OBJECTIVE: To estimate whether young women taking the first pill on the day of prescription had higher continuation rates and lower pregnancy rates than women who waited until menses to start the oral contraceptive pill (OCP).
METHODS: We recruited 1,716 women aged younger than 25 years seeking to initiate the oral contraceptive at three publicly funded family planning clinics, and randomly assigned them to conventional initiation of the pill (conventional start) or immediate, directly observed ingestion of the first pill (quick start) during the clinic visit. Women underwent follow-up interviews at 3 and 6 months.
RESULTS: Sixty percent of participants discontinued the pill, and 8% became pregnant during follow-up. Women who took the first pill in the clinic were more likely to continue to the second OCP pack (odds ratio 1.5, 95% confidence interval 1.0–2.1.); however, the Quick Start approach did not improve OCP continuation rates at 3 and 6 months. Those assigned to Quick Start were slightly less likely to become pregnant within 6 months from the time they started the pill (hazard ratio 0.90, 95% confidence interval 0.64—1.25). Eighty-one percent of women rated the Quick Start approach as acceptable or preferable to waiting. Rates of serious adverse events were low and similar in the two groups.
CONCLUSION: Protocols that require a woman to wait until the next menses to start hormonal contraceptives are an obstacle to contraceptive initiation. Directly observed, immediate initiation of the pill improves short-term continuation.
CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov, www.clinicaltrials.gov, NCT00068848
LEVEL OF EVIDENCE: I
Immediate initiation of oral contraceptives without waiting for menses is safe and increases continuation to the second pack.
From the1 Department of Obstetrics and Gynecology, Columbia University, New York, New York; 2Department of Obstetrics and Gynecology, University of Texas—Southwestern, Dallas, Texas; 3Adolescent Health Center, Mt. Sinai School of Medicine, New York, New York; 4Department of Obstetrics and Gynecology, Emory University, Atlanta, Georgia; and 5Heilbrunn Department of Population and Family Health, Mailman School of Public Health, Columbia University, New York, New York.
See related editorial on page 1266.
Funded by National Institutes of Health grant number HD42413.
Corresponding author: C. Westhoff, 630 West 168 Street, New York, NY 10032; e-mail: email@example.com.
Financial Disclosure Dr. Zieman consulted for or received honoraria from pharmaceutical companies who manufacture contraceptives (Organon, Roseland NJ; Duramed, Woodcliff Lake, NJ; and Ortho, Shrewsbury, NJ). Dr. Carolyn Westhoff consulted for or received honoraria from pharmaceutical companies who manufacture contraceptives (Organon, Roseland, NJ; Duramed, Woodcliff Lake, NJ).
Oral contraceptives (OCs) remain the most popular birth control method in the United States, but access to OCs is limited, particularly for the youngest women who are likely to lack medical insurance and a regular source of care.1 Even after receiving a prescription, a young woman is likely to have an additional delay before beginning the OC. Patients generally begin all other prescription medications immediately after receiving the prescription, but conventional OC starting instructions require waiting until the next menses to begin the OC.2 Delays in initiating the OC are associated with a risk of pregnancy.
When OCs were first introduced, instructions to start with the next menses were justified by the desire to avoid giving an OC to a woman with an early pregnancy, because we had little information about their teratogenic potential and lacked quick, highly sensitive pregnancy tests. Subsequently, studies of inadvertent exposure to OCs during pregnancy revealed a lack of teratogenicity,3 so the main reason to delay OC initiation until menses is no longer relevant.
Family planning clinic studies have found that, after receiving the first pack, up to 24% of adolescents never take the first pill,4,5 mainly due to confusion about starting instructions, waning motivation, or becoming pregnant while waiting to start the OC. To minimize these barriers to effective contraception, we ask patients with a negative urine pregnancy test to swallow the first pill under direct observation during the clinic visit, regardless of menstrual cycle day, calling this approach “Quick Start.” We simultaneously administer emergency contraception whenever indicated.6 We also recommend back-up contraception with condoms during the first week of OC use, and perform a repeat pregnancy test if there is no withdrawal bleed at the end of the first OC pack.
Our initial studies suggested that women taking the first pill in the clinic were more likely to continue to the second pack of pills (odds ratio [OR], 2.8, 95% confidence interval [CI] 1.1–7.3),6 and 97% of women preferred to start the same day rather than to wait.7 Quick Start was associated with acceptable bleeding patterns8,9 and, despite prevalent recent unprotected intercourse, only 2% experienced “window” pregnancies during the first OC cycle, suggesting a contraceptive benefit from immediate initiation.6–11
To further evaluate the effects of the Quick Start approach we have carried out a large, multicenter, randomized trial. Our objectives were to estimate whether young women taking the first pill on the day of prescription had higher continuation rates and lower pregnancy rates than women who waited until menses to start the OC.
