Study setting and population

Mombasa is the second largest city in Kenya with a population of over one million people. It is an important economic and trade hub, with a large international airport and seaport serving Kenya and neighbouring countries. It also has a long-standing tourist industry, attracting large numbers of visitors each year. The study focussed on two divisions (Kisauni and Changamwe) in Mombasa county, which were purposively selected based on previous experience working in these areas and proximity to the research NGO. The main economic activity in coastal Kisauni revolves around tourism, whilst Changamwe, encompassing the harbour, airport and the start of the Trans East African highway, is an important route for trucks and private vehicles. It is estimated that some 9,300 FSWs live and work in Mombasa[3].

Data collection and assessments

Our mixed-methods study consisted of a prospective cohort and serial embedded qualitative inquiries. Detailed outline of the methodology of the cohort study is provided elsewhere[26, 27]. In brief, trained field workers were assigned to 11 zones within Kisauni and Changamwe, and were responsible for inviting women to enrol in the study and maintaining contact with them throughout the study. Convenience sampling was used and FSWs were recruited in their homes, guesthouses and streets.

To be eligible, FSWs had to be aged 16 years or older, provide adequate locator information for tracing, be HIV negative based on laboratory testing, and not currently be pregnant as assessed by self-report and confirmed by negative pregnancy screening test. Women aged 16 or 17 years who were eligible for the study were considered to be mature minors and able to consent for participation in the study. Given the sensitive and stigmatising nature of the subject matter, next of kin, caretakers, or guardians were not asked to consent on behalf of 16- and 17-year-old participants. All participants, including those aged less than 18 years, had to be able and willing to provide written informed consent for study participation. Illiterate volunteers could provide a thumbprint witnessed by an independent person of their choosing, who was able to affirm that informed consent had taken place without coercion.

Those planning to travel or relocate from the study areas, or participating in other HIV intervention studies, were excluded from participation. Participants could voluntarily withdraw from the study at any time; site investigators could also withdraw volunteers from the study in order to protect their safety or if they were unable or unwilling to comply with study procedures. External monitoring was conducted, verifying available source documentation with study-specific clinical record forms.

Visits for screening and enrolment were followed by quarterly study visits until 12 months post-enrolment. Study visits took place at a primary health centre in Changamwe and an FSW drop-in centre in Kisauni. Informed consent, collection of baseline characteristics including demographics and STI symptoms, and testing for HIV and pregnancy was performed at screening (visit 1).

A structured questionnaire was administered at each of the five study visits to collect data on socio-demographics (baseline), sexual behaviour, relationship characteristics, pregnancy intention and family planning use. Local staff translated the English questionnaires into Swahili, which were field tested before use. Questionnaires were administered in English or Swahili depending on the preference of the study participant. At least four attempts were made to locate participants who defaulted on follow up.

Pregnancy and HIV testing was done at each quarterly visit. A single positive urine test was considered diagnostic of pregnancy. Contraceptive and risk-reduction counselling were also provided at each visit. Contraceptives, including male and female condoms, were offered free of charge. The study did not include provision of access to abortion services, or specific counselling for women who had an unintended or unwanted pregnancy. Gynaecological examination and sexually transmitted infection (STI) testing were done at baseline and 12 months, as well as when clinically indicated. STI and HIV testing and management procedures have been described elsewhere[26].

Qualitative inquiries aimed to explore FSWs’ attitudes towards and experiences of family planning, unintended pregnancy and abortion. Four focus group discussions were held with 6–10 randomly selected cohort participants, separated by age group (16–20, 20–24, 25–29 and 30–34 years, consistent with Kenyan Demographic and Health Survey methodology[28]). FSWs who had become pregnant during the course of the cohort study were purposively sampled by the study team to take part in 20 in-depth interviews. Participants were selected based on whether the research team felt they could contribute interesting perspectives. Participants were invited to partake by female peer educators who were part of the study team, and well known in the local sex work community. focus group discussions and interviews were held at the research clinic and drop-in centres based within the community. A moderator and note taker employed by ICRH and experienced in qualitative research conducted each focus group discussion, prompted by pilot-tested guides. In addition, five key informant interviews were scheduled with one clinician, one nurse, one research assistant and two community field workers. The number of key expert interviews was chosen to allow for a detailed examination of the main themes that emerged during interviews and group discussions. No repeat interviews were done, and transcripts were not shared with the participants.

