High Human Immunodeficiency Virus Incidence and Prevalence and Associated Factors Among Adolescent Sexual Minority Males—3 Cities, 2015

Abstract

Background

Much has been written about the impact of human immunodeficiency virus (HIV) among young (13–24) sexual minority men (SMM). Evidence for concern is substantial for emerging adult (18–24 years) SMM. Data documenting the burden and associated risk factors of HIV among adolescent SMM (<18 years) remain limited.

Methods

Adolescent SMM aged 13–18 years were recruited in 3 cities (Chicago, New York City, and Philadelphia) for interview and HIV testing. We used χ² tests for percentages of binary variables and 1-way analysis of variance for means of continuous variables to assess differences by race/ethnicity in behaviors. We calculated estimated annual HIV incidence density (number of HIV infections per 100 person-years [PY] at risk). We computed Fisher’s exact tests to determine differences in HIV prevalence by selected characteristics.

Results

Of 415 sexually active adolescent SMM with a valid HIV test result, 25 (6%) had a positive test. Estimated annual HIV incidence density was 3.4/100 PY; incidence density was highest for blacks, followed by Hispanics, then whites (4.1, 3.2, and 1.1/100 PY, respectively). Factors associated with higher HIV prevalence included black race; ≥4 male partners, condomless anal sex, and exchange sex in the past 12 months; and a recent partner who was older, black, HIV-infected, or had ever been in jail or prison (P < .05).

Conclusions

HIV-related risk behaviors, prevalence, and estimated incidence density for adolescent SMM were high, especially for minority SMM. Our findings suggest that initiating intervention efforts early may be helpful in combating these trends.

Young (13–24 years) sexual minority males (SMM)—those who identify as gay, bisexual, or who have sexual contact with persons of the same or both sexes—are disproportionally affected by human immunodeficiency virus (HIV) infection [1, 2]. The evidence for concern is substantial for the emerging adult segment of the young SMM population. In 2015, 81% of HIV diagnoses among young SMM were in those aged 20–24 years, and among all diagnoses of HIV among SMM, 22% were in this age group. The risk of HIV was even more notable for minority SMM [3]. There are also concerns about high HIV incidence among emerging adult SMM [4, 5]. In terms of the driving forces behind these statistics, there is a growing consensus that individual-level risk behaviors cannot adequately explain the HIV-related racial disparities observed. Rather, relationship and partner-level factors as well as sociocultural and structural determinants of health, including community, political, educational, and economic factors are now understood to play a key role in HIV risk [6–9].

In contrast to what is known about emerging adult SMM, the data documenting the burden and associated risk factors of HIV among adolescent SMM remain extremely limited. The numbers and rates of diagnoses of HIV infection in this group are small. Among all HIV diagnoses attributed to male-to-male sex, just 5% were among those aged 13–19 years [3]. However, the timing of infection relative to diagnosis is often unclear; if asymptomatic, many individuals are not diagnosed until years after they are infected. The high prevalence of diagnoses of HIV infection among SMM aged 20–24 years suggests that at least some of these infections occurred during adolescence [10].

Although estimates of HIV incidence are an essential element for monitoring the epidemic, very few studies to date have included adolescent SMM. A notable exception comes from Garofalo et al [11], who conducted a longitudinal study of HIV risk among 450 SMM (aged 16–20 years at baseline) in Chicago. HIV incidence was 5.2/100 PY for participants aged 16–17 years and 3.8/100 PY for those aged 18–20 years. In addition, a study of SMM attending New York City sexually transmitted disease clinics reported that HIV incidence was actually higher among males aged <20 years (5.7/100 PY) compared with other age groups [12]. In both studies, incidence was higher among minorities [11, 12].

The characteristics and behaviors associated with HIV risk and acquisition also remain largely unexplored for adolescent SMM. Research indicates that prevalence of risky sexual behavior is high [11, 13] and that this population is particularly vulnerable to risk associated with partner selection, partner characteristics, and other sociocultural and structural determinants [11, 13–15]. The 2015 National Youth Risk Behavior Survey highlighted the disparity in HIV-related risk behaviors among sexual minority students (grades 9–12) compared to heterosexual students. Among male students, the prevalence of having had sexual intercourse before age 13 years was higher and condom use at last sex was lower among those who had sexual contact with males than among those who had sexual contact with only females. Also, SMM students were more likely to have ever injected any illegal drug than their heterosexual counterparts [16].

