University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Contraception among HIV concordant and discordant couples in Zambia: a randomized controlled trial
Global Health Sciences Literature Digest
Published January 24, 2008
Journal Article

Mark KE, Meinzen-Derr J, Stephenson R, Haworth A, Ahmed Y, Duncan D et al. Contraception among HIV concordant and discordant couples in Zambia: a randomized controlled trial. J Women's Health (Larchmt). 2007 Oct;16(8):1200-10.


To evaluate the impact on non-barrier contraceptive use and pregnancy of an intervention to provide family planning services and improved contraceptive access among HIV concordant and discordant couples already provided with condoms for HIV prevention.

Study Design

A randomized, controlled trial with two intervention and one control arm.


The study took place in Lusaka, Zambia at the Zambia-Emory HIV Research Project (ZEHRP) counseling and testing center. Zambia has high HIV prevalence and its total fertility rate of 5.9 births per woman is among the highest in the world. The country's contraceptive prevalence rate is relatively low, with only 23% of women of reproductive age reporting using a modern contraceptive method, and 27% reporting an unmet need for family planning.


Couples participating in studies at the ZEHRP counseling and testing center and visiting the center for a study follow-up visit from September 9,1996, through August 31,1997 were eligibility screened for this trial with a urine pregnancy test and the administration of a brief questionnaire during which the husband and wife were interviewed separately. Couples were ineligible if the woman was >38 years old or the man was >60 years old, if the woman was pregnant, <1 month postpartum, or currently using an intrauterine contraceptive device (IUD) or hormonal method of birth control, or if the couple was considered infertile. Infertility was defined as prior surgical sterilization, self-report of infertility on a prior medical questionnaire, or a 24-month period in the last 5 years without using an IUD, hormonal method of birth control, or breastfeeding without a resultant pregnancy. Couples ineligible for a nonpermanent reason (i.e., pregnancy or <1 month postpartum) could be rescreened for eligibility at subsequent study visits. Of 608 screened, 251 couples were randomized.


Men and women separately completed identical questionnaires about fertility decision making and knowledge of, attitudes about, and prior use of contraceptives. A trained nurse or clinical officer then gave all groups a 20-30-minute educational talk on contraceptive methods and the concept of family planning, risks of unprotected sex, and vertical transmission of HIV. Using both barrier and non-barrier contraceptive methods was emphasized. The control group then received a 20-30- minute talk on STDs to control for greater time spent in the clinic by the other two groups and was referred to the Lusaka Planned Parenthood Association of Zambia (PPAZ) clinic for family planning methods. After the family planning presentation, the intervention 1 group was offered non-barrier contraceptives free of charge at the ZEHRP clinic. Non-barrier methods included oral contraceptives (OCs) (combined estrogen-progestogen pills and progestogen-only pills), injectables (Noristerat or Depo-Provera), and IUDs. The intervention 2 group received the same talks and services as the intervention 1 group, with the addition of participation in a group presentation about Zambian inheritance law and will preparation, as it was hypothesized that couples planning for the death of one or both partners might make an informed choice to limit the size of their families. It was also hypothesized that instructing women on statutory inheritance laws might free them from spousal and familial pressure to conceive. Condoms and spermicides were provided to all couples in all three arms but were not included in the non-barrier contraception outcome measure based on prior data in discordant couples showing no association between condom use and pregnancy. Women in either intervention group who desired Norplant or surgical sterilization were referred to the University Teaching Hospital, and transport and service fees were paid. The PPAZ clinic (control group) provided the same methods as the research center clinic, with the addition of Norplant insertions on site. Staff training and family planning service provision was standardized between the two sites. Follow-up data collected every 3 months through November 30, 1998, included urine pregnancy testing and information on current contraceptive use. Self-reported contraceptive data for user-independent methods (injectables, Norplant, IUDs, and surgical sterilization) were verified with PPAZ and research center charts, but compliance with OCs was not assessed. Women who missed a follow-up visit received a hand-delivered reminder letter and subsequently a nurse home visit to encourage follow-up or ascertain the reason for lack of follow-up. These procedures were discontinued after two consecutive missed follow-up visits.

Primary Outcomes

Primary outcomes were initiation of a non-barrier contraceptive within 3 months of randomization and incident pregnancy.