The Kenyatta National Hospital Ethics and Research Committee reviewed and gave approval for the full study, including for the collection of consent from minors without guardian consent on their behalf (P199/11/2005).

Study measures

A FSW was defined as a woman reporting to have had sexual intercourse at least once in the past three months, and having received money in exchange for it as part of her livelihood in the last six months. Pregnancy intention was assessed at each visit for each type of sexual partner: emotional partners such as boyfriends and husbands; regular clients; and casual clients. Unintended pregnancy was defined as any positive pregnancy test among women who reported no intention to become pregnant in that period.

Contraceptive use was assessed and described by type. Modern method use is defined as oral contraceptives, implants, injectables, intra-uterine devices (IUD), male and female condoms (self-reported consistent use with all partners over the past 3 months), sterilization, Lactational Amenorrhoea Method, and emergency contraceptives, as per the World Health Organization definition[29]. IUDs and implants were further categorised as long-acting reversible contraceptives (LARCs). Adherence to user-dependent contraceptive methods was not assessed. The term traditional contraceptive methods was used to describe the use of withdrawal or fertility awareness methods such as the use of menstrual cycle, and symptom-based methods such as monitoring cervical mucus and body temperature. Dual method use was defined as the consistent use of condoms concurrently with another modern method of contraception. Having an abortion was defined as having had either a miscarriage or any method of pregnancy termination.

Emotional partner was defined as a person with whom the FSW has a sexual relationship and an emotional attachment and is not dependent on the exchange of money for sex every time, e.g. a boyfriend or husband. FSWs’ subjective experiences of being controlled by an emotional partner was assessed by an adaptation[30] of the Relationship Control Subscale from the Sexual Relationship Power Scale[31]. This consisted of a 12-item questionnaire with a four-point Likert scale for each item, ranging from 1 (strongly agree) to 4 (strongly disagree). Total cumulative scores were categorized as low relationship control (score 12–24), medium control (score 25–36) and high control (37–48)[27].

Alcohol used was assessed using the Alcohol Use Disorders Identification Test (AUDIT) criteria and categorized as abstinence; low-risk drinking (AUDIT score 1–7); hazardous drinking (AUDIT score of 8–15); and harmful drinking or alcohol dependence (AUDIT score of ≥16)[32].

Data management and analysis

Data were double entered by separate clerks and analysed using Stata SE 12.0 (Stata Corporation, College Station, TX, USA). Descriptive analysis of population characteristics assessed the distribution of continuous variables and the frequency distribution of categorical variables in contingency tables, stratified by current use of modern contraceptives. Associations were assessed between current modern contraceptive use, and demographic, sex work and reproductive health characteristics at cohort entry. Chi-square tests were used for analysis of categorical variables, while for continuous variables we used Student’s t test (variables with a Normal distribution) or Wilcoxon rank sum tests (non-Normal variables). Using multivariable logistic regression, we investigated whether socio-demographic or reproductive predictors were independently associated with the period prevalence of unintended pregnancy during the 12-month follow-up.

Data were collected from in-depth interviews and focus group discussions via note-taking and audio-recording. Tape recordings were transcribed and translated verbatim. Data were typed in plain text and imported into Nvivo 7 for thematic analysis. One experienced female qualitative researcher coded the data, drawing initial categories (nodes) and emerging themes from the transcripts and notes. Pseudonyms were assigned to each participant in the qualitative study, and are reported in this manuscript, preserving participants’ anonymity. Interviews were analysed thematically, following a grounded-theory approach. This process involved initial open coding of transcripts, then identification of key concepts that best summarised and explained the participants’ experiences. Key concepts were then used to guide the remainder of the analysis and writing up of the data. The conceptual framework was cross-checked by other members of the study team as the findings were being written up, to build a degree of consistency in overall interpretation.