Research with adolescent SMM is imperative to fully understand and address the HIV epidemic among SMM as a whole. It is insufficient to infer results of studies of older SMM to adolescent SMM [17–20]. In light of the shortage of research among adolescent SMM relative to other SMM, in 2015, the Centers for Disease Control and Prevention (CDC) conducted an HIV behavioral surveillance project, the National HIV Behavioral Surveillance for Young Men Who Have Sex with Men (NHBSYMSM), to ascertain the prevalence of risk behaviors and HIV infection among this population. In this paper, we describe the sociodemographic characteristics of the sample, examine racial-ethnic variations in sexual partnerships and HIV-related risk behaviors, and provide estimates of HIV incidence and HIV prevalence among this group of adolescent SMM.

METHODS

Three cities implemented NHBS-YMSM: Chicago, New York City, and Philadelphia. Participants were eligible to enroll in the study if they met the following criteria: (1) between the ages of 13 and 18 years; (2) birth sex male and currently living as a male; (3) a resident of the Metropolitan Statistical Area; (4) able to complete the survey in English or Spanish; (5) reported any sexual contact with another male or self-identified as gay or bisexual or indicated same-sex attraction.

One of the original objectives of the NHBS-YMSM was to determine the most effective and feasible sampling method(s) to reach adolescent SMM. Therefore, 3 sampling methods were used to recruit participants into the project: venue-based sampling in New York City only, respondent-driven sampling in all 3 participating cities, and Facebook sampling in Philadelphia and Chicago. Venue-based sampling is a sampling strategy that utilizes venues (eg, clubs, organizations, street locations) within the project area to obtain the desired sample [21]. Respondent-driven sampling is a chain recruitment method that begins with a set of “seeds” who recruit members of their social networks to participate in project activities, who in turn recruit other members of their social networks [22, 23]. Facebook sampling employs targeted banner ads to identify and recruit adolescent SMM into the study.

Regardless of the sampling method, all NHBS-YMSM consenting participants underwent an in-person eligibility screening and, if eligible, completed the behavioral assessment with a trained interviewer. Anonymous HIV testing was offered to all participants regardless of self-reported HIV infection status. HIV testing was performed on blood specimens with rapid or laboratory-based testing and confirmed with Western blot or immunofluorescence. A nonreactive rapid test was considered a definitive negative result; a reactive (preliminary positive) rapid test result was considered a definitive positive result only when confirmed by supplemental laboratory testing (eg, Western blot or immunofluorescence assay). HIV-infected participants were assisted by project staff members for linkage to medical care and other supportive services.

Individuals were compensated for their participation in project activities, receiving approximately $25 for the behavioral assessment and $25 for HIV testing. NHBS-YMSM activities were approved by local institutional review boards in each participating city and approved by CDC. A waiver of documentation of informed consent was requested and received for all 3 locations. In addition, a waiver of parental permission for participants <18 years of age was requested and received in New York City (approved for 13–17 years of age), Philadelphia (approved for 14–17 years of age), and Chicago (approved for 16 and 17 years of age). The age categories included in the waiver of parental permission were determined by the local institutional review boards.

Data Analysis and Measures

Data were analyzed in aggregate as a convenience sample. Participants were included in this analysis if they had a completed, valid survey and reported ever having vaginal or anal sex with male or female partners (ie, sexually active). First, we described sociodemographic characteristics of adolescent SMM and examined racial-ethnic variations in sexual partnerships and HIV-related risk behaviors. During the NHBS-YMSM survey, participants were asked about characteristics of partnerships and behaviors that occurred over several time periods: ever in their life, during the past 12 months, and during the past 3 months. Participants were also asked about their 3 most recent sex partners in the past 3 months. If a participant had not had sex with anyone in the past 3 months, he was asked about the last person he had sex with in the past 12 months. For the current analysis, we utilized the information about the most recent sex partners to examine whether a participant had at least 1 sexual partnership (among 3 possible partnerships) with the characteristic or behavior of interest. Financial instability was assessed with the question, “was there a time when there was not enough money in your house for rent, food, or utilities such as gas, electric, or phone?” Exchange sex was defined as exchanging sex for something such as money or drugs. Partner concurrency was defined as the respondent or their partner probably or definitely having sex with other people during the time of a sexual relationship. We computed percentages for binary variables and means, standard errors, and ranges for continuous variables. We determined the statistically significant differences (P < .05) between the racial/ethnic groups using χ² test for percentages of binary variables and 1-way analysis of variance for means of continuous variables.

