Expanded access to antiretroviral therapy (ART) in Africa offers HIV-positive women the opportunity to live longer, resume sexual relationships, and become pregnant with substantially reduced risk of HIV transmission to HIV-negative partners and infants.1–4 Unmet need for contraception and unintended pregnancy among HIV-positive women has been described,5,6 with increased attention to fertility desire and need for safer conception.7 Recent population-based surveys suggest that 7%–40% of HIV-positive women in Africa desire more children.5,8,9 Fertility desire, without adequate counseling for safer conception, could be an important risk factor for secondary HIV transmission if women engage in more condomless sex despite having a detectable viral load (VL).
Women who are HIV positive and engage in condomless sex can greatly reduce the risk of HIV transmission through consistent adherence to ART and resultant viral suppression.10,11 Fertility desire could be associated with higher rates of viral suppression if women believe that taking ART in pregnancy will reduce their risk transmitting HIV to their sexual partner and their baby.12 Conversely, fertility desire could increase the risk of detectable VL if women interrupt HIV treatment because of concerns about medication toxicities to the fetus13 or fear of abandonment by a partner on learning her HIV status.14,15
Female sex workers (FSWs) in Africa are a key population at high risk for HIV acquisition and for HIV transmission to sexual partners.7,16 In Kenya, 14% of HIV transmission is due to sex work,17 with potentially higher rates with regular partners because of lower condom use.18 This population has a high frequency of unintended pregnancy and risk of vertical transmission.8 The use of nonbarrier methods of contraception among FSWs has been low in several studies,8,19ranging from 12% to 54%.19,20 The most common nonbarrier methods that FSWs use are hormonal injectables and pills.7,19,20 Combination prevention programs for HIV-positive FSWs, which emphasize consistent condom use and high adherence to ART to prevent onward transmission, focus on reducing their risk of transmitting HIV to sex partners.8,21 Although pregnancy prevention is emphasized, women's fertility preferences are typically not addressed.22 Despite stigma related to sex work and HIV status, many HIV-positive FSWs desire more children.22–24 Studies exploring fertility desire in relation to risk factors for HIV transmission in FSWs are lacking.23 To address this knowledge gap, we conducted a prospective cohort study of fertility desire as a risk factor for condomless sex (using prostate-specific antigen [PSA] test as a biologic marker), detectable plasma VL, and transmission potential (visits when both PSA detection and a detectable VL were present), in HIV-positive FSWs among Mombasa, Kenya.
Women were recruited for this prospective cohort study between October 2012 and December 2015 from an ongoing cohort at Ganjoni Clinic, a primary venue for FSW sexually trans- mitted infections (STIs) testing and treatment in Mombasa for more than 3 decades. Ongoing recruitment of new participants occurred through outreach at bars 2–8 times per month. Eligible women were laboratory-confirmed HIV positive, ≥18 years old, had initiated ART, and reported ever exchange of sex for cash or in-kind payment. Detailed methods for this cohort have been published.25 Women were excluded from this analysis if they were pregnant or postmenopausal. At enrollment, women completed a standardized face-to-face interview to collect sociodemographic, health, and behavioral data. A study clinician conducted a physical examination including a speculum-assisted pelvic examination to collect genital swabs for laboratory testing. Participants returned monthly for behavioral data collection. Physical examinations and genital sample collection for detection of semen and STI were repeated quarterly, and blood samples were collected for VL testing every 6 months. Participants received free medical care at the research clinic, counseling on consistent adherence to ART and condom use for health promotion and prevention of onward HIV transmission, demonstrations on correct condom use, if needed, nonbarrier contraceptive options (oral contraceptives, hormonal injections, intrauterine contraceptive devices), STI treatment, and ART according to Kenyan National Guidelines. At each visit, participants were compensated 250 Kenyan shillings (about $2.50). This study was approved by the ethics committees of Kenyatta National Hospital and the University of Washington. All participants provided written informed consent.
