural (villages, hamlets, towns, cities under 10,000 people).”

Social Variables

Experienced ageism Experienced ageism was measured using an investigator-adapted version of Wight et al.’s (2015) Ageism scale. The adaptation to Wight et al. (2015) scale changed the wording to ask about the past 12 months. The scale assessed if participants had any occurrence within the past 12 months of the following acts or impressions attributed to one’s age: “bullied,” “made fun of by a stranger/strangers,” “ignored by others,” “called a derogatory name,” “rejected by younger people,” “not taken seriously,” and “treated like a child” with “yes” or “no” responses. The scale score for experienced ageism was taking the sum score of seven items from the Ageism Scale (Wight et al., 2015). Answers of “yes” were given a score of 1 and answers of “no” a score of 0. The actual range was a minimum score of 0 and a maximum score of 7, with higher scores indicating more experienced ageism. Cronbach’s alpha was (α = 0.73) with a skewness of 1.65.

Internalized homophobia The scale used to measure internalized homophobia was the Revised Internalized Homophobia Scale (Herek et al., 2009). Example statements include, “I wish I weren’t gay” and “If someone offered me the chance to be completely heterosexual, I would accept the chance.” Response options were on a 5-point Likert scale and ranged from 1 (disagree strongly) to 5 (agree strongly). The computed scale score took the sum of the five items (Herek et al., 2009) and divided by the total number of items, with higher scores indicating more internalized homophobia. The final score was set to “missing” if there were any missing values for the five items. The computed scale had a Cronbach’s alpha of (α = 0.80) and skewness of 1.88. The theoretical score range was 1–5 and actual score range was 1–4.40.

Comfortable with health provider Comfortability with health provider was measured by asking “How comfortable are you discussing your sexual health and well-being with your health provider?” Responses included “not comfortable”, “somewhat comfortable,” and “very comfortable.”

Health Status Variables

HIV status HIV status was measured by asking participants, “Have you ever been told by a health care provider that you had HIV and/or AIDS?” with “yes” or “no” response options. These response options were investigator adapted from “either or both diagnoses” and “none” from a study that explored HIV disparities among older gay and bisexual men (Emlet et al., 2020).

Overall health Overall health was measured using the Self-Rated Health Measure (Turner et al., 2016). Participants were asked how much they agreed with four statements such as, “You seem to get sick a little easier than other people” and “In general, your health is excellent.” Responses ranged from 1 (definitely true) to 5 (definitely false). Overall health was computed by summing the four items, with higher scores indicating better health. The scale was adapted to ask respondents to indicate their agreement within the last 12 months. The final score was set to “missing” if there were any missing values for the four items. The scale had a theoretical range of 4–20 and an actual range of 6–20. Cronbach’s alpha was (α = 0.71) with a skewness of −0.25.

High blood pressure A 2011 Behavioral Risk Factor Surveillance System (Centers for Disease Control & Prevention, 2011) question was used to measure high blood pressure. It asked, “Have you EVER been told by a doctor, nurse, or other health professional that you have high blood pressure?” The response options for this question consisted of “yes,” “no,” “told borderline high or pre-hypertensive,” and “don’t know/not sure.” The high blood pressure analytic variable was collapsed into a binary of “yes” or “no” with “borderline high or pre-hypertensive” and “don’t know/not sure” coded as “no.”

Depression Depression was measured using the items from the Patient Health Questionnaire-2 scale (Kroenke et al., 2003) and an adapted depression question from the 2011 Behavioral Risk Factor Surveillance System (BRFSS; Centers for Disease Control & Prevention, 2011). The BRFSS question was, “Has a doctor, nurse, or other professional ever told you that you had a depressive disorder (including depression, major depression, dysthymia, or minor depression)?” with adapted response options of “yes,” “no,” or “don’t know/not sure” (“refused” was removed as a response option). Participants who scored a 3 or above on the PHQ-2 scale or answered “yes” to the BRFSS question were considered clinically diagnosed with depression or had depression symptoms. Those who answered “no” and scored less than 3 on the PHQ-2 scale were considered not to be clinically diagnosed or have reported symptoms.