Socio-demographic characteristics

Mean age of participants at enrolment was 25.1 years (SD = 5.2; Table 1). The majority of women (70.1%) had not attended secondary school. The average age of first sex was 17 years, and women had been sex workers for a median 4 years. Half the participants (49.8%) reported only part-time sex work. Weekly income from sex work was ≤500 KSHs ($5.65 USD) for 23.9% of women.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Socio-demographic characteristics of 400 female sex workers in a prospective cohort study in Mombasa, Kenya, stratified by use of modern contraceptives at baseline.

https://doi.org/10.1371/journal.pone.0162871.t001

Single women made up the majority of the cohort (72%), with only 2.8% married or cohabiting (11/400), and 25.3% (101/400) separated, divorced or widowed. However, 71.6% (285/398) of women reported currently having an emotional partner. Over one third (37%) of those with an emotional partner had low levels of control in their relationship. Most participants (80%) had one or more living children.

Contraceptive use and attitudes towards use

The majority (96%) of women reported having ever used contraceptives. The most common contraceptive reported was condoms (93%), followed by injectables (55%). Seven percent of women reported having ever used a LARC. At enrolment, just over half (57%; 228/399) were using modern contraception, of which 12 women (5%) reported using a LARC. Only 10% of the women reported current use of dual contraception for prevention of pregnancy and HIV/STIs. A large minority of women were inconsistently using condoms, either alone (36%) or in combination with traditional methods (9 women; 2%).

In the qualitative component of the study, participants generally expressed positive views towards family planning and described the decision to use it as “wise”, an important strategy to plan the size of one’s family and provide for one’s dependents, as well as a way to avoid community ridicule for bearing (more) children as sex workers. Participants often noted that they were the sole providers for their children, and that increasing the burden by having more children would be imprudent, as this exchange in a focus group discussion illustrates:

Dalila: “…With FP [Family Planning] you are able to take care of the children you already have. Then you are able to give yourself time or a chance to plan your life.”

Fathiya: “…It is good to plan because maybe you only have enough resources to take care of two children. Then if a pregnancy occurs, you know like us FSWs having a small child really affects our work since one can’t work then. You have to take care of the small child and hence the others suffer since you have no one to help you.”

A number of participants cited side-effects as a reason for not using contraception. Side-effects described in the focus group discussions included perceptions of "spoiling the uterus", becoming infertile, having backache, "causing itching in the private parts", expansion of growths, rapid pulse, the vagina becoming very wet or smelling badly, and having no libido, while one in-depth interview participant noted that babies that are conceived after using contraception would be born with deformities. Participants in the in-depth interviews and focus group discussions expressed fears about how these side-effects would impact on their work, and were particularly fearful of frigidity, as this dialogue in a focus group discussion illustrates:

Mercy: “…from what she has said, if you plan your family and you are a FSW, and you use either the pills, if you sleep with a client he will tell you: “you are ‘cold’ (frigid).” So on my side, I think it is good that one doesn’t use FP methods because business will dwindle.”

Awiti provided another justification: “…For example, since I am a FSW, then I get injected, then I have sex with a client, he won’t enjoy it like he would with a FSW who doesn’t use FP methods. Hence men will run away from you and you won’t have good business. You will be watery and frigid.”

One participant implicitly evoked how sex workers have a competitive advantage over wives who have lost their libido apparently due to contraception, and thus that sex workers would not want to be placed in the same position:

Esther: “…Some say when a lady uses injection family planning anakuwa baridi (laughter)… One of my clients is married and I asked him the reason for moving around with women and he told me that his wife is ‘baridi’.”