Next, we derived estimates of annual HIV incidence density [4, 24], among participants who reported ever having anal sex with a male partner and had a valid NHBS-YMSM positive or negative NHBS-YMSM HIV test result (n = 371), by including the number who tested positive for HIV infection as the numerator and the total number of person-years at risk as the denominator. Person-years at risk was calculated by subtracting age at first anal sex with a male from first positive HIV test (for persons who reported having previously been diagnosed with HIV infection) or current age minus age at first anal sex with a male (for all others). About 20% of sexually active adolescent SMM who reported anal sex with a male were missing the age at first anal sex. Multiple imputation based on age, race, and ethnicity was used to impute the missing age at first anal sex for these participants. Using this method, we calculated estimated annual incidence density for the entire group, as well as by racial/ethnic group (black, Hispanic, white/other; note that participants of white and other races were grouped together). Finally, we present HIV prevalence overall and by selected characteristics among the adolescent SMM who had a valid positive or negative NHBS-YMSM HIV test result. We computed Fisher’s exact tests to determine statistically significant differences (P < .05).

RESULTS

The NHBS-YMSM total sample consisted of 569 participants with valid and complete interviews; of these, 453 (80%) reported being sexually active. Among sexually active participants, mean age was 16.7 years, 42% identified as black, 38% as Hispanic, and 16% as white, and 4% reported other or multiple race groups. Almost half (45%) grew up in a 2-parent household. Of the sample, 26% reported household financial instability during the past 12 months and 30% reported ever being kicked out of the house or running away. The majority of the adolescent men identified as gay (64%) or bisexual (33%). More than half (60%) had ever been tested for HIV and 54% had done so in the past 12 months. Ninety percent reported visiting a healthcare provider in the past 12 months (Table 1).

Table 1.

Selected Characteristics of Sexually Active^a Adolescent Sexual Minority Males, National Human Immunodeficiency Virus Behavioral Surveillance for Young Men Who Have Sex with Men, 2015

Characteristic	No. (N = 453)	%
Age, y
13–15	44	10
16–17	267	59
18	142	31
Race/ethnicity
Black	188	42
Hispanic	174	38
White	71	16
Otherb	19	4
Family structurec
2 parents	204	45
Single parent	194	43
Other relatives	41	9
Other living situation	14	3
Household financial instabilityd, past 12 mo
No	335	74
Yes	116	26
Ever kicked out of the house or run away
No	319	70
Yes	134	30
Sexual identity
Gay	287	64
Bisexual	148	33
Heterosexual	13	3
Binge drinkinge, past 30 d
No	319	71
Yes	133	29
Marijuana use, past 12 mo
No	213	47
Yes	240	53
Stimulant usef, past 12 mo
No	387	85
Yes	66	15
Ever injection drug use
No	450	99
Yes	3	1
Health insurance
No	25	6
Yes	421	94
Visited healthcare provider, past 12 mo
No	45	10
Yes	408	90
Ever had oral sex with a male partner
No	29	6
Yes	424	94
Ever tested for HIV
No	182	40
Yes	270	60
Tested for HIV, past 12 mo
No	208	46
Yes	243	54
City
Chicago	190	42
New York City	176	39
Philadelphia	87	19
Recruitment methodg
Venue-based sampling	130	29
Respondent-driven sampling	241	53
Facebook sampling	82	18
Total	453	100

Numbers might not add to total because of missing data. Abbreviation: HIV, human immunodeficiency virus.

^aParticipants who ever had anal or vaginal sex.

^bIncludes persons who indicated American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, or other race.

^cFamily structure for most of respondent’s childhood. Two parents could be biological, adoptive, or a stepparent and biological/adoptive parent; single parent could be biological or adoptive; other living situation could include foster parent(s), another adult only, such as a stepparent, or other living situation such as a group home.