Couples randomized to all three groups were similar in age, educational level, prior contraceptive use and knowledge, fertility desires, current number of children, baseline condom use, and HIV status. Follow-up rates and death rates were not significantly different among the three groups. Fifty-one (19%) women had used a hormonal contraceptive or IUD before study entry; OC had been used by 45, injectables by 4, and an IUD by 2 women. The majority of enrolled couples (92%) were HIV discordant (124 M+/F-, 107 F+/M-), but 16 (6%) were concordant positive couples and 4 (2%) were concordant negative couples. Approximately half the couples eventually wanted to have more children. Fertility decisions were most often perceived as being made by the husband alone (44% of men and 54% of women) or the couple together (46% of men and 36% of women) and only very rarely by the wife alone (3% of men and 5% of women). Baseline knowledge of contraceptive methods, particularly condoms and OCs, was high among couples in all arms of the RCT. Within 3 months of the intervention, 33% (27) of control couples, 80% (89) of intervention 1 couples, and 76% (44) of intervention 2 couples had adopted a non-barrier contraceptive method (p<0.001 for comparison of each intervention group with the control group, but p=0.52 comparing intervention group 1 to group 2.) In each arm, OCs and injectables were the most commonly chosen non-barrier contraceptive method.

At 12 months, pregnancies had occurred in 22% of control, 22% of intervention 1, and 16% of intervention 2 group couples. Survival analysis for time to pregnancy showed a slight trend toward decreasing pregnancy incidence across the three arms, but the difference between the control group and the intervention groups did not achieve significance (p=0.24) nor did it differ significantly by HIV status. Contraceptive switching, stopping, and user failure all contributed to pregnancies among women who initiated non-barrier contraception. Of 83 women who initiated OC use at entry, only 30 (36%) were still using OC without a pregnancy at the end of their follow-up period. Of 73 baseline injectable users, only 28 (38%) were still using injections without a pregnancy at the end of the follow- up period. The pregnancy rate in women who did not initiate a non-barrier method use at enrollment was 43% (38 of 91). Compared with nonusers, pregnancy incidence was significantly lower among women who initiated injectable contraception at baseline (p=0.02), but not among baseline OC users. Stratifying the analysis by HIV status showed that among HIV-negative women, the choice of baseline contraceptive method was not significantly associated with incident pregnancy (p=0.37); however, among HIV-positive women, pregnancy incidence was significantly lower among those who adopted injectable contraceptives (p=0.02). Additionally, rates of injectable discontinuation were lower among HIV-positive women: 27% of HIV positive women who adopted injectable contraceptives discontinued use compared with 39% of HIV-negative users. Self-reported perfect (100%) condom use during all coital acts in each 3-month follow-up interval did not differ by study arm or non-barrier method: 63% of control group follow-up intervals, 62% of intervention 1 group intervals, and 58% of intervention 2 group intervals. Similarly, perfect condom use was reported in 61% of intervals during which OCs were used and 59% of intervals when injectables were used.


Use of injectables lowered pregnancy rates but use of oral contraceptives did not lower rates compared to using condoms alone. The addition of a session on inheritance law did not lower pregnancy rates. The authors conclude that reducing provider bias against injectables and other barriers to injectable use might allow more HIV-positive women to freely chose injections as their contraceptive method and decrease unplanned pregnancy and pediatric AIDS.

Quality Rating

Using the Jadad quality assessment scale for RCTs, this study received a low rating. Although randomized, the randomization process led to a discrepancy in the number of couples allocated to each arm, as the randomization was done by day rather than by couple. In addition, due to the nature of the intervention, double blinding was not possible. Also, because this was a pilot study, the sample size was constrained by the need to recruit couples only from those already involved in other studies at the research center. This may have resulted in insufficient statistical power to detect differences between groups when differences actually did exist. The recruitment exclusively of couples who were prior participants in the Center's HIV research studies may restrict the generalizability of the results.

In Context

Previous concerns(1,2) that the use of hormonal contraceptives may increase a woman's risk of HIV infection or the risk of transmission to her partner was not observed in this study, as the use of non-barrier contraceptives was not associated with decreased condom use, and the reporting of condom use was consistent across all three RCT arms. In addition, the finding that pregnancy incidence was significantly lower among injectable users than among those who used condoms has been shown previously in only one other study.(3)

Programmatic Implications

Although further study is needed to confirm the effectiveness of injectables, reducing provider bias against injectables and other barriers to injectable use might allow more HIV-positive women to freely chose injections as their contraceptive method and, therefore, decrease unplanned pregnancy and pediatric AIDS. The integration of HIV and family planning services can be effective in promoting dual-method use, but further research is needed to assess ways to translate increased method uptake into lower pregnancy incidence.


  1. Mostad SB, Overbaugh J, DeVange DM Welch MJ, Chohan B, Mandaliya K et al. Hormonal contraception, vitamin A deficiency, and other risk factors for shedding of HIV-1 infected cells from the cervix and vagina. Lancet 1997 Sep 27;350(9082):922-7.
  2. Morrison C, Schwingl PJ. Oral contraceptive use and infectivity of HIV-seropositive women. JAMA 1993 Nov 17; 270(19):2298.
  3. Darney PD, Callegari LS, Swift A, Atkinson ES, Robert AM. Condom practices of urban teens using Norplant contraceptive implants, oral contraceptives, and condoms for contraception. Am J Obstet Gynecol 1999 Apr;180(4):929-37.