Gilliam, Melissa L. MD, MPH; Neustadt, Amy MPH; Kozloski, Michael MS; Mistretta, Stephanie BA; Tilmon, Sandra MPH; Godfrey, Emily MD, MPH
Each year, more than 3 million women in the United States experience an unintended pregnancy.1 Although many unintended pregnancies occur among women who are not using contraception, almost half occur among women who reported using contraception during the month of conception.1 Oral contraceptive pills (OCPs) are one of the most widely used contraceptive methods; however, more than 1 million unintended pregnancies result from OCP method failure, misuse, and discontinuation.2
The largest number of unintended pregnancies occur among women aged 20 to 24 years.1 Although studies of unintended pregnancy focus on women in general, for many women, these years coincide with attending college. Studies have shown that college and graduate students are vulnerable to unintended pregnancy because of high rates of sexual activity and difficulties with contraceptive adherence. Although the rate of unintended pregnancy among college students is not known, in one study, only 52% of university OCP users were consistent pill takers.3 Other studies have found inaccurate beliefs about the risks and underestimation of the health benefits of OCPs among college students.4,5
Few investigators have examined whether contraceptive adherence in college and university students can be improved with nondaily contraceptive methods. Prior studies suggest that students have favorable attitudes toward trying the contraceptive vaginal ring. In a survey of 691 college students, respondents preferred a nondaily (monthly) method over a daily method, and those most concerned about the risks of hormonal contraception were willing to try the contraceptive vaginal ring.6 In a randomized controlled trial (RCT) of adolescents and young adults from the general population, acceptability of the contraceptive vaginal ring was higher compared with OCPs.7 However, there is a lack of substantive information on college and university students' attitudes toward and adherence with a nondaily method such as the contraceptive vaginal ring.
The Acceptability of the NuvaRing (Merck, Whitehouse Station, NJ) (ACCEPT) study was an RCT to consider adherence to and acceptability of the contraceptive vaginal ring compared with OCPs among female college and university students over a 3-month period. We hypothesized that students would better adhere to a nondaily contraceptive method and find it more acceptable than a daily method. Our objective was to compare satisfaction with and adherence to the contraceptive vaginal ring and a daily low-dose OCP among college and graduate students by using a novel method of electronic data collection.
MATERIALS AND METHODS
We recruited local participants for the ACCEPT study at two Midwest university-based sites through passive (online and subway advertisements) and active recruitment (referrals from local clinics and health care providers). We enrolled healthy women aged 18 to 45 years who were full-time college or university students and remained so throughout the duration of the study. Participants were required to have daily access to a working electronic mail address to fulfill tracking objectives and had to agree to avoid vaginal product use for the duration of the study unless instructed by a physician or study clinician. Exclusion criteria included known or suspected pregnancy, pregnancy within the 2 months preceding trial initiation, plans for a pregnancy within the next 6 months, contraindications to a combined hormonal contraceptive method, allergy or hypersensitivity to either drug intervention, any investigational drug use within the 2 months preceding trial initiation, family or personal history of blood clots, prior use of the contraceptive vaginal ring, contraceptive patch or oral contraceptive use within 1 month or injectable contraceptive use within 6 months of trial initiation, and any other contraindication as indicated in the medication labeling. This study was reviewed and approved by the institutional review boards of both participating institutions: the University of Chicago and the University of Illinois at Chicago. The University of Chicago served as the sponsoring institution. The clinical trial was registered with ClinicalTrials. gov. The Consolidated Standards of Reporting Trials guidelines for reporting an RCT were followed and documented. All participants provided written consent before participating in this study.
After obtaining written informed consent, participants completed demographic, medical, sexual, and contraceptive history questionnaires and underwent a physical examination that included blood pressure, height and weight, a urine test for pregnancy and a nucleic acid amplification test for Neisseria gonorrhea and Chlamydia trachomatis. A positive N. gonorrhea or C. trachomatis result was not an exclusion criterion. Pap tests were offered if clinically indicated (as a service to the participant, not an exclusion criterion). Participants were questioned regarding symptomatic vaginal discharge; if present, study clinicians evaluated discharge by wet mount preparation. Study physicians treated abnormal laboratory findings as indicated. Participants were then randomly assigned to either the contraceptive vaginal ring (NuvaRing; Schering-Plough Corporation, Kenilworth, NJ) or OCPs (Ortho Tri-Cyclen Lo; Ortho-McNeil-Janssen Pharmaceuticals, Titusville, NJ). This pill was chosen because it was a popular pill that was available and affordable through the university formulary.