^dIn the past 12 months, was there a time where there was not enough money in your house for rent, food, or utilities such as gas, electric, or phone?

^eFive or more alcoholic drinks in 1 sitting.

^fCocaine, ecstasy, poppers (amyl nitrate), or methamphetamine.

^gVenue-based sampling was conducted in New York City; respondent-driven sampling was conducted in Chicago, New York City, and Philadelphia; and Facebook sampling was conducted in Chicago and Philadelphia.

The mean age of first anal sex with a male was 15 years and 82% had ever had anal sex with a male partner. In the past 12 months, 51% reported condomless anal sex with a male and 38% had ≥4 oral or anal male sex partners. Black adolescent SMM were more likely than white/other adolescent SMM to have ever had anal sex with a male partner (86% vs 74%), to have engaged in exchange sex in the past 12 months (7% vs 1%), to report sexual concurrency within a recent relationship (55% vs 40%), and to have a lower mean age of first vaginal or anal sex with a female partner (13.6 years vs 15.1 years) and less likely to report alcohol or drugs before or during last sex (23% vs 34%). Black adolescent SMM were more likely than Hispanic adolescent men to have an older partner (73% vs 60%). Black adolescent SMM were more likely than Hispanic and white/other adolescent SMM to have a black partner (76% vs 25% and 11%, respectively), and had a lower mean age of first oral sex with a male partner (14.0 years vs 14.6 years and 15.2 years, respectively). Both black adolescent SMM and Hispanic adolescent SMM were more likely than white/other SMM to have a partner who had ever been in prison or jail (20% and 16% vs 6%, respectively), whereas black and Hispanic SMM were less likely than white adolescent SMM to report sex within 1 week of meeting a new partner (24% and 19% vs 42%, respectively). Black adolescent SMM and Hispanic adolescent SMM had a lower mean age of first anal sex with a male partner (14.6 years and 15.0 years) compared with white/other adolescent SMM (15.9 years) (Tables 2 and 3).

Table 2.

Sexual Partnerships and Human Immunodeficiency Virus (HIV)–Related Risk Behaviors Among Sexually Active^a Adolescent Sexual Minority Males, by Race/Ethnicity, National HIV Behavioral Surveillance for Young Men Who Have Sex With Men, 2015

Characteristic	Total		Black		Hispanic		White/Otherb
	(N = 453)		(n = 188)		(n = 174)		(n = 91)
	No.	%	No.	%	No.	%	No.	%
Ever had vaginal or anal sex with a female partner
No	285	63	114	61	108	62	63	70
Yes	168	37	74	39	66	38	27	30
Ever had anal sex with a male partner
No	82	18	26	14	33	19	23	26
Yes	371	82	162	86c	141	81	67	74c
Condomless anal sex with a male partner, past 12 mo
No	222	49	94	50	80	46	48	53
Yes	231	51	94	50	94	54	42	47
≥4 male partnersd, past 12 mo
No	283	62	117	62	109	63	57	63
Yes	170	38	71	38	65	37	34	37
Exchange sexe, past 12 mo
No	434	96	175	93	169	97	89	99
Yes	19	4	13	7c	5	3	1	1c
Exclusive receptive anal sex with male partners, past 3 mo
No	274	67	105	62	111	71	57	70
Yes	135	33	64	38	46	29	25	31
3 most recent partnersf
Older partnerg
No	121	33	35	27	63	40	23	30
Yes	247	67	95	73h	96	60h	55	71
Black partner
No	254	56	45	24	130	75	79	89
Yes	197	44	143	76c,h	43	25h	10	11c
HIV-infected partner
No	440	97	180	96	170	98	89	99
Yes	13	3	8	4	4	2	1	1
Partner of unknown HIV status
No	286	63	122	65	114	66	49	54
Yes	167	37	66	35	60	35	41	46
Met partner onlinei
No	280	62	120	64	105	60	55	62
Yes	172	38	68	36	69	40	34	38
Sexual concurrencyj
No	227	50	85	45	88	51	54	60
Yes	226	50	103	55c	86	49	36	40c
Partner ever injected drugs
No	416	92	172	92	159	91	84	93
Yes	37	8	16	9	15	9	6	7
Alcohol or drugs before or during last sex
No	336	74	145	77	131	75	59	66
Yes	117	26	43	23c	43	25	31	34c
Partner ever in jail or prison
No	382	84	150	80	146	84	85	94
Yes	71	16	38	20c	28	16k	5	6c,k
Serious partnerl
No	186	41	77	41	65	37	43	48
Yes	267	59	111	59	109	63	47	52
Sex within a week of meeting
No	336	74	142	76	141	81	52	58
Yes	117	26	46	24c	33	19k	38	42c,k

Numbers might not add to total because of missing data. Abbreviation: HIV, human immunodeficiency virus.