Fertility desire was assessed quarterly by asking a question adapted from the Kenyan Demographic and Health Survey: “Do you want any/more children?”26 Women who responded yes were asked about fertility intent by asking: “Are you trying to get pregnant now?”
This study had 2 primary outcomes. Semen was detected in vaginal secretions by PSA test, collected at quarterly examination visits (ABACard, West Hills, CA). This biomarker is most sensitive for detecting semen within 24–48 hours after condomless sex.27 Detectable plasma VL, examined every 3 months, was defined as ≥180 copies per milliliter (Hologic, Gen-Probe, San Diego, CA). This cut-point was higher than the assay lower limit of detection (<30 copies per milliliter) because some 100-mL samples had to be diluted 6-fold, to a final volume of 600 mL before testing. Because results were not available in real time, women did not know their VL status during the study.
HIV transmission potential was defined as occurring at visits when both semen detection by PSA test and detectable plasma VL were present. The relationship between fertility desire and HIV transmission potential was evaluated as an exploratory outcome.
A secondary analysis was conducted to evaluate fertility desire as a risk factor for self-reported condomless sex. Sexual behavior data were collected monthly. Any condomless sex in the past week was defined as the number of total sex acts exceeding the number of sex acts with a condom, as in our previous studies.28,29 Because this outcome was highly skewed, we created a binary variable to represent any versus no condomless less. Sensitivity analyses were performed with additional self-reported sexual behavioral outcomes (Supplemental Digital Content Table 1, http://links.lww.com/QAI/B139). In a separate secondary analysis, fertility desire was evaluated as a risk factor for poor adherence to ART. This variable was defined as >48 hours late for a scheduled refill based on research pharmacy data (“late refill,” hereafter).30
Covariate data were collected at enrollment and time updated at different intervals depending on the measure. Characteristics included age; years in sex work; early sexual debut (<15 years vs. ≥15); highest education level (<8 years vs. ≥8); workplace (bar, nightclub, home/other); number of live births; postpartum status (≤9 months since last birth); and use of any nonbarrier contraceptive method. Modern nonbarrier contraception, collected monthly, was defined as methods with a typical-use failure rate of less than 10% per year (injectable hormonal contraception, oral contraceptives, and intrauterine device) compared with condoms only or no method.31 Women were asked quarterly to identify whether they had a regular male sex partner (boyfriend or husband), who was not a client or casual partner (yes/no). Women were asked annually about that regular partner's expected reaction to a future pregnancy (excited, neutral, and upset). Exposure to controlling behaviors (yes to ≥1 of 7 acts ever) and to any intimate partner violence by the regular partner (yes to ≥1 of 13 acts past 12 months) were measured annually as in our previous studies,25,32 using items adapted from a global violence against women survey.33
Depressive symptoms in the past 2 weeks were assessed every 6 months by Patient Health Questionnaire-9 [0–4 (minimal), 5–9 (mild), 10 or higher (moderate or severe)].34 Alcohol use in the past year was assessed annually by the Alcohol Use Disorders Identification Test [(AUDIT), 0 (nondrinkers), 1–6 (minimal), 7–15 (moderate), and ≥16 (possible alcohol use disorder)].35 Disclosure of HIV status was assessed every 6 months by asking whether women had ever shared their results with someone, and if so, with whom.36
Women contributed survey and laboratory data to this analysis from October 2012 until administrative censoring (December 31, 2015). Fertility desire status was carried forward for intervening visits until the next quarterly assessment to permit analyses with outcomes and covariate data that occurred between quarterly visits.
The primary analyses tested the hypotheses that fertility desire was associated with higher risk of semen detection by PSA test and, separately, with higher risk of detectable VL.13 We also explored whether fertility desire was associated with higher risk of HIV transmission potential events.