PARTICIPANTS AND METHODS
This trial was conducted between March 2003 and February 2005 in family planning clinics at Emory University, Atlanta; Mt. Sinai Medical Center, New York; and the University of Texas Southwestern, Dallas. These clinical sites are all funded in part by Title X, as well as other federal and state grants. Clinical and Coordinating Center Institutional Review Boards reviewed and approved the study, which also had an external Data Safety and Monitoring Board.
The study intervention was random assignment to immediate OC initiation at the time of study enrollment (Quick Start) or conventional initiation (Conventional Start). Young women requesting OCs were the target population. After completion of usual care, clinicians referred women to study staff for screening according to the following criteria: 1) aged younger than 25 years, 2) a current negative pregnancy test, 3) sexually active, 4) no use of an OC within 7 days (“Current use”) or Depo-Provera within 6 months, 5) no desire for pregnancy within the next 6 months, and 6) no lactational amenorrhea. The Texas Institutional Review Board mandated exclusion of postpartum or postabortion women if aged younger than 18 years.
After written informed consent, subjects participated in a baseline interview to collect demographic characteristics; sexual, contraceptive, and reproductive history; and detailed contact information. Interviewers were bilingual, and study materials were available in English and Spanish. After baseline interviews, subjects were randomly assigned using stratified blocks. The stratification variables were study site, age (younger than 18 years, 18 or older), and new compared with experienced OC user. A random number generator produced treatment assignments in a 1:1 allocation ratio. The Coordinating Center generated the allocation schedule, printed treatment assignments, and distributed these in numbered opaque envelopes.
Treatment assignments were Quick Start or Conventional Start. Subjects assigned to Quick Start opened the pack and swallowed the first pill under direct observation. Subjects assigned to Conventional Start received instructions to take the first pill during their next period. Subjects received standardized pill-taking instructions, a take-home instruction sheet, and at least one pack of pills at enrollment. Clinician preferences dictated the brand of OC and the provision of additional pill packs or prescriptions. We instructed participants to use condoms as backup during the first week of OC use.
Subjects participated in 3- and 6-month telephone interviews to assess OC continuation, pregnancy, sexual activity (risk of pregnancy), and adverse events. Women who became pregnant were excluded from additional follow-up. With subject permission, we sought clinical follow-up to confirm adverse events and to date the onset of reported pregnancies. Pregnancy dating relied mainly on ultrasonography. Women whose pregnancies began before study enrollment were excluded from follow-up. We used medical records to identify pregnancy in the 96 subjects who missed both follow-up interviews. The Coordinating Center reviewed all clinical information and made final determinations regarding all adverse events and pregnancy dating.
Investigators underwent centralized training to standardize enrollment and interview techniques. The Coordinating Center edited a 5% stratified sample of questionnaires for completion and accuracy, and found no patterns of errors, falsifications, or missing data. Site visits took place to confirm implementation of all aspects of the standard protocol.
The main study hypothesis was that immediate initiation of OCs in a population at high risk of discontinuation and pregnancy would lead to higher continuation rates and to lower pregnancy rates. The sample size was chosen to have more than 90% power to identify an increase in OC continuation from 50% to 60% at 6 months (with a 2-tailed alpha=0.05). This resulted in only 63% power to identify a decrease in the pregnancy rate as large as from 11% to 7%. We used univariable and multivariable logistic regression analyses to estimate odds ratios and confidence intervals for Quick Start compared with Conventional Start and the effect of baseline characteristics on OC continuation. We used Cox proportional hazards regression to compare the risk of pregnancy with Quick Start compared with Conventional Start. Analyses were performed using SPSS 13.0 software (SPSS Inc., Chicago, IL).
We enrolled 1,720 women, but excluded four women after randomization due to violations of inclusion criteria. We also excluded 33 pregnant women who had a negative pregnancy test on the date of study enrollment, but whose estimated date of conception (based on ultrasonogram) preceded the enrollment date, leaving 1,683 women eligible for the 3-month follow-up. Fifty-one additional pregnancies began before the 3-month interview; these women were included in all analyses, but we excluded them from further follow-up, so that 1,632 women were eligible for the 6-month interview (Fig. 1). Overall, 89% of eligible subjects completed the 3-month interview (Quick Start, 90%; Conventional Start, 88%). Eighty-four percent of eligible subjects completed the 6-month interview (Quick Start, 84%; Conventional Start, 83%). Ninety-six subjects (6%) missed both follow-up interviews; these subjects were less likely to enroll at the Texas clinics, but they otherwise resembled subjects who had follow-up interviews.