^aParticipants who ever had anal or vaginal sex.

^bIncludes persons who indicated American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, or other race.

^cIndicates significantly different values (P < .05) between black non-Hispanic and white non-Hispanic/other participants.

^dHad oral or anal sex.

^eHad sex in exchange for something such as money or drugs.

^fParticipants were asked about their 3 most recent sex partners in the past 3 months. If the participant did not have sex with anyone in the past 3 months, they were asked about the last person he had sex with in the past 12 months. The “yes” category indicates that a participant had at least 1 sexual partnership (among a possible 3 partnerships) with the characteristic or behavior of interest.

^gDue to mandated reporting requirements, Philadelphia did not ask about partner’s age. Chicago and New York City asked whether the partner was younger, older, or the same age as the participant.

^hIndicates significantly different values (P < .05) between black non-Hispanic and Hispanic participants.

ⁱMet partner on the internet or through a mobile app (eg, Grindr, Jack’d, Scruff).

^jDuring the time of a sexual relationship, the participant was having sex with other people and/or the partner was probably or definitely having sex with other people.

^kIndicates significantly different values (P < .05) between Hispanic and white non-Hispanic/other participants.

^lA male with whom the participant has dated for a while and feels close to and may call a boyfriend or partner.

Table 3.

Sexual Partnerships and Human Immunodeficiency Virus (HIV)-Related Risk Behaviors Among Sexually Active^a Adolescent Sexual Minority Males, by Race/Ethnicity, National HIV Behavioral Surveillance for Young Men Who Have Sex with Men, 2015

Characteristic	Total				Black				Hispanic				White/Otherb
	(N = 453)				(n = 188)				(n = 174)				(n = 91)
	No.	Mean	(Range)	SD	No.	Mean	(Range)	SD	No.	Mean	(Range)	SD	No.	Mean	(Range)	SD
Age at first oral sex with a male partner	343	14.5	(4–18)	2.3	123	14.0c,d	(4–18)	2.6	147	14.6d	(4–18)	2.0	73	15.2c	(8–18)	2.1
Age at first vaginal or anal sex with a female partner	139	14.1	(6–18)	2.1	56	13.6c	(6–18)	2.5	58	14.3	(10–18)	1.7	25	15.1c	(12–18)	1.5
Age at first anal sex with a male partner	298	15.0	(6–18)	2.1	111	14.6c	(6–18)	2.4	127	15.0e	(6–18)	2.0	60	15.9e	(12–18)	1.5
No. of male anal sex partners, past 12 mo	452	3.8	(0–250)	12.7	187	3.9	(0–40)	5.8	174	3.7	(0–250)	19.0	91	3.7	(0–37)	6.4
No. of condomless male anal sex partners, past 12 mo	452	1.2	(0–22)	2.4	187	1.4	(0–22)	2.7	174	1.1	(0–12)	1.8	91	1.3	(0–17)	2.6
No. of anal sex acts, past 3 mo	371	7.3	(0–205)	17.2	151	7.3	(0–205)	20.4	142	7.8	(0–120)	16.7	78	6.4	(0–46)	10.2
No. of condomless sex acts, past 3 mo	304	7.7	(0–145)	16.8	125	7.0	(0–145)	18	119	8.4	(0–120)	17.8	60	7.9	(0–60)	11.4

Numbers might not add to total because of missing data. Abbreviation: SD, standard deviation.

^aParticipants who ever had anal or vaginal sex.

^bIncludes persons who indicated American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, or other race.

^cIndicates significantly different values (P < .05) between black non-Hispanic and white non-Hispanic/other participants.

^dIndicates significantly different values (P < .05) between black non-Hispanic and Hispanic participants.

