Volume 102, Number 1, July 2003

Original Research

July 2003

Volume 102, Number 1
Pages 8 - 16

Advance Supply of Emergency Contraception: Effect on Use and Usual Contraception—A Randomized Trial

Rebecca A. Jackson, MD,^a Eleanor Bimla Schwarz, MD, MS,^b Lori Freedman, MS,^a and Philip Darney, MD, MSc^a

Objective: To evaluate whether advance provision of emergency contraception increases its use and/or adversely affects usual contraceptive practices.

Methods: We performed a randomized controlled trial comparing advance provision of emergency contraception with usual care in 370 postpartum women from an inner-city public hospital. Participants were followed for 1 year; 85% were available for at least one follow-up session. All participants received routine contraceptive education. The intervention group received a supply of emergency contraception (eight oral contraceptive pills containing 0.15 mg of levonorgestrel and 30 µg of ethinyl estradiol) and a 5-minute educational session. We compared use of emergency contraception and changes in contraceptive behaviors between groups.

Results: Women provided with pills were four times as likely to have used emergency contraception as women in the control group over the course of the year (17% versus 4%; relative risk [RR] 4.0; 95% confidence interval [CI] 1.8, 9.0). Women were no more likely to have changed to a less effective method of birth control (30% versus 33%; RR 0.92; 95% CI 0.63, 1.3), or to be using contraception less consistently (18% versus 25%; RR 0.74; 95% CI 0.45, 1.2). About half of each group reported at least one episode of unprotected intercourse during follow-up, but women who received emergency contraception were six times as likely to have used it (25% versus 4%; RR 5.8; 95% CI 2.1, 16.4).

Conclusion: Advance provision of emergency contraception significantly increased use without adversely affecting use of routine contraception. It is safe and appropriate to provide emergency contraception to all postpartum women before discharge from the hospital.

Financial Disclosure: This study was partially funded by an unrestricted grant from the Packard Foundation. The Packard Foundation is a nonprofit organization. They provided funds for supplies and oral contraceptive pills. Salary support was not provided.

^aCenter for Reproductive Health Research and Policy and Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco and San Francisco General Hospital, San Francisco, California, USA
^bDivision of General Internal Medicine, Department of Medicine, University of California, San Francisco and San Francisco General Hospital, San Francisco, California, USA

The United States has one of the highest rates of abortion of any developed country.¹ Emergency contraception has proven to be safe and effective^2-6 and has the potential to decrease abortion by up to 50%.⁷ However, use remains limited,^8-10 despite extensive nationwide public education campaigns^11-13 and the introduction of two dedicated emergency contraception products. Knowledge about emergency contraception by both patients and providers remains insufficient,^8,9,14-18 and access remains limited because of difficulties obtaining prescriptions¹⁹ or unavailability of emergency contraception pills at local pharmacies.^13,20

Whether emergency contraception can fulfill its potential for decreasing unintended pregnancies depends on women's ability to easily obtain and use it. The majority of US women remain unfamiliar with emergency contraception. Of those reporting any familiarity, fewer than 25% know how to obtain pills and that they must be used within 72 hours of unprotected intercourse.^8,9,12,14,21 Even those who know how to obtain emergency contraception may not be able to secure a prescription and find a pharmacy that stocks it within 72 hours.

Potential strategies to increase access include pharmacist dispensation, over-the-counter availability, and advance provision by health care providers. Advance provision is superior to the other options in that it allows for concurrent contraceptive and sexually transmitted diseases education by medical providers and has been shown to be cost-saving in the United States and Canada.^22,23 In addition, the efficacy of emergency contraception is improved the earlier it is used,^24,25 and women with emergency contraception pills on hand have been shown to start therapy sooner.^26,27 Randomized clinical trials and controlled cohort studies in different countries have shown increased use when a pill supply is provided,^2,26-28 and the only controlled trial in the United States found a three-fold increase in emergency contraception use at 4 months of follow-up in a high-risk adolescent population.¹⁰

