The prevention of unintended pregnancies has important consequences for women and their families; it serves to lengthen birth intervals and to reduce infant and maternal morbidity and mortality. Additional benefits of family planning among HIV-infected women include reductions in the number of HIV-infected infants; the number of orphans; and the familial consequences of maternal illness, death or both. Although the prevention of unintended pregnancies among HIV-infected women is recognized as a critical, cost-effective strategy in the prevention of mother-to-child transmission, it is widely underutilized [1–5].
The effectiveness of a contraceptive method to prevent pregnancy is contingent upon the method's safety, degree of protection afforded and both consistent and correct use. Most HIV-infected women can use many modern contraceptive methods safely. The WHO Medical Eligibility Criteria for Contraceptive Use (WHO MEC) do not place restrictions on the use of hormonal methods, including combined oral contraceptives (COCs) and depot medroxyprogesterone acetate (DMPA), on the basis of HIV infection; however, special consideration may be needed for women taking antiretroviral (ARV) drugs [6,7]. Intrauterine devices (IUDs) are also considered safe for HIV-infected women and women with AIDS who are clinically well on antiretroviral therapy (ART) [6,7]. Condoms can be highly effective in reducing HIV transmission  and can also reduce transmission of other sexually transmitted infections (STIs) [9,10]. Therefore, condoms are recommended for all HIV-infected women to reduce risk of HIV transmission and STI acquisition or transmission. The prevention of unintended pregnancy also depends on the continued use of contraception; when women are given their choice of contraceptive method, continuation rates are higher . To increase the use of methods that are highly effective under typical use (hormonal methods and IUDs) among HIV-infected women wishing to avoid pregnancy, it is, therefore, important to both estimate what proportion are medically eligible to use these methods and to identify other factors that may influence method choice. Although studies in general populations report that a range of factors such as method effectiveness, childbearing goals, relationship characteristics, method-related experiences, provider-related attitudes and service access and provision are associated with contraceptive choice , little is known about factors that may influence contraceptive choice among HIV-infected women.
This study examined patterns of contraceptive choice among HIV-infected women and assessed whether choice was associated with sociodemographic characteristics, reproductive history and sexual behaviors. We explored these issues using baseline data from an ongoing prospective cohort study among HIV-infected women being conducted in Russia, where the HIV epidemic continues to grow , the burden of HIV is increasing among women  and rates of child abandonment by HIV-infected women are high .
Study population and procedures
Data for this cross-sectional analysis were derived from the baseline visit of the Prospective Contraceptive Effectiveness and Safety Study (PROCESS). PROCESS is an ongoing prospective cohort study examining the safety, effectiveness and acceptability of contraceptive methods among HIV-infected women. Women are enrolled into one of four contraceptive method groups of their choice: COCs (30 μg ethinylestradiol/150 μg desogestrel) along with condoms, DMPA along with condoms, copper IUD (copper surface area 375 mm2) along with condoms, or male condoms alone. The study was approved by the institutional review boards of the Centers for Disease Control and Prevention (CDC) and of the collaborating institutions in Russia. All participants provided written informed consent.
Participants were recruited in the beginning of October 2007 from three sites that provide routine care and gynecologic/obstetric services for HIV-infected women in St. Petersburg, Russia. Prior to this study, family planning services were not routinely provided in these locations. Eligible participants were aged 16–45 years with documented confirmation of HIV infection, sexually active in the last year, not pregnant, not breastfeeding or intending to breastfeed and desired to avoid pregnancy for at least 12 months. Women were ineligible if they had a hysterectomy or were otherwise sterile, or if they had used DMPA, oral contraceptives or an IUD within the previous 3 months.
At baseline, eligible women received contraceptive counseling on all methods approved for use in Russia and counseling on the necessity of condom use for the prevention of HIV transmission and STI acquisition and transmission. After counseling, contraceptive preferences were assessed by asking women to select from 14 methods all that they would consider using if provided free of cost. Among methods they would consider, women were asked to select those that were preferred (more than one method could be chosen). Women willing to consider using one of the study methods were eligible to continue with enrolment procedures.