In the analysis of fertility desire as a risk factor for semen detection by PSA test, effect modification by nonbarrier contraceptive use was evaluated based on the hypothesis that the relationship between fertility desire and condomless sex would differ depending on whether women reported using a modern nonbarrier contraceptive method. An interaction term between fertility desire and nonbarrier contraceptive use was entered into a model that included the outcome, fertility desire, and any nonbarrier contraceptive use. If the P value for the interaction term was <0.10 using a Wald test, we performed stratified analyses.
Log-binomial generalized estimating equations with independence working correlation structure and robust standard errors were used to generate relative risks (RRs) and 95% confidence intervals (CIs).37 Multivariate model building evaluated prespecified confounding factors based on previous studies.9,23,38–40 Variables that changed the relative risk estimate by >10% in univariate analyses were retained in the multivariate model.41 Conceptual diagrams of the proposed relationships between variables are provided in a supplementary digital appendix (Supplemental Digital Content Figs. 1 and 2, http://links.lww.com/QAI/B137, http://links.lww.com/QAI/B138). Age (continuous) and number of live births (continuous) were prespecified potential confounding factors included in the multivariate model.40 Additional variables were each tested for association with the outcome in a separate univariate model. These factors included sociodemographic characteristics, age of first sex, alcohol use, depressive symptoms, intimate partner violence in the past year, HIV disclosure to anyone other than a health care provider, and disclosure to a husband or boyfriend. Variables were entered into a regression model in descending order of the size of their RRs in unadjusted analyses. In the event of collinearity, the more plausible confounding factor was retained in the model. For subgroup analyses by nonbarrier modern contraceptive use, model building was performed separately in each stratum. Additional exploratory analyses were conducted using fertility intent as an exposure.
To evaluate the association between fertility desire and detectable VL in every 6 months visits with VL data, we included age and education level as prespecified confounding factors based on our conceptual diagram. Secondary analyses of fertility desire and late refill, a marker for poor adherence, used the final adjusted model from the VL analysis.
The exploratory analysis of fertility desire and HIV transmission potential included visits with both PSA and VL data, with age as the only prespecified confounding factor because age was the common confounding factor in the PSA and VL analyses. Because the combined outcome was less frequent than the individual outcomes (31/981 visits, 3.1%), these analyses used generalized estimating equation linear regression with robust standard errors to estimate the absolute risk (risk difference) of HIV transmission potential events, comparing visits with and without fertility desire.
Missing exposure and covariate data were <2%, and missing outcomes were <10% (PSA, <1%, VL data, <7%). Therefore, a complete case analysis was performed. All analyses were conducted in STATA 13.0 (College Station, TX).
Power calculations were performed separately for each primary aim, accounting for correlation of observations within women. A sample of 2000 visits for the PSA analysis was expected to provide >80% power to detect an RR ≥1.6 overall and >80% power to detect an RR ≥1.8 for subgroup analyses of PSA detection. A target of 990 visits in the VL analysis was required to provide >80% power to detect an RR ≥1.8 for detectable VL.
Overall, 255 women contributed 475 person-years to the analyses. The median number of follow-up visits per women was 19 (interquartile range 7–33). Baseline characteristics of participants are presented in Table 1. Median age was 38 years (33–42), and median time in sex work was 9 years (5–14). Most women reported having a regular emotional partner (ie, boyfriend or husband; 208/255, 82.0%). One-quarter of women reported fertility desire (65/255, 25.5%). Fertility desire was modestly higher at visits when women reported having a regular partner [370/1534 visits (24.3%) compared with 92/487 (18.9%); χ2 = 6.05, P value = 0.014]. About one-third of women reported using a modern nonbarrier contraceptive method (79/255, 30.9%).
