RESEARCH AND PRACTICE

Black Men Who Have Sex Prevention. The Brothers y Hermanos study
recruited Black and Latino MSM to examine
male in the past 3 months, receptive unpro-
tected anal intercourse with a male in the past

With Men and the factors associated with HIV risk behavior and
HIV infection. Men were enrolled from May
3 months, or unprotected anal or vaginal
intercourse with a female in the past 3 months.
Association of Down-Low 2005 to April 2006 in Philadelphia, Los
Angeles, and New York City. Candidates had to
Descriptive analyses and the c2 test were
used to compare demographic and sexual risks
Identity With HIV Risk be male (and identify as such); be 18 years or
older; self-identify as Black, African American,
of the full sample of down-low MSM and
nondown-low MSM. For all analyses, unad-
Behavior or Latino; and report sexual intercourse (oral,
anal sex, or mutual masturbation) with a man in
justed bivariate relationships were examined
first to test for differences between down-low–
Lisa Bond, PhD, Darrell P. Wheeler, PhD, MPH, the past 12 months. Participation was open to identified and nondown-low–identified MSM.
Gregorio A. Millett, MPH, Archana Bodas men who were HIV positive, HIV negative, or Variables included in multivariate models were
LaPollo, MPH, Lee F. Carson, MSW, and of unknown serostatus. Respondent-driven those found to be statistically significant in
Adrian Liau, PhD sampling11,12 was used to recruit participants. bivariate analyses (P £ .10) or relevant based on
Participants completed an audio computer- a priori knowledge. Multivariate logistic re-
assisted self-interview followed by HIV testing. gression assessed the contribution of down low
Black men ‘‘on the down low’’
All participants, except those who disclosed dur- identity to each of the sexual risk behavior
have been considered prime agents
of HIV transmission in the Black ing screening that they had previously been outcomes among MSM reporting sex with male
community despite little empirical diagnosed as HIV positive, received pretest or female partners in the past 3 months, while
evidence. We assessed the relation- counseling and were tested for HIV antibodies adjusting for age, sexual orientation, HIV
ship between down-low identifica- using a rapid, oral fluid HIV antibody test. serostatus, study site, sex trade, income, and
tion and sexual risk outcomes among Participants who preliminarily tested positive on type of sexual partners.
1151 Black MSM. Down-low Identifi- the rapid test and those who disclosed during
cation was not associated with un- screening that they were HIV positive provided a RESULTS
protected anal or vaginal sex with blood specimen for confirmatory testing through
male or female partners. Future HIV Western blot assay. Our analysis was restricted One third of participants (31%) identified
prevention programs and research
to data collected from 1151 Black MSM in New with the term ‘‘on the down low.’’ Down-low
should target sexual risk behaviors
York City and Philadelphia. MSM were more likely than were nondown-
of Black men, irrespective of identity,
and not focus on the ‘‘down low.’’ Identification with ‘‘down low’’ was assessed low MSM to report their sexual identity as
(Am J Public Health. 2009;99:S92– by the question ‘‘Do you identify with any of bisexual and to report that it was ‘‘very im-
S95. doi:10.2105/AJPH.2007.127217) the following terms? (check all that apply): portant’’ to keep their same-sex behaviors se-
queer; same-gender loving; in the life; on the cret (Table 1). One fourth of down-low MSM
down low; none.’’ Men who checked ‘‘on the also reported a gay sexual identity, and 11.1%
The term ‘‘on the down low’’ has been used down low’’ were categorized as ‘‘identify with reported being straight or heterosexual. Re-
to describe Black men who identify as hetero- the down low’’; all other men were categorized garding the meaning of ‘‘down low,’’ compara-
sexual yet put their female partners at risk as ‘‘do not identify with the down low.’’ The ble proportions of down-low MSM and
for HIV infection by secretly having homosex- meaning of down low was also asked of re- nondown-low MSM selected similar character-
ual sex.1–7 Black men on the down low have spondents and worded as follows: ‘‘Recently istics that ‘‘best describe’’ men who are down
been implicated in the popular press as prime people have been talking more about the down low. Half of the men (irrespective of down-low
agents of HIV transmission in the Black com- low or DL, but this may mean different things identity) did not characterize the down low as
munity.1,8–10 However, the extent to which the to different people. We would like to know having a wife or a girlfriend.
down low is associated with HIV risk behavior what DL means to you. Which of the following A majority of participants had previously
has never been quantitatively examined among a do you feel best describes men who are DL or been tested for HIV and there were no signif-
large sample of Black men who have sex with on the down low? (check all that apply): ex- icant differences between down-low MSM and
men (MSM). We sought to assess whether or tremely masculine; always tops (men who nondown-low MSM in HIV testing history.
not down low identity is associated with greater penetrate during anal or oral sex); less likely to Similar proportions of down-low MSM and non
sexual risk behavior with male or female partners have HIV; have fewer male sex partners than down-low MSM tested HIV positive and were
of Black MSM. gay identified men; DL only refers to Black unaware of their infection, but down-low MSM
men; have wives or girlfriends; only have male were significantly less likely to be HIV positive
METHODS sex partners; I don’t agree with any of these; I overall.
never heard of the term down low or DL.’’ Similar proportions of down low and non-
Data were collected as part of a study fun- Risk behavior outcomes (yes or no) included down-low men had recent sexual intercourse
ded by the Centers for Disease Control and insertive unprotected anal intercourse with a with a male partner. Among these participants,