Next, study clinicians assessed potential participants' eligibility to use COCs, DMPA and the IUD on the basis of the WHO MEC [6,7]. Although postpartum women can safely have IUDs inserted within 48 h after delivery, it was not logistically feasible to offer this option in our study. Thus, postpartum women who had given birth within the previous 4 weeks were not considered eligible to use the IUD but could return and enroll in the IUD group 4 weeks after delivery. All women were eligible to use male condoms. If women reported latex allergy, they were given polyurethane condoms. After collecting medical histories, conducting physical and pelvic examinations, and collecting biologic specimens, study clinicians determined whether participants were eligible to use COCs, DMPA, or the IUD.
To enrol in the study, participants were asked to choose one of the study methods to use for at least 1 year from among the methods they were eligible to use. Contraceptive methods were provided free of charge; all women, including those choosing COCs, DMPA or IUDs, were given male condoms. Postpartum women were instructed to begin COCs 21 days after delivery; all other women began COCs immediately. DMPA injections were given at the baseline visit and every 3 months thereafter. IUDs were either inserted at baseline or within 3 weeks due to clinician preferences to perform insertions during or soon after menses. Women were considered enrolled when they received their chosen method. After choosing their contraceptive method, women were interviewed to collect demographic, reproductive, behavioral and HIV-related information.
Enrolment in the study is ongoing; once sample size goals are reached in any study group, enrolment in that group is discontinued. The goals are to enrol approximately 200 women in the condom alone group and approximately 400 women in hormonal methods (COCs or DMPA). Because of few women choosing the IUD, there are no sample size goals for this group; enrolment in the IUD group will continue for programmatic purposes until sample size goals are reached in the other groups. As of January 2009, the sample size goal was reached in the condom alone group and enrolment in that group was closed. In this report, in which method of choice is the main outcome of interest, we limited our analysis to those enrolled while each of the study groups remained open, so as to include all women who were able to potentially choose from all four study methods.
The main outcome in our analysis was contraceptive choice, defined as the method group participants chose at enrolment (i.e., COCs along with condoms, DMPA along with condoms, IUD along with condoms, or condoms alone). As the outcomes of interest were common in the study sample, prevalence ratios and 95% confidence intervals were estimated using multivariable Poisson regression with robust variance estimates [16,17]. The first analysis examined factors associated with choosing COCs along with condoms compared with condoms alone and with choosing DMPA along with condoms compared with condoms alone. Results were similar when using multinomial logistic regression. The second analysis examined factors associated with choosing a highly effective method (COCs, DMPA or IUD) along with condoms versus condoms alone. Highly effective methods were defined as methods with failure rates during typical use of less than 10% . Variables were included in the model on the basis of a priori decisions, their univariate association with outcomes, or both. Models were checked for overdispersion and multicollinearity. Analyses were performed with SAS version 9.1 (SAS Institute Inc., Cary, North Carolina, USA), using two-sided statistical inferences and a significance level of a P value of 0.05 or less.
Prior to enrolment closing in the condom group, 4760 HIV-infected women aged 16–45 years presented for care at the study sites and were screened for study eligibility. Of these women, 3581 were not at risk for pregnancy (were currently pregnant, sterile or not sexually active) and were thus ineligible. An additional 13 women were ineligible due to breastfeeding, 455 women were ineligible due to planning to become pregnant in the next 12 months and 260 women were ineligible for other reasons (e.g., not agreeing to study procedures and intoxication). Thus, 451 women were eligible to participate in the study of whom 435 (96%) enrolled. The median age of study participants was 26 years and the majority were married (42%) or in a nonmarital union (41%) (Table 1). Most had one sexual partner in the last year (89%) and one or no sexual partners in the last 30 days (97%). The majority of women (91%) had ever been pregnant, 62% had at least one abortion, 77% had at least one live birth and over half (53%) did not desire a future pregnancy. Most women (96%) previously used condoms, 29% used oral contraceptives, none used DMPA and 4% used copper IUDs. In the previous year, 37% of women reported always using condoms, whereas 11% reported never using condoms. No women reported being told they were having AIDS-related symptoms at baseline and 19% were currently taking ARV drugs. Of those not taking ARV drugs, 8% (n = 27) had CD4 cell counts below 200 cells/μl and were referred for care.