Advance provision has remained controversial, however, because some believe it may encourage repeated use, improper use, or contraceptive and sexual risk taking.^16,17,29 A few studies have examined the impact of prophylactic provision of emergency contraception on the use of other contraceptive methods. In a study of adolescents, Raine et al found a statistically significant increase in the use of less effective birth control methods in the advance provision group compared with the education-only group but no difference in overall rates of unprotected intercourse.¹⁰ The largest study of advance provision, performed in Scotland, found that type of contraceptive method was not different between the groups, but consistency of use was not measured.³⁰ Ellertson et al recently published a report from a low-risk population of condom users in India comparing advance provision with education alone showing no difference in condom use.²⁸ However, only 6% of participants reported unprotected intercourse during the study. A thorough evaluation of the impact of emergency contraception provision on both consistency of contraception use and method choice in a US population with adequate follow-up has not been published.

We conducted a randomized controlled trial to examine the effect of advance provision of emergency contraception on contraceptive behaviors, emergency contraception use, and knowledge in a group of high-risk women over the course of 1 year. Our hypotheses were that advance provision would increase knowledge and use of emergency contraception but would not decrease use of effective contraception.

Materials and Methods

We have previously reported on the eligibility criteria, survey instruments used, and baseline findings of the enrolled participants.⁹ Briefly, we invited consecutive postpartum women at a public inner-city hospital from September 1998 through March 1999 to participate. Women were eligible if they had had a live birth, spoke English or Spanish, and would be available for follow-up in 1 year (Figure 1). The majority of exclusions were made because participants did not speak English or Spanish (n = 91), had undergone a postpartum tubal ligation (n = 57), or would be unavailable for follow-up (n = 92). An additional 28 women refused to participate, and 69 were eligible but not approached for enrollment because of an unanticipated early discharge. One woman was enrolled twice on different dates into different groups. For analysis purposes, she was assigned to the emergency contraception group because she had received emergency contraception education and pill supply. Women who were trying to become pregnant were excluded from final analysis. All women gave written consent before participation. The University of California, San Francisco, Committee on Human Research approved the trial.

Figure 1. Study enrollment and follow-up. Jackson. Emergency Contraception. Obstet Gynecol 2003.

Subjects were enrolled on the day of discharge after routine contraceptive counseling had been provided by the medical staff. After determining eligibility and obtaining informed consent, trained bilingual researchers interviewed participants about their knowledge of emergency contraception, previous contraceptive use, pregnancy experiences, and demographic characteristics. Details of the instrument have been previously reported.⁹ Briefly, the instrument was adapted from a Kaiser Family Foundation questionnaire.⁸ Questions regarding emergency contraception were taken verbatim from the Kaiser Family Foundation survey. We added additional questions regarding contraceptive use, safe sex practices, and pregnancy experiences. The questionnaire was administered in person at baseline and by telephone at 6 and 12 months. Surveys required approximately 10 minutes to complete.

After completion of the baseline questionnaire, women assigned to the emergency contraception group were given a standardized, 5-minute educational session, a supply of emergency contraception pills with verbal and written instructions, and an educational brochure. The regimen consisted of one emergency contraception treatment: eight oral contraceptive pills containing 0.15 mg of levonorgestrel and 30 µg of ethinyl estradiol. Included in the educational materials were instructions for obtaining additional emergency contraception pills if needed. Control group subjects received only the routine contraceptive counseling provided to all women before discharge. This counseling was performed by the patient's medical provider and did not usually include a discussion of emergency contraception.