S92 | Research and Practice | Peer Reviewed | Bond et al. American Journal of Public Health | Supplement 1, 2009, Vol 99, No. S1
 RESEARCH AND PRACTICE

TABLE 1—Demographic, HIV Status, and Risk Behavior Comparison of Black MSM Who Do and
Do Not Identify With the Term ‘‘Down Low’’: Philadelphia and New York City, 2005–2006 (N = 1151)

Do Not Identify With Down Identify With Down

Low, No. (%) (n = 790) Low, No. (%) (n = 361) P

Age .353
18–29 111 (14.1) 47 (13.1)
30–39 174 (22.1) 65 (18.1)
40–49 364 (46.1) 175 (48.6)
‡ 50 140 (17.7) 73 (20.3)
Gross annual income .680
< $5,000 264 (34.4) 131 (36.8)
$5,000–$9,999 183 (23.8) 85 (23.9)
$10,000–$19,999 163 (21.2) 66 (18.5)
$20,000–$29,999 72 (9.4) 39 (11.0)
‡ $30,000 86 (11.2) 35 (9.8)
Sexual identity <.001
Homosexual or gay 443 (56.2) 101 (28.1)
Bisexual 225 (28.6) 200 (55.6)
Heterosexual or straight 83 (10.5) 40 (11.1)
Other 37 (4.7) 19 (5.3)
What best describes men on the down lowa
Have wives or girlfriends 400 (51.0) 176 (48.8) .489
Extremely masculine 142 (18.1) 87 (24.1) .018
Always the top partner 131 (16.7) 82 (22.7) .015
Have fewer male partners than gay men 152 (19.4) 82 (22.7) .191
Don’t agree with any of above statements 119 (15.2) 48 (13.3) .406
Never heard of the ‘‘DL’’b 25 (3.2) ... ...
Ever tested for HIV 726 (92.0) 324 (89.8) .206
HIV status <.001
Negative 343 (43.8) 202 (56.3)
Positive (infected, aware)c 362 (46.2) 124 (34.5)
Positive (infected, unaware)c 79 (10.1) 33 (9.2)
Type of sexual intercourse, past 3 months
Traded sex with a male 160 (20.3) 106 (29.4) .001
Any anal intercourse with male 555 (70.3) 265 (73.4) .273
Insertive UAI with male (n = 820)d 263 (47.4) 138 (52.1) .209
Receptive UAI with male (n = 820)d 205 (36.9) 74 (27.9) .011
Any vaginal/anal intercourse with female 182 (23.0) 167 (46.3) <.001
Unprotected vaginal or anal intercourse with female (n = 349)e 108 (59.3) 98 (58.7) .901
Importance of keeping sex with men secret .001
Not at all important 273 (34.7) 50 (13.9)
Somewhat/little important 328 (41.3) 145 (40.2)
Very important 188 (23.9) 166 (46.0)

