© Lippincott Williams & Wilkins
Migration, ethnicity and environment: HIV risk factors
for women on the sugar cane plantations of the
Toye H. Brewer‡, Julia Hasbun*, Caroline A. Ryan, Stephen E. Hawes,
Samuel Martinez†, Jorge Sanchez, Martha Butler de Lister*,
Jose Constanzo*, Jose Lopez* and King K. Holmes
Objective: To determine risk factors for HIV infection among women living in the
sugar cane plantation communities (bateyes) of a large private sugar cane company
in the Dominican Republic.
Design: Cross-sectional study of sexually active female volunteers living in the
Methods: Of 98 bateyes, 23 were randomly selected and visited by a mobile
medical unit, to interview, examine and test volunteers for seroreactivity to HIV and
Results: The 490 subjects ranged in age from 16 to 72 years (median, 37 years);
53% were born in Haiti, 36% in Dominican Republic bateyes, and 12% elsewhere in
the Dominican Republic; 58% had no formal education; and 87% had no income.
HIV seropositivity was found in 28 women (5.7%), including 8.8% of those aged
< 35 years. By logistic regression analysis, HIV infection was independently
associated with age < 35 years [odds ratio (OR), 4.5; P < 0.01), being single with
children (OR, 4.3; P < 0.01), more than one lifetime sex partners (OR, 3.4;
P = 0.06), engaging in sex during menses (OR, 3.2; P = 0.02), and self-description
as a prostitute (OR, 4.4; P = 0.05)]. For Haitian women, those coming to the
Dominican Republic alone were more likely to have HIV infection than those
coming with a male partner. Less than 4% of women reported condom use at last
Conclusions: Women in the bateyes have a much higher rate of HIV infection than
that estimated for women in the general population of Dominican Republic and a
rate comparable to that of female sex workers in the Dominican Republic. AIDS
prevention in the bateyes should address condom education and distribution as well
as employment opportunities and education for women.
© 1998 Lippincott Williams & Wilkins
AIDS 1998, 12:1879–1887
Keywords: HIV, women, development, migration, poverty, epidemiology,
From the University of Washington, Seattle, Washington, USA, *AIDSCAP, Santo Domingo, Dominican Republic, and the
†University of Connecticut, Storrs, Connecticut, USA. ‡Present address: University of Miami, Miami, Florida, USA.
Note: The results of this study were presented in part at the XI International Conference on AIDS, Vancouver, July 1996. This
study does not necessarily reflect the views or policies of USAID.
Sponsorship: Supported by NIAID Training Grant A107044; NIAID Center for AIDS Research Grant AI27757; the Fogarty
International Center IARTP FIC T22 TW00001; a grant from the Bristol-Myers Squibb Corporation; and supported in part by
the US Agency for International Development (USAID) as part of Family Health International’s AIDS Control and Prevention
(AIDSCAP) Project (623-02380A-00-4031-00).
Requests for reprints to: Dr King K. Holmes, Center for AIDS and STD, University of Washington, 1001 Broadway, #215,
Seattle, WA 98122, USA.
Date of receipt: 12 December 1997; revised: 4 June 1998; accepted: 8 July 1998.
AIDS 1998, Vol 12 No 14
Migration represents one of many social factors con-
tributing to the global AIDS pandemic .
International truck drivers and the commercial sex
workers (CSW) who service them have been a source
of HIV dissemination in East Africa [2,3]. In Europe,
HIV-2 is most common in Portugal and France, former
colonizers of the western African countries where
HIV-2 has highest prevalence . Internal seasonal
migration is a risk for HIV infection among rural
Senegalese workers [4,5], Ugandan workers , and
South African miners [7,8]. Although tourism, war and
commercial travel have all played a substantial role in
the dissemination of HIV, the migration of the young,
rural poor, both on a seasonal and long-term basis, has
probably been of greater importance in developing
countries . Several studies in Africa have linked sea-
sonal migration to increased HIV seroprevalence [4–8],
citing the social disruption caused by the migration of
young men without their partners, skewed sex ratios,
and the exchange of sex for subsistence goods by
impoverished women as the keys to this link.
Since the mid-1980s, residents of plantation-based
communities for sugar cane workers (‘bateyes’) have
had the highest rates of HIV infection in the
Dominican Republic, comparable to rates among
CSW, men who have sex with men (MSM), and pris-
oners . These high rates in bateyes have been widely
attributed to the large number of Haitian immigrants
who have historically resided in the bateyes. The
Dominican Republic occupies the eastern two-thirds
and Haiti the western one-third of the island of
Hispanola. Haiti lags behind the Dominican Republic
in nearly every social and economic index (Table 1).