Interviewer: “Baridi to mean?”

Esther: “To mean that anytime they have sex, the wife is never turned on even if he touches her so he gets disappointed and he has been persuading her not to use it but she can't because she doesn’t want children.”

It is clear that not all participants shared the belief in the range of side-effects associated with contraception as Fathiya remarks in her focus group discussion:

“…Others do not use [contraception] because of hearsay. This is a group [FSWs] that has never used FP and they hear stories about the side-effects the FP methods have caused on them. Obviously they would fear to fall in the same trap and that is why they opt not to use them.”

Predictors of current use of modern contraceptive

FSWs reporting current use of modern contraceptives were on average older (25.9 years compared with 24.1 years, P<0.001) and more likely to have children (P<0.001), compared to those not currently using modern contraception (Table 1). About two thirds of the 120 women who had secondary or tertiary education used contraception (n = 79), compared to just over half of the 169 women who did not complete primary school (n = 88; P = 0.020). Compared to women who were using a modern contraceptive, those not doing so were more likely to report harmful or dependent alcohol use (14% vs 7%; p = 0.058), and to have an emotional partner (79% vs 66%; P = 0.005).

Qualitative inquiries revealed that control of men over sex workers’ lives and over social norms in the deeply patriarchal society of Mombasa’s urban slums cannot be overstated. Participants often noted how they could not use condoms with their emotional partners. Mukondi in her in-depth interview notes the following:

Mukondi: “…Because we had reconciled with the father of my son and we were together, so I could not tell him to use a condom.”

Interview: “What fear did you have from telling him to use a condom?”

Mukondi: “…He would have perceived me wrong.”

One participant’s partner perceived a request for condoms as an attack on his masculinity and pride:

Nafula: “…You find a man and tell him to wear a condom. Aaah, he will tell you; ‘I can't buy a sweet and eat it with its wrapper! Don’t take me like a small kid. Usinifanye mimi mshamaba’ (don’t make me look uncivilized).”

Another participant employed a variety of strategies to hide her contraception, and had to negotiate the ever-present risk of violence from her partner:

Shani: “It reached a time when he… when he sees them since he wanted me to bear him a child… I was also monitoring his behaviour that was not good…so he used to throw them [contraceptives] away whenever he found them in my handbag until I used to hide them, even if it was at my neighbour’s house so that when I came in the evening I would take and sleep.

Interviewer: “What were his behaviours?”

Shani: “Just for something minor, he would beat you… beat you…it’s like his wishes were that one of fighting…”

Zawadi also noted the risks of disclosing family planning use in her in-depth interview:

Interviewer: “…Okay. If a partner is unaware of FP use what could happen if he somehow got to know about it?”

Zawadi: “He will withdraw his support because he will see that you are using FP, so there’s no importance of him being with you.

Aluna noted how her boyfriend made her stop contraception as a precondition to continuing their relationship, and she noted her prospects of marriage served as a counter to contraception use:

Aluna: “I am not using any FP method because I got a man who is serious with marriage until he came to our home, knew my mother, so in the state of being together, he wanted me to get pregnant, but we have normally have sex once in a month because he travels a lot.”

Unintended pregnancies

At enrolment, some 82% (321/400) of sex workers reported to have one (38%) or more (42%) children (Table 1). During the 12-month follow up, a total of 103 pregnancies occurred in the 400 FSWs enrolled in the study. At enrolment, only 2 reported an intention to become pregnant with their emotional (n = 1) or casual partner (n = 1). During the 12-month follow-up, a total of 14 women reported a pregnancy intention at one or more time points. All but one of the women who intended pregnancy wanted their emotional partner to father the child. Of the 14 women wanting pregnancy, 9 became pregnant (data not shown). Therefore, unintended pregnancies occurred among a quarter of participants (94/386).

