Initiation of the contraceptive implant does not appear to reduce women’s use of condoms, at least in the short term. According to a randomized controlled trial conducted among female clinic clients in Jamaica, condom use—as objectively measured by a biomarker of recent semen exposure—during the three-month study period was no lower among women who had begun using the contraceptive implant than among those who had not.¹ However, implant users were more likely than others to report having had unprotected sex in the two days or the week prior to being surveyed (prevalence ratios, 1.5–1.8).

To examine whether use of the contraceptive implant is associated with a reduction in condom use, researchers conducted a randomized controlled trial among 414 women recruited in 2012–2013 from seven maternal and child health or family planning clinics in Kingston, Jamaica. Half of the women were selected to receive a contraceptive implant immediately, and half to delay insertion until after the three-month study period. Women were eligible to participate in the study if they were aged 18–44, had not had a hysterectomy or were not planning to have one in the next three months, did not report being HIV positive, were not currently using a long-acting reversible contraceptive method or planning to use one in the next three months, and had no contraindications for use of the implant. At baseline, and at one-month and three-month follow-up visits, participants answered questions about their demographic characteristics, sexual activity and condom use. A urine sample was taken at each time point for pregnancy testing; women who had a positive pregnancy test result at baseline were not eligible to participate in the study. In addition, a vaginal swab was taken at each time to be tested for prostate-specific antigen (PSA)—a biomarker for semen that is detectable within 48 hours of unprotected sex. An intent-to-treat analysis using generalized estimating equations was used to examine PSA positivity across study groups.

The baseline characteristics of women in the immediate insertion and delayed insertion groups were similar. On average, women in each group were 25 years old and had had two children. In each group, 69% of women had at least a high school education, and 98% had never used a contraceptive implant; more than nine out of 10 respondents in each group had never been married, had not had four or more alcoholic drinks in the past week and had never received money or gifts in exchange for sex. Among the immediate insertion group, 58% had used male condoms in the past month, 9% the injectable, 13% the pill, 17% other methods and 26% no method; those proportions among the delayed insertion group were 52%, 8%, 21%, 18% and 23%.

The proportion of women at baseline who had a positive PSA test result was 23% in the immediate insertion group and 27% in the delayed insertion group; at the one-month follow-up visit, those proportions were 26% and 20%, respectively, and at the three-month follow-up visit, they were 26% and 23%. Changes in PSA positivity over time were not significantly different between the study groups.

At baseline, the proportions of women in the immediate insertion group who reported having had unprotected sex within the two days and the week prior to interview were 17% and 23%, respectively; the proportions were 22% and 31% at the one-month follow-up, and 17% and 27% at the three-month follow-up. The proportions of women in the delayed insertion group who reported at baseline having unprotected sex within the two days and the week prior to interview were 14% and 23%, respectively; at the one-month follow-up, those proportions were 12% and 21%, and at the three-month follow-up, they were 11% and 18%. Women’s reporting unprotected sex within the two days prior to interview was greater for the immediate insertion group than for the delayed insertion group at the one-month follow-up visit (prevalence ratio, 1.8), and women’s reporting unprotected sex within the week prior to interview was greater for the immediate insertion group than for the delayed insertion group at both follow-up visits (1.5 each).

The authors comment that the self-reported data may be inaccurate because of such issues as social desirability and recall bias, and that their use of PSA as an objective biological marker of semen exposure "circumvents the problems of relying only on self-report." They note, however, that their study had several limitations, including that it evaluated condom use for only the three months following implant insertion and that PSA is detectable only within 48 hours of unprotected sex; they suggest that future studies using a longer evaluation period and a longer-lasting biomarker for semen exposure would be useful. The authors conclude that their study "suggests that it is possible to introduce contraceptive implants into a population at high risk for unintended pregnancy and STI without a large decrease in the use of condoms, in the short term."—J. Rosenberg