The income gap between the two countries has long
encouraged mainly young Haitian men from poor rural
families to cross the border in search of income oppor-
tunities in the Dominican Republic . The two
countries are culturally distinct: the Creole-speaking
Haitians are proud of their African heritage, while the
Spanish-speaking Dominicans emphasize their Hispanic
ancestry . Since the early 1900s, sugar plantations in
the Dominican Republic have recruited Haitian men as
seasonal sugar cane workers, employment regarded as
too arduous and poorly paid for the more prosperous
Over the decades, bateyes have become communities
of year-round workers and their families, accommodat-
ing a surge in the adult male population of 20–40%
during the 5–6-month harvest when approximately
30 000 young Haitian men cross the border to cut
cane. Women, with fewer economic opportunities,
account for less than 10% of migrants. Whereas most
cane workers return to Haiti after the harvest, and
many return in subsequent years, a substantial minority
remain longer in the Dominican Republic, perhaps for
a lifetime. Because many of both sexes remain in the
Dominican Republic, batey residents include perma-
nent and seasonal Haitians, along with Dominicans of
Haitian descent, persons of mixed Dominican-Haitian
ancestry, and poor Dominicans.
As the poorest country in the northern hemisphere,
Haiti also bears a stigma as the country hit earliest and
hardest by the HIV epidemic. Careful epidemiologic
investigations by Johnson and Pape  indicated that
the first probable case of AIDS in Haiti occurred in
1978. The first Dominican case was not noted until
1983 . In both countries, most cases initially
occurred among men, and MSM probably contributed
significantly to the initiation of the epidemic (42.7% of
initial AIDS cases in the Dominican Republic versus
> 25% in Haiti) [16,17], but the epidemics have subse-
quently been fuelled by heterosexual transmission.
However, HIV seroprevalence rates have been up to
10-fold higher in Haiti than in the Dominican
Republic, affecting up to 10% of periurban slum-
dwelling women attending prenatal clinics near Port au
Prince compared with 1.25% of lower socioeconomic
status women in the Dominican Republic [18,19].
Rates among CSW in Port au Prince are nearly 50%,
whereas prostitutes in Santo Domingo and Puerto Plata
have estimated rates of infection of below 6% [16,20].
Table 2 compares reported HIV seroprevalence rates
for these and other populations in both countries.
The higher prevalence of HIV infection in Haiti has
led many in the Dominican Republic to view the HIV
epidemic in the bateyes as a ‘Haitian problem’.
However, the highest HIV seroprevalences have been
found in urban Haiti, whereas rates among rural adults
Table 1. Socioeconomic comparisons between Haiti and the
Population density per square mile
Racial mix (%)
Per capita gross domestic product (US$)
Telephones per capita
Life expectancy (years)
Physicians per capita
Infant mortality per 1000
Source: World Almanac and Book of Facts (New Jersey: Funk and
1 per 141 per 126
1 per 6083
1 per 934
Migration, women and HIV risk factors Brewer et al.
(the origin of virtually all cane workers and year-round
batey residents) were only about 2–3% up until the late
1980s (Table 2), when rates of 10% and higher were
being reported in the bateyes. The present study sug-
gests that while some Haitians are undoubtedly already
infected when they arrive in the Dominican Republic,
the social disruption inherent in a migratory cycle
fuelled by poverty, and the resulting strategies for eco-
nomic survival, are also of primary importance in the
transmission of HIV within the bateyes. This makes
increased HIV awareness and preventive measures of
the utmost importance.
Thus, bateyes represent a unique setting in which to
evaluate the determinants of HIV infection. Because of
the high rates of HIV seroprevalence previously
reported in batey residents and workers, together with
conflicting data on risk determinants, and a lack of
detailed information on sexual behavior, this study
examined the ecological determinants of HIV infection
among female batey residents, particularly those related
to patterns of migration and sexual behaviors. Because
this study was performed in the bateyes during the ‘off’
season, this analysis includes only long-term and per-
manent migrants (i.e., non-seasonal), their descendants,
and native Dominicans.
Materials and methods
Study population and study design
The company collaborating in this study encompasses
98 bateyes in the southeastern sugar belt, ranging in
population from 21 to 1700 per batey, with a total
population of about 35 000 during the harvest of 1995
(December 1994–May 1995) with approximately 7800
women over the age of 15 years. During the study,
conducted in July–August 1995, it was estimated that
there were 20–30% fewer adult men than were present
during the harvest season. During the 1995 harvest,
38% of the population was aged below 15 years, and
for those aged above 15 years the sex ratio was 1.7 : 1.
Assuming a 20% decrease in men aged above 15 years,
the sex ratio would have been about 1.4 : 1. The com-
pany provided a mobile medical unit equipped with an
examining table, portable generator, and medications.
Of 98 bateyes, 23 were randomly selected for study.
During a 6-week period the van went without advance
notice to each selected batey accompanied by two
physicians, two laboratory technicians and two inter-
viewer/health promoters. Women who appeared for
consultation were invited to participate in a study of
women’s health. Women aged at least 16 years, sexu-
ally active within the past year, not pregnant and not
currently menstruating were eligible to participate. The
study procedure was explained, and 99.3% women
gave informed consent to participate in the study and
undergo HIV testing. Typically, 20 women were
examined daily on a first-come-first-served basis.
Depending on population, bateyes were visited for up
to three consecutive days.