When asked what contraceptive methods they would consider using if made available at no cost, almost all women selected male condoms (95%) (Table 2); over half would consider using oral contraceptives (59%); 44% would consider a hormonal patch; 43% would consider DMPA; and 22% would consider the copper IUD. Almost two-thirds (65%) would consider using some hormonal method (oral contraceptives, DMPA, vaginal ring or patch). When asked to select which methods would be preferred methods if available at no cost, almost three quarters of women (74%) selected male condoms, 34% selected oral contraceptives, 20% selected DMPA and 4% selected the copper IUD. Over half of women (56%) considered a hormonal method to be a preferred method and over one-third (35%) selected a dual method preference, meaning they selected as preferred methods both condoms and a highly effective contraceptive method (i.e., condoms along with hormonal method or condoms along with IUD).
The majority of women were judged by the study clinicians to be eligible to use COCs (89%) and DMPA (94%) (Table 3). The most common reasons for ineligibility for COCs were 35 years of age or older and history of migraine headaches (3% of all women), hypertension (2%) and taking medication for tuberculosis (TB) or seizures (1%). The most common reasons for ineligibility for DMPA included severe hypertension (1%) and taking medication for TB or seizures (1%). Among all women, 63% were judged to be eligible to use the IUD; among those known to not be postpartum (n = 308), 87% were eligible to use the IUD. In addition to being postpartum (29%), the most common reasons for ineligibility for the IUD included having a uterus incompatible with IUD insertion (9%) and having AIDS and not being clinically well on ART (2%). Paralleling reported preferences, condoms alone was the most commonly chosen contraceptive method (47%), followed by COCs along with condoms (29%), DMPA along with condoms (20%), and IUD along with condoms (4%). Of those eligible to use COCs, 32% chose COCs along with condoms; of those eligible to use DMPA, 22% chose DMPA along with condoms; and, of those eligible to use the IUD, 6% chose the IUD along with condoms. Almost all women chose a method they considered to be a preferred method (n = 404, 93%).
We examined factors independently associated with choice of COCs along with condoms versus condoms alone and DMPA along with condoms versus condoms alone among women who were eligible to use both COCs and DMPA (Table 4). Women more likely to choose COCs along with condoms than condoms alone included those who desired (prevalence ratio = 2.1) or who were uncertain if they desired (prevalence ratio = 1.6) a future pregnancy, those who previously used oral contraceptives (prevalence ratio = 1.5) and those who reported never (prevalence ratio = 3.7) or sometimes (prevalence ratio = 2.4) using condoms in the last year. Significant predictors of choosing DMPA along with condoms compared with condoms alone included having two or more live births (prevalence ratio =1.6), postpartum enrolment (prevalence ratio = 3.1) and recent injection drug use (prevalence ratio = 1.6).
Over half (55%) of women eligible to use a highly effective method (COCs, DMPA, or IUD) along with condoms chose one of these methods (Table 5). This proportion was over 80% in those who reported never using condoms over the last year (88%), in those who recently used injection drugs (82%) and in postpartum women (85%). Independent predictors of choosing a highly effective contraceptive method along with condoms versus condoms alone included having two or more live births (prevalence ratio = 1.4), postpartum enrolment (prevalence ratio = 1.3), desiring (prevalence ratio = 1.4), or uncertainty about desiring (prevalence ratio = 1.3) a future pregnancy, prior oral contraceptive use (prevalence ratio = 1.3), recent injection drug use (prevalence ratio = 1.3) and never (prevalence ratio = 2.3) or sometimes (prevalence ratio = 1.9) using condoms in the last year.