Earning a slightly higher weekly income from sex work of 1001–2000 KShs [US$10–19] was associated with a higher risk of unintended pregnancy (AOR = 2.4, 95%CI = 1.2–4.8; P = 0.017) than those earning under 500 KShs (Table 2). Having an emotional partner (AOR = 1.9, 95%CI = 1.0–3.5; P = 0.045) and being of a younger age (AOR = 3.0, 95%CI = 1.4–6.5; P = 0.008 comparing ≤24 years versus 30 years and older) were also significant independent predictors of unintended pregnancy (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Predictors of unintended pregnancy during 12-months follow-up among 386 Kenyan female sex workers who did not report an intention to become pregnant.

https://doi.org/10.1371/journal.pone.0162871.t002

In the qualitative discussions, participants constructed marriage as a source of security and financial assurance–indeed, if a sex worker was married, then an unintended pregnancy was not viewed as particularly problematic as the husband was assumed to provide. Barongo in her focus group discussion mentioned:

“FSWs are of many types. There are FSWs like me who have husbands. Therefore giving birth is not an issue, he will take care of the baby, isn’t it? So we are different, like me, I will give birth for my husband because he wants children.”

Moreover, use of only male or female condoms (AOR = 2.7, 95%CI = 1.6–4.7; P<0.001) or the use of traditional or no methods (AOR = 2.9, 95%CI = 0.8–9.9; P = 0.096) at enrolment were also predictors of unintended pregnancy during follow-up. Due to the nature of these coitus- and user-dependent contraceptives, ‘forgetfulness’ was a frequently cited reason for failure. Shani in her in-depth interview said:

“I was using [contraception], but sometimes you can forget just like the way I am explaining. You could sometimes forget and realize you have involved yourself in a sexual act”

Aluna, another sex worker noted:

“But as for me, I can say it was foolishness on my side because I was having a method to avoid pregnancy but I never gave it a follow-up you see. It was available, but I had never given it a follow-up. That is why it [unintended pregnancy] happened like so I can say to me I was a fool…Pure foolishness.”

Participants in the in-depth interviews and focus group discussions gave the following examples of incorrect use or method failure that could have led to unintended pregnancies: bursting of condoms, clients who removed condoms just before sex, contraceptive pills taken only on the days that sex takes place, and the belief that the injection's protective efficacy lasts long after the date that she has to return to the clinic. Another frequent explanation was that the pregnancy was “bad luck”–thus suggesting that it was something beyond the participant’s control and subject to chance or (mis)fortune.

Consequences of unintended pregnancy

Unintended pregnancy posed a range of risks to participants' work as this focus group discussion exchange indicates:

Alice: “To get pregnant is to ask for trouble”

Naomi: “Problems, you won’t be able to take care of the children, your means of livelihood will diminish, since by the time you give birth, all clients will run away from you, even if you have just one client, he might become malicious and stop supporting you since he knows you are an FSW and had many partners, he’s not sure the pregnancy is his. Others will laugh at you for being pregnant and others will ask you why you don’t take care of the kids you have instead of getting others…”

Zawadi also noted some negative direct effects of pregnancy which impacted on her work:

Zawadi: “…and also when pregnant, I always felt moody and like not having sex so I had to stop my sex work. Also when I had sex I bled a lot.”

Interviewer: “Mmmmm…you mean because of morning sickness?

Zawadi: “Yes, I always felt like throwing up always, so I could not serve my clients and that meant I didn’t get money.”

In a context where sex work is highly stigmatised, a visibly pregnant sex worker would attract additional social scorn and judgement. Many participants noted that pregnancy precluded generating an income. Grace in her in-depth interview stated:

Grace: “I will just have to sit idle as I will not be doing my job again”

Several participants noted that clients were offended by pregnancy "because no client will come to you when he sees your tummy bulging" (Shani). In contrast, one participant noted that pregnancy increased her attractiveness to clients: “When I was pregnant I used to get them a lot because they claimed I was very hot” (Janet).