Note. MSM = men who have sex with men; DL = down low; UAI = unprotected anal intercourse. Numbers may not add to totals due to missing data.
a
These statements were checked by participants as best describing men who are down low. Other statements not shown were checked by less than 15% of participants in the DL or
non-DL samples.
b
Zero participants who identified with the term ‘‘down low’’ reported that they had ‘‘never heard of the DL.’’
c
Infected-aware men self-reported that they were HIV-positive at the time of study enrollment, and their positive status was confirmed through study testing. At enrollment,
infected-unaware men self-reported they were HIV-negative, did not know their status, did not return for result of their last HIV test, or last test result was indeterminant. These men
subsequently tested HIV-positive through study testing.
d
Includes only participants who had a male sex partner in the past 3 months (n = 820; 265 DL men; 555 non-DL men).
e
Includes only participants who had a female sex partner in the past 3 months (n = 349; 167 DL men; 182 non-DL men).

Supplement 1, 2009, Vol 99, No. S1 | American Journal of Public Health Bond et al. | Peer Reviewed | Research and Practice | S93
 RESEARCH AND PRACTICE

comparable proportions of down-low and additional analyses we conducted (not shown) (irrespective of their HIV serostatus) engaged
nondown-low MSM reported insertive unpro- in which we restricted the sample to HIV- in similar sexual risks with male and female
tected anal intercourse, whereas down-low positive men, we found no significant differ- partners as nondown-low–identified MSM. Al-
MSM were significantly less likely than were ences in the proportions of down-low identified though down-low MSM were more likely than
nondown-low MSM to report receptive unpro- and nondown-low identified men reporting were nondown-low MSM to report a female sex
tected anal intercourse. Down-low MSM were unprotected sex with female partners who were partner, both groups of men engaged in com-
more likely to report sex trade activities with HIV negative or whose HIV status was un- parable sexual risks with female partners. Sec-
men and more likely than were nondown-low known (N = 72; c2 = .937; P = .333) or male ond, identifying with the down low did not
MSM to have sexual intercourse with women. partners (N = 365; c2 = .016; P = .898). always imply having female sex partners. Half
However, one fourth of nondown-low men also of the down-low men in our study reported no
reported recent sex with a woman, and down-low DISCUSSION recent sex with a woman, whereas 23% of
MSM and nondown-low MSM reported compa- nondown-low men reported a recent female
rable rates of unprotected sex with women. Our study suggests several important find- sex partner. Third, our study challenges pre-
Table 2 shows the results of 3 multivariate ings. First, contrary to what has been reported vailing assumptions that men on the down low
logistic regression models for each risk behav- in the popular press, down-low identity was not primarily identify as heterosexual. Down-low
ior outcome. Controlling for other covariates, associated with engaging in greater sexual risk MSM were much more likely to identify as
down-low identity was not significantly as- behavior with female or male partners. Black bisexual or homosexual than heterosexual, and
sociated with any sexual risk outcome. In MSM who identified with the down low nondown-low MSM were just as likely as

TABLE 2—Multivariate Logistic Regression Models of the Association Between Down Low
Identity and Unprotected Intercourse with Male and Female Partners, Among Black MSM
Recruited in New York City and Philadelphia, 2005-2006

Unprotected Insertive Anal Unprotected Receptive Anal Unprotected Vaginal or Anal

Intercourse With a Male in Intercourse With a Male in the Intercourse With a Female in
the Past 3 Monthsa,c (n = 801) Past 3 Monthsa,c (n = 801) the Past 3 Monthsb,c (n = 342)
AOR (95% CI) P AOR (95% CI) P AOR (95% CI) P