The prevention of unintended pregnancies among HIV-infected women is highly dependent on use of safe and effective contraceptive methods. In this study, which integrated family planning services into HIV clinical care settings in St. Petersburg, Russia, the majority of HIV-infected women were judged to be eligible to use contraceptive methods that are highly effective under typical use (COCs, DMPA and IUD) and over half of women chose to use one of these methods in combination with male condoms.
Some of our findings regarding contraceptive preferences and choice among HIV-infected women reflect contraceptive use patterns among Russian women in general. A survey of women at risk for unintended pregnancy in Russia found that at most recent intercourse, condoms were the most frequently used method (42%), with fewer women using oral contraceptives (12%) or IUDs (9%) . However, in contrast to the finding from this survey that no women used other hormonal methods (injectable, implant, patch or ring) , we found that a number of HIV-infected women preferred these methods and one-fifth chose to use DMPA. Consistent with our findings, studies of HIV-infected women in the United States  and France  have also noted that barrier methods are the most commonly used method, hormonal methods are less commonly used and IUD use is rare. In contrast to the low (<5%) proportion of women in our study who would prefer tubal sterilization, U.S. studies [22,23] have reported high rates (>25%) of tubal sterilization among HIV-infected women. The low preference for tubal sterilization in our study likely reflects Russian policy limitations based on age and parity as to who can undergo the procedure. Although tubal sterilization may be an appropriate option for HIV-infected women, some may experience poststerilization regret, particularly if the procedure is performed at a young age and some may undergo expensive attempts at reversal. The IUD is similar to tubal sterilization in that it is highly effective, long lasting and user independent, but unlike tubal sterilization, it is easily reversible. Future studies may wish to examine reasons for the low preference for the IUD among HIV-infected women.
We found that women most likely to choose a contraceptive method that is highly effective during typical use included those who reported never or sometimes using condoms in the last year. Women who used condoms inconsistently, whether because they or their partner elected not to use them, were likely to choose a highly effective female-controlled method to improve protection against pregnancy. As condoms, as typically used, are not as effective in preventing pregnancy as are hormonal methods or IUDs , it is preferable for HIV-infected women wishing to avoid pregnancy to consider using a more effective method of contraception in combination with condoms to provide dual protection against both unintended pregnancy and STI/HIV transmission/acquisition. Nonetheless, if condoms are used consistently and correctly, they can be an effective method to prevent pregnancy, with an estimated failure rate of 2% . We observed that women who reported always using condoms in the last year were more likely to choose to continue using condoms alone for pregnancy prevention. For these women, it is possible that condoms alone could be an effective contraceptive method if they continue with consistent and correct use. In general, however, reported condom use in the study sample prior to enrolment was low. Using follow-up data, we will be able to evaluate frequency of condom use among all study groups and to assess whether use is lower among women using highly effective contraceptive methods, as previously reported in a study of HIV-infected women .
Postpartum women, who are at risk of short interpregnancy intervals and related adverse effects on maternal and infant health [25,26], were more likely than nonpostpartum women to select a highly effective method, particularly DMPA. The demand for effective contraception may be greater soon after delivery; many postpartum women may have chosen DMPA because it could be started immediately. COCs could not be started until 21 days after delivery and providers were not able to offer immediate postpartum insertion of IUDs in this study; however, the feasibility and acceptability of offering postpartum IUDs may be an important area for future research.
Several other factors were also associated with choice of a highly effective contraceptive method. Women who desired or who were uncertain about desiring future pregnancy were more likely than their counterparts to choose a highly effective method. Although somewhat paradoxical, given all study participants wished to avoid pregnancy for at least 12 months, this may reflect a greater desire among women with future pregnancy intentions to plan their future pregnancies. Women who previously used oral contraceptives were more likely than those who had not to choose a highly effective method, particularly COCs. Although we do not have data on satisfaction with previously used contraceptive methods, this association is likely to reflect a preference among these women for hormonal methods. Nonetheless, almost 30% of women who previously used oral contraceptives and were eligible to use both COCs and DMPA chose condoms alone, suggesting possible displeasure with COCs based on prior experiences. It may be useful to inquire about previous method experiences as part of contraceptive counseling in order to address any related concerns.