Given the close proximity of patients to each other on Labor and Delivery, education provided to an emergency contraception subject could potentially be overheard by control subjects. To prevent contamination of the control group, we enrolled all women on a given day to the same group. The unit of randomization was, therefore, date of discharge. The daily group assignment was made using a random number generator by a separate researcher such that the individual enrolling subjects on a given day would be unable to predict to which group subjects would be assigned. Researchers conducting baseline interviews also provided the emergency contraception educational session and were thus not masked to group assignment. However, questionnaires were multiple-choice instruments, and a standard script was provided. Study personnel blinded to group assignment performed the follow-up questionnaires, data entry, and analyses.

The primary outcome was self-reported use of emergency contraception. Secondary outcomes included change in use of other contraceptive methods and knowledge about emergency contraception. Contraceptive and sexual behaviors were assessed by asking about types of contraception used and consistency of use. General knowledge about emergency contraception was evaluated with two questions from the Kaiser Family Foundation survey. First, we asked: "If a woman just had sex and thinks she might become pregnant, is there anything she could do in the next few days to prevent pregnancy, or not?" A "yes" response was recorded only if the subject could then correctly name or describe emergency contraception or "morning-after" pills. Familiarity was assessed with the question, "Have you ever heard of morning-after pills, also called emergency contraceptive pills, or not?" Women who had heard of emergency contraception were then asked seven additional questions about the correct timing for the use of emergency contraception, perceptions about safety, mechanism of action, effectiveness, availability, and prior use.

Secondary outcomes involved use and change in use of contraception. Routine use of contraception was defined as self-reported contraceptive use all or most of the time (1 or 2 on a 5-point scale). Less consistent use was defined as a change from more frequent to less frequent use using the 5-point scale or an unplanned pregnancy during the 6- or 12-month follow-up. Contraceptive methods were classified as being very effective (intrauterine device, depot medroxyprogesterone acetate, interval sterilization, levornorgestrel implants, and oral contraceptives) versus poorly effective (barrier, withdrawal, rhythm, and none). Less effective method was defined as a change from very to poorly effective methods during follow-up.

We calculated our sample size based on the hypothesis that advance provision of emergency contraception would increase its use from a baseline of 1.5% to 10%. These numbers were based on reported use in the United States of 1-5%⁸ and increased use with advance provision in the United Kingdom from a baseline of 27% to 47%.³⁰ Detecting an increase to 10% from 1.5% with a two-tailed alpha of 0.05 and 80% power would require a sample size of 140 per group. Assuming 30% loss to follow-up, we planned for 182 per group.

Our secondary hypothesis was that reliable contraceptive use would not be markedly affected by advance provision of emergency contraception. Given a prior study showing that 37% of women who started oral contraceptive pills switched to less reliable methods within 6 months,³¹ we estimated 25-40% would switch to less reliable contraception within 1 year. Using this as our baseline, a sample size of 140 per group would allow us to detect a 20% change in the proportion switching to less reliable use with 80% power.

We examined differences between groups and differences within each group over time. Differences between the emergency contraception and standard care groups were analyzed using the Fisher exact test or Student t test. Trends within groups were analyzed using the McNemar test for binary outcomes and with the paired t test for continuous outcomes. Analyses were performed with the individual subject as the unit of analysis. We repeated the analyses taking into account cluster sampling by date of discharge. We used a logistic normal likelihood ratio test using the Huber/White robust estimate of variance^32,33 taking into account clustered sampling.^34,35 We compared odds ratios calculated with and without adjusting for cluster sampling. Variables were considered independently associated with the outcome if significant at P = .05. Results are presented as relative risks (RR) with 95% confidence intervals (CI). A risk ratio of greater than one indicates an increased likelihood of the outcome in the intervention group.

All analyses were conducted using STATA Statistical Software 6.0 (StataCorp, 2000, College Station, TX).

Results

We enrolled 370 postpartum women; 184 were assigned to the emergency contraception group and 186 to the usual care (control) group (Figure 1). At 6 months postpartum, follow-up was available for 78% and at 1 year, 69%. Overall, 85% were available for at least one follow-up session. There were no differences between groups in the proportion lost to follow-up; nor were there differences in baseline traits between those lost to follow-up and those who completed the study.

