Levels of alcohol use and history of HIV testing among female
sex workers in Mombasa, Kenya
Angela M. Bengtson
a,*
, Kelly L’Engle
b
, Peter Mwarogo
c
, and Nzioki King’ola
d
a
Department of Epidemiology, University of North Carolina, Chapel Hill, NC, USA
b
Social and Behavioral Health Sciences Department, FHI 360, Research Triangle Park, NC, USA
c
FHI 360, Nairobi, Kenya
d
International Centre for Reproductive Health, Mombasa, Kenya
Abstract
HIV testing is a critical first step to accessing HIV care and treatment, particularly for high-risk
groups such as female sex workers (FSWs). Alcohol use may be a barrier to accessing HIV
services, including HIV testing. We analyzed data from a cross-sectional survey of 818 FSWs in
Mombasa, Kenya, and estimated the association between different levels of alcohol use and
having never tested for HIV. In multivariable analyses, higher levels of alcohol consumption were
associated with having never tested for HIV (PR 1.60; 95% CI: 1.07, 2.40). Future interventions
should explore whether reducing harmful drinking improves HIV testing among FSWs.
Keywords
HIV testing; alcohol use; female sex workers; HIV prevention
Introduction
HIV testing is the critical first step to accessing HIV care and treatment (Dilernia et al.,
2013; Kilmarx & Mutasa-Apollo, 2013). Timely linkage to care and early initiation of
antiretroviral therapy (ART) are essential to achieve virologic suppression and reduction in
HIV transmission (Cohen et al., 2011; Govindasamy et al., 2011; MacPherson et al., 2012;
Rosen & Fox, 2011). Test-and-treat HIV prevention strategies are increasingly being
considered for a number of high-risk populations, including female sex workers (FSWs;
Delva et al., 2012). For test-and-treat strategies to be effective, barriers to HIV testing need
to be addressed.
FSWs in sub-Saharan Africa (SSA) are widely recognized as a critical population to target
for HIV testing and prevention services (Braunstein et al., 2011; World Health Organization,
2005). The prevalence of HIV among FSWs in Mombasa is estimated between 30–35%
(Luchters et al., 2010; van der Elst et al., 2009), compared with 4.3% among the general
© 2014 Taylor & Francis
*
Corresponding author. abengtso@live.unc.edu.
NIH Public Access
Author Manuscript
AIDS Care. Author manuscript; available in PMC 2015 July 21.
Published in final edited form as:
AIDS Care. 2014 ; 26(12): 1619–1624. doi:10.1080/09540121.2014.938013.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
population (National AIDS and STI Control Programme, Ministry of Health, Kenya, 2013).
Nearly 14% of new HIV infections are estimated to be attributable to sex with FSWs in
Kenya (Kenya National AIDS Control Council, 2009). Despite the need for HIV testing
among FSWs, voluntary counseling and testing has been reported as low as 7%
(Abdelrahim, 2010). In Kenya, nearly 60% of FSWs report not knowing their HIV status
(Luchters et al., 2008). Younger age, less education, and not having a regular sexual partner
have been associated with decreased HIV testing among FSWs in Asia (Hong et al., 2012;
Xu et al., 2011). Less is known about factors associated with HIV testing among FSWs in
SSA.
Alcohol use among FSWs is high and has long been identified as a contributor to risky
sexual behavior (Asiki et al., 2011; Kalichman, Simbayi, Kaufman, Cain, & Jooste, 2007;
World Health Organization, Department of Mental Health and Substance Dependence,
2000; Zablotska et al., 2006). In Kenya, 30% of FSWs report drinking daily (Chersich et al.,
2007). Alcohol use has been associated with an increased number of sexual partners,
unprotected sex, inconsistent condom use, and sexual violence (Coldiron et al., 2008;
Pitpitan et al., 2012; Weiser et al., 2006). Outcomes across the HIV care continuum,
including sub-optimal ART adherence and loss to follow-up from HIV care, have also been
linked to alcohol use (Deribe, Hailekiros, Biadgilign, Amberbir, & Beyene, 2008; Kenya et
al., 2013; Nakimuli-Mpungu et al., 2012; Ohl et al., 2013). Alcohol use has previously been
associated with HIV testing behavior and may play a role in failure to seek HIV testing
(Fatch et al., 2012; Luseno & Wechsberg, 2009). However, little is known about the
relationship between alcohol use and HIV testing among FSWs in SSA.