Diabetes Diabetes was measured using an adapted question from the 2005–2006 CDC Diabetes Questionnaire (National Health and Nutrition Examination Survey, 2007). Participants were asked, “Have you ever been told by a doctor or other health professional that you have diabetes?” with answers of “yes,” “no,” “pre-diabetes or borderline diabetes,” “no,” and “don’t know/not sure.” The words “sugar diabetes” were removed from the question stem. Diabetes was collapsed into a binary variable of “yes” or “no” with “pre-diabetes or borderline diabetes” “no,” and “don’t know/not sure” coded as “no.”

Prostate cancer Prostate cancer was measured by asking, “Has a doctor ever told you that you had prostate cancer?” with responses of “yes,” “no,” and “don’t know/not sure.” A binary variable was created with “yes” and “no” categories and “don’t know/not sure” was considered “no.” This question was adapted from 2019 Behavioral Risk Factor Surveillance Survey questions about other diseases (Centers for Disease Control & Prevention, 2019).

Health Behavior Variables

Physical activity level To obtain information about a participant’s level of physical activity, a question from the Physical Activity Measure was used (Brown, 2011). The question asked, “In general how often do you participate in moderate or intense physical activity for at least 30 min? Moderate physical activity will cause a slight increase in breathing and heart rate such as brisk walking.” Response options were “not at all,” “less than once a week,” “1–2 times a week,” “3 times a week,” “more than three times a week but not every day,” and “every day” with higher scores indicating more physical activity.

Smoking status Two questions from the 2019 BRFSS (Centers for Disease Control & Prevention, 2019) were used to assess smoking status. The first question asked, “Have you smoked at least 100 cigarettes in your life?” with responses of “yes” and “no.” Response options refused and don’t know/not sure were removed. Those who selected “yes” were asked, “Do you smoke cigarettes every day, some days, or not at all?” Response options included “every day,” “some days,” “not at all,” and “don’t know/not sure.” The final analytic variable combined responses from both questions to create three categories of smoking status: current, former, and never. Those who answered “yes” to the first question and “not at all” or “don’t know/not sure” were coded as former smokers. Participants who answered “yes” to the first question and “some days” or “every day” to the second question were considered current smokers. Participants who answered “no” to the first question and “not at all” or “don’t know/not sure” to the second question were considered never smokers.

Phosphodiesterase inhibitor use Phosphodiesterase-5 (PDE5) inhibitor use was measured by asking participants, “In the past 12 months, have you used PDE5 inhibitors (such as Viagra or Cialis) for sexual encounters?” with “yes” or “no” response options.

Alcohol use before or during sex Alcohol use during sex was measured using one question from the Substance Use Measure (Knyazev et al., 2004), which was adapted to ask about alcohol use immediately before or during sex. The original question was “Have you used alcohol?” The revised question was, “In the past 12 months, have you used alcohol immediately before or during sex?” Response options were “yes” or “no.”

Illicit drug use before or during sex Illicit drug use before or during sex was measured by using an adapted question item that asked participants about drug use in general (Knyazev et al., 2004). The original question stated, “Have you ever tried drugs?” The adapted question was, “In the past 12 months, have you used illicit drugs immediately before or during sex (e.g., marijuana, ketamine, poppers, crystal meth, heroin, etc.)?” with adapted response options of “yes” or “no.”

Statistical Analysis

All statistical analyses were conducted using IBM SPSS V.28 (IBM SPSS Statistics for Windows, 2021). Descriptive analyses were conducted to summarize all variables of interest. Means and standard deviations were measured for continuous variables, and frequencies and percentages were calculated for categorical variables. Bivariate analyses were conducted using non-parametric Mann–Whitney tests to test the relationship between categorical independent variables with two groups and the outcome variable of ED, and Kruskal–Wallis tests were conducted to test the relationship between categorical independent variables with three or more groups and ED. For continuous variables that were non-normally distributed, Spearman’s r test was conducted.

A mediation analysis was conducted between internalized gay ageism and ED with a potential mediator of blood pressure. However, at the bivariate level internalized gay ageism and ED were not significantly associated; therefore, a hierarchical linear regression was chosen. A four-stage hierarchical linear regression was conducted to explore the relationship between predictor variables that were significant p < .10 level in the bivariate analysis and ED. To ensure sufficient power to detect significant differences in the outcome variable we previously suggested guidelines for regression modeling (Harrell, 2001). An alpha of 0.10 for the bivariate analysis was selected to limit the number of variables in the multivariate model. Internalized gay ageism, although not significant, was included for theoretical reasons. In the first model, sociodemographic characteristics were entered (i.e., age, race and ethnicity, and residence). In the second model, health variables were entered (i.e., overall health, HIV status, and blood pressure). In the third model, health behavior variables were entered (i.e., physical activity, alcohol use before and during sex). In the final model, internalized gay ageism was entered.