Risk factors and clinical assessment
Each woman was interviewed using a structured
questionnaire, delivered either in Spanish or Haitian-
Creole, to obtain data on sociodemographic characteris-
tics, medical and sexual history, partner characteristics
and knowledge and attitudes about AIDS and sexually
transmitted diseases (STD). Interviews were conducted
by two Afro-Dominican women who were born, raised
and residing in the bateyes; these women were fluent in
Spanish and Creole, and were employed by the state
public health sector as health promoters. The interview-
ers underwent 2 days of intensive training, after which
the questionnaire was revised to better reflect the
bateyes’ popular vernacular. The questionnaire was
Table 2. Reported HIV seroprevalence rates in the Dominican Republic and Haiti.
First author [reference] Group under studyYearn % Seropositive
Capellan  Bateyes, Santo Domingo
(male and female)
Dominican Republic prenatal screen
Female sex workers
Men having sex with men
Female sex workers
Gomez  1992
11.0Tabet  365
Urban slum prenatal
Healthy rural adults
(male and female)
Female sex workers
(male and female)
Urban slum prenatal
Pape  5/10
Behets  1995663 8.4
AIDS 1998, Vol 12 No 14
pretested in 20 women, revised, and then administered
to each participant in a private setting.
Each woman underwent examination of the external
genitalia, speculum and bimanual examination, and
microbiologic studies for genital infection.
Observations were made on (i) quantity and character-
istics of vaginal discharge, including pH (ColorpHast
strips, Boehringer Mannheim, Indianapolis, Indiana,
USA); (ii) appearance of endocervical discharge; (iii)
cervical characteristics including bleeding induced by
swabbing; and (iv) findings on bimanual examination.
Cervical swabs were cultured for Neisseria gonorrhoeae
and tested for Chlamydia trachomatis by enzyme
Fifteen milliliters of blood were obtained from each
participant for serologic testing for syphilis and HIV
infection. Trained laboratory technicians used a
portable generator-supplied microscope to examine
wet mounts of vaginal discharge for Trichomonas vagi-
nalis, fungal hyphal forms, and clue cells. Amine-like
odor was detected after adding a 10% KOH solution.
For isolation of N. gonorrhoeae, cervical specimens were
inoculated onto modified Thayer–Martin medium
(Baltimore Biological Laboratories; Becton Dickinson,
Cockeysville, Maryland, USA) in the mobile unit,
stored immediately in candle jars and kept at ambient
temperature in the mobile unit (~30°C) until the unit
returned from the field each day. In the laboratory,
plates were incubated at 36°C and inspected at 24 and
48 h. Colonies suspicious for N. gonorrhoeae were repli-
cated and sent by courier to the Institute for
Dermatology and Sexually Transmitted Diseases in
Santo Domingo for Gram’s stain, oxidase testing, fer-
mentation testing (ApiQuad Ferm, Analytab Products,
Plainview, New York, USA) and susceptibility testing.
Endocervical specimens for chlamydial antigen
detection were stored in iced coolers in the field until
transported to the laboratory at the end of the day for
storage at 4°C. Micro Trak EIA (Syva, Behring
Diagnostics, Inc., San Jose, California, USA) were run
weekly according to manufacturer’s instructions to
detect chlamydial antigen.
Sera reactive in the Venereal Disease Research
Laboratory (VDRL) slide test were titrated to endpoint
reactivity, and confirmed by fluorescent treponemal
antibody absorbed (FTA-Abs) test. Sera repeatedly
reactive to HIV-1 enzyme-linked immunosorbent assay
(ELISA; Pasteur Diagnostics, Marnes-la-Coquette,
France) were confirmed HIV-positive by Western blot.
Women found to have a curable STD were given free
treatment, and encouraged to send their partners for
treatment. All women received colored, illustrated
educational materials in Haitian-Creole/Spanish
concerning HIV and STD risk reduction, free condoms
and education on condom use. HIV test results were
delivered to an independent HIV social service organi-
zation, which provided notification and counseling.
Data were entered into EpiInfo version 6.0 (Centers
for Disease Control and Prevention, Atlanta, Georgia,
USA), and analyzed using EpiInfo and SAS software
(SAS Institute, Cary, North Carolina, USA). Records
with missing data for specific variables were excluded
for particular analyses, and therefore denominators
varied. Risk factors for HIV infection identified by
univariate analysis were analyzed by multiple logistic
During the 6-week period, 513 women approached the
mobile unit in 23 different bateyes, 509 agreed to par-
ticipate, and 503 completed interviews and physical
examinations from whom 490 (with completed ques-
tionnaires, physical examination and HIV serology)
were included in analyses of HIV risk factors.
Characteristics of the study population
Mean age was 36.3 years, with 75% aged between
22 and 50 years (Table 3). Fifty-four per cent of the
women were born in Haiti, 34% in the Dominican
Republic bateyes, and 12% elsewhere in the
Dominican Republic. Mean years of education was
only 1.4 years, with 58% having received no formal
education. Only 10.8% of women were considered sin-
gle, with 74.4% being in consensual unions, 14.2% in
legally recognized unions, and the remainder widowed
or separated. Only 13.4% reported independent gener-
ation of income. Although the mean number of preg-
nancies per subject was five, only 20.8% had ever had a
pelvic examination; 99 (20%) reported current contra-
ception (43 tubal ligation, 37 oral contraception, four
hormonal injections, and two each condom/intrauter-
ine device use). Nearly 20% had a history of exchang-
ing sex for money/goods, including 2.6% who defined
themselves as prostitutes. Only 11% perceived them-
selves at any risk for having an STD, although 18.3% of
the sample were VDRL- and FTA-Abs-positive and
35% had either a positive ELISA test for C. trachomatis,
positive culture for N. gonorrhoeae from endocervical
specimens, or motile T. vaginalis on microscopy. Only
3.2% reported condom use at last intercourse. When
asked, ‘Have you ever used a condom and it broke?’,
89.2% answered that they had never used a condom; of
the remainder, one said, ‘yes’ and the rest said, ‘no’.