Factors other than individual characteristics, such as contraceptive costs and provider practices, undoubtedly influence contraceptive choice. We attempted to minimize the effects of both of these factors by providing methods free of cost and conducting standardized training for providers on contraceptive counseling. However, we cannot rule out the possibility that providers had some influence on choice. We were unable to control for such possible influence, as certain characteristics were highly correlated with study site, for example, almost all postpartum women in the sample (98%) were seen by one provider at one site. Nonetheless, given the high correlation observed between contraceptive preference and choice, the influence of providers on choice was likely minimal.
Given the cross-sectional nature of our analysis, we could not examine adherence to contraceptive methods; we expect to examine this in the future using prospectively collected data. Many of the factors we examined such as reproductive history, sexual history and drug use were based on self-report. Although there may be some misclassification due to inaccurate recall or reporting, we have no reason to believe this would differ by method choice. The generalizability of our findings to other populations of HIV-infected women is unclear. Women enrolled in our study were relatively healthy, without signs of advanced disease. It is possible that HIV-infected women with more advanced disease or in other settings may have more contraindications to highly effective contraceptive methods and may have different contraceptive preferences. Finally, we examined contraceptive choice within the context of offering only four contraceptive methods; nonetheless, these methods represent some of the most commonly used methods worldwide.
In this study of HIV-infected women, over half who were eligible to use methods that are highly effective under typical use (COCs, DMPA or copper IUD) chose to use one of these methods in combination with condoms. Reassuringly, several characteristics that may place women at greater risk for unintended pregnancy and its adverse consequences were associated with choice of highly effective contraceptive methods along with condoms; these characteristics include having two or more live births, being postpartum, recently using injection drugs and using condoms inconsistently during the last year. Our findings may aid in the development of targeted interventions to increase the uptake of effective contraception among HIV-infected women. For example, as the postpartum period may be a time during which women are particularly receptive to contraceptive uptake , this period may be an optimal consideration for targeted interventions. Additionally, our findings regarding method preferences may be useful in determining the best possible contraceptive method mix consistent with HIV-infected women's preferences and the goal to increase utilization of highly effective methods. Ultimately, to successfully meet the contraceptive needs of HIV-infected women, additional factors influencing choice, adherence, or both need to be addressed, including elimination of barriers to cost and access, examination of possible side effects, interaction with other medications, influence of partner involvement and service delivery strategies that maximize prevention of both pregnancy and infection.
This study was funded by United States Agency for International Development, Russia through an Interagency Agreement with the CDC.
The authors thank the study participants as well as past and present members of the PROCESS study team for their important contributions to this study.
M.W. performed the analysis and drafted the manuscript; M.W., D.K., K.C., P.M., D.J., N.R., and S.H. contributed to study concept and design; S.A., N.A., and M.M. were responsible for clinical care of patients at the study centers, data collection, or both; all authors reviewed, edited, and approved the manuscript.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.
Conflicts of interest: None.
1. Reynolds HW, Janowitz B, Homan R, Johnson L. The value of contraception to prevent perinatal HIV transmission. Sex Transm Dis 2006; 33:350–356.
2. Reynolds HW, Janowitz B, Wilcher R, Cates W. Contraception to prevent HIV-positive births: current contribution and potential cost savings in PEPFAR countries. Sex Transm Infect 2008; 84(Suppl 2):ii49–ii53.
3. Sweat MD, O'Reilly KR, Schmid GP, Denison J, de Z I. Cost-effectiveness of nevirapine to prevent mother-to-child HIV transmission in eight African countries. AIDS 2004; 18:1661–1671.