Fertility Desire and Semen Detection by PSA Test
The analysis of fertility desire as a risk factor for semen detection by PSA test included 255 women and 2011 quarterly examination visits. Desire for more children was common (453/1985, 22.8% of visits in 90 women). Semen detection by PSA test occurred at 373/2011 (18.6%) visits in 152 women. We found evidence of effect modification by nonbarrier contraceptive use (interaction P value = 0.004), so results are presented separately in visits where women reported use of modern methods and visits when they did not (Table 2). At visits where women reported not using nonbarrier contraception, fertility desire was associated with higher risk of semen detection by PSA test (82/385, 21.3% vs. 158/1007, 15.7%; RR 1.36, 95% CI: 0.97 to 1.90). This association increased, and became statistically significant, in the model adjusted for age and number of live births [adjusted relative risk (aRR) 1.58, 95% CI: 1.12 to 2.23]. In exploratory analyses, fertility intent was also associated with higher risk of positive PSA when women reported no modern contraception [94/340 (27.7%) vs. 196/1,232 (15.9%); aRR 1.95, 95% CI: 1.32 to 2.90]. By contrast, at visits when women reported using modern nonbarrier contraception, fertility desire was associated with significantly lower risk of semen detection by PSA in unadjusted analysis (10/77, 13.0% vs. 121/536, 22.6%; RR 0.58, 95% CI: 0.36 to 0.92), and in the final model adjusting for age, number of live births, and age of first sex (aRR 0.58, 95% CI: 0.35 to 0.94) (Table 3).
In analyses of self-reported sexual behavior, fertility desire was associated with higher risk of self-reported condomless sex in the past week in the subgroup of visits where women reported not using modern nonbarrier contraception, after adjusting for age and number of live births (aRR 1.58, 95% CI: 1.12 to 2.23). Interestingly, and in contrast to results from the PSA analysis, in the subgroup of visits when women were using a nonbarrier method, fertility desire was also associated with higher risk of self-reported condomless sex, after adjusting for age, number of live births, and age of first sex (aRR 2.55, 95% CI: 1.23 to 5.73).
Fertility Desire, Detectable Viral Load, and Poor Adherence to ART
Overall, 240 women contributed 995 visits (386 person-years) to the VL analysis, and 212 women contributed 4104 visits to the adherence analysis (28 women had missing outcomes). Detectable plasma VL occurred at 159/995 (16.0%) visits in 66 women. Fertility desire was not associated with detectable VL in the unadjusted model or after adjusting for age and education level (31/219, 14.2% vs. 128/776, 16.5%; aRR 0.82, 95% CI: 0.46 to 1.44) (Table 3). Although late refills were common (815/4104 monthly visits, 19.9%), fertility desire was not associated with late refill (aRR 1.08, 95% CI: 0.82 to 1.41).
Fertility Desire and HIV Transmission Potential
HIV transmission potential events occurred at only 31/987 (3.1%) visits in 19 women. Fertility desire was not associated with significantly higher absolute risk of HIV transmission potential events in the unadjusted analysis (10/218, 4.6% vs. 21/769, 2.7%, risk difference 0.02, 95% CI −0.017 to 0.050) or in the final model adjusted for age and male partner controlling behaviors (aRD 0.011, 95% CI −0.031 to 0.050).
In this prospective study of HIV-positive FSWs enrolled in HIV treatment in Kenya, fertility desire was associated with higher risk of biological evidence of condomless sex only when women reported not using nonbarrier contraception. Fertility desire was not associated with higher risk of detectable plasma VL. Episodes of HIV transmission potential were infrequent regardless of fertility desire. These results suggest that combination prevention, focused on condom use and ART adherence to reduce transmission risk, may be effective even in FSWs who report that they desire more children.
Results from this prospective cohort study expand on findings from previous cross-sectional studies by showing a temporal relationship between reports of fertility desire and unprotected sex, using biomarker evidence of condomless sex, and examining differences in this association by nonbarrier contraceptive use. Cross-sectional studies of HIV-positive FSWs in West Africa and general population women in Uganda have found that fertility desire was associated with a higher proportion of self-reported condomless sex.23,42 However, a cohort study in general population HIV-positive women in Uganda did not find an association between fertility desire and self-reported condomless sex.40 These differing findings may be due to differences in measures and study populations.