Therefore, 10.8% at most had ever used a condom,
since some women who answered ‘no’ had probably
never used one.
Migration, women and HIV risk factors Brewer et al.
Factors associated with HIV infection
Of all women, 5.7% were HIV-seropositive, including
8.8% of those aged below 35 years, and 23% of 13
women who described themselves as prostitutes. HIV
seropositivity was found in 7.4% of women born in
Haiti, compared with 3.9% of those born in the
Dominican Republic [odds ratio (OR), 2.0; P = 0.08].
Amongst the latter, HIV seropositivity was found in
4.6% of women born in bateyes, and 1.7% of those
born elsewhere in the Dominican Republic.
Factors associated with HIV infection by univariate
analysis (P < 0.1 or OR >2 or <0.5) included age
below 35 years, coming to the Dominican Republic
without a partner, being single with children, having
more than one lifetime sex partner, engaging in sex
during menses, and self-description as a prostitute
(Table 4). Potential protective factors included being
born in the Dominican Republic (excluding bateyes)
and considering oneself Dominican. Compared with
women who migrated from Haiti with a sex partner,
those who migrated to the Dominican Republic with-
out a partner were at increased risk for HIV infection.
Factors found not to be associated with HIV infection
included history of ever returning to Haiti, receiving a
blood transfusion, history of genital ulcer disease, hav-
ing a most recent partner who is Haitian, or who cuts
cane or seasonally migrates, history of anal sex, legal
versus consensual union, informal exchange of sex for
money/goods, or having positive VDRL and FTA-Abs
on serologic testing. Although 20% of women
acknowledged experiencing religious trances, this was
not associated with HIV infection.
Multiple logistic regression analysis, adjusting for factors
identified as associated with HIV by univariate analysis,
revealed significant independent associations of HIV
seropositivity with age below 35 years, being single
with children, coming to the Dominican Republic
without a partner, engaging in sexual intercourse
during menses, having more than one lifetime sexual
partner, and self-description as a prostitute.
Although prostitution was a significant risk factor for
HIV infection, only three (11%) HIV-infected women
were self-described prostitutes. Three additional
HIV-positive women gave a history of exchange of sex
for money. Thirteen (46%) out of 28 infected women
acknowledged only one or two lifetime sex partners.
Of women from Haiti, 55% came to the Dominican
Republic with a partner and 45% without a partner,
including women who either came alone or with a
family member. Those coming without a partner were
at significantly increased risk of HIV infection.
Characteristics of Haitian women with and
without HIV infection
To assess why Haitian women who came to the
Dominican Republic with a partner had lower rates of
Table 3. Demographic and behavioral characteristics and current
evidence of sexually transmitted disease (STD) among 503 women
Mean (range) age (years)
Place of birth [n/total (%)]
Elsewhere in the Dominican Republic
Mean (SD) years of education
Single [n/total (%)]
Generating income [n/total (%)]
Head of own household [n/total (%)]
Ever had a gynecologic examination
Mean (SD) no. pregnancies
Mean (SD) age at first intercourse (years)
Condom use last intercourse [n/total (%)]
Exchanged sex for money [n/total (%)]
More than one sex partner in past month
Three or fewer lifetime sex partners
Self-perceived risk for STD [n/total (%)]
Positive for syphilis and treponemal antibody*
Positive for chlamydia, gonorrhea,
or trichomonas* [n/total (%)]
*Syphilis by Venereal Disease Research Laboratory test, Treponema
pallidum by fluorescent treponemal antibody absorbed test, and
Chlamydia trachomatis by enzyme-linked immunosorbent assay,
Neisseria gonorrhoeae by culture, or Trichomonas vaginalis by
Table 4. Univariate and multivariate analyses of risk factors associated with HIV infection.
[OR (95% CI)]
[OR (95% CI)] Risk factor
Age < 35 years
Born in Dominican Republic
Came to Dominican Republic†
Single with children
Multiple lifetime sex partners
Sex with menses
Describe self as prostitute
*Multiple logistic regression, adjusting for the other factors. †Versus reference category of born in the Dominican Republic. OR, Odds ratio;
CI, confidence interval.
AIDS 1998, Vol 12 No 14
HIV infection than women who came without a part-
ner, we further compared these two groups (Table 5).