The goal of the present analysis was to investigate the association between levels of alcohol
use and never having tested for HIV among FSWs who use alcohol in Mombasa, Kenya.
Methods
Data for the present analysis come from 818 women who completed the baseline interview
of a longitudinal intervention to reduce alcohol use among FSWs in Mombasa, Kenya
between October 2011 and October 2012. FSWs were enrolled from three community drop-
in centers that provide condoms and routine HIV/sexually transmitted infection (STI) testing
through US Agency for International Development’s (USAID) AIDS, Population and Health
Integrated Assistance Plus (APHIAplus) program. Participants were randomized to either a
brief alcohol-reduction intervention or nutritional counseling and followed for 12 months.
The goal of the intervention was to reduce alcohol consumption among FSWs; therefore,
only women who regularly drank alcohol but were not alcohol dependent, self-reported
being a FSW, were ≥18 years of age, and lived in Mombasa were eligible for inclusion in
the intervention and present analysis. Participants received HIV and STI testing as part of
the intervention at baseline, 6 and 12 months. Women were asked about HIV testing
frequency regardless of HIV status. HIV incidence was a primary endpoint of the main
study; therefore, lab-confirmed HIV status was prioritized and information on knowledge of
HIV status was not collected. Baseline HIV status was used as a surrogate for knowledge of
HIV status. All participants provided written informed consent and the study was approved
Bengtson et al.
Page 2
AIDS Care. Author manuscript; available in PMC 2015 July 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
by the ethical review boards of the Kenyatta National Hospital in Kenya and FHI360 in the
USA
Measures
Alcohol use during the past year was measured using the WHO-endorsed 10-item Alcohol
Use Disorders Identification Test (AUDIT; Babor & Higgins-Biddle, 2001), which has been
previously validated in Kenya (Saunders, Aasland, Babor, de la Fuente, & Grant, 1993).
Women enrolled in the intervention had AUDIT scores from 7 to 19. In the present analysis,
we used previously established definitions (Babor & Higgins-Biddle, 2001; MacAskill et al.,
2011; Miller, Zweben, DiClemente, & Rychtarik, 1992) to categorized alcohol use into
higher “harmful” alcohol use (AUDIT score 16–19), compared to lower “hazardous” alcohol
use (AUDIT score 7–15; referent). Whether a participant had tested for HIV prior to
enrolling in the study was self-reported at baseline. Confounders of interest included
baseline age, education, number of children, number of sexual partners in the last 7 days,
number of years as a FSW, and HIV status. Categorization of age and number of years as a
FSW was based on quartiles. Very few women reported being currently married (3%) or
living with a partner (8%); therefore, these variables were not included as confounders.
Statistical analysis
We used log-binomial and Poisson models with a robust variance estimator (Barros &
Hirakata, 2003) to estimate the association between alcohol use and having never tested for
HIV. Multivariable estimates were obtained using a Poisson model, due to convergence
issues with the log-binomial model. In addition to all confounders, multivariable estimates
included adjustment for the drop-in center attended. We conducted a sensitivity analysis
restricted to HIV-uninfected women at baseline since knowledge of being HIV-infected
could impact both testing behavior and alcohol consumption. All analyses were conducted
using Stata 11 (StataCorp, College Station, TX).
Results
Women included in the study overall were young in age (30% aged 18–23); however, a
higher proportion of women who had never tested for HIV were aged 31–54 (38% compared
to 23%). The majority of women had little education (55% never attended school or only
primary school) and had at least one child (81%). A higher proportion of women with no
prior HIV testing reported >7 years of sex work, compared to women who had tested for
HIV (31% compared to 24%). A higher proportion of HIV-infected women had never tested
for HIV (29% compared to 19% of women who had tested; Table 1).