The Quick Start and Conventional Start groups were well balanced with regard to baseline characteristics. Subjects were mainly Latina, young, and poor; approximately 40% reported previous OC use (Table 1). Thirty-one percent of study interviews took place in Spanish. During follow-up, we identified 27 serious adverse events during the study. The number and type of serious adverse events were similar in the Quick Start and Conventional Start groups (Table 2).
Overall, 88% of subjects completed at least one pack and began a second pack of OCs (Quick Start 90%, Conventional Start 86%); 60% of subjects were still taking the OC at their 3-month interview and 40% at their 6-month interview. Quick Start was associated with an increase in completing at least one pack of pills and beginning a second pack (adjusted OR 1.5, 95% CI 1.0–2.1). At 3 months and 6 months, the Quick Start and Conventional Start groups had similar OC continuation rates (OR 1.1, 95% CI 1.0–1.1). Thus, the Quick Start effect on OC continuation is positive, but is brief as might be expected (Table 3). The number needed to benefit is 33. The effect of the Quick Start intervention was similar in subgroups defined by age, study site, and previous OC experience. There were 536 postpartum women who entered the study (500 in Texas); in these women the Quick Start intervention had no effect (OR 1.0 for continuation to the second pack, 95% CI 0.4–2.3).
Participants were 10% more likely to continue the OC with each additional year of age (Table 3). A baseline plan to use the OC for more than 6 months was significantly associated with OC continuation throughout follow-up. Being very certain about wanting to use the OC at baseline was also associated with continuation to the second pack, but this association waned somewhat at 3 and 6 months. Finally, partner awareness of the subject's plan to use the OC was associated with OC continuation throughout follow-up.12 Analyses shown in Table 3 were adjusted for study site and for each other.
Most of the other characteristics shown in Table 1 were not associated with OC continuation in the multivariable analyses. Even though “no desire for pregnancy” was a study inclusion criterion, 99 subjects responded, during the baseline interview (after enrollment and consent), that they would be “very happy” if pregnant in the next six months. In this subgroup 12 women became pregnant during the study; 10 in the Conventional Start group and 2 in the Quick Start group (P<.01), suggesting that immediate OC initiation may have a particular benefit among women who are ambivalent about pregnancy. At the 3-month interview, 64% of women allocated to Quick Start preferred this approach, 18% would have preferred to start later, and 17% had no preference. In the Conventional Start group, 25% would have preferred immediate initiation, 55% preferred waiting (their assigned treatment), and 19% had no preference (P<.001).
Subjects received their OC packs directly from the clinic. To assess the effect of OC supply on continuation we asked each subject how many packs she received at enrollment, along with how many packs she received by prescription. Packs provided ranged from one to 14, although 92% of subjects received three or four packs. The number of packs provided was unrelated to subject characteristics. Only 43 women in this study received prescriptions for refills. For each subject we calculated whether she received enough OC packs on the day of enrollment to last until the day her first follow-up interview occurred. The scheduled 3-month interview sometimes occurred slightly before or after 3 months; therefore, we used exact dates of enrollment and follow-up to determine whether a subject would still have OC supplies on the date of her interview, or whether she would have needed to return to the clinic for a refill before that date. Women who had not run out of OCs before the 3-month interview were more likely to be continuing users at 3 months and even at 6 months than the women who would have run out of pills sooner (Table 3, OC supplies at 3-month interview).