A unique feature of this study was our finding that the association between fertility desire and condomless sex differed by nonbarrier contraceptive use. Notably, fertility desire was associated with lower risk of semen detection in visits where women were using nonbarrier methods, which may suggest that these women were actively preventing pregnancy in the near term while still desiring a future pregnancy.
In this era of combination HIV prevention, an important finding in this study was the low number of HIV transmission potential events. Although viral suppression was below the United Nations' target of 90%, it was higher than reported in previous studies in FSWs in Africa.43,44 Although the subgroup of women who were not using nonbarrier contraception and desired an additional child had increased risk of biologic evidence of condomless sex, we observed no increased risk of events suggestive of HIV transmission potential. This indicates that the combination of ART and condom use was effective at minimizing episodes of concurrent unprotected sex and detectable plasma VL.45 A more deliberate effort to provide preconception counseling for these women could help to support safer conception. This approach could emphasize confirmation of VL suppression, followed by timed condomless sex during periods of peak fertility, to minimize risk of HIV transmission and STI acquisition.8,22
In this study, we found substantial underreporting of unprotected sex compared with that of semen detection by PSA test. This finding is consistent with our previous study comparing biologic and self-reported measures of condomless sex in this cohort.46 Self-reported condomless sex measures recalled behavior and willingness to report it. Interestingly, in women not using nonbarrier contraception, both PSA and self-reported data suggested that fertility desire was associated with more condomless sex. By contrast, among women using nonbarrier contraception, PSA results suggested less condomless sex, whereas self-report suggested more. This striking difference in PSA versus self-reported behavior indicates that HIV-positive women using nonbarrier contraceptives may be more comfortable reporting condomless sex compared with women who do not use any method besides condoms.47
This study had several strengths. The longitudinal cohort design with time-updated exposure and outcome measurements permitted evaluation of temporal relationships between fertility desire and condomless sex, VL, and episodes representing HIV transmission potential. Biomarker endpoints added rigor to these findings. A standardized measure of fertility desire was used, enhancing comparability with previous studies.39,40,48 An ongoing complementary qualitative study in this cohort will help to clarify pathways linking fertility desire, contraceptive use, and sexual behavior. Finally, this study focused on HIV-positive FSWs, which could inform tailoring of combination HIV prevention programs for this key population.22
This study also had limitations. First, we evaluated fertility desire with a single question, which may not have fully captured women's fertility preferences.49 Qualitative studies in HIV-positive women reveal that women may experience varying degrees of fertility desire, depending on their relationships, HIV-status disclosure, and concern about perceived risk of HIV transmission.15,38 Second, fertility desire is a sensitive topic in HIV-positive women and is subject to social desirability bias. Fertility desire may have been over-reported because of social expectations for women to bear children,14,50,51 or underreported if some women feared being judged by the provider for wanting more children despite the risk of HIV transmission.21,38 Condomless sex may have differed by partner type. Our sexual behavior measures were not partner specific. Finally, the sample included HIV-positive FSWs receiving care at a research clinic, with median age of 38. The relationship between fertility desire and HIV transmission risk indicators may differ in FSWs who are not receiving HIV care tailored to this key population and in younger women. Specifically, FSWs desiring more children who are not on ART likely have elevated risk of HIV transmission potential.3
Fertility desire was common in this cohort of HIV-positive FSWs on ART but did not seem to increase HIV transmission risk. Even in women who desired more children and were not using contraception other than condoms, events representing HIV transmission potential were rare. This finding suggests that combination prevention emphasizing consistent condom use and ART adherence is effective in this key population. Assessment of fertility desire among HIV-positive FSWs in HIV services should be routine. Understanding women's fertility desires, whether and when they want more children in the future, would help to tailor appropriate contraceptive options and safer conception counseling. These services could help women to realize their reproductive goals while also protecting their own health and the health of their partners and future children.52,53
The authors are grateful to the study participants and our research, clinical, laboratory, outreach, and administrative staff for making this study possible.