^eIndicates significantly different values (P < .05) between Hispanic and white non-Hispanic/other participants.

Overall, the estimated annual HIV incidence density was 3.4/100 PY. Incidence was 4.1/100 PY for blacks, 3.2/100 PY for Hispanics, and 1.1/100 PY for white/other (Figure 1). Table 4 examines characteristics associated with being HIV infected among 415 adolescent SMM who had a valid positive or negative NHBS-YMSM HIV test result. Of these, 25 (6%) were HIV infected. HIV prevalence was higher among black than white/other SMM (9% vs 1%) and among participants from Philadelphia as compared with Chicago (12% vs 4%). HIV prevalence was higher among adolescent SMM who reported ≥4 male partners (14% vs 4%), condomless anal sex with a male partner (10% vs 2%), and exchange sex (32% vs 5%) in the past 12 months compared to those who did not. Examining recent sexual partnerships, HIV prevalence was higher among adolescents who had at least 1 older partner (6% vs 1%), black partner (10% vs 3%), HIV-infected partner (25% vs 5%), and a partner who had ever been in jail or prison (15% vs 4%). Also HIV prevalence was higher among those who had sex within a week of meeting their sex partner (11% vs 4%).

Figure 1.

Estimated human immunodeficiency virus (HIV) incidence density among adolescent sexual minority males, National HIV Behavioral Surveillance for Young Men Who Have Sex with Men (NHBS-YMSM), 2015. Analyses among participants who reported ever having anal sex with a male partner and had a valid positive or negative, NHBS-YMSM test result. Other race includes persons who indicated American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, or other race. Estimated HIV incidence density was 3.4 per person-years (PY) at risk (24 HIV infections among 702 PY), 4.1/100 PY among black (15 infections among 367 PY), 3.2/100 PY among Hispanic (8 HIV infections among 247 PY), and 1.1/100 PY among white and other adolescent males (1 HIV infection among 87 PY). Abbreviation: HIV, human immunodeficiency virus.

Table 4.

Human Immunodeficiency Virus (HIV) Infection^a Among Sexually Active^b Adolescent Sexual Minority Males, by Selected Characteristics, National HIV Behavioral Surveillance for Young Men Who Have Sex with Men, 2015

Characteristic	Total No. (n = 415)	HIV-Infected
		No.	%	P Value
		Age, y
13–15	37	2	5	.9
16–17	245	14	6
18	133	9	7
Race/ethnicity
Black	172	15	9	.05c
Hispanic	160	9	6
White/otherd	82	1	1
Household financial instabilitye, past 12 mo
No	307	17	6	.48
Yes	106	8	8
City
Chicago	187	8	4	.06f
New York City	143	7	5
Philadelphia	85	10	12
Sexual identity
Gay or bisexual	399	24	6	.3
Heterosexual	11	1	9
Anal sex before age 13 y
No	167	12	7	.41
Yes	248	13	5
≥4 male partnersg, past 12 mo
No	291	13	4	.01
Yes	81	11	14
Condomless anal sex with a male partner, past 12 mo
No	173	4	2	.003
Yes	218	21	10
Exchange sexh, past 12 mo
No	396	19	5	<.0001
Yes	19	6	32
3 most recent partnersi
Female partner
No	367	23	6	.75
Yes	48	2	4
Older partnerj
No	105	1	1	.04
Yes	227	14	6
Black partner
No	227	7	3	.01
Yes	186	18	10
HIV-infected partner
No	403	22	5	.03
Yes	12	3	25
Partner of unknown HIV status
No	260	17	7	.67
Yes	155	8	5
Met partner onlinek
No	250	11	4	.09
Yes	164	14	9
Sexual concurrencyl
No	200	8	4	.1
Yes	215	17	8
No	382	21	5	.13
Yes	33	4	12
Alcohol or drugs before or during last sex
No	307	14	5	.06
Yes	108	11	10
Partner ever in jail or prison
No	348	15	4	.003
Yes	67	10	15
Serious partnerm
No	166	9	5	.83
Yes	249	16	6
Sex within a week of meeting
No	301	12	4	.01
Yes	114	13	11

Numbers might not add to total because of missing data. Abbreviation: HIV, human immunodeficiency virus.

^aParticipants who had a valid positive or negative National HIV Behavioral Surveillance for Young Men Who Have Sex with Men test result are included in this analysis.

^bParticipants who ever had anal or vaginal sex.

^cIndicates significantly different values (P < .05) between black non-Hispanic and white non-Hispanic/other participants.

^dIncludes persons who indicated American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, or other race.

^eIn the past 12 months, was there a time where there was not enough money in your house for rent, food, or utilities such as gas, electric, or phone?

^fIndicates significantly different values (P < .05) between Chicago and Philadelphia.

^gHad oral or anal sex.

^hHad sex in exchange for something such as money or drugs.

ⁱParticipants were asked about their 3 most recent sex partners in the past 3 months. If the participant did not have sex with anyone in the past 3 months, they were asked about the last person he had sex with in the past 12 months. The “yes” category indicates that a participant had at least 1 sexual partnership (among a possible 3 partnerships) with the characteristic or behavior of interest.

^jDue to mandated reporting requirements, Philadelphia did not ask about partner’s age. Chicago and New York City asked whether the partner was young, older, or the same age as the participant.

^kMet partner on the internet or through a mobile app (eg, Grindr, Jack’d, Scruff).

^lDuring the time of a sexual relationship, the participant was having sex with other people and/or the partner was probably or definitely having sex with other people.

^mA male with whom the participant has dated for a while and feels close to and may call a boyfriend or partner.

DISCUSSION

Overall, HIV prevalence and estimated incidence density for adolescent SMM in this study were high, driven primarily by the rates for black and Hispanic participants. Both individual behavior and partner characteristics were associated with being HIV infected. Similar to previous research [11, 14, 16], reported prevalence of several key HIV-related risk behaviors, including condomless anal sex with a male partner in the past 12 months, ≥4 oral or anal male sex partners in the past 12 months, and sexual concurrency with at least 1 recent partner, was high across all racial/ethnic groups in our sample.

We did find some racial/ethnic differences in HIV-related risk behaviors. Black adolescent SMM reported an earlier age of sexual debut, which provides an extended window of risk opportunity and could be a possible contributing factor in the disproportionate HIV rates by race/ethnicity [25, 26]. Black adolescent SMM were also more likely than their white/other adolescent SMM counterparts to have ever had anal sex with a male partner, have engaged in exchange sex in the past 12 months, and report sexual concurrency within a recent relationship. Consistent with previous literature [5, 27, 28], we also found differences in sexual partner characteristics and sexual network factors by race/ethnicity. In particular, black adolescent SMM were more likely than the other groups to have a black partner. Previous studies have proposed that the background prevalence rates in the population paired with racially homogenous sexual networks is a contributing factor to why HIV rates are disproportionately high among black SMM [8, 28, 29]. White/other adolescent SMM were more likely than black adolescent SMM to report alcohol or drugs before or during last sex and were more likely than both black and Hispanic SMM to report sex within 1 week of meeting a new partner. Although these behaviors place white/other adolescent SMM at risk for HIV and other sexually transmitted infections, the lower HIV prevalence in their partner pool may lower the impact of these risks for HIV acquisition.

This study has several limitations. First, the study is a convenience sample and the analyses do not account for the sampling design. Participants may not be representative of all SMM 13–18 years of age, and data are not weighted. Second, data were self-reported and may have associated biases. Third in the calculation of PY at risk for incidence density, we used age at diagnosis to approximate age at HIV infection, which likely inflates the denominator by increasing the estimate of years at risk. However, as this is a young population, we expect the amount of inflation to be small. On the other hand, only participants who reported ever having anal sex with a male partner were included for calculating PY at risk; if this behavior was misreported, we could have underestimated the denominator.

The findings from this study, including the early onset of sexual activity, high rates of HIV-related risk behavior, and elevated HIV prevalence and estimated incidence density, suggest that effective HIV prevention strategies, including “test and treat” and preexposure prophylaxis (PrEP), may be helpful in reducing HIV risk among adolescent SMM. Unfortunately, there is currently a dearth of HIV prevention interventions that focus on this population specifically, and to date, PrEP is not approved for use below the age of 18 [11, 30]. Therefore, it is imperative that future HIV prevention research examine interventions that not only include adolescent SMM but do so in a way that encompasses the developmental needs of the adolescent life stage [31–33].