The randomization sequence was created using a computer-generated randomization algorithm in blocks of 10. An individual at the University of Chicago not associated with the clinical portion of this study created and sealed the randomization envelopes. Each center received sets of opaque sealed envelopes containing the randomly allocated contraceptive method assignment. When a student was determined to be eligible for enrollment, the clinical research staff at each corresponding research center opened the next sequentially numbered opaque envelope and gave the participant a 3-month supply of the contraceptive method indicated within the envelope. In addition, participants received U.S. Food and Drug Administration–approved patient instructions for the assigned medication; written instructions for medication initiation; and information about how to contact the study staff or a 24-hour on-call physician, should questions arise. The research staff, study physicians, statisticians, and participants were unblinded to the contraceptive method group assignment.
Research staff instructed the participant to begin her assigned hormonal contraceptive method on day 4 or 5 of her next menses and to call the research office on the day of method initiation. Once a start date was confirmed, study coordinators registered each new participant in an online database system administered by the Center for the Advancement of Distance Education at the University of Illinois at Chicago. Participant registration included name, e-mail address, group assignment, and medication start date (as confirmed by each participant). The Center for the Advancement of Distance Education sent online, Internet-based diaries and surveys (baseline, 3-month, and 6-month surveys) to participants to complete throughout the study. Surveys and diaries were developed in collaboration with the Center for the Advancement of Distance Education and the Survey Research Laboratory at University of Illinois at Chicago and were beta tested by five research staff members who checked the accuracy of questions and verified skip patterns before the start of the study. Participants were given 2 weeks to complete diary entries or surveys.
Once the participant started her method, she received an email notification from the Center for the Advancement of Distance Education to complete the baseline survey, which included demographic (race, sex, age, sexual orientation, current relationship status, and sexual and pregnancy history) and lifestyle/social information (living situation, class schedule, credit hours, work responsibilities, and sleep schedule). The participant subsequently completed daily online diaries regarding adherence to and satisfaction with her assigned contraceptive method; participants were notified of the daily diary through an e-mail message from the Center for the Advancement of Distance Education at 1 am stating, “Hi, this is a friendly reminder to complete your diary for today,” which linked to the online survey. In the online diary, the participant was asked whether she took the pill (yes/no) or whether the ring was in or out each day. After three consecutive menstrual cycles of method use, she completed an online questionnaire regarding satisfaction with and intention to continue her assigned method. Six-month online surveys served to determine whether the participant continued contraception and, if so, which method she was using. For each week of completed daily dairies, participants earned $5 and were not compensated for weeks with incomplete daily diaries. After completing the 3- and 6-month online follow-up surveys, they were paid an additional $25. Participants could receive up to $90 in reimbursement if they completed all of the diary weeks and both the 3- and 6-month follow-up surveys.
The trial sample size of 300 participants total (150 participants each group) was based on two-sided 5% significance testing with 80% power to detect a difference of 10% in adherence between students in the contraceptive vaginal ring group and OCP group. Because of enrollment time constraints, only 276 of the anticipated 300 participants were enrolled.
The primary aim of the ACCEPT study was to compare rates of perfect method use during the first 3 months of contraceptive use. “Perfect use” was defined as reporting never missing a pill or never removing the contraceptive vaginal ring for more than 2 hours during days 1–21 of all three monthly cycles; perfect method use was analyzed using the participants' daily diaries to determine whether a participant missed a pill or removed the contraceptive vaginal ring during days 1–21 of her monthly cycle. We did not track the time of day that the pill was taken. Perfect use analysis was completed based on an intent-to-treat population where all participants who discontinued early, were lost to follow-up (defined as not completing the 3-month survey and making no diary entries within the third month of the trial), or did not complete the 3-month survey were considered “imperfect users.” Secondary aims were to compare rates of satisfaction and intention to continue the contraceptive method at 3 months. To understand poststudy behaviors, we considered continuation rates at 6 months (ie, 3 months after the end of the study period). To measure satisfaction, each participant was asked to rate her level of agreement on a five-point scale to the statement “I am satisfied with using the contraceptive vaginal ring/taking the pill.” The scale was a Likert scale with 5 denoting “strongly agree” and 1 denoting “strongly disagree.” Data are presented in three categories collapsing “agree” with “strongly agree” and “disagree” with “strongly disagree” because of small numbers in each category. This analysis of the secondary aims comprised only the per-protocol population, and all participants who were lost to follow-up, discontinued early, or did not complete any part of the survey were excluded (Fig. 1).
