Factors associated with use of long-acting reversible contraception among female sex workers operating in Gulu district, Northern Uganda – a cross-sectional study


 Background: Long-acting reversible contraception (LARC) are the most effective and reliable contraceptives for female sex workers (FSWs) and require periodic users’ involvement only at the time of application or re-application. However, information on LARC use among FSWs in Uganda is scarce. To fill this gap, we determined the prevalence of LARC use among FSWs and examined factors associated with LARC use among FSWs operating in Gulu district, Northern Uganda.Methods: Across-sectional study was conducted among 300 FSWs aged 18 years and above and operating in Gulu district. Semi-structured questionnaires were used to measure factors associated with the use of LARC: intrauterine device (IUD), Implants, and injectables. Data analyses were conducted using STATA 14.0 and restricted to 280 non-gravid adult FSWs aged 18-49 years who were not on permanent contraception method. To examine factors associated with LARC use, prevalence ratios (PR) with robust standard errors were computed using Poisson regression.Results: Among the participants, the mean age (SD, range) was 26.5 (5.9, 18 - 45) years, 53.2% never married, 66.1% reported consistent condom use independent of LARC, 58.9% had unintended pregnancy during a lifetime, 48.6% had at least one unintended pregnancy during sex work, and 37.4% had at least one induced abortion. The prevalence of LARC use was 58.6%; the majority were using Implants (48.2%), followed by injectables (42.7%), and IUDs (9.1%). Independent factors associated with LARC use included: longer duration of sex work [≥ two years] (adjusted PR=1.44, 95% CI: 1.03-2.02), higher parity [≥ two] (adjusted PR=1.13, 95% CI: 1.02-1.26), history of unintended pregnancy during sex work (adjusted PR=1.24 CI: 1.01-1.51), and being a brothel/lodge-based FSWs (adjusted PR=1.28, 95% CI: 1.01-1.63).Conclusions: There is a big gap in LARC use with only 58.6% of FSWs using LARC. LARC use was associated with longer duration of sex work, higher parity, history of sex work-related unintended pregnancy, and being a brothel/lodge-based FSW. Therefore, interventions to improve LARC use should intensively target the newly recruited FSWs, FSWs with low parity, and FSWs not based in brothels or lodges.


Background
In 2019,218 million women of reproductive age in low-and middle-income countries (LMICs) had an unmet need for modern contraception (1). Almost half (49%) of all the pregnancies among women in LMICs were unintended resulting in several negative impacts on maternal and child health (1). A metaanalysis indicates minority groups like female sex workers (FSWs) are at greater risk of unmet need for contraception and unintended pregnancy due to a high frequency of intercourse, the high number of sexual partners, and the high level of stigmatization from the community and health care workers (2).
Furthermore, the incidence of unintended pregnancy among FSWs in LMICs is as high as 27.1 per 100 person-years (2). This is due to the low uptake of reproductive health services like the long-acting reversible contraception (LARC) among FSWs in the LMICs (3).
Nevertheless, the high unmet need for modern contraception (39.0%) among women of reproductive age in Gulu district (4) is an indication that the family planning indicators may be worse among FSWs in the region. Moreover, FSWs in Uganda were noted to have high rates of unplanned pregnancies (5)(6)(7)(8) and post-abortion complications (8)(9)(10). These further suggest a big gap in LARC use among FSWs and threaten future fertility of FSWs who still desire to have children (11).
Majority of the key population programs in LMICs target HIV services with little attention and resources accorded to the sexual reproductive health of FSWs (12)(13)(14)(15). Subsequently, HIV preventive packages like condoms have seen a drastic improvement in availability and coverage. Condom use protects against both sexually transmitted infections and unintended pregnancies. However, because of its relatively lower effectiveness, inconsistent use, and breakage, condom use alone does not provide complete protection against unintended pregnancies among the FSWs (13)(14)(15)(16)(17)(18)(19). On the other hand, LARC are well known to be effective, completely reversible, more cost-effective to other reversible methods and extremely reliable methods of contraception especially for FSWs since they all require users' involvement only periodically at the time of application or re-application (20)(21)(22)(23)(24).
Despite the many circumferential evidence of suboptimal use of LARC among FSWs in Uganda, not much is known about factors that may in uence the use of LARC in this population, especially among FSWs operating in Gulu district. Closing this gap is very crucial for improving LARC uptake (25) in order to prevent unplanned pregnancies among FSWs that may results in unsafe abortion and its complication.
The current study aimed at determining factors associated with LARC use among FSWs in the Gulu district in order to generate evidence needed by the ministry of health and development partners to inform appropriate interventions to promote LARC use among FSWs.