Results

Participant Characteristics

Among the 181 participants, the mean age was 65.29 years (SD = 9.32). The majority had an undergraduate degree (34.8%) or graduate degree (45.9%) degree and were White (85.6%). Slightly over half of participants were single or widowed compared to being in some form of a relationship (50.3% versus 49.2%). Nearly half of participants resided in an urban setting (urban 48.1% versus suburban 34.8%, small city 11.6%, and rural 5.0%). Most participants were HIV negative (81.2%) and a little more than half indicated they had high blood pressure (50.3%). About half had been clinically diagnosed or indicated symptoms of depression (47.5%). The mean average of rated overall health was 14.8 (SD = 3.58) and a majority of participants reported that they exercised at least one or two times a week or more. The mean internalized gay ageism score was 2.21 (SD = .55), the mean experienced ageism score was 1.06 (SD = 1.49), and mean internalized homophobia score was 1.41 (SD = .61). Most did not use PDE5 inhibitors 66.9%, illicit drugs 74.6%, or alcohol 64.6% before/during sex. The mean ED score was just over 2 (M = 2.23, SD = 1.5) meaning on average participants reported having ED “once or twice” in the past 12 months (see Table 1).

Bivariate Analysis

In bivariate analyses, the variables that were associated with ED at the p < .10 were alcohol use before/during sex (p = .052), physical activity level (p = .039), and race/ethnicity (p = .035; see Table 2), HIV status (p = .084) and high blood pressure (p = .098; Table 2).

Table 2 Bivariate analyses of Midwestern gay men 50 years or older and erectile dysfunction (N = 181)
Full size table

Age was positively associated with ED (r = 0.28, p < .001), while overall health was negatively associated with ED (r = −.35, p < .001). Notably, internalized gay ageism (p = .811) and experienced ageism (p = .825) were not significantly associated with ED.

Hierarchical Linear Regression Analysis

Model 1

In Model 1, age positively predicted ED, β = 0.287, t = 3.43, p < .001. Age, race and ethnicity, and residence accounted for 10.2% variation in ED and contributed significantly to the model F(6, 134) = 3.66, p = 0.002 (data not shown).

Model 2

In Model 2, age remained significant, β = 0.264, t = 3.26, p = .001. Overall health negatively predicted ED, β = −0.286, t = −3.69, p < .001 (see Table 3). Introducing overall health, HIV status, and blood pressure explained an additional 19.3% of variation in ED and significantly contributed to the model F(3, 131) = 6.05, p < .001 (data not shown).

Table 3 Hierarchical regression analysis of predictors of erectile dysfunction
Full size table

Model 3

In Model 3, age remained significant, β = 0.246, t = 3.05, p = .003. Overall health remained significant in Model 3, β = −0.231, t = −2.90, p = .004 (see Table 3). Adding physical activity and alcohol use before and during sex explained an additional 21.8% of the variance in ED and significantly contributed to the model F(2, 129) = 3.03, p = .052 (data not shown).

Model 4

In Model 4, age (β = 0.224, t = 2.70, p = .008) and overall health (β = −0.247, t = −3.05, p = .003) remained significant. Neither physical activity nor alcohol use before and during sex remained significant in Model 4. Finally, adding internalized gay ageism to the final model explained 0.07% of variance of ED and was insignificant to the model F(1, 128) = 1.25, p = .265 (data not shown). In terms of effect size, R2 for Model 4 is 0.286, and adjusted R2 was 21.9%.

Discussion

This study helps to fill in the gap in knowledge about the sexual health of older gay men by examining a wide range of established and unique potential predictors of ED among a sample of gay men aged 50 years or older in the Midwest. Among this sample, participants reported experiencing ED once or twice in the past 12 months on average, and before controlling for other factors, ED was associated with age, overall health, physical activity, and alcohol use before and during sex. In line with previous research among the general population of men (Laumann et al., 2007), we found that increasing age and decreasing overall health were both predictors of ED.

