Sociodemographic, Sexual, and HIV and Other Sexually Transmitted Disease Risk Profiles of Nonhomosexual-Identified Men Who Have Sex With Men

Abstract

I examined sociodemographic, sexual, and HIV and other sexually transmitted disease risk differences among homosexual- and nonhomosexual-identified men who have sex with men (MSM) in the United States. Non-Mexican Latino ethnicity, marriage or cohabitation, religiosity, and incarceration history were positively associated with being nonhomosexual identified. Being nonhomosexual identified was associated with some risk (e.g., more sexual intercourse while intoxicated) and protective (e.g., fewer male partners) behaviors. Probabilistic sampling strategies may be useful in future research and intervention efforts.

Does being nonhomosexual-identified affect sexual behaviors and susceptibility to HIV and other sexually transmitted diseases (STDs) in men who have sex with men (MSM)? Risk and protective factors have been found in relation to nonhomosexual identities, and nonhomosexual-identified and homosexual-identified MSM likely have different sociodemographic profiles.^1–13 However, because few studies use probability-based sampling methods, the scientific community's understanding of MSM may be limited.^5,6,13,14 In this study, I used a nationally representative, probabilistic sample to examine sociodemographic, sexual, and HIV or STD risk differences among nonhomosexual-identified and homosexual-identified MSM (Tables 1 and 2).

TABLE 1.

Results of Logistic Regression Models of Nonhomosexual Identity Regressed on Sociodemographic Covariates: United States, National Survey of Family Growth, 2002

	Unweighted No.		Weighted %		Odds Ratioa (95% Confidence Interval)
	Nonhomosexual	Homosexual Identified	Nonhomosexual Identified	Homosexual Identified	Unadjusted	Adjustedb
Race/ethnicity
African American	27	14	19.0	10.7	2.85** (1.06, 7.66)	1.39 (0.50, 3.82)
Mexican	15	5	14.5	3.3	7.13† (2.54, 20.00)	2.21 (0.67, 7.35)
Non-Mexican Latino	17	11	13.9	6.0	3.74*** (1.44, 9.73)	2.95** (1.17, 7.44)
Other	2	2	4.4	2.3	3.11 (0.24, 40.98)	0.83 (0.03, 23.46)
White (Ref)	44	65	48.2	77.7	1.00	1.00
Age, y	…	…	29.5	32.0	0.96* (0.93, 1.00)	0.97 (0.93, 1.01)
Education, y	…	…	12.8	13.7	0.88* (0.78, 1.01)	0.98 (0.82, 1.16)
Income (14-category ordinal variable)c	…	…	7.7	9.5	0.89*** (0.82, 0.98)	0.98 (0.88, 1.09)
Married or cohabiting with a woman	14	4	20.2	1.8	13.45† (3.77, 48.04)	5.05** (1.26, 20.32)
Religiosityd	…	…	5.7	4.1	1.45† (1.21, 1.75)	1.33*** (1.07, 1.64)
Small town or rural residence	12	5	15.7	3.6	5.03† (1.98, 12.79)	2.30 (0.70, 7.59)
Incarceration (lifetime)	35	18	39.2	21.1	2.42** (1.06, 5.52)	2.39** (1.05, 5.45)
Foreign birth	19	8	14.5	5.4	2.97* (0.90, 9.81)	2.34 (0.61, 8.98)

Note. All analyses include the full sample (N = 202). Ellipses indicate the data were not applicable.

Source. Data are from the National Center for Health Statistics.¹⁵

^aHomosexual-identified men who have sex with men (MSM) were the reference category.

^bMultivariate analyses were adjusted for all sociodemographic predictors.

^cThe median income for nonhomosexual-identified MSM was between $20 000 and $24 999. The median income for gay-identified MSM was between $30 000 and $34 999.

^dReligiosity was measured with an 8-category index based on attendance at religious events and the importance of religion.

*P < .10; **P < .05; ***P < .01; †P < .001.

TABLE 2.