1. Litwin LE, Makumbi FE, Gray R, et al. Impact of availability and use of ART/PMTCT services on fertility desires of previously pregnant women in rakai, Uganda: a retrospective cohort study. J Acquir Immune Defic Syndr. 2015;69:377–384.
2. Kaida A, Laher F, Strathdee SA, et al. Childbearing intentions of HIV-positive women of reproductive age in Soweto, South Africa
: the influence of expanding access to HAART in an HIV hyperendemic setting. Am J Public Health. 2011;101:350–358.
3. Cohen MS, Chen YQ, McCauley M, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
4. Day S, Graham SM, Masese LN, et al. A prospective cohort study of the effect of depot medroxyprogesterone acetate on detection of plasma and cervical HIV-1 in women initiating and continuing antiretroviral therapy. J Acquir Immune Defic Syndr. 2014;66:452–456.
5. Ngugi EW, Kim AA, Nyoka R, et al. Contraceptive practices and fertility desires among HIV-infected and uninfected women in Kenya: results from a nationally representative study. J Acquir Immune Defic Syndr. 2014;66(suppl 1):S75–S81.
6. Mak J, Birdthistle I, Church K, et al. Need, demand and missed opportunities for integrated reproductive health-HIV care in Kenya and Swaziland: evidence from household surveys. AIDS. 2013;27(suppl 1):S55–S63.
7. Schwartz S, Papworth E, Thiam-Niangoin M, et al. An urgent need for integration of family planning services into HIV care: the high burden of unplanned pregnancy, termination of pregnancy, and limited contraception use among female sex workers in Cote d'Ivoire. J Acquir Immune Defic Syndr. 2015;68(suppl 2):S91–S98.
8. Schwartz SR, Papworth E, Ky-Zerbo O, et al. Reproductive health needs of female sex workers and opportunities for enhanced prevention of mother-to-child transmission efforts in sub-Saharan Africa
. J Fam Plann Reprod Health Care. 2015;43:50–59.
9. Taulo F, Berry M, Tsui A, et al. Fertility intentions of HIV-1 infected and uninfected women in Malawi: a longitudinal study. AIDS Behav. 2009;13(suppl 1):20–27.
10. Khademi A, Anand S, Potts D. Measuring the potential impact of combination HIV prevention in sub-Saharan Africa
. Medicine (Baltimore). 2015;94:e1453.
11. Bekker LG, Johnson L, Cowan F, et al. Combination HIV prevention for female sex workers: what is the evidence? Lancet. 2015;385:72–87.
12. Sofolahan YA, Airhihenbuwa CO. Cultural expectations and reproductive desires: experiences of South African women living with HIV/AIDS (WLHA). Health Care Women Int. 2013;34:263–280.
13. Ngarina M, Tarimo EA, Naburi H, et al. Women's preferences regarding infant or maternal antiretroviral prophylaxis for prevention of mother-to-child transmission of HIV during breastfeeding and their views on Option B+ in Dar es Salaam, Tanzania. PLoS One. 2014;9:e85310.
14. Smith DJ, Mbakwem BC. Life projects and therapeutic itineraries: marriage, fertility, and antiretroviral therapy in Nigeria. AIDS. 2007;21(suppl 5):S37–S41.
15. Kisakye P, Akena WO, Kaye DK. Pregnancy decisions among HIV-positive pregnant women in Mulago Hospital, Uganda. Cult Health Sex. 2010;12:445–454.
16. Baral S, Beyrer C, Muessig K, et al. Burden of HIV among female sex workers in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12:538–549.
17. Gelmon L, Kenya P, Oguya F, et al. Kenya: HIV Prevention Response and Modes of Transmission Analysis. Nairobi, Kenya: National AIDS Control Council; 2009.
18. Voeten HA, Egesah OB, Varkevisser CM, et al. Female sex workers and unsafe sex in urban and rural Nyanza, Kenya: regular partners may contribute more to HIV transmission than clients. Trop Med Int Health. 2007;12:174–182.
19. Decker MR, Yam EA, Wirtz AL, et al. Induced abortion, contraceptive use, and dual protection among female sex workers in Moscow, Russia. Int J Gynaecol Obstet. 2013;120:27–31.
20. Sutherland EG, Alaii J, Tsui S, et al. Contraceptive needs of female sex workers in Kenya—a cross-sectional study. Eur J Contracept Reprod Health Care. 2011;16:173–182.
21. Dhana A, Luchters S, Moore L, et al. Systematic review of facility-based sexual and reproductive health services for female sex workers in Africa
. Glob Health. 2014;10:46.
22. Schwartz SR, Papworth E, Ky-Zerbo O, et al. Safer conception needs for HIV prevention among female sex workers in Burkina Faso and Togo. Infect Dis Obstet Gynecol. 2014;2014:296245.
23. Aho J, Koushik A, Rashed S. Reasons for inconsistent condom use among female sex workers: need for integrated reproductive and prevention services. World Health Popul. 2013;14:5–13.
24. Duff P, Shoveller J, Feng C, et al. Pregnancy intentions among female sex workers: recognising their rights and wants as mothers. J Fam Plann Reprod Health Care. 2015;41:102–108.
25. Wilson KS, Deya R, Masese L, et al. Prevalence and correlates of intimate partner violence in HIV-positive women engaged in transactional sex in Mombasa, Kenya. Int J STD AIDS. 2015;27:1194–1203.
26. Kenya National Bureau of Statistics, National AIDS Control Council, National AIDS/STD Control Programme, National Public Health Laboratory Services, Kenya Medical Research Institute, National Coordinating Agency for Population and Development. Kenya Demographic and Health Survey 2008–09. Calverton, MD: Macro I; 2009.
27. Hochmeister MN, Budowle B, Rudin O, et al. Evaluation of prostate-specific antigen (PSA) membrane test assays for the forensic identification of seminal fluid. J Forensic Sci. 1999;44:1057–1060.
28. McClelland RS, Richardson BA, Wanje GH, et al. Association between participant self-report and biological outcomes used to measure sexual risk behavior in human immunodeficiency virus-1-Seropositive female sex workers in Mombasa, Kenya. Sex Transm Dis. 2011;38:429–433.
29. Wilson KS, Deya R, Yuhas K, et al. A prospective cohort study of intimate partner violence and unprotected sex in HIV-positive female sex workers in Mombasa, Kenya. AIDS Behav. 2016;20:2054–2064.
30. Graham SM, Masese L, Gitau R, et al. Antiretroviral adherence and development of drug resistance are the strongest predictors of genital HIV-1 shedding among women initiating treatment. J Infect Dis. 2010;202:1538–1542.
31. Trussell J. Contraceptive failure in the United States. Contraception. 2011;83:397–404.
32. Wilson KS, Wanje G, Yuhas K, et al. A prospective study of intimate partner violence as a risk factor for detectable plasma viral load in HIV-positive women engaged in transactional sex in Mombasa, Kenya. AIDS Behav. 2016;20:2065–2077.
33. Garcia-Moreno C, Jansen HA, Ellsberg M, et al. Prevalence of intimate partner violence: findings from the WHO multi-country study on women's health and domestic violence. Lancet. 2006;368:1260–1269.
34. Omoro SA, Fann JR, Weymuller EA, et al. Swahili translation and validation of the Patient Health Questionnaire-9 depression scale in the Kenyan head and neck cancer patient population. Int J Psychiatry Med. 2006;36:367–381.