Demographic characteristics were similar between the groups at enrollment (Table 1) and among those who followed up (data not shown). The majority of subjects were Latina (72%). The average age was 25.6 years; 18% were teens. Most were married or living as married (73%). Thirty-eight percent reported a prior unwanted pregnancy and 17% a prior elective abortion. Two-thirds stated that their most recent pregnancy was unplanned; 29% of these resulted from a contraceptive failure.

At baseline, 11 women (3%) had reported use of emergency contraception in the past (Table 2). Although 36% of the population had heard of emergency contraception or morning-after pills, only 19% could name or describe a method to prevent pregnancy after sex, and only 7% knew the correct timing for its use.

Click for Table 2

Women who received advance provision of emergency contraception were significantly more likely to have used it during follow-up (Table 2). Over the course of the year, 17% of women in the intervention group used emergency contraception compared with 4% in the control group (RR 4.0; 95% CI 1.8, 9.0). In total, there were 23 new users of emergency contraception in the emergency contraception group compared with five in the control group (RR 4.9; 95% CI 1.9, 12.6). Only five women used multiple doses of emergency contraception over the 12-month period, and three were in the advance provision group. Among women who reported at least one episode of unprotected intercourse during the year, 25% used emergency contraception in the advance provision group compared with 4% in the control group (RR 5.8; 95% CI 2.1, 16.4). There were 16 (10%) unplanned pregnancies reported in the control group compared with 11 (7%) in the emergency contraception group (P = .16). Emergency contraception use was reported by four of the women who became pregnant (all in the emergency contraception group); one may have had an emergency contraception failure, two used it incorrectly, and in one, we had insufficient information to determine if it had been used correctly.
Significant improvement in contraceptive use was observed during follow-up in both the control and intervention groups (Table 3). Three-quarters of participants changed birth control methods over 1 year with over half in each group switching to methods that were more effective. In addition, there were substantial improvements in the consistency of contraceptive use by women in both groups. At baseline, only 35% of each group reported routine use of contraception at a time when they did not desire pregnancy, whereas at follow-up, more than 80% did (P < .05). Notably, there were no differences detected between the advance provision and usual care groups on any measure of consistency of contraceptive use, type of method, or change in use over the study period. Among exclusive condom users, there was no decrease in the use of condoms in the emergency contraception group. Over half the women in both groups reported at least one episode of unprotected intercourse during the year with no significant differences over time or between groups.

Click for Table 3

General and specific knowledge about emergency contraception was significantly better in the advance provision group for nearly all the measures examined. Risk ratios for knowledge measures ranged from 1.3 to 3.8 (Table 2). For both groups, knowledge measures were better at 6 months compared with at baseline (P < .05). In the emergency contraception group, however, knowledge plateaued between 6 and 12 months, whereas in the control group, most knowledge indicators continued to show improvement (P < .05). For example, at baseline, 22% in the control group knew a prescription was required to obtain emergency contraception. At 6 months, it increased to 30% and at 1 year, 41%. Neither willingness to use emergency contraception nor objection to its use differed between the groups nor changed over time.
Primary analyses were performed with the individual subject as the unit of analysis. We repeated the analyses taking into account cluster sampling by date of discharge (Table 4) and found no appreciable difference in odds ratios before and after adjustment for clustered sampling.