We identified 172 pregnancies during the study and assigned a clinically validated last menstrual period date for 171, usually based on prenatal ultrasonogram results. Thirty-three last menstrual periods were at least 14 days before study entry: 17 in the Quick Start group and 16 in the Conventional Start group. Most of these “window” pregnancies were diagnosed by 8 weeks of gestation (Quick Start, 86%; Conventional Start, 81%): In the Quick Start group two women appeared for care at 14 and 22 weeks of gestation, and in the Conventional Start group three women appeared for care at 13, 14, and 15 weeks of gestation. The 17-year-old gravida 1 who appeared for care at 22 weeks had discontinued OCs earlier and had no menses since study enrollment; we have no further information about her delay in entry into care; she delivered a healthy infant at 39 weeks of gestation. Two hundred thirty-seven subjects received emergency contraception at enrollment; only seven of these women (3%) had one of the early pregnancies. The Quick Start and Conventional Start subjects had 138 well-dated pregnancies that began during the study: 66 of 802 (8.2%) in Quick Start subjects and 72 of 788 (9.1%) in Conventional Start subjects. Pregnancies in the Quick Start group were slightly less frequent than those in the Conventional Start group (Fig. 2); thus, the brief increase in OC continuation in the Quick Start group may have translated into a modest decrease in the risk of pregnancy (hazard ratio 0.90, 95% CI 0.64–1.25) within 6 months of starting the pill; however, this study did not have adequate statistical power to distinguish such small differences in the proportion of women who became pregnant. The hazard ratio did not change after adjustment by site or age. Only 99 women enrolling in this study said they would be “very happy” if pregnant in the next 6 months (Table 1). Not surprisingly, these 99 women were more likely to become pregnant during the study (12.1% compared with 7.9%, P<.02). Of the 12 women in this group who became pregnant, 10 were assigned to Conventional Start (P<.05), hinting that the Quick Start approach might be useful among women who are ambivalent about pregnancy at time they start the oral contraceptive.
We observed high rates of OC discontinuation and unintended pregnancy among study participants. Within 6 months of initiation, 60% discontinued the OC and nearly 8% became pregnant. Only one woman who continued the OC for 6 months became pregnant during the study, reminding us that the OC is a highly effective contraceptive if it is continued as long as it is needed. All pregnancies in the study were unintended. The 6-month pregnancy rate observed here is consistent with the characteristics of this population, with clinical experience, and with past studies of young OC users.5,13
This randomized controlled trial enrolled young, poor, mainly Latina women aged younger than 25 years. Our simple intervention was applied in busy urban clinics in Texas, New York, and Georgia. Whether our results are applicable to other populations is unknown; however, oral contraceptive discontinuation and unintended pregnancy are also common in the U.S. among women who are white and nonpoor.
The Quick Start intervention was acceptable, in fact preferable, to the women in this study (81%) and in our previous studies.6–11 The Quick Start intervention briefly increased OC continuation, consistent with our previous observational study6; however, there was no difference between the groups at 3 or 6 months. In contrast, in our similar clinical trial using the Quick Start approach for the initiation of Depo–Provera, continuation rates and pregnancy rates were significantly better at 6 months after immediate compared with delayed initiation.14 This further supports the importance of long-acting contraceptive methods in reducing unintended pregnancy. The direct effect of Quick Start on OC continuation was small; therefore, the main benefit of immediate initiation may prove to be decreasing clinic time, because this approach eliminates the need for detailed patient counseling on how and when to start the OC.15 The Quick Start approach is only available in settings where clinicians dispense contraceptive supplies (or at least the first cycle) directly.
Other factors important for OC continuation in this study suggest useful interventions: At baseline, 96% of our subjects wanted to use the OC for more than 6 months, but only 1.4% received enough OC supplies for 6 months or more. Thus, participants needed to make additional visits within a few months to continue the OC. Nearly all study subjects needed a refill before the end of the study, but those with enough OC packs for more than 3 months were substantially more likely to continue longer. Providing more OC packs may have an additional up-front cost to the clinic; however, having less frequent visits may be cost saving and time saving for the patient and the clinic. Analyses from the California Family Planning Program showed that women who received 13 cycles at the first visit were more likely to have OCs on hand after one year than women who received only 3 cycles.16 We already know that frequent revisits for OC users are not medically necessary. Pending more data, providing a greater number of OC packs to poor women beginning the pill may be an easy way to substantially improve continuation, and thus to reduce unintended pregnancy. Providing a limited number of OC packs does not serve the expressed needs of the patient.
Reducing unintended pregnancy is the focus of many interventions,17–19 but few programs have any effect on OC continuation rates or pregnancy rates. What has been missing is an examination of obstacles within clinical services; this study focused on a specific obstacle to initiation of oral contraception—having the woman wait to start until her next menstrual cycle. Waiting to start the pill made sense 40 years ago. Now we know there is no teratogenic potential,3 now we have emergency contraception, and now we have quick and sensitive pregnancy tests; thus, waiting to start is no longer necessary. Rather, waiting to start is an obstacle for sexually active people who are requesting hormonal contraception. Similarly, dispensing a small number of OC packs or otherwise requiring frequent revisits is an obstacle to the continued use of this safe and extensively studied contraceptive.