Methods
The aim, design, and setting of the study The aim of this study was to determine the prevalence of LARC use among FSWs and examined factors associated with LARC use among FSWs operating in Gulu district, Northern Uganda. To achieved the study, aim, we conducted a cross-sectional study among non-gravid adult FSWs of reproductive age (18-49 years) who were operating in Gulu district, Northern Uganda. Gulu district is located in Northern Uganda. More than 80% of the inhabitants of Gulu districts are low-income earners who practice subsistence farming (26). The district has an estimated 1425 FSWs (27) living and working mainly within Gulu municipality. Unpublished program data showed that more than 1300 fully mapped-out FSWs receive HIV treatment or HIV preventive services from TASO Gulu.

Sample size and sampling
Cross-sectional data were collected among 300 FSWs between March and June 2020 and data analysis was restricted to 280 FSWs of reproductive age (18-49 years) who were not pregnant and not on permanent contraception [bilateral tubal ligation (BTL)] at the time of the survey. The sample size was determined using the Cochran [1963,1975] (28) formula as part of a project that studied depression among FSWs in Gulu district and presented elsewhere (29). Study participants were randomly selected from an updated database of FSWs at TASO Gulu.

Data collection
Data was collected on socio-demographic characteristics, sex work-related characteristics, obstetric history, utilization of modern contraceptives, HIV status, and condom use through face-to-face interviews using a pre-tested semi-structured questionnaire. The questionnaire was developed in English language and translated into Acholi language, the predominant local language in Gulu. The rst author and a trained female research assistant collected data in either Acholi or English language guided by the participant's literacy level and preference.

Data management and statistical analysis
Data were entered and cleaned using EPI INFO 7 and exported to STATA 14.0 for analysis. Univariate results were reported using frequencies with corresponding proportions or means with standard deviations (SD) as appropriate. To examine the associations between LARC use and independent variables, prevalence ratios (PR) were obtained using bivariate and multivariable Poisson regression with robust variances as the outcome variable was highly prevalent (>10%) since odds ratios tend to overestimate the effect size in such instances (30,31). Results were presented using unadjusted and adjusted PR respectively along with their corresponding 95% con dence intervals (CI) and p-values. From bivariate analyses, statistically signi cant independent variables (p<0.05) were checked for multicollinearity using Pearson correlation (r) analysis and variables were considered multicollinear if r ≥ 0.4. Finally, all the eligible variables were included in the multivariable analysis with a stepwise backward elimination technique until a nal "best t" model was derived. Finally, predictors from both bivariate and multivariate analyses whose CI did not include the null value (1.0) were considered statistically signi cant.

Results
Socio-demographic, obstetric, and sex work-related characteristics of participants Among the 380 selected participants, 302 were successfully traced, 300 consented, 280 were considered at the analysis stage since 20 participants were ineligible for nal analyses:17 were pregnant, two were on permanent contraception using BTL, and one was menopausal The mean age of the participants was (SD, range) 26.5 (± 5.9, 18-45) years, more than two-thirds (67.9%) were below 30 years of age, 60.7% had primary or no education, almost one-third (31.6%)of the participants had an average monthly income within the rst income quartile (<$28), 53.2% never married, 92.6% had at least one previous pregnancy, and 48.9% had at least one unintended pregnancy during sex work, 95.4% used condoms during their last sexual encounters, but only two-thirds (66.1%) reported using condoms consistently (Table 1).