The prevalence of HIV among all FSWs in the study was 20%. Approximately, 11% of
FSWs reported never having been tested for HIV prior to enrolling in the study. Among
FSWs who had ever been tested, 45% reported testing every 3 months or more, 15%
reported testing every 6 months, 12% reported testing every 1–3 years, and 16% reported
testing only once.
In multivariable analyses, women who reported harmful alcohol consumption were 1.60
(95% CI: 1.07, 2.40) times as likely to have never tested for HIV, compared to women with
Bengtson et al.
Page 3
AIDS Care. Author manuscript; available in PMC 2015 July 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
hazardous alcohol consumption. Being age 31 or older was independently associated with
never having tested for HIV (PR 1.98; 95% CI: 1.06, 3.70), while having at least one child
(PR 0.39; 95% CI: 0.24, 0.64) was associated with previous HIV testing (Table 2). In the
sensitivity analysis restricted to HIV-uninfected women, the association between harmful
drinking and having never tested for HIV was similar (PR 1.64; 95% CI: 1.00, 2.69).
Discussion
In this population of FSWs who consume alcohol, women who reported higher harmful
drinking were significantly more likely to have never tested for HIV, compared to women
who reported lower hazardous drinking. These results are consistent with previous evidence
suggesting that alcohol use is associated with HIV testing behavior for women in SSA
(Fatch et al., 2012; Luseno & Wechsberg, 2009). While the cross-sectional nature of our
study precludes attributing causality, our results suggest that FSWs with harmful alcohol use
may be more likely to have never tested for HIV.
Increasingly, alcohol use has been recognized as an important factor for a range of HIV
outcomes, including HIV testing (Fatch et al., 2012; Gari et al., 2013; Peltzer & Mlambo,
2010), accessing ART(Arasteh & Des Jarlais, 2009) and ART adherence (Kenya et al.,
2013; Lyimo et al., 2012; Nakimuli-Mpungu et al., 2012; Ohl et al., 2013). Brief
interventions to reduce alcohol use and risky sexual behaviors, paired with HIV testing, have
shown promise in the short term (Edelman et al., 2012). However, interventions with longer
follow-up are needed to determine whether reducing alcohol use impacts HIV testing and
sexual risk behaviors over time.
Older age has previously been associated with having received an HIV test among women in
Kenya (Cherutich et al., 2012). In our study, older age was independently associated with
having never tested for HIV, suggesting that efforts to test FSWs may be missing older
FSWs. Having a child was associated with prior HIV testing, likely due to being tested
during antenatal care (Cherutich et al., 2012; MacPhail, Pettifor, Moyo, & Rees, 2009). The
relationship between alcohol use and HIV testing did not differ meaningfully when analyses
were restricted to HIV-uninfected women, which may reflect the fact that HIV-infected
women were unaware of their status prior to study enrollment.
Our study has several strengths and limitations. Strengths include the large sample of a high-
risk population of FSWs attending drop-in centers in Kenya and the use of a validated
measure of alcohol use. Limitations include the restricted range of alcohol use (AUDIT 7–
19), the fact that alcohol use in the previous 12 months may not impact lifetime HIV testing
behavior, inability to draw causal inference due to the cross-sectional design and use of HIV
status as a surrogate for knowledge of HIV status.
HIV testing is a critical early step to accessing HIV care and treatment and is of particular
importance for high-risk populations, such as FSW. In order to be effective, interventions to
improve HIV testing need to target individuals least likely to get tested. In our analyses,
individuals with higher levels of drinking were more likely to have never been tested for
Bengtson et al.
Page 4
AIDS Care. Author manuscript; available in PMC 2015 July 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
HIV. Future interventions to improve HIV testing among FSWs should consider targeting
individuals with higher levels of alcohol use.