Results of Logistic Regression Models of Sexual Behavior Patterns and HIV or Sexually Transmitted Disease (STD) Risks Regressed on Non-Gay Identity: United States, National Survey of Family Growth, 2002

	Unweighted No.		Weighted %		Odds Ratioa (95% Confidence Interval)
	Nonhomosexual Identified	Homosexual Identified	Nonhomosexual Identified	Homosexual Identified	Unadjusted	Adjustedb
Sexual behavior patterns
Sexual contact with women, past y	55	6	56.7	6.8	17.89† (5.73, 55.86)	17.66† (5.18, 60.17)
≥ 4 female partners, lifetime	63	18	68.3	38.1	3.50*** (1.50, 8.17)	4.14*** (1.56, 10.97)
≥ 4 male partners, lifetime	53	82	57.9	90.8	0.14† (0.05, 0.36)	0.14† (0.05, 0.36)
≥ 2 female partners, past yc	28	1	52.3	40.3	1.62 (0.14, 18.65)	1.49 (0.13, 17.42)
≥ 2 male partners, past y	50	60	57.7	64.3	0.76 (0.37, 1.55)	0.90 (0.42, 1.93)
Oral sex only with men, lifetime	29	3	26.5	2.8	12.41† (3.12, 49.27)	12.13† (2.77, 53.02)
Insertive anal intercourse only with men, lifetimed	19	10	23.1	11.8	2.24* (0.86, 5.88)	3.16** (1.08, 9.19)
Receptive anal intercourse only with men, lifetime	11	4	26.6	3.0	11.90† (3.26, 43.44)	11.55† (2.68, 49.85)
Insertive and receptive anal intercourse, lifetime	46	80	50.3	85.2	0.18† (0.08, 0.41)	0.15† (0.06, 0.35)
HIV or STD risks
STD treatment, past y	16	12	20.4	10.6	2.15 (0.61, 7.57)	1.81 (0.51, 6.38)
Chlamydia, past ye	7	3	40.1	23.5	2.18 (0.24, 19.95)	0.45 (0.02, 8.75)
Gonorrhea, past y	8	5	41.7	31.3	1.57 (0.19, 13.27)	0.31 (0.02, 4.80)
Genital warts, lifetime	12	12	12.1	9.7	1.29 (0.50, 3.33)	1.39 (0.52, 3.66)
Herpes, lifetime	13	9	12.8	11.4	1.15 (0.33, 3.96)	1.08 (0.29, 4.08)
Syphilis, lifetime	7	8	13.4	5.0	2.93* (0.88, 9.76)	3.23* (0.93, 11.30)
Condom use with last woman, lifetime	55	43	58.5	42.8	1.89* (0.96, 3.72)	1.79* (0.90, 3.57)
Condom use with last man, lifetime	48	42	49.0	38.1	1.56 (0.81, 3.01)	1.62 (0.81, 3.22)
HIV test, any, lifetime	75	87	68.5	85.5	0.37** (0.15, 0.90)	0.49 (0.18, 1.31)
STD test, past y	36	40	36.2	40.4	0.84 (0.41, 1.72)	0.81 (0.40, 1.66)
Sexual intercourse while intoxicated with drugs or alcohol (at least half the time), past y	27	17	28.2	16.0	2.06* (0.93, 4.54)	2.22** (1.05, 4.66)
Intravenous drug use, past y	4	2	5.9	1.1	5.56 (0.61, 50.65)	4.08 (0.60, 27.76)
Sex in exchange for money or drugs, past y	14	5	15.0	8.8	1.81 (0.49, 6.78)	1.82 (0.47, 7.09)

Source. Data are from the National Center for Health Statistics.¹⁵

^aHomosexual-identified men who have sex with men (MSM) were the reference category.

^bOdds ratios were adjusted for age, race/ethnicity, and educational level.

^cAnalyses for 2 or more female partners only included men who had sex with a woman in the past year.

^dMeasures for insertive, receptive, and both anal intercourse roles indicate lifelong behavior patterns among those who ever had anal intercourse. All 3 patterns are mutually exclusive.

^eHistory of chlamydia and gonorrhea were assessed only for MSM treated for an STD in the previous year.

*P < .10; **P < .05; ***P < .01; †P < .001.

METHODS

Data were from the 2002 National Survey of Family Growth, a nationally representative, stratified-cluster sample of 4928 household-abiding males aged 15 to 44 years.¹⁵ The response rate was 78%. I selected the subsample of 202 sexually active MSM, all with at least 1 past-year episode of anal intercourse or oral sex with a man, for inclusion in my study.