For those who arrived without a partner, the age at the
arrival was significantly younger, a significantly higher
percentage reported sex for money, the reported num-
ber of sex partners was somewhat higher, and the mean
length of residence in the bateyes was longer. For
Haitian women aged below 35 years, the mean dura-
tion of residence in the Dominican Republic was 14
years for those with and 8.8 years for those without
HIV infection; amongst those aged below 35 years, the
prevalence of HIV seropositivity was two (7.4%) out of
27 with 0–3 years in the bateyes, five (17%) out of 29
with 4–7 years, two (9%) out of 23 with 8–11 years,
and four (40%) out of 10 with ≥ 12 years in the bateyes
(age-adjusted P = 0.05). About half of the HIV-posi-
tive women migrating from Haiti without a partner
came to the Dominican Republic in the same year they
initiated sexual intercourse or before their first sexual
experience. These data, taken together, are most con-
sistent with acquisition of HIV infection by many
Haitian women in the Dominican Republic, rather
than in Haiti prior to migration.
Overall, Haitian women were significantly less likely
than women born in the bateyes or other parts of the
Dominican Republic to report exchanging sex for
money/goods [OR, 0.5; 95% confidence interval (CI),
0.3–0.8; P < 0.001]. For women coming from Haiti
with a partner, the mean lifetime number of sex part-
ners was five, whereas the corresponding number was
eight for those coming without a partner, and 10 for
women born in the Dominican Republic (P = 0.006
comparing all Haitian women with women born in the
Characteristics of self-described prostitutes and
women exchanging sex for goods
When asked, ‘What kind of work do you do?’,
13 women spontaneously described themselves as
prostitutes. By univariate analysis, these women were
significantly more likely than other women to consider
themselves heads of their family (OR, 28.4; P ≤ 0.01),
and non-Catholic (OR, 4.0; P = 0.03), to have
increased perceived risk of STD (OR, 5.3; P = 0.01),
and a history of genital ulcer disease (OR, 5.2;
P = 0.02) or knowledge of having had a positive blood
test for syphilis (OR, 6.0; P = 0.03). They reported a
median of 65 partners during their lifetime and 24 dur-
ing the past year. For 22 other women who acknowl-
edged exchanging sex for money/goods more than
10 times during the past year but not self-described as
prostitutes, the prevalence of HIV infection was not
increased. The median lifetime number of partners
reported was 48 with a median of nine partners over
the last year. For 83 women who reported having ever
exchanged sex for money/goods at least once, the
median lifetime number of partners reported was 15,
with a median of one partner during the past year.
Women who had exchanged sex for money/goods
more than 10 times in the past year did not differ sig-
nificantly from prostitutes in other characteristics
except in more often considering themselves Catholics.
Factors associated with reported condom use
Only 16 (3.2%) women acknowledged condom use at
last intercourse, including 8% of those reporting more
than one partner in the last 3 months and 11% of single
women. Factors significantly associated with condom
use at last intercourse include age ≤ 35 years (OR, 3.1;
95% CI, 1.0–11.3; P = 0.04), single marital status (OR,
5.5; 95% CI, 1.7–17.5; P = 0.01), and self-description
as a prostitute (OR, 6.2; 95% CI, 0.85–32.3; P = 0.07).
Variations among bateyes
The HIV seroprevalence ranged from 0 to 14% in the
23 bateyes studied. Although HIV prevalence, condom
use and formal/informal exchange of sex varied consid-
erably by batey, varying rates of HIV seroprevalence
could not be ascribed to recorded differences between
bateyes, such as size, the presence of bars or self-defined
prostitutes on a given batey.
Table 5. Demographic and behavioral characteristics and HIV seroprevalence for women born in Haiti who migrated to the Dominican
Republic with or without a partner.
Born in Haiti,
(n = 116)
[OR (95% CI)]*
Born in Haiti,
came with partner
(n = 146)Characteristic
Mean ± SD age at arrival (years)
Exchanged sex for money [n/total (%)]
Describe self as prostitute [n/total (%)]
Mean lifetime no. partners
More than 10 lifetime partners [n/total (%)]
Mean residence in Dominican Republic (years)
HIV-positive [n/total (%)]
< 35 years
≥ 35 years
*Odds ratios (OR) and 95% confidence intervals (CI) for association of the characteristic in Haitian women coming to the Dominican
Republic with versus without a partner.
26.3 ± 8.9
19.8 ± 10.4
Migration, women and HIV risk factors Brewer et al.
The mean HIV seroprevalence of 5.7% in sexually
active women in these bateyes was much higher than
the overall rate amongst other women in published
reports from the Dominican Republic, and roughly
equal to that of CSW in the Dominican Republic.
While the seropositive female population is heteroge-
neous, women most likely to be infected were those
below 35 years of age, single with children, had
engaged in sex during menses, had more than one life-
time sexual partner, or were self-described prostitutes.
Amongst those migrating from Haiti, migrating with-
out a partner, and for those aged below 35 years,
duration of residence in the bateyes were risk factors
for HIV infection.