Identify as ‘‘down low’’ 1.12 (0.80, 1.55) .511 0.87 (0.60, 1.27) .474 0.84 (0.53, 1.34) .471
Age, y 0.99 (0.98, 1.01) .517 0.99 (0.97, 1.00) .090 1.00 (0.97, 1.02) .843
Sexual identity
Heterosexual or straight Ref Ref Ref
Homosexual or gay 0.97 (0.50, 1.90) .938 7.36 (2.68, 20.22) <.001 1.07 (0.40, 2.88) .894
Bisexual 1.03 (0.54, 1.99) .921 2.27 (0.83, 6.22) .112 0.94 (0.53, 1.67) .826
Other 0.59 (0.22, 1.54) .279 2.27 (0.62, 8.37) .218 0.85 (0.29, 2.47) .763
HIV status, self-reported
Negative Ref Ref Ref
Positive 0.97 (0.69, 1.37) .870 1.42 (0.97, 2.07) .074 0.57 (0.31, 1.05) .073
Unknown 1.09 (0.69, 1.70) .719 1.61 (0.98, 2.64) .058 0.94 (0.48, 1.83) .853
Traded sex with a male past 3 months 1.95 (1.36, 2.78) <.001 2.11 (1.43, 3.13) <.001 1.17 (0.69, 1.99) .552
Type of sex partner(s) in past 3 monthsd
Main partner only Ref Ref Ref
Casual partner only 0.96 (0.65, 1.41) .836 0.64 (0.41, 0.99) .043 1.07 (0.65, 1.78) .780
Both main partner and casual partner 2.16 (1.40, 3.31) <.001 1.65 (1.05, 2.60) .030 3.93 (1.97, 7.85) <.001
Study site
Philadelphia, PA Ref Ref Ref
New York City, NY 1.29 (0.94, 1.76) .114 1.08 (0.77, 1.53) .648 0.65 (0.39, 1.09) .104

Note. MSM = men having sex with men; AOR = adjusted odds ratio; CI = confidence interval.
a
Participants reporting no sex in the past 3 months with a male partner are excluded from this analysis.
b
Participants reporting no sex in the past 3 months with a female partner are excluded from this analysis.
c
Models adjusted for income level of participants.
d
Type of partner refers to male partners in models predicting insertive and receptive unprotected anal intercourse, and female partners in model predicting unprotected vaginal and anal intercourse.
Main partner was defined as a ‘‘committed relationship.’’

S94 | Research and Practice | Peer Reviewed | Bond et al. American Journal of Public Health | Supplement 1, 2009, Vol 99, No. S1
 RESEARCH AND PRACTICE

down-low MSM to report being heterosexual. The authors would like to thank the entire Brothers y 13. Ford CL, Whetten KD, Hall SA, Kaufman JS,
Hermanos study team, community-based partners in Thrasher AD. Black sexuality, social construction, and
Last, compared with nondown-low MSM, down-
Philadelphia and New York City for their collaboration research targeting ‘‘the down low’’ (‘‘the DL’’). Ann Epi-
low–identified MSM reported fewer instances of on the study, and volunteer members of local community demiology. 2007;17:209–216.
receptive unprotected anal intercourse with male advisory boards who guided the design and implemen- 14. Millett GA, Malebranche D, Mason B, Spikes P.
tation of the study.
partners and were less likely to test HIV positive. Focusing ‘‘down low’’: bisexual Black men, HIV risk and
heterosexual transmission. J Nat Med Assoc. 2005;97:
Our study has its limitations. First, the survey
52S–59S.
did not differentiate primary identification as Human Participant Protection
This study was approved by the institutional review 15. Pathela P, Hajat A, Schillinger J, Blank S, Sell R,
down low from other identities that down-low boards of the Public Health Management Corporation, Mostashari F. Discordance between sexual behavior and
men may have chosen. Second, the men in our the New York City Department of Health and Mental self-reported sexual identity: a population-based survey
Hygiene, Hunter College, and the Centers for Disease of New York City men. Ann Intern Med. 2006:145;416–
sample were predominantly low income and
Control and Prevention. All HIV testing information was 425.
were recruited in metropolitan areas in the collected with signed informed consent. 16. Wheeler DP. Exploring HIV prevention needs of
Northeast; they may not be representative of nongay-identified Black and African American men who
Black MSM overall. Third, we relied on self- have sex with men: a qualitative exploration. Sex Transm
Dis. 2006;33(suppl 7):S11–S16.
reports to assess sexual risk behaviors, al- References
17. Wolitski RJ, Jones KT, Wasserman JL, Smith JC. Self-
though use of audio computer-assisted self- 1. Denizet-Lewis B. Double lives on the down low. New
identification as ‘‘down low’’ among men who have sex
interviews likely reduced underreporting York Times Magazine. August 3, 2003. Available at http://
with men (MSM) from 12 US cities. AIDS Behav.
query.nytimes.com/gst/fullpage.html?sec=health&
because of social desirability. res=9F0CE0D61E3FF930A3575BC0A9659C8B63.
2006;10:519–529.
Findings from our study underscore the Accessed July 1, 2007.
limitations of using identity rather than sexual 2. Harris EL. Passing for straight: talking to brothers on
behavior to predict sexual risk.13–17 Interven- the down low. Essence. June 8, 2004. Available at http://
www.essence.com/essence/bodyandsoul/relationships/
tions that focus on men who identify as down 0,16109,648878,00.html. Accessed July 1, 2007.
low will fail to reach significant numbers of 3. King JL. On the Down Low: A Journey Into the Lives of
nondown-low–identified men who are sexually ‘‘Straight’’ Black Men Who Sleep with Men. New York, NY:
involved with women. The scientific community Broadway Books; 2004.