35. Saunders JB, Aasland OG, Babor TF, et al. Development of the alcohol use Disorders identification test (AUDIT): WHO Collaborative Project on early detection of persons with harmful alcohol Consumption–II. Addiction. 1993;88:791–804.
36. Peltzer K, Chao LW, Dana P. Family planning among HIV positive and negative prevention of mother to child transmission (PMTCT) clients in a resource poor setting in South Africa
. AIDS Behav. 2009;13:973–979.
37. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121–130.
38. Wekesa E, Coast E. Fertility desires among men and women living with HIV/AIDS in Nairobi slums: a mixed methods study. PLoS One. 2014;9:e106292.
39. Wagner GJ, Wanyenze R. Fertility desires and intentions and the relationship to consistent condom use and provider communication regarding childbearing among HIV clients in Uganda. ISRN Infect Dis. 2013:2013.
40. Homsy J, Bunnell R, Moore D, et al. Reproductive intentions and outcomes among women on antiretroviral therapy in rural Uganda: a prospective cohort study. PLoS One. 2009;4:e4149.
41. Maldonado G, Greenland S. Simulation study of confounder-selection strategies. Am J Epidemiol. 1993;138:923–936.
42. Wanyenze RK, Matovu JK, Kamya MR, et al. Fertility desires and unmet need for family planning among HIV infected individuals in two HIV clinics with differing models of family planning service delivery. BMC Womens Health. 2015;15:5.
43. Huet C, Ouedraogo A, Konate I, et al. Long-term virological, immunological and mortality outcomes in a cohort of HIV-infected female sex workers treated with highly active antiretroviral therapy in Africa
. BMC Public Health. 2011;11:700.
44. Konate I, Traore L, Ouedraogo A, et al. Linking HIV prevention and care for community interventions among high-risk women in Burkina Faso–the ARNS 1222 “Yerelon” cohort. J Acquir Immune Defic Syndr. 2011;57(suppl 1):S50–S54.
45. Mayer KH, Venkatesh KK. Interactions of HIV, other sexually transmitted diseases, and genital tract inflammation facilitating local pathogen transmission and acquisition. Am J Reprod Immunol. 2011;65:308–316.
46. Norwood MS, Hughes JP, Amico KR. The validity of self-reported behaviors: methods for estimating underreporting of risk behaviors. Ann Epidemiol. 2016;26:612–618 e612.
47. Surie DYK, Wilson K, Masese L, et al. Modern contraceptive use and unprotected sex in high-risk HIV-positive women in Kenya. Conference on Retroviruses and Opportunistic Infections, 13–16 February, 2017. Seattle, WA; 2017.
48. Maier M, Andia I, Emenyonu N, et al. Antiretroviral therapy is associated with increased fertility desire, but not pregnancy or live birth, among HIV+ women in an early HIV treatment program in rural Uganda. AIDS Behav. 2009;13(suppl 1):28–37.
49. Moreau C, Trussell J, Bajos N. Contraceptive paths of adolescent women undergoing an abortion in France. J Adolesc Health. 2012;50:389–394.
50. Todd CS, Stibich MA, Laher F, et al. Influence of culture on contraceptive utilization among HIV-positive women in Brazil, Kenya, and South Africa
. AIDS Behav. 2011;15:454–468.
51. Beyeza-Kashesya J, Kaharuza F, Mirembe F, et al. The dilemma of safe sex and having children: challenges facing HIV sero-discordant couples in Uganda. Afr Health Sci. 2009;9:2–12.
52. Salamander Trust. Building A Safe House on Firm Ground: Key Findings From a Global Values and Preferences Survey Regarding the Sexual and Reproductive Health and Human Rights of Women Living With HIV. Geneva, Switzertand: World Health Organization; 2014.
53. World Health Organization. Consolidated Guidelines on HIV Prevention, Diagnosis, Treatment and Care for Key Populations. Geneva, Switzerland: World Health Organization; 2014.