In previous reports from the bateyes, Koening et al.
 and Capellan et al.  found HIV seroprevalence
rates of 3–14% and 9.3%, respectively, on bateyes in
different areas of the country. Capellan et al.  also
found that HIV infection among Haitians was associ-
ated with prolonged residence in the Dominican
Republic, but reported that participation in voodoo-
related trances with ‘probable unconscious sexual inter-
course and exchange of blood’ was significantly
associated with HIV infection, a finding not supported
by our data, or by numerous studies carried out among
Haitians in Haiti [16,18,25]. Unlike Capellan et al. ,
we found a significant increase of HIV infection among
Our study is the first to focus on the influence of
migration history among women in the bateyes, and to
document the particular risk of migrating without a
partner. Such women were nearly six times more likely
than other migrating Haitian women to have HIV
infection. Although those migrating without a partner
might have been more likely to have HIV infection
before leaving Haiti, the evidence suggests otherwise.
Nearly half of these women came to the Dominican
Republic before 1978 or in the same year they initiated
sexual activity, or prior to initiation of sexual activity,
making infection in Haiti highly unlikely.
The influence of migrating without a partner is proba-
bly related to the dependence of single women in the
bateyes on exchange of sex for money/goods for sur-
vival. Haitian women were also more likely than
women born elsewhere in the Dominican Republic to
have Haitian male partners; the seroprevalence of HIV
in the bateyes has been found to be higher for Haitian
men than for Dominican men, with the rate for men of
mixed Dominican-Haitian ancestry being intermediate
. Women born in the bateyes were more likely
than Haitian women to give a history of exchange of
sex for money/goods, and were also more likely to
identify themselves as prostitutes. It is possible that
being a ‘daughter of the batey’ , born and raised in
an environment where the exchange of sex for money
is common, might predispose to such activity.
Haitian women, while poor, have been described as the
most economically autonomous in the Caribbean,
noted for their prominent roles in the marketplace .
Despite the relative wealth of the Dominican Republic
in comparison to Haiti, women’s autonomy is severely
limited on the bateyes due to more stereotyped roles
for women . Because agricultural work is consid-
ered men’s work in the Dominican Republic, women
have few economic opportunities on the bateyes,
which helps sustain the unequal sex ratio; they have
been described as an ‘economic burden for cane work-
ers’ [14,15]. In the ‘off’ season, we found only 13% of
women acknowledging current income-generating
activities, similar to the 15% found in 1986 . Such
activities are undoubtedly more common during the
harvest season, when increased company wages put
more money into circulation locally, making it possible
for women to offer the cane cutters a variety of infor-
mal services, including preparing meals, doing laundry,
and offering sexual services.
Recent field studies carried out in the bateyes by social
scientists suggest that women play an important role in
supplementing the meager wages men earn in the cane
fields by pursuing a variety of petty commercial activi-
ties, and that at any given time most batey women are
not engaged in sex work [13,27]. Yet these findings
may be more representative of women firmly settled on
the sugar estates than of those who have only recently
arrived. When Haitian women first enter the sugar
estates, many are in desperate poverty, cross the border
without a husband, and at times cannot immediately
establish contact with friends of family to aid them in
settling and finding work. Single women arriving in the
bateyes are not eligible to receive housing either as
individuals or in groups. Newly arriving women with-
out relatives or friends for access to housing ‘almost
always take up union with a cane worker immediately
on arrival’ . However, most of these unions are
short-lived and represent ‘an interim survival strategy’
Conversation with the respondents after the structured
questionnaire disclosed a distinction between prostitu-
tion and the exchange of sex for money/goods in the
bateyes. For example, one woman in her late thirties
had been left alone since her last partner migrated else-
where in the Dominican Republic; since she was single
her survival strategy involved regularly exchanging sex
for money/goods with three different men, yet this was
not prostitution in her mind and she could imagine no
other way for a single woman to survive in the bateyes.
Another young women said she had exchanged sex for
money on numerous occasions over the past year but
AIDS 1998, Vol 12 No 14
acknowledged only one sex partner. When questioned
about this ‘inconsistency’ she responded that she
receives money for sex only from her common-law
husband. This illustrates the common occurrence of
exchange of sex for money as the basis of some consen-
sual unions in the bateyes. Only the professional who
accepts cash for sex from any man as the principal
source of income is considered a prostitute in the
bateyes. These attitudes are in keeping with rural
Haitian sexual mores reported by anthropologists .
The fact that more than 40% of women with a history
of exchange of sex for money/goods had not done so
in the past year suggests that the exchange of sex for
money/goods is an interim survival strategy for many
As in epidemiologic studies carried out in other parts of
the world where HIV transmission is primarily hetero-
sexually transmitted , we found that many women
infected with HIV had not engaged in ‘high risk’
behavior (i.e., prostitution or exchange of sex for
money, or having more than two lifetime sex partners).
Aral et al.  have pointed out ‘a strong ecological
component to an individual’s risk for exposure to sexu-
ally transmitted infections,’ independent from one’s
own behavior. Thus, a study in Haiti of 38 HIV-posi-
tive men with HIV-positive spouses  found 43% of
the men had more than 10 sex partners in the past year,
whereas none of 136 female sex partners interviewed
had multiple sex partners. Batey women probably have
had a greater lifetime number of sex partners on aver-
age, and a higher frequency of exchanging sex for
money/goods, than most women in Haiti. For exam-
ple, Behets et al.  reported that only 0.7% of
women in a shanty town outside of Port au Prince
reported exchanging sex for money/goods.