LARC use among FSWs in Gulu district
Fifty-eight-point six percent (58.6%) of the study participants were using LARC. The most commonly LARC method used by the study participants was the Implants (48.2%) and IUD (9.1%) was the least commonly used LARC (Figure 1).    (32)(33)(34). The current level of use of LARC is sub-optimal yet only two-thirds of FSWs were consistently using condoms. Many women in this region avoid using LARC because of their desire to have medium to large numbers of children (35). Therefore, many FSWs tend to depend on condoms even though literature showed that within just one month, nearly half (47.5%) of FSWs experience condom failure (36) which exposes them to the risk of unintended pregnancy in case they were not on the more reliable and effective LARC. Among the users of LARC, the majority were on Implant (48.2%) and injectables (42.7%). This is in agreement with the trend in the utilisation of LARC among general population of women of reproductive age in Northern Uganda whereby Injectables and Implants were also found to be the commonest LARC methods (35).
Secondly, LARC use was 28% higher among brothel-/lodge-based FSWs compared to their counterparts who were not brothel/lodge-based. Previous HIV prevention programs targeting commercial sex workers in developing countries showed that FSWs in brothels/lodges were more likely to use LARC due to the high level of organization and easier access to health care services (37). Besides, participants who had worked longer as a sex worker (two or more years) had a 44% increase in the utilization of modern contraceptives compared to their counterparts who had been a sex worker for periods of less than two years. This nding is however contrary to observations in China and Russia indicating that longer duration of sex work did not affect the use of LARC (16,17). Whereas access to modern contraceptives is not limited in China and Russia, in Northern Uganda, access to modern contraceptives is a challenge especially for the new FSWs who also tend to be younger in age and thus face parental and societal restriction from accessing family planning methods. In fact, in this study, we found that those who had had two or more children had a higher prevalence of LARC use. Chinese and Russian FSWs could have a higher level of LARC awareness, better attitude towards LARC use, and desire of smaller numbers of children like the one child policy in China which could explain the lack of effect of sex work duration on LARC use.
We also found that LARC use was associated with a history of unintended pregnancy. This could further explain the higher use among older women. Those who had ever had unintended pregnancy during sex work were 24% more likely to use LARC compared to their colleagues who never experience unwanted pregnancy during sex work. Further analysis revealed that FSWs who experienced unintended pregnancy during sex work were more likely to terminate their pregnancies (p<0.001). A similar observation was made elsewhere (10). FSWs with unintended pregnancy during sex work and terminated their pregnancies are more likely to use LARC (16). The increased likelihood of LARC use among FSWs with induced abortion is because women who ever had induced abortion may use LARC to avoid the previous bad experience of pregnancy (38,39).
Lastly, FSWs with higher parity [≥ two] had a signi cant increase (22%) in LARC use compared to their counterparts who had one or no previous pregnancy. Similar ndings where lower parity negatively affected LARC use were reported among FSWs in Swaziland (13), India (40), and Tanzania (41). Besides, majority of women (93.4%) in Northern Uganda desire three or more children (35).Therefore, this desire to give birth to more children among the nulliparous and low parity FSWs, made them less likely to use LARC compared to their high parity counterparts. This highlights the need for further sensitization and improvement in the ease of access to LARC, especially among the newer and nulliparous FSWs.

Strengths and limitations of this study
This study provides the evidence needed by health care systems to strengthen family planning policies and practices that would increase the accessibility and uptake of LARC among FSWs. Likewise, unlike most previous studies on FSWs that used non-probability sampling methods, we utilised a simple random sampling technique to select our participants providing more generalizable ndings for this group of FSWs and those from similar settings.
The study was cross-sectional and only elicited associations but not causations between LARC use and the various factors investigated. Also, the information collected may have been in uenced by recall bias since we asked FSWs about to recall their personal experience. However, most of the information asked were for events within two weeks prior to data collection and thus the possibility for recall bias was reduced. Besides, we collected some sensitive information relating to sex work that might have been di cult to answer and could have resulted in information bias. However, the participants had friendly interview experience with the PI and the female research assistant who had developed close working relationships with the FSWs while providing HIV care, treatment and prevention services to the FSWs in region. Therefore, chances of information bias were greatly reduced

Conclusions
There is a considerable gap in the use of LARC since only less than six in ten FSWs were using LARC. LARC use was associated with: longer duration of sex work, higher parity, history of sex work-related unintended pregnancy, and being a brothel/lodge-based FSW. Therefore, interventions to improve LARC use should intensively target the younger or newly recruited FSWs while the FSWs who do not work in organized places like brothels and lodges.

Declarations
Ethics approval and consent to participate We obtained ethical clearance for this study from the Makerere University School of Public Health Higher Degrees, Research, and Ethics Committee. Each participant provided written informed consent. We maintained participants' privacy and con dentiality throughout the different processes of participant's enrollment into the study and data collection and analysis.

Consent for publication
Not applicable Availability of data and materials The datasets used and analyzed during the current study are available from the corresponding author on reasonable request to oumasimple@gmail.com.

Competing interests
The authors declare that they have no competing interests.

Funding
None.
37. Overs C. An analysis of HIV prevention programming to prevent HIV transmission during commercial sex in developing countries.  Figure 1