Discussion
Our study confirms prior research showing increased use of emergency contraception in women given a prophylactic supply.^10,26-28,30 Although we found a three- to six-fold increased use of emergency contraception in the advance provision group, the actual rates of use were lower than those reported by other authors. Approximately 10% of women in the advance provision group used emergency contraception during each 6-month period compared with 2-3% in the control group. Similarly, among women reporting unprotected intercourse during the study period, those given an emergency contraception supply were six times more likely to have used them. However, this corresponds to emergency contraception use by only 25% of those who reported unprotected intercourse.
These low rates of use suggest that ready access is not the only issue. A lack of recognition of pregnancy risk has been shown to limit emergency contraception use.³⁶ In our population of postpartum women, who may have been breast-feeding and/or amenorrheic, this is certainly a possibility. Analysis of our baseline data found increased willingness to use emergency contraception among those who knew that it is safe and that it is not an abortifacient.⁹ Education about emergency contraception should, therefore, emphasize the method's safety and that its mechanism of action is through ovulation delay and prevention of implantation and not abortion.³⁷ Further study needs to focus on what motivates and/or prevents women from appropriately using emergency contraception.
We found no evidence for less effective contraceptive use in women given advance provision of emergency contraception. We examined a variety of measures of contraceptive method choice and consistency of use and found no differences between groups for any measure. Although more than three-quarters of participants changed contraceptive methods over the year-long study, switching to a less effective birth control method was not more common in the advance provision group, as was described in a US study of adolescents.¹⁰ On the contrary, as reported by others,^27,30,38 consistency of contraception use and method selection improved in both groups after educational intervention or emergency contraception use. This improvement is likely a result of a number of factors such as increased exposure to health care during pregnancy and postpartum, routine contraceptive education and prescription given at the time of discharge, and increased motivation to prevent or delay further pregnancies. Given the large proportion of women who change birth control methods, a strategy of distributing emergency contraception only to women who use barrier methods would be inadequate. Finally, very few women in the treatment group used emergency contraception more than once. However, they were provided only one emergency contraception treatment. Therefore, our study does not address whether providing women with easy access, such as making emergency contraception available over the counter, might cause women to use it repeatedly.
As expected, women in the emergency contraception group demonstrated greater knowledge about it at follow-up. However, only one-quarter of women in the emergency contraception group could state the correct timing for use. In a similar population of inner-city women at a family planning clinic given an advance supply of emergency contraception, only 18% used it correctly.²¹ This reinforces the need for specific education about timing of emergency contraception use, as timing has clearly been shown to affect efficacy.
Unlike other studies of advance provision wherein the comparison group received emergency contraception education, our comparison group received no specific emergency contraception education. This allowed us to estimate the effect of public and provider education programs on emergency contraception knowledge. We found a marked increase in general awareness of emergency contraception in the control group over the course of 1 year, but detailed knowledge improved only modestly: 40% knew a prescription was needed, but only 10% knew the correct timing, and 5% knew both. Unfortunately, this increase in knowledge did not translate into an increase in use.
There are limits to the generalizability of the study findings. Our population of mostly Latina, low-income, postpartum women is not representative of the US population. However, given that two-thirds had unplanned pregnancies and that about half continued to practice unsafe sex in the year after delivery, they are a population at high risk for unintended pregnancy. In addition, discussion of birth control routinely occurs before discharge providing a convenient time for discussion and/or distribution of emergency contraception. A second potential limitation is that personnel who administered the baseline questionnaire were not masked to group assignment. Given the standard questionnaire and interviewing script, this should have had little effect on results. Furthermore, individuals masked to study group assignment performed all subsequent interviews, data collection, and analysis. We rely on self-reported data, which can be erroneous, and might not reflect actual behaviors. The most reliable way to assess whether emergency contraception provision affects contraception use or sexual behaviors would be to measure unintended pregnancy and sexually transmitted diseases infection rates. Although the incidence of unintended pregnancy was lower among women who had emergency contraception available at home, the difference did not reach statistical significance. A much larger study with longer follow-up time would be required to assess these outcomes.
The requirement for a prescription to obtain emergency contraception poses a formidable barrier to its prompt, effective use. This study supports the feasibility and safety of advance provision of emergency contraception in postpartum women at the time of hospital discharge. With the recent introduction of a progestin-only method, Plan B(r), side effects are decreased and efficacy is increased in comparison with the Yuzpe method (high doses of combined oral contraceptives). Given that emergency contraception meets customary criteria for over-the-counter use,³⁹ over-the-counter availability may eventually become a reality. Until then, an emergency contraception supply and appropriate education should routinely be offered to all postpartum women before discharge from the hospital.