Emotional partners were reported to not be very reliable and liable to abandon women when they found out that they were pregnant. This placed women in a particularly precarious position as use of contraception may lead to abandonment, while paradoxically, becoming pregnant may elicit a similar response. In such a context, it may be pragmatic for FSWs to not use contraception, or not attach significant value to using it consistently, and to find means to abort the pregnancy when they ascertain that their partners do not support the pregnancy sufficiently:

Faith: “[my partner] was very happy because he wanted a baby.”

Interviewer: “So he was happy?”

Faith: “Very much and hoped that it could be twins!”

Interviewer: “Was he supportive of you financially or emotionally?

Faith: “At first he was very happy but later on, it became a burden to him as in he thought twice and thought it could become a burden, especially when I started asking him for financial assistance.”

Interviewer: “Mmmm…so what kind of support did he offer you?

Faith: “Okay, he supported me in buying food, feeding the other children and also for me. But still not sufficiently. About finances I had to cater for my other needs, so in the early months I stayed in ‘ubangaizaji’ (sex work) and because I saw I wouldn’t make it, I decided to abort.”

Abortion

In the prospective cohort, reported lifetime abortion rate was 21% (84/400).

Many in-depth interview and focus group discussion participants expressed strongly negative views on abortion including interpretations of abortion as ‘a sin’, ‘rejecting God’s will for a child’, and ‘fearing disapproval or rejection’ by family and emotional partners. Some participants noted the dangers of abortion and in particular, the risk of death and the subsequent implications for remaining dependents:

Janet: “Let's say I wanted to abort, but my mother disapproved of it and told me that many people die because of abortions and you could also destroy your “kizazi” [lineage] that's when I decided to just persevere [with the pregnancy].”

Others, like Aluna, decided that abortion was worth the risk in the longer-term:

Aluna: “I think it is something…it is something I had never done before and I pray not to do it again because it disgusts me a lot (ilinisinya sana), because I wouldn't have liked to do such a thing, but I think I didn't have the ability to keep that pregnancy…and at that time a lot of thoughts about life, you see. (in-depth interview 20)

Several participants saw abortions through despite their partners’ resistance and the risk of losing their support.

Interviewer: “…and what role did your boyfriend play in facilitating the abortion or against?”

Shani:” …He was very much against me having the abortion, I told you he is poor, but I love him. He told me if I aborted I would die, but for me I told myself, I had to try my best because a baby was an added burden, so there was no way I would keep the pregnancy.”

Esther kept the abortion a secret in the fear that her partner would retaliate:

Interviewer: “Okay…and do you know which partner made you pregnant?”

Esther: “Yes.”

Interviewer: “Did you tell him?”

Esther:”No, because he would have brought chaos and violence thinking that I had aborted his baby, so I did not see any reason of telling him, but he sent me money for hospital bills.”

Participants were asked to describe methods of abortion that they had heard of. Some vividly described injections by biomedical doctors or “sucking machines”. Other strategies included inserting sharp objects (such as pipes, needles, or an umbrella spike) into a pregnant woman’s womb (usually by the woman herself), ingesting malaria drugs or an overdose of contraceptive pills, or taking”local means” or substances prepared by”local people” that included bitter herbs such as subili and neem, as Maggie described:

Interviewer: “Okay, and how do women not wanting to keep pregnancies procure abortion?”

Maggie: “Some go to the hospital, also others do it at home where they use ‘subili’, ‘muarubaini’ (neem), and others use Omo [brand of washing powder].”

Interviewer: “Mmmm…any other method?.

Maggie: “Others buy medicines from chemists and use an overdose. Others use long sharp objects like sweater needles, they insert which then causes the baby to come out.”

Interviewer: “…and what about cost?

Maggie: “At the hospital, definitely you have to pay large amounts of money depending on the age of the pregnancy. But at home there are no costs unless you bleed to death or other complications.”