Statistical analysis was completed using SPSS 17.0 software (SPSS Inc., Chicago, IL). We used descriptive statistics to present the demographic and behavioral characteristics, 3-month continuation intentions, and 6-month method use by study group. We performed univariable analyses using the χ2 test for categorical variables, sample t tests to compare means for continuous variables, and the Mann-Whitney U test to compare medians for continuous variables to determine statistical differences between the contraceptive vaginal ring and OCP groups.
Between March 2006 and December 2008, we screened 276 students, of whom 273 were eligible (Fig. 1). We randomly assigned 136 to the contraceptive vaginal ring group and 137 to the OCP group. Shortly after randomization, two participants withdrew consent and did not ever start their assigned method. One OCP-assigned participant reported feeling nervous about starting OCPs, and one contraceptive vaginal ring–assigned participant reported depression. Both groups were similar at baseline, with significant differences found only for past pregnancies (Tables 1 and 2). Despite falling short of our goal of enrolling 300 women, post hoc analysis revealed an observed power of 0.91 to determine a significant difference between the two groups.
The women in the sample led busy daily lives, with 76% reporting at least some weekly paid employment in addition to full-time schoolwork and 90% reporting at least some weekly extracurricular time commitment. They slept less during the week than on weekends, sleeping a median of 7 hours per night on weeknights and 8 hours per night on weekends.
A total of 121 women allocated to the contraceptive vaginal ring group and 126 allocated to the OCP group were included in the 3-month analysis. We excluded 24 participants from the analysis at 3 months. Among those excluded from the contraceptive vaginal ring group, two participants did not complete the survey, eight were considered lost to follow-up, and four discontinued early: one citing a lack of sexual activity, two for adverse events (breakthrough bleeding/cramping, mood swings and increased acne breakouts), and one following a physician's request due to a preexisting unrelated illness. Among those excluded from the OCP group, one did not complete the survey, eight were considered lost to follow-up, and one discontinued early citing a lack of sexual activity.
Perfect method use was assessed following inclusion of an intent-to-treat population (n=135 in the contraceptive vaginal ring group and n=136 in the OCP group). Perfect use for the 3 months of contraceptive use varied significantly between contraceptive vaginal ring users (57%) and OCP users (45%). Adherence was better for the contraceptive vaginal ring compared with OCPs during the first 2 months of use (P=.03 and P=.05, respectively); however, results were not statistically significant in the third month (Table 3). Overall diary completion rates were similar between groups: 75% of contraceptive vaginal ring users and 80% of OCP users completed at least 90% of the daily diaries.
Satisfaction rates were nearly equal across both treatment arms: 82 (68%) of 121 contraceptive vaginal ring users and 89 (71%) of 126 OCP users reported being satisfied with their method (P=.50). Similarly, 67 (55%) of 121 contraceptive vaginal ring users and 78 (62%) of 126 OCP users believed the method to which they were assigned was the best method for family planning (P=.36). Three-month plans to continue their assigned contraceptive methods also did not differ significantly between the groups; 52 (43%) of 121 contraceptive vaginal ring users and 65 (52%) of 126 OCP users reported plans to continue their method (P=.16) (Table 3).
At 6 months, 117 (86%) of 136 contraceptive vaginal ring participants and 123 (90%) of 137 OCP participants completed the protocol. Three participants from the contraceptive vaginal ring group and four participants from the OCP group did not complete the 6-month survey and thus were excluded from the analysis. Six-month method continuation rates between contraceptive vaginal ring users and OCP users did not differ significantly: 31 (26%) of 117 contraceptive vaginal ring users and 36 (29%) of 123 OCP users continued the contraceptive vaginal ring or OCPs, respectively (P=.61) (Table 4). Additionally, at 6 months, we found that nearly half of the women in both groups (46% of contraceptive vaginal ring users and 48% of OCP users) used condoms only or no method (P=.88) (Table 4).
The ACCEPT study compared adherence to and satisfaction with the contraceptive vaginal ring compared with OCPs in a cohort of undergraduate and graduate students from Midwestern colleges and universities. Identifying strategies for increasing contraceptive adherence is extremely important because unintended pregnancy carries personal and social consequences. College-aged women typically fall within the age range that has the highest rates of unintended pregnancy compared with women in other age groups. While this population-based rate of unintended pregnancy refers to all women aged 20–24 years, we studied students as a population in which an unintended pregnancy may be particularly problematic. Thus, we sought to compare whether college students could more successfully use a nondaily method compared with a daily method. Our findings indicate that more college and university students randomly assigned to the contraceptive vaginal ring demonstrated perfect use compared with those randomly assigned to OCPs. Although our findings showing high rates of missed pills are consistent with prior studies, we demonstrate that a nondaily method promotes adherence.3,8,9
Women found both methods equally acceptable. They were highly satisfied with whichever method they were assigned to and also thought that the method to which they were assigned offered the best pregnancy prevention. Given that contraceptive vaginal ring users reported more perfect use than OCP users and were equally satisfied with their method, providers should be encouraged to recommend the contraceptive vaginal ring to female undergraduate and graduate students who are interested in a nondaily method. The difference in adherence was noted in the first 2 months of method use; this finding might suggest that young women might need additional support in the early months of pill use. We also considered behavior after the intervention period in an effort to obtain a further understanding of satisfaction and adherence. However, at 6 months, the majority of participants were not using a hormonal method of contraception. This finding is particularly troubling because this population expressed high rates of acceptability at 3 months. Future research could explore potential reasons why college and graduate students enroll in a contraceptive clinical trial and stop using hormonal contraception after the trial.
The limitations of this study must be considered. First, this study is an RCT offering internal but not necessarily external validity. Thus, we do not know whether these findings would be replicable in other populations (eg, among other students in other settings). The study cohort consisted of women willing to be randomly assigned to either method of contraception, which may represent a different population than those who are committed to using only the contraceptive vaginal ring or OCPs. The fact that participants in this study received free contraception and compensation for participation also may limit the generalizability of the study findings and affect continuation rates after study completion, because cost of contraception has been shown to affect contraceptive choice and continuation.10 Similarly, because this study was randomized, it did not realistically reflect real-life situations in which young women choose a method after being counseled about all options. However, despite randomization, both methods were found to be highly acceptable. Second, we only studied the first 3 months of use, but even in this short period, we demonstrate a difference in contraceptive use behaviors. Data from a longer time period would be of interest. Third, we considered perfect pill use (missing no pills during the 21-day cycle) but not effective use (taking a pill every day for 21 days or missing only one pill). Because a single missed pill does not necessarily result in an increased risk of pregnancy, it is important not to overinterpret the findings. Fourth, although the groups had similar characteristics at baseline, the group of women assigned to the contraceptive vaginal ring was twice as likely to have been pregnant as the women assigned to OCPs; they had also experienced a greater number of live births than the women assigned to OCPs. It is possible that these women might have been more motivated to use their method perfectly than the group assigned to OCPs. Finally, this study was funded by the manufacturer of the contraceptive vaginal ring. However, the company only reviewed the manuscript and did not advise on the design, conduct, statistical analysis, or interpretation of the findings for this study. No modifications to the analysis or interpretation were made after the manuscript review by the manufacturer.
In conclusion, college and university students in this study randomly assigned to either the contraceptive vaginal ring or OCPs were equally satisfied with their assigned hormonal contraceptive method. Overall, contraceptive vaginal ring users seemed to use their method more perfectly compared with OCP users. Rates of use of either method at 3 months after the intervention period were low for both groups. The ACCEPT study provides preliminary data suggesting that students may find nondaily methods easier to adhere to when initiating use than daily methods. However, the findings of nonuse at 6 months suggest that studies of long-acting reversible methods (the intrauterine device and implants) are also needed in this population. Longer studies are needed to determine whether perfect use among the two groups would have remained the same or significantly differed past the 3-month time period.