Sociodemographic measures consisted of self-reported nonhomosexual identity, race/ethnicity, age, years of education, annual household income, heterosexual marriage or cohabitation, religiosity, foreign birth, incarceration history, and small town or rural residence. Sexual behaviors included past-year contact with women, having 4 or more lifetime female or male partners, having 2 or more past-year female or male partners, having oral sex only (no anal intercourse) with men, and lifetime anal intercourse roles among those who had anal intercourse with men. Risk or protective behaviors consisted of STD history, HIV or STD testing, sexual intercourse while intoxicated with drugs or alcohol, sex in exchange for money or drugs, intravenous drug use, and condom use with most recent partners.

I used logistic regression analyses to model the relation of nonhomosexual identity to sociodemographic factors. I also used logistic regression models to predict the odds of possessing behavioral and risk characteristics as a function of being nonhomosexual identified. Sampling and design weights allowed the sample to represent MSM in the United States by adjusting for oversampling, nonresponse, noncoverage, and the stratified-cluster design.¹⁶ I used SAS version 9.1.3 (SAS Institute Inc, Cary, NC) to perform all analyses.

RESULTS

Table 1 presents sociodemographic predictors of being nonhomosexual identified. Of the 202 MSM, 105 (52%) were nonhomosexual identified. Non-Mexican Latino ethnicity, heterosexual marriage or cohabitation, religiosity, and incarceration history were positively predictive of being nonhomosexual identified.

Table 2 contains results for behavioral and risk analyses. Relative to homosexual-identified MSM, nonhomosexual-identified MSM had statistically greater odds of past-year opposite-gender contact, 4 or more lifetime female partners (among those having opposite-gender contact), and only oral sex with men during all of their same-gender encounters. The odds of nonhomosexual-identified MSM having 4 or more lifetime male partners and both insertive and receptive anal intercourse with men during their lifetime were substantially lower than the odds for homosexual-identified MSM.

The only significant difference in risk behaviors between nonhomosexual-identified and homosexual-identified MSM occurred for being intoxicated with drugs or alcohol during past-year sexual encounters, with nonhomosexual-identified MSM having a higher odds compared with homosexual-identified MSM.

DISCUSSION

Recent studies have provided scientists with a better understanding of non homosexual-identified MSM's sociodemographic, sexual, and risk profiles.^{6–9,17–22} Yet the sentiment that non homosexual-identified MSM are at greater risk for contracting HIV or STDs than are homosexual-identified MSM persists.^1,2 Supporting this belief is the notion that homosexual-identified MSM—identifiable via homosexual communities through which HIV and STD interventions are channeled—more readily receive preventive information.⁵ Although this may be true, my results suggest that the sexual and risk profiles of nonhomosexual-identified and homosexual-identified MSM are comparatively complex.

Compared with homosexual-identified MSM, nonhomosexual-identified MSM had a higher odds of having 4 or more lifetime female partners, sexual intercourse while intoxicated with drugs or alcohol, and only receptive anal intercourse. However, nonhomosexual-identified MSM's greater likelihood of having only oral sex during their lifetimes, a lower number of male partners, and a lack of both insertive and receptive anal intercourse with men is consistent with the results of some studies.^2,6 The latter may explain why some have found nonhomosexual-identified MSM to have lower HIV rates than do homosexual-identified MSM.^10,23

This study confirmed findings from existing studies that non-Mexican Latino ethnicity,^2,12 heterosexual marriage or cohabitation,⁶ religiosity,²⁴ and incarceration history²⁵ are correlates of nonhomosexual identification. The finding that African American ethnicity was not independently associated with being nonhomosexual identified emerged in multivariate modeling; researchers sometimes rely on bivariate tests when discussing African American MSM's inclination toward nonhomosexual identification.^2,12 Differences in sociodemographic and sexual profiles of nonhomosexual-identified and homosexual-identified MSM should be acknowledged to maximize the effectiveness of appropriate interventions.

Studies of nonhomosexual-identified MSM could be enhanced with laboratory-based HIV or STD test results, multiple and thorough measures of condom use, and larger sample sizes. These were limitations of this study.

Nonetheless, the use of probabilistic, nationally representative data sheds new light on the sociodemographic, sexual, and risk profiles of nonhomosexual-identified MSM. These data also confirmed findings from opportunistic samples, which suggests that such samples have provided relatively accurate depictions of nonhomosexual-identified MSM. This is encouraging, given researchers’ doubts regarding the utility of nonprobabilistic data.^5,14 Nonetheless, when feasible, probabilistic sampling at national and local levels may be most appropriate for research efforts and interventions designed to curb the spread of HIV or STDs in this vulnerable population.