Previous research suggests that men in the general population are more likely to receive tests for prostate cancer than for CRC.5
In contrast, this analysis found that more men received CRC than PSA tests; among gay/bisexual men, the percentage ever receiving CRC tests approached 80%. Although this finding suggests that gay/bisexual men have better access to CRC testing than do heterosexuals, the difference between these groups may partly be due to the fact that colonoscopy, sigmoidoscopy, and proctoscopy can be used to diagnose problems associated with receptive anal sex.45
However, using these procedures to evaluate sexually related symptoms would likely lead to the “serendipitous detection” of any cancerous lesions that may be present.46
Because colonoscopy is used to diagnose HIV-related malignancies such as lymphoma, the higher prevalence of HIV among gay/bisexual men may partially account for their difference from heterosexuals in the use of this procedure.
Compared with heterosexuals, gay/bisexual men had lower adjusted odds of having up-to-date PSA testing (). The bivariate analysis, however, showed no difference in this outcome by sexual orientation (). To explain this inconsistency, we performed a sensitivity analysis to examine how the sexual orientation finding was affected after adjusting for other variables (). Sexual orientation became significantly (p<0.05) associated with up-to-date PSA testing when adjusted for race/ethnicity, education, or language proficiency (results not shown). It appears that the negative association of sexual orientation with up-to-date PSA testing was masked in the unadjusted analysis because gay/bisexual men were more likely than heterosexuals to be White, college educated, and monolingual English speakers – each of which had positive associations with having an up-to-date test. It is possible that gay/bisexual men had lower odds of up-to-date PSA testing than did heterosexuals because of a greater knowledge of the controversy surrounding the PSA test. The ambiguities in both the PSA and CRC test findings suggest that future research should elicit more specific information about the circumstances surrounding the use of these procedures by gay/bisexual men.
A consistent finding across models was the decreased odds of testing for APIs compared with Whites. There is evidence that APIs have lower risk of CRC, which may be related to racial/ethnic differences in dietary fat intake; soy proteins that are common in some Asian diets have also been shown to inhibit the growth of prostate cancer cells.47
Given such data, APIs may be less likely to perceive a need for CRC and PSA testing than do other groups. Previous work has shown that Latinos who are less acculturated to the U.S. are less likely to receive CRC screening than are others;48
this study found decreased odds of CRC testing among Latinos even after accounting for acculturation with English language proficiency. Although African Americans did not differ from Whites in the multivariate models, the low percentage of PSA testing among gay/bisexual African Americans in the bivariate analysis suggests that defining prostate cancer disparities in terms of race/ethnicity alone may neglect subgroups of high-risk minority men.
Even without detailed information on interpersonal relationships, the interaction of sexual orientation and living alone shows that basic facts about home life can reveal important differences in service use. In the health outcomes literature, living alone has been associated with adverse events such as myocardial infarction and subsequent cardiac death.49,50
Although the main effect for this variable was decreased odds of cancer testing, it appears that living alone represents an advantage for gay/bisexual men in obtaining PSA tests. Gay/bisexual men who live alone may have a greater number of friends and acquaintances than do other men, and these relationships may facilitate cancer testing specifically through referrals to providers and more generally through the modeling of positive health behaviors. To test this hypothesis in future work, surveys measuring sexual orientation should include variables that more fully characterize the social networks of gay/bisexual men.
Because cancer risk increases with age, the higher odds of testing with every 5-year increase in age suggests that the provision of these services is driven to some degree by clinical preventive guidelines. Early detection is most likely to benefit individuals with both an increased risk of cancer and a reasonable expectation of being alive in the next decade.46
Thus, the negative association between testing and the relatively larger values of the squared age term may reflect providers’ knowledge of the decreasing mortality benefit of asymptomatic cancer detection in late life.
Having less than a college education was associated with lower odds of PSA testing, perhaps due to limited access to information resources that support healthcare decision-making.32
Men no longer in the workforce may generally use more health services, including PSA and CRC tests, because of chronic conditions that need ongoing care. Compared with private insurance coverage, public-sector coverage may have been associated with up-to-date CRC testing because of the relatively generous benefits of Medicaid and Medicare. The lower odds of PSA testing for men with no regular source of care underscore the importance of continuity of care.
Current smokers’ lower odds of ever receiving PSA and CRC tests is cause for some concern, because smoking is associated with both CRC and prostate cancer. Former smokers may have had higher odds CRC testing than did men who never smoked because of a stronger sense of health-related self-efficacy, which can help sustain a difficult behavioral change such as smoking cessation. Higher odds of testing among non-binge drinkers suggest that moderate alcohol consumption is a better indicator of healthy behavior than is abstinence. The finding that obesity was associated with greater odds of ever receiving a CRC test is promising, as it suggests that providers are attending to this cancer risk factor.
These analyses were restricted largely to men age 50 and over, when screening is recommended regardless of family history. Still, the lack of data on men with a family history is a study limitation, because some guidelines recommend that this group begin screening at age 45.28
The small number of gay/bisexual minority men in this sample prevented us from representing these groups separately with interaction terms in the multivariate models. A related limitation is the difficulty of interpreting the “other” race/ethnicity findings, which represent an aggregation of small but culturally diverse groups such as Alaska Natives and American Indians. The survey questions on CRC testing were limited by the inclusion of proctoscopy (which is not recommended for screening) and the exclusion of barium enema. Further, colonoscopy may be considered up-to-date for as long as 10 years after a normal test. Although consistent with previous research,6
our use of a 5-year cut-off for defining an “up-to-date” CRC test may have misclassified some respondents who had a normal colonoscopy 6–10 years earlier.
These population-based estimates of the association of sexual orientation with cancer testing have several implications for program planning and research. Planners seeking to reduce racial/ethnic disparities in prostate cancer may need to consider sexual orientation when developing culturally specific screening programs for high-risk subgroups of men, such as African Americans. Because living alone may not represent a lack of social support among gay/bisexual men, more detailed information on social networks is needed to assess the potential impact of support on cancer testing in this population. Finally, future research in this area should collect data that can help explain why being gay or bisexual is positively associated with the use of such widely recommended procedures as colonoscopy or sigmoidoscopy, but negatively associated with a procedure as controversial as the PSA test.