Nonetheless, amongst a small sample of 38 men
recruited in the bateyes for HIV testing during the pre-
sent study, the mean lifetime number of female sex
partners was 28 (median, 20), far greater than the life-
time number of male partners for women in the study
(mean, eight; median, two).
This study had certain limitations as well as strengths.
First, the sample comprised women who voluntarily
approached the mobile unit, met eligibility criteria and
agreed to participate in the study. Although women
were eager to participate and take advantage of the
opportunity to receive a pelvic examination from
physicians and free HIV testing, some high-risk women
may have avoided the mobile unit, not wanting to
know their HIV status or worrying about confidential-
ity. However, few women could have known their
HIV status, since serologic testing has been nearly inac-
cessible in this setting. Second, the prevalences of HIV
varied considerably amongst the bateyes sampled, and
this was not explained. However, the random selection
of bateyes from a large sugar cane company, and the
use of mobile vans for outreach, both contributed to
making this a representative sample that may have gen-
eralizability to other bateyes. Although the rates of pay
and living conditions were somewhat better on this
estate than in the bateyes of the state-owned Consejo
Estatal de Azucar, the seasonal influx of unaccompanied
men and lack of housing and economic opportunities
for women are constants that suggest that similar condi-
tions would be found in the public sector.
The two interviewers usually established an excellent
rapport with the subjects who showed little reluctance
to openly discuss sexual histories. However, the low
level of education of the subjects was a limitation; many
women found it difficult to state their age, or estimate
their lifetime number of sex partners, the number of
times they had exchanged sex for money, and the age at
which various life events had occurred, for example.
Some women answered ‘yes’ to questions about con-
dom use until it was discovered that they were trying to
conceal their unfamiliarity with condoms.
Globally, HIV has increasingly become a disease of
poor women as nearly all countries evolve towards a
heterosexual HIV epidemic. As heterosexual spread
increases, women are at greatest risk not only because
the virus may be spread more efficiently from
male-to-female than from female-to-male in some set-
tings, but because of multiple factors that may limit
women’s ability to negotiate safe sexual behavior, such
as those described here. Short-term objectives for HIV
prevention in this population and in similar settings in
the Dominican Republic should entail education pro-
jects in Spanish and Creole for men and women on the
magnitude of the problem of HIV infection, methods
of transmission, and the use of condoms as prevention.
Condoms must be made widely available in all bateyes
at the lowest possible price. Algorithms for syndromic
management of STD, and improved access to STD ser-
vices are also essential, as is health education promoting
healthcare-seeking behavior for men and women.
Improved syndromic management of STD has been
shown to reduce heterosexual HIV transmission .
In addition, women need greater economic opportuni-
ties in the bateyes. Dominican economist Moya Pons
concluded that women must be brought into the for-
mal economic sector of the bateyes to both improve
the quality of life for all, and to utilize a large pool of
untapped labor. While the short-term objectives are of
immense immediate importance, only by tackling the
larger issues of structural barriers to housing and
employment of women in the bateyes can this problem
be meaningfully addressed.
Decosas et al.  have noted that the contribution of
migration to HIV dissemination is not as straightfor-
ward as it might seem, in that, ‘it is not the origin, or
the destination of migration, but the social disruption
Migration, women and HIV risk factors Brewer et al.
which characterizes certain types of migration which
determines vulnerability to HIV’. Although the origin
of the migrants is undoubtedly a contributing factor to
the HIV epidemic in the bateyes, unequal sex ratios,
conjugal separation, poverty and poor economic
opportunities for women cannot be ignored as impor-
tant contributors. Amelioration of these conditions, as
well as education and promotion of condoms, is imper-
ative to the control of HIV in this setting.
The authors thank Margarita Rosado de Quinones and
her staff at the Institute of Dermatology and Sexually
Transmitted Diseases in Santo Domingo for collabora-
tion with laboratory analysis of specimens collected for
this study; and Mary Catlin for organizing logistic sup-
port for the study.
1. Quinn TC: Population migration and the spread of types 1
and 2 HIV. Proc Natl Acad Sci USA 1994, 91:2407–2414.
Bwayo JJ, Omari AM, Mutere AN, et al.: Long distance truck
drivers: 1. Prevalence of STDs. East Afr Med J 1991,
Serwadda D, Wawer MJ, Musgrave SD, Sewankambo NK,
Kaplan JE, Gray RH: HIV risk factors in three geographic strata
of rural Rakai District, Uganda. AIDS 1992, 6:983–989.
Pison G, Le Guenno B, Lagarde E, Enel C, Seck C: Seasonal
migration: a risk factor for HIV infection in rural Senegal.
J Acquir Immune Defic Syndr 1993, 6:196–200.
Lagarde E, Pison G, Enel C: A study of sexual behavior change
in rural Senegal. J Acquir Immune Defic Syndr 1996,
Nunn AJ, Wagner HU, Kumali A, Kengeya-Kayondo JF, Mulder
DW: Migration and HIV-1 seroprevalence in a rural Ugandan
population. AIDS 1995, 9:503–506.