4. Wilcher R, Petruney T, Reynolds HW, Cates W. From effectiveness to impact: contraception as an HIV prevention intervention. Sex Transm Infect 2008; 84(Suppl 2):ii54–ii60.
5. Bongaarts J, Sinding S. A response to critics of family planning programs. Int Perspect Sex Reprod Health 2009; 35:39–44.
6. WHO. World Health Organization medical eligibility criteria for contraceptive use
(3rd ed.). Geneva: WHO; 2004.
8. Weller S, Davis K. Condom effectiveness in reducing heterosexual HIV transmission. Cochrane Database Syst Rev
9. Manhart LE, Koutsky LA. Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? A meta-analysis. Sex Transm Dis 2002; 29:725–735.
10. Warner L, Stone KM, Macaluso M, Buehler JW, Austin HD. Condom use and risk of gonorrhea and Chlamydia
: a systematic review of design and measurement factors assessed in epidemiologic studies. Sex Transm Dis 2006; 33:36–51.
11. Pariani S, Heer DM, Van AM Jr. Does choice make a difference to contraceptive use? Evidence from east Java. Stud Fam Plann 1991; 22:384–390.
12. Frost JJ, Darroch JE. Factors associated with contraceptive choice and inconsistent method use, United States, 2004. Perspect Sex Reprod Health 2008; 40:94–104.
13. UNAIDS. Report on the global HIV/AIDS epidemic
14. Twigg J. HIV/AIDS in Russia: commitment, resources, momentum, challenges. Washington, District of Columbia: Center for Strategic and International Studies; 2007.
15. Khaldeeva N, Hillis SD, Vinogradova E, Voronin E, Rakhmanova A, Yakovlev A, et al
. HIV-1 seroprevalence rates in women and relinquishment of infants to the state in St Petersburg, Russia, 2002. Lancet 2003; 362:1981–1982.
16. Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005; 162:199–200.
17. Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159:702–706.
18. Trussell J. Contraceptive failure in the United States. Contraception 2004; 70:89–96.
19. Regushevskaya E, Dubikaytis T, Nikula M, Kuznetsova O, Hemminki E. Contraceptive use and abortion among women of reproductive age in St. Petersburg, Russia. Perspect Sex Reprod Health 2009; 41:51–58.
20. Massad LS, Evans CT, Wilson TE, Golub ET, Sanchez-Keeland L, Minkoff H, et al
. Contraceptive use among U.S. women with HIV. J Womens Health 2007; 16:657–666.
21. De Vincenzi I, Jadand C, Couturier E, Brunet JB, Gallais H, Gastaut JA, et al
. Pregnancy and contraception in a French cohort of HIV-infected women. SEROCO Study Group. AIDS 1997; 11:333–338.
22. Lindsay MK, Grant J, Peterson HB, Willis S, Nelson P, Klein L. The impact of knowledge of human immunodeficiency virus serostatus on contraceptive choice and repeat pregnancy. Obstet Gynecol 1995; 85:675–679.
23. Stanwood NL, Cohn SE, Heiser JR, Pugliese M. Contraception and fertility plans in a cohort of HIV-positive women in care. Contraception 2007; 75:294–298.
24. Diaz T, Schable B, Chu SY. Relationship between use of condoms and other forms of contraception among human immunodeficiency virus-infected women. Supplement to HIV and AIDS Surveillance Project Group. Obstet Gynecol 1995; 86:277–282.
25. Zhu BP. Effect of interpregnancy interval on birth outcomes: findings from three recent US studies. Int J Gynaecol Obstet 2005; 89(Suppl 1):S25–S33.
26. Conde-Agudelo A, Rosas-Bermudez A, Kafury-Goeta AC. Effects of birth spacing on maternal health: a systematic review. Am J Obstet Gynecol 2007; 196:297–308.
27. Cwiak C, Gellasch T, Zieman M. Peripartum contraceptive attitudes and practices. Contraception 2004; 70:383–386.