Acknowledgments
The authors thank Debra H. Weiner, Betsy Tolley, and Allison Prickett for their helpful comments in the
development of this manuscript.
Funding
This research was supported by the President’s Emergency Plan for AIDS Relief (PEPFAR) through the USAID
under the terms of [grant number AID-623-A-11-00007]. Ms Bengtson was supported by a fellowship through the
University of North Carolina, Chapel Hill and FHI 360. The views expressed in this document are those of the
authors and do not necessarily reflect those of FHI 360 or the funding agencies.
References
Abdelrahim MS. HIV prevalence and risk behaviors of female sex workers in Khartoum, north Sudan.
AIDS (London, England). 2010; 24(2):S55–S60.10.1097/01.aids.0000386734.79553.9a
Arasteh K, Des Jarlais DC. HIV testing and treatment among at-risk drinking injection drug users.
Journal of the International Association of Physicians in AIDS Care (Chicago, IL: 2002). 2009;
8:196–201.10.1177/1545109709336222
Asiki G, Mpendo J, Abaasa A, Agaba C, Nanvubya A, Nielsen L, Kamali A. HIV and syphilis
prevalence and associated risk factors among fishing communities of lake Victoria, Uganda.
Sexually Transmitted Infections. 2011; 87:511–515.10.1136/sti.2010.046805 [PubMed: 21835763]
Babor, TF.; Higgins-Biddle, JC. Brief intervention for hazardous and harmful drinking. A manual for
use in primary care. 2001. Retrieved July 30, 2013, from http://whqlibdoc.who.int/hq/2001/
who_msd_msb_01.6b.pdf
Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: An empirical
comparison of models that directly estimate the prevalence ratio. BMC Medical Research
Methodology. 2003; 3(1):21.10.1186/1471-2288-3-21 [PubMed: 14567763]
Braunstein SL, Ingabire CM, Geubbels E, Vyankandondera J, Umulisa MM, Gahiro E, van de Wijgert
JH. High burden of prevalent and recently acquired HIV among female sex workers and female
HIV voluntary testing center clients in Kigali, Rwanda. PloS One. 2011; 6:e24321.10.1371/
journal.pone.0024321 [PubMed: 21949704]
Chersich MF, Luchters SMF, Malonza IM, Mwarogo P, King’ola N, Temmerman M. Heavy episodic
drinking among Kenyan female sex workers is associated with unsafe sex, sexual violence and
sexually transmitted infections. International Journal of STD & AIDS. 2007; 18:764–
769.10.1258/095646207782212342 [PubMed: 18005511]
Cherutich P, Kaiser R, Galbraith J, Williamson J, Shiraishi RW, Ngare C. KAIS Study Group. Lack of
knowledge of HIV status a major barrier to HIV prevention, care and treatment efforts in Kenya:
Results from a nationally representative study. PloS One. 2012; 7:e36797.10.1371/journal.pone.
0036797 [PubMed: 22574226]
Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N. HPTN 052 Study
Team. Prevention of HIV-1 infection with early antiretroviral therapy. The New England Journal of
Medicine. 2011; 365:493–505.10.1056/NEJMoa1105243 [PubMed: 21767103]
Coldiron ME, Stephenson R, Chomba E, Vwalika C, Karita E, Kayitenkore K, Haworth A. The
relationship between alcohol consumption and unprotected sex among known HIV-discordant
couples in Rwanda and Zambia. AIDS and Behavior. 2008; 12:594–603.10.1007/
s10461-007-9304-x [PubMed: 17705032]
Delva W, Eaton JW, Meng F, Fraser C, White RG, Vickerman P, Hallett TB. HIV treatment as
prevention: Optimising the impact of expanded HIV treatment programmes. PLoS Medicine.
2012; 9:e1001258.10.1371/journal.pmed.1001258 [PubMed: 22802738]
Bengtson et al.
Page 5
AIDS Care. Author manuscript; available in PMC 2015 July 21.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript