We conducted in-depth, semistructured individual interviews with 30 contraceptive users 17–19 years old in Atlanta, Georgia. To recruit participants, we actively approached individuals and posted fliers in publicly funded clinics that provide family planning services and an academically-affiliated outpatient adolescent medicine clinic, as well as university and other community settings. Additionally, snowball sampling was used whereby enrolled individuals were invited to refer friends.

We screened young women for eligibility in person or via phone, depending on the recruitment approach. Individuals were eligible to participate if they: (1) were 17–19 years of age; (2) had vaginal sex with a male sex partner in the previous 6 months; and (3) used a long-acting (IUD or implant) or short-acting (oral contraceptives, DMPA, patch, or ring) contraceptive method at last sex. We purposively enrolled a stratified sample of contraceptive users on the basis of their method’s length of action. Specifically, we recruited young women using LARC (n = 10) for which no action is needed over a 3–10 year period after insertion; DMPA, ring, or patch, for which quarterly, monthly, or weekly action must be taken, respectively (n = 10); and oral contraceptives that must be taken daily (n = 10). We included oral contraceptive users as a distinct group because of previous findings from a nationally representative sample of high school students that condom use prevalence was lower among LARC users compared with oral contraceptive users but not DMPA, patch, or ring users.¹⁰ Although patch and ring users were eligible, none of the individuals enrolled in the study were using these methods.

We conducted all individual, in-depth interviews between June 2017 and January 2018, according to procedures approved by the Emory University institutional review board. Depending on the participant’s preference, either an in-person (n = 21) or telephone interview (n = 9) was scheduled. Participants completed a brief quantitative survey that assessed demographic characteristics, contraceptive and condom use, health-seeking behaviors, and relationship factors. The qualitative interview used a semistructured interview guide that covered pregnancy prevention and STI prevention goals and strategies, condom use behavior and motivations, and health services experiences (Table 1). Interviews lasted approximately 40 minutes and were digitally audio-recorded and professionally transcribed. Each participant received $25 for completing the study.

We conducted a thematic analysis, following guidelines from Braun and Clark.²⁵ Deidentified interview transcripts, which were reviewed for accuracy by the first author, were uploaded to MAXQDA version 12.3 (VERBI Software). We then developed a codebook, starting with deductive codes on the basis of the interview guide. Two coders independently reviewed a subset of transcripts (n = 10) to identify inductive codes, or those emerging from the data (eg, condom error, defined as references to a condom breaking or slipping off during sex, perceived risk, personal experience) and met to discuss potential additions. Including inductive and deductive codes is fairly standard for thematic analysis.²⁶ Each coder then independently coded the remaining transcripts so that all transcripts were double coded; we resolved any discrepancies through discussion. Additionally, each coder created a case-level meta-matrix to summarize condom use behavior, motivations, and influencing factors for each participant. The 2 coders discussed these summaries to ensure consistent interpretation of the data. The first author developed additional matrices to examine coded text across the entire sample and within and between the 3 strata of contraceptive users. Themes emerged relevant to 2 over-arching areas: (1) description of condom use motivations; and (2) explanation of condom use motivations. In addition to salient quotes, we provide descriptive statistics as appropriate to support the qualitative findings.

Overall, the mean age was 18.5 years, and more than half of the 30 participants were black/African-American (60%; n = 18; Table 2). Most participants had used condoms with their current contraceptive method (83%; n = 25), either consistently (every time they had sex) or inconsistently (not every time they had sex), on the basis of participants’ qualitative descriptions. Taking into account consistent and inconsistent use, 8 LARC (80%), 8 DMPA (80%), and 9 oral contraceptive (90%) users also used condoms. Five participants (17%), including 2 LARC users, 2 DMPA users, and 1 oral contraceptive user, had previously been diagnosed with an STI; 1 DMPA user had an unplanned pregnancy.

Two primary themes, reflecting variation according to contraceptive type, describe condom use motivations among the 25 individuals using condoms with their current, more effective contraceptive method (either consistently or inconsistently on the basis of participants’ qualitative descriptions). Specifically: (1) oral contraceptive users seemed to primarily use condoms for pregnancy prevention; and (2) LARC users seemed to primarily use condoms for STI prevention. DMPA users’ current condom use motivations were more mixed and are not explicitly discussed.

Most oral contraceptive users who were also using condoms cited pregnancy prevention as a motivator, either primarily or in addition to STI prevention. For example, “I use a condom to prevent from STDs [sic] and then that’s my second for preventing pregnancy.” In fact, 18 of the 25 (72%) individuals who used condoms with their current contraceptive method indicated that additional protection against unintended pregnancy was an important reason for their condom use (Table 3). Among participants not using condoms with their current contraceptive method, most had used them previously for pregnancy prevention.