Abdool-Karim Q, Abdool-Karim SS, Singh B, Short R, Ngxongo S:
Seroprevalence of HIV infection in rural South Africa. AIDS
Jochelson K, Mothibeli M, Leger JP: Human immunodeficiency
virus and migrant labor in South Africa. Int J Health Serv 1991,
AIDSCAP: The Assessment of AIDS in the Dominican Republic: A
Report to USAID. Santo Domingo: AIDSCAP; 1993.
10. Lundahl M: The Haitian Economy: Man, Land and Markets. New
York: St Martin’s; 1983:111–150.
11. Charles C: La raza: una categoria significativa en el proceso de
insercion de los trabajadores haitianos en Republica
Dominicana. In La Cuestion Haitiana en Santo Domingo. Edited
by Lozano W. Santo Domingo: Facultad Latinoamericana de
Ciencias Sociales, Programa Republica Dominicana;
12. del Castillo J: La Inmigracion de Braceros Azucareros en la
Republica Dominicana 1900–1930. Santo Domingo: Universidad
Autonoma de Santo Domingo; 1978:54–58.
13. Martinez S: Peripheral Migrants: Haitians and Dominican Sugar
Plantations. Nashville: University of Tennessee; 1995:118–130.
14. Murphy M: Dominican Sugar Plantations: Production and Foreign
Labor Integration. New York: Prager; 1991:33–39, 106–107,
15. Moya Pons F, et al.: El Batey: Estudio Socioeconomico de los
Bateyes del CEA. Santo Domingo: Fondo para el Avance de las
Ciencias Sociales; 1986:93–101.
16. Johnson WD, Pape J: AIDS in Haiti. In AIDS: Pathogenesis and
Treatment. Edited by J. Levy. New York: Marcel Dekker;
17.Programa Control de Enfermedades Transmitidas Sexualmente y
SIEDA (Procets): Epidemiological study of HIV/AIDS. SIDA Bol
Epidemiol 1994, 6:2.
18. Boulos R, Halsey N, Holt E, et al.: HIV-1 in Haitian Women
1982–1988. The Cité Soleil/JHU AIDS Project Team. J Acquir
Immune Defic Syndr 1990, 3:721–728.
19. Gomez E, Ramirez A, Pena C, et al.: Sentinel seroprevalence
surveys for HIV-1 in the DR. VIII International Conference on
AIDS. Amsterdam, July 1992 [abstract Po-C4066].
20. Capellan M, de Moya EA, Koenig E, et al.: Desarrollo de la epi-
demia de VIH y SIDA en bateyes de distrito nacional 1990.
SIDA Bol Epidemiol 1990, 2:7.
21. Tabet SR, De Moya EA, Holmes KK, et al.: Sexual behaviors and
risk factors for HIV infection among men who have sex with
men in the Dominican Republic. AIDS 1996, 10:201–206.
22. Pape JW, Stanbuck ME, Pamphile M, et al.: Prevalence of HIV
infection and high-risk activities in Haiti. J Acquir Immune
Defic Syndr 1990, 3:995–1001.
23. Behets FM, Desmormeaux J, Joseph D, et al.: Control of sexu-
ally transmitted diseases in Haiti: results and implications of
a baseline study among pregnant women living in Cité Soleil
shantytowns. J Infect Dis 1995, 172:764–771.
24. Koenig RE, Pittaluga J, Bogart M, et al.: Prevalence of antibodies
to human immunodeficiency virus in Dominicans and
Haitians in the Dominican Republic. JAMA
25. Farmer P: AIDS and Accusation: Haiti and the Geography of
Blame. Berkeley: UC Press; 1992:220–228.
26. Mintz S: Economic role and cultural traditions. In The Black
Woman Cross Culturally. Edited by Steadt FC. Cambridge:
27. Jansen S, Nfillan C: Genero, Trabajo y Etnia en los Bateyes
Dominicanos. Santo Domingo: Instituto Tecnologico de Santo
28.Lowenthal IP: Labor, sexuality and the conjugal contract in
rural Haiti. In Haiti – Today and Tomorrow. Edited by Foster
CR, Valdman A. Lanham: University Press of America;
29. Allen S, Lindan C, Serufilira A, et al.: Human immunodeficiency
virus infection in urban Rwanda. Demographic and behav-
ioral correlates in a representative sample of childbearing
women. JAMA 1991, 266:1657–1663.
30. Aral SO, Fullilove RE, Coutinho RA, Van Den Hoek JAR:
Demographic and societal factors influencing risk behaviors.
In Research Issues in Human Behavior and Sexually Transmitted
Diseases in the AIDS Era. Edited by Wasserheit JN, Aral SO,
Holmes KK. Washington, DC: ASM; 1991:161–176.
31. Grosskurth H, Mosha F, Todd J, et al.: Impact of improved
treatment of sexually transmitted diseases on HIV infection in
rural Tanzania: randomized controlled trial. Lancet 1995,
32. Decosas J, Kane F, Anarfi JK, SodJi KD, Wagner HU: Migration
and AIDS. Lancet 1995, 346:826–828.