should address more salient, and scientifically 4. Williams S. My downlow husband. Black Star News.
March 11, 2005. Available at: http://www.blackstarnews.
valid, contributors to sexual risk among Black com/?c=135&a=1394. Accessed July 1, 2007.
MSM than identifying as ‘‘down low.’’ j 5. A secret sex world: living on the ‘‘down low’’
[television broadcast]. Oprah Winfrey Show. CBS televi-
sion. April 16, 2004.
6. Johnson JB. Secret gay encounters of Black men
About the Authors could be raising women’s infection rate. San Francisco
Lisa Bond, Archana LaPollo, and Lee Carson are with Chronicle. May 1, 2005. Available at: http://sfgate.com/
the Public Health Management Corporation, Philadelphia, cgi-bin/article.cgi?file=/c/a/2005/05/01/DOWNLOW.
PA. Darrell Wheeler is with the Hunter College School of TMP&type=health. Accessed July 1, 2007.
Social Work, City University of New York, New York, NY.
Gregorio Millett and Adrian Liau are with the Centers for 7. National Association of State and Territorial AIDS
Disease Control and Prevention, Atlanta, GA. Directors. Black Men Who Have Sex with Men. Issue Brief
Requests for reprints should be sent to Lisa Bond, Public No. 1; February 2006. Available at: http://www.nastad.
Health Management Corporation, 260 South Broad St, org/Docs/Public/Publication/2006221_BlackMSM_
Philadelphia, PA 19102-5085 (lisab@phmc.org). IssueBrief_FINAL.pdf (PDF). Accessed July 1, 2007.
This brief was accepted on March 9, 2008. 8. Wright K. The great down-low debate. A new Black
sexual identity may be an incubator for AIDS. The Village
Voice. June 6–12, 2001. Available at: http://www.
Contributors villagevoice.com/news/0123,wright, 25340,5.html.
L. Bond initiated and completed analyses of all data and Accessed July 1, 2007.
led the writing of the article. D. P. Wheeler contributed
to conceptualizing the article, interpreting the findings, 9. Sternberg S. The danger of living ‘down low’; Black
and editing the article. G. A. Millett contributed to con- men who hide their bisexuality can put women at risk.
ceptualizing the article, interpreting the findings, and USA Today. March 15, 2001:1D.
editing the article. A. B. LaPollo contributed to analyzing 10. Vargas JA. HIV-positive, without a clue: Black men’s
the data, interpreting the findings, and editing the article. hidden sex lives imperiling female partners. The Wash-
L. F. Carson assisted with collecting data and editing the ington Post. August 4, 2003. Available at: http://www.
article. A. Liau assisted with interpreting the findings and washingtonpost.com/wp-dyn/content/article/2006/11/
editing the article. All authors were involved in study 27/AR2006112700669.html. Accessed July 1, 2007.
design, reviewing drafts of the article, and providing 11. Heckathorn DD. Respondent-driven sampling: a
approval of the final version. new approach to the study of hidden populations. Soc
Probl. 1997;44:174–199.
Acknowledgments 12. Heckathorn DD, Semaan S, Broadhead RS, Hughes
The Brothers y Hermanos study was funded through a JJ. Extensions of respondent-driven sampling: a new
cooperative agreement from the US Centers for Disease approach to the study of injection drug users aged 18–
Control and Prevention. 25. AIDS Behav. 2002;6:55–67.

Supplement 1, 2009, Vol 99, No. S1 | American Journal of Public Health Bond et al. | Peer Reviewed | Research and Practice | S95