Contrary to our hypothesis, internalized gay ageism was not a significant predictor of ED in this sample. There are explanations for this null finding. First, participants reported low levels of internalized gay ageism on average. In addition, our study consisted of individuals who were highly educated, mostly White, and of high socioeconomic status, and who therefore may have access to health or mental health services and other advantages that could lead to decreased experiences of both internalized gay ageism and ED (Williams & Cooper, 2019), potentially influencing a reduction in both internalized gay ageism and ED. Thus, future studies may benefit from exploring the association between internalized gay ageism and ED with a larger and more diverse sample of older gay men. Lastly, a recent study found that self-directed ageism was associated with sexual functioning among older adults (Gitliz & Ayalon, 2023). Although a nonsignificant result was found in our study for similar variables, the effect sizes were in the same direction and similar magnitude; self-directed ageism and sexual function (β = −0.18, p = .02) (Gitliz & Ayalon, 2023) and internalized gay ageism and ED (β = −0.089, p = .265). Therefore, it is possible that our study was underpowered. Alternatively, older sexual minorities may be better prepared to handle bias. For example, a previous pilot study found that older sexual minorities reported having lower perceived ageism compared to older straight adults, potentially due to resilience factors associated with being a sexual minority (Flesia et al., 2023). Therefore, our finding of low internalized gay ageism and the lack of association between internalized gay ageism and erectile function could be attributed to an explanation similar to the one given by Flesia et al. (2023) for perceived ageism.

Alcohol use is known to be more prevalent among sexual minorities compared to heterosexual populations (Hughes et al., 2016b), and while participants who reported alcohol use before or during sex had lower ED scores compared to older gay men who did not, the relationship was non-statistically significant. In the short term, alcohol use has stress-reducing effects (Pohorecky, 1981; Sillaber & Henniger, 2004) and may be the motive for older gay men to use before or during sex to focus on the sexual activity and not their sexual anxiety (Pyke, 2020). Although alcohol use before or during sex did not predict ED when other factors were controlled, future studies might examine this factor further in relation to ED among older gay men.

Internalized homophobia was included in the multivariate model, as it is a unique stressor and commonly experienced by gay men (Herek et al., 1998). However, no relationship was found between internalized homophobia and ED in this sample. Our null finding is opposite of a prior study that examined internalized homophobia and sexual dysfunction among lesbian, bisexual, and gay adults (Kuyper & Vanwesenbeeck, 2011). The different finding in our study may be explained by the focus on older gay men aged 50 and older. Additionally, participants in our study reported experiencing low levels of internalized homophobia, which may contribute to the nonsignificant association between internalized homophobia and ED.

In terms of race, participants who identified as other, multiracial, or multiethnic had lower ED compared to Black participants, which is interesting as a previous study found that Black men had higher prevalence of ED than White men (Selvin et al., 2007). However, race and ethnicity were no longer significant when physical activity and alcohol use before or during sex were controlled. Thus, racial differences may at least partially be explained by health behavior disparities such as physical activity level or alcohol use before or during sex, but future studies should continue to explore the role of intersectional identities and ED risk.

Implications

These findings suggest the importance of overall health and sexual functioning as older gay men age. Clinicians and educators may find these results help them to better understand the contributing factors of ED among their older gay men clients or patients. A holistic assessment, such as a biopsychosocial assessment, may be vital to assess older gay men’s sexual health and determine possible interventions as the current study confirmed that sociodemographic characteristics are known to be associated with ED.

Limitations

Although this study represents a novel exploration into factors associated with ED among older gay men in the Midwestern United States, the results should be viewed with caution. The sample was predominantly White, highly educated, and with high socioeconomic status, and was restricted to older gay men in the Midwest. Social acceptability and laws regarding LGBTQIA + issues vary from region to region and may have influenced how participants responded. Furthermore, the current study utilized a cross-sectional design, which does not allow for causal inferences to be made; future studies might use longitudinal designs to address this limitation. Additionally, health items were assessed using participant self-reporting for several items that assessed health, which increases risk of recall bias or misreporting. Finally, future studies may consider collecting a larger sample to reduce potential type II error.

Conclusion

Age and overall health were found to be associated with ED among older gay men in the Midwestern United States. While this trend is similar to general populations of men, this study reinforces the importance of health as an integral piece of healthy sexual functioning among older gay men. Thus, clinicians and educators must highlight the importance of health-promoting behaviors, especially among older adult clients who present with sexual dysfunction. Additionally, it is essential to understand the potential for social and structural facets that influence health and how these may affect a multitude of avenues of gay men’s lives. Future studies should continue to investigate the factors associated with ED in older gay men and examine this phenomenon among a more diverse sample.