References
1. Henshaw SK. Abortion incidence and services in the United States, 1995-1996. Fam Plann Perspect 1998;30:263-70, 87.
2. Glasier A. Safety of emergency contraception. J Am Med Womens Assoc 1998;53(5 Suppl 2):219-21.
3. Trussell J, Rodriguez G, Ellertson C. Updated estimates of the effectiveness of the Yuzpe regimen of emergency contraception. Contraception 1999;59:147-51.
4. Espinos JJ, Senosiain R, Aura M, Vanrell C, Armengol J, Cuberas N, et al. Safety and effectiveness of hormonal postcoital contraception: A prospective study. Eur J Contracept Reprod Health Care 1999;4:27-33.
5. Vasilakis C, Jick SS, Jick H. The risk of venous thromboembolism in users of postcoital contraceptive pills. Contraception 1999;59:79-83.
6. Creinin MD. A reassessment of efficacy of the Yuzpe regimen of emergency contraception. Hum Reprod 1997;12:496-8.
7. Trussell J, Stewart F, Guest F, Hatcher RA. Emergency contraceptive pills: A simple proposal to reduce unintended pregnancies. Fam Plann Perspect 1992;24:269-73.
8. Delbanco SF, Mauldon J, Smith MD. Little knowledge and limited practice: Emergency contraceptive pills, the public, and the obstetrician-gynecologist. Obstet Gynecol 1997;89:1006-11.
9. Jackson R, Schwarz E, Freedman L, Darney P. Knowledge and willingness to use emergency contraception among low-income post-partum women. Contraception 2000;61:351-7.
10. Raine T, Harper C, Leon K, Darney P. Emergency contraception: Advance provision in a young, high-risk clinic population. Obstet Gynecol 2000;96:1-7.
11. Trussell J, Bull J, Koenig J, Bass M, Allina A, Gamble VN. Call 1-888-NOT-2-LATE: Promoting emergency contraception in the United States. J Am Med Womens Assoc 1998;53(5 Suppl 2):247-50.
12. Delbanco SF, Stewart FH, Koenig JD, Parker ML, Hoff T, McIntosh M. Are we making progress with emergency contraception? Recent findings on American adults and health professionals. J Am Med Womens Assoc 1998;53(5 Suppl 2):242-6.
13. Coeytaux F, Pillsbury B. Bringing emergency contraception to American women: The history and remaining challenges. Womens Health Issues 2001;11:80-6.
14. van Royen AR, Calvin CK, Lightner CR. Knowledge and attitudes about emergency contraception in a military population. Obstet Gynecol 2000;96:921-5.
15. Sills MR, Chamberlain JM, Teach SJ. The associations among pediatricians' knowledge, attitudes, and practices regarding emergency contraception. Pediatrics 2000;105:954-6.
16. Sherman CA, Harvey SM, Beckman LJ, Petitti DB. Emergency contraception: Knowledge and attitudes of health care providers in a health maintenance organization. Womens Health Issues 2001;11:448-57.
17. Golden NH, Seigel WM, Fisher M, Schneider M, Quijano E, Suss A, et al. Emergency contraception: Pediatricians' knowledge, attitudes, and opinions. Pediatrics 2001;107:287-92.
18. Beckman LJ, Harvey SM, Sherman CA, Petitti DB. Changes in providers' views and practices about emergency contraception with education. Obstet Gynecol 2001;97:942-6.
19. Trussell J, Duran V, Shochet T, Moore K. Access to emergency contraception. Obstet Gynecol 2000;95:267-70.
20. American College of Obstetricians and Gynecologists. Pharmacists limit women's access to emergency contraception, press release. Washington, DC: American College of Obstetricians and Gynecologists, 1999
21. Endres LK, Beshara M, Sondheimer S. Experience with self-administered emergency contraception in a low-income, inner-city family planning program. J Reprod Med 2000;45:827-30.