In this study and others, there were no harms associated with immediate initiation of the hormonal contraceptive. The study did not have statistical power to detect differences between the groups in rare serious problems; however, there was no a priori reason to expect such differences. The Quick Start approach has no additional costs over usual care, and may prove to be cost saving due to a decrease in time needed to educate a patient regarding when to start her contraceptive. Patients in this and other studies preferred the Quick Start approach.7,10,11,20 Therefore, this approach to OC initiation should be an option for all patients in any setting where OC supplies are directly dispensed. As with other interventions, however, this approach is not a quick fix for the complex problem of unintended pregnancy,16 and additional innovations in service delivery are needed. Finally, our results suggest that increased pack supply may be useful in increasing OC continuation.
1. Alan Guttmacher Institute. Sex and America's teenagers. New York (NY): Alan Guttmacher Institute; 1994.
2. Williams-Deane M, Potter L. Current oral contraceptive use instructions: an analysis of patient package inserts. Am Plann Perspect 1992;24:111–5.
3. Bracken MB. Oral contraception and congenital malformation in offspring: a review and meta-analysis of the prospective studies. Obstet Gynecol 1990;76:552–7.
4. Polaneczky M, Slap G, Forke C, Rappaport A, Sondheimer S. The use of levonorgestrel implants (Norplant) for contraception in adolescent mothers. N Engl J Med 1994;331:1201–6.
5. Oakley D, Sereika S, Bogue EL. Oral contraceptive pill use after an initial visit to a family planning clinic. Fam Plann Perspect 1991;23:150–4.
6. Westhoff C, Kerns J, Morroni C, Cushman LF, Tiezzi L, Murphy PA. Quick start: a novel oral contraceptive initiation method. Contraception 2002;66:141–5.
7. Morroni C, Grams M, Tiezzi L, Westhoff C. Immediate monthly combination contraception to facilitate initiation of the depot medroxyprogesterone acetate contraceptive injection. Contraception 2004;70:19–23.
8. Westhoff C, Morroni C, Kerns J, Murphy PA. Bleeding patterns after immediate vs. conventional oral contraceptive initiation: a randomized, controlled trial. Fertil and Steril 2003;79:322–9.
9. Westhoff C, Osborne LM, Schafer JE, Morroni C. Bleeding patterns after initiation of the contraceptive pill and vaginal ring. Obstet Gynecol 2005;106:89–96.
10. Sneed R, Westhoff C, Morroni C, Tiezzi L. A prospective study of immediate initiation of depo medroxyprogesterone acetate contraceptive injection. Contraception 2005;71:99–103.
11. Schafer JE, Osborne LM, Davis AR, Westhoff C. Acceptability and satisfaction using Quick Start with the contraceptive vaginal ring versus an oral contraceptive. Contraception 2006;73:488–92.
12. Kerns J, Westhoff C, Morroni C, Murphy PA. Partner influence on early discontinuation of the pill in a predominantly Hispanic population. Perspect Sex Reprod Health. 2003 35:256–60.
13. Finer LB, Henshaw SK. Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspectives on Sexual and Reproductive Health 2006;38:90–6.
14. Rickert VI, Tiezzi L, Lipshutz J, Leon J, Vaughan RD, Westhoff C. Depo Now: preventing unintended pregnancies among adolescents and young adults. J Adolesc Health 2007;40:22–8.
15. Rickert VI, Berenson AB, Williamson AJ, Wiemann CM. Immediate recall of oral contraceptive instructions: implications for providers. Am J Obstet Gynecol 1999;180:1399–406.
16. Foster DG, Parvataneni R, de Bocanegra HT, Lewis C, Bradsberry M, Darney P. Number of oral contraceptive pill packages dispensed, method continuation, and costs. Obstet Gynecol 2006;108:1107–14.
17. Kirby D. No easy answers: research and findings on programs to reduce teen pregnancy. Washington (DC): National Campaign to Prevent Teen Pregnancy; 1997.
18. Halpern V, Grimes DA, Lopez L, Gallo MF. Strategies to improve adherence and acceptability of hormonal methods for contraception. Cochrane Database Syst Rev 2006;1:CD004317.
19. Jaccard J. Adolescent contraceptive behavior: the impact of the provider and the structure of clinic-based programs. Obstet Gynecol 1996;88:57S–64S.
© 2007 The American College of Obstetricians and Gynecologists
20. Westhoff CL, Heartwell S, Edwards S, Zieman M, Stuart G, Cwiak C, et al. Oral contraceptive discontinuation: do side effects matter? Am J Obstet Gynecol 2007;196:412.e1–6.