Participants further described their motivation to use condoms for pregnancy prevention in a specific way that we’ve termed “the safe side,” referring to participants’ desires for additional protection (ie, “back-up plan”) “just in case” or to be “safe than sorry” (see Table 4 for salient examples). Oral contraceptive users in particular recognized that their contraceptive methods were not 100% effective and described condoms as a way to be on the safe side for pregnancy prevention, as shown in Table 4. A few indicated concern about method effectiveness and motivation to use condoms as back-up pregnancy prevention at specific times, including when initiating a new method, starting a new pill pack, missing pills, ovulating, or when “paranoia is more active.”

Several participants described hormonal contraceptive methods as a back-up to condoms rather than the other way around. A young woman who considered her implant as a back-up method said, “I feel a lot like more calm in knowing that I have a back-up plan and I’m safe if anything happened.” In general, participants were clear that using condoms alone did not provide sufficient protection against pregnancy. When asked why she does not solely use condoms, a DMPA user said, “I don’t want 1 type of protection. I’d rather have 2, to kind of break the odds of getting pregnant.” An oral contraceptive user specifically cited concern about condoms breaking as a reason for using pills as a “second layer of protection.” Most participants were aware of the potential for condom errors, and a number had experienced them.

Most LARC users also using condoms explained their use primarily in relation to preventing STIs (Table 3). An implant user who had previously used oral contraceptives highlighted how her motivations changed on the basis of her contraceptive method: when using pills it “was like 60/40 at that point. It was like 60 like oh my God, STIs, but the 40 was like I don’t know how consistent [with pills] I am right now.” Now using the implant “it’s very much like 100 STI, 0 pregnancy.” Another LARC user who was using condoms consistently for pregnancy prevention stated an intention to discontinue her condom use. Moreover, 2 of the LARC users not using condoms had previously used them for pregnancy prevention (alone and with oral contraceptives). Some LARC users noted their decision to use condoms as a way to provide extra STI protection in the context of their relationship with a serious boyfriend, but in general, descriptions of condom use in relation to being on the safe side for STI prevention were uncommon.

Although participants’ motivations seemed to vary according to contraceptive type, the themes identified that explain those motivations were similar across the contraceptive categories and included: (1) sexual health education; (2) personal awareness/personal experience; and (3) perceived consequences and risks. These themes are summarized in the following sections, with salient quotations provided in Table 5.

Multiple participants who were motivated to use condoms as an additional contraceptive noted that their sexual health education in school emphasized pregnancy prevention, but not STI prevention. Likewise, for participants motivated to use condoms for pregnancy prevention, conversations with parents tended to emphasize pregnancy prevention over STI prevention. In contrast, several participants motivated to use condoms for STI prevention described school-based sexual health education that clearly addressed this prevention goal. Several LARC users who reported using condoms consistently for STI prevention described conversations with their parents about STIs, and one noted that her mother offered to provide condoms. However, sexual health education did not always correspond to condom use motivations and behavior. As an example, one DMPA user summarized a conversation in which her mother promoted condoms to be on the safe side for preventing STIs, but she did not feel she needed to use condoms because she was in what she perceived to be a monogamous relationship and had been tested, although she did not know if her partner had been tested as well.

More than half of participants across contraceptive methods had parents, other family, and/or friends who experienced an unintended pregnancy. Many of these contraceptive users cited pregnancy prevention as a reason for using condoms with their method. Several participants described pregnancy scares among friends, which made them particularly concerned about preventing pregnancy. However, a couple of young women had close, personal experiences with STIs that shaped their motivations to use condoms for STI prevention. For example, a pill user who was using condoms for pregnancy and STI prevention described being “super cautious” since her ex-boyfriend notified her that he had been diagnosed with herpes. In contrast to personal awareness aligning with condom use motivations, a young woman whose mother had an STI was more concerned about pregnancy than STI prevention.

Personal experience also seemed to be important with regard to STI prevention. Of the 5 participants previously diagnosed with at least 1 STI, 4 described using condoms for STI prevention. Likewise, the 1 participant who reported an unintended pregnancy and was currently using DMPA noted that she thought about pregnancy more than sexually transmitted diseases “because I’ve been pregnant.”

For many participants motivated to use condoms for additional pregnancy prevention, it seemed the consequences of pregnancy outweighed those of STIs. For many, having a baby would negatively affect future goals and also have shorter-term consequences, including financial hardship and disappointing their parents. Several of these participants struggled to articulate why STI prevention was important to them or viewed it merely as an inconvenience. In contrast, participants who were motivated to use condoms for STI prevention described specific consequences of STIs that they perceived as severe, including stigma, infertility, potential incurability, and transmission to partners.

A few participants perceived the consequences of pregnancy and STIs to be equally concerning. However, their perceived risk of STIs was low because they were in perceived stable, mutually monogamous relationships. Overall, perceived STI risk was low to moderate, often largely on the basis of partnership characteristics, and many of these participants were not using condoms or were using them for pregnancy prevention only.