22. Trussell J, Koenig J, Ellertson C, Stewart F. Preventing unintended pregnancy: The cost-effectiveness of three methods of emergency contraception. Am J Public Health 1997;87:932-7.
23. Trussell J, Wiebe E, Shochet T, Guilbert E. Cost savings from emergency contraceptive pills in Canada. Obstet Gynecol 2001;97:789-93.
24. Regulation TFoPMoF. Randomised controlled trial of levonorgestrel versus the Yuzpe regimen of combined oral contraceptives for emergency contraception. Lancet 1998;352:428-33.
25. Piaggio G, von Hertzen H, Grimes DA, Van Look PF. Timing of emergency contraception with levonorgestrel or the Yuzpe regimen. Task Force on Postovulatory Methods of Fertility Regulation. Lancet 1999;353:721.
26. Skibiak J, Ahmed Y, Ketata M. Testing strategies to improve access to emergency contraception pills: Prescription vs. prophylactic distribution. New York: Population Council, 1999.
27. Lovvorn A, Nerquaye-Tetteh J, Glover EK, Amankwah-Poku A, Hays M, Raymond E. Provision of emergency contraceptive pills to spermicide users in ghana. Contraception 2000;61:287-93.
28. Ellertson C, Ambardekar S, Hedley A, Coyaji K, Trussell J, Blanchard K. Emergency contraception: Randomized comparison of advance provision and information only. Obstet Gynecol 2001;98:570-5.
29. Gold MA, Schein A, Coupey SM. Emergency contraception: A national survey of adolescent health experts. Fam Plann Perspect 1997;29:15-9, 24.
30. Glasier A, Baird D. The effects of self-administering emergency contraception [see comments]. N Engl J Med 1998;339:1-4.
31. Rosenberg MJ, Waugh MS. Oral contraceptive discontin-uation: A prospective evaluation of frequency and reasons. Am J Obstet Gynecol 1998;179:577-82.
32. Huber PJ. The behavior of maximum likelihood estimates under non-standard conditions. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, 1967, Berkeley, CA
33. White H. Maximum likelihood estimation of misspecified models. Econometrica 1982;50:1-25.
34. Rogers WH. Regression standard errors in clustered samples. Stata Tech Bull 1993;13:19-23.
35. Donner A, Klar N. Analysis of binary outcomes. Design and analysis of cluster randomization trials in health research. London: Arnold, 2000:92-5.
36. Sorensen MB, Pedersen BL, Nyrnberg LE. Differences between users and non-users of emergency contraception after a recognized unprotected intercourse. Contraception 2000;62:1-3.
37. Swahn ML, Westlund P, Johannisson E, Bygdeman M. Effect of post-coital contraceptive methods on the endometrium and the menstrual cycle. Acta Obstet Gynecol Scand 1996;75:738-44.
38. Rowlands S, Devalia H, Lawrenson R, Logie J, Ineichen B. Repeated use of hormonal emergency contraception by younger women in the UK. Br J Fam Plann 2000;26:138-43.
39. Grimes DA, Raymond EG, Scott Jones B. Emergency contraception over-the-counter: The medical and legal imperatives. Obstet Gynecol 2001;98:151-5.

Address reprint requests to: Rebecca A. Jackson, MD, San Francisco General Hospital, Department of Obstetrics and Gynecology, #6D, 1001 Potrero Avenue, San Francisco, CA 94110; E-mail: jacksonr@obgyn.ucsf.edu

Received October 28, 2002.
Received in revised form March 26, 2003.
Accepted April 2, 2003.

Copyright © 2003 by The American College of Obstetricians and Gynecologists
Published by Elsevier Science Inc.
Visit Obstetrics & Gynecology online at http://www.greenjournal.org