Despite the high efficacy and safety of early medical abortion using a combined regimen of mifepristone plus misoprostol, clinic follow-up remains an integral part of standard medical abortion service delivery globally.1,2 In-person clinic visits can be costly and time-consuming for both the woman and the health care system. In many clinics, a transvaginal ultrasound examination is performed routinely. Some women find that the additional follow-up visit decreases the acceptability of medical abortion1 and it is not uncommon for women to fail to return for the 2-week follow-up appointment.
Preprocedure and postprocedure system of serum human chorionic gonadotropin (hCG) measures are accurate in identifying ongoing pregnancies after medical abortion.3–5 Quantitative serum hCG testing to confirm successful medical abortion is commonly used in the United States and Europe and recommended as a means of follow-up by the Royal College of Obstetricians and Gynecologists and the World Health Organization.6,7
Urine pregnancy tests offer another way to measure hCG levels after medical abortion.8–10 Semiquantitative pregnancy tests have been used successfully in two subsequently published medical abortion studies in the United States and Vietnam. In both studies, the test correctly identified all of the ongoing pregnancies among study participants.11,12
The goal of this study was to test the efficacy and feasibility of a service delivery protocol that replaces the routine clinic visit with ultrasonographic examination with sequential semiquantitative pregnancy tests and a standardized self-administered follow-up questionnaire administered over the phone.
MATERIALS AND METHODS
This multicenter study enrolled women seeking early medical abortion at four hospitals (Hung Vuong Hospital, National Ob-Gyn Hospital, Hanoi Ob-Gyn Hospital, and HocMon General District Hospital) in Vietnam. The most common treatment regimen used for early medical abortion services at the hospitals consists of oral 200 mg mifepristone followed in 24–48 hours by 800 micrograms buccal misoprostol administered at home. Women presenting for early medical abortion at one of these four facilities were screened for participation and those meeting the eligibility criteria were enrolled. Eligibility criteria included intrauterine pregnancy less than or equal to 63 days of gestation; no known contraindications to abortion with mifepristone, misoprostol, or both; general good health; ability to provide informed consent and willingness to return for follow-up; literate and able to complete an at-home symptom checklist; and possessing a working, personal phone number. Women not eligible or choosing not to participate in the study received standard clinic follow-up.
Participants were randomly assigned to one of two study groups: 1) clinic follow-up (eg, standard care); or 2) phone follow-up. Standard care entailed a clinic visit 2 weeks after mifepristone administration. At the visit, each participant was interviewed about her clinical condition and underwent bimanual examination and transvaginal ultrasonography to confirm her abortion status. In the event of an ongoing pregnancy, immediate surgical evacuation was offered. Women with incomplete abortion at follow-up were given the option of immediate surgical completion or an additional dose of misoprostol followed by a clinic visit 1 week later. All additional clinic visits, additional treatment received, or both were recorded on study forms. Once complete abortion was ascertained, participants were asked about their future preferences and satisfaction and acceptability with the follow-up procedure and released from the study.
Women randomized to phone follow-up were scheduled for a telephone call 2 weeks after mifepristone administration. Women assigned to phone follow-up were asked to use a urine-based semiquantitative pregnancy test before swallowing mifepristone to locate the range of their hCG before the abortion (baseline). In contrast to typical urine-based pregnancy tests, the semiquantitative pregnancy test provides estimates of hCG levels by individual strips reading sequential ranges of hCG: at least 25 milli-international units/mL, at least 100 milli-international units/mL, at least 500 milli-international units/mL, at least 2,000 milli-international units/mL, and at least 10,000 milli-international units/mL. A study nurse explained how to perform the semiquantitative pregnancy test and provided each participant with an additional test, a urine sample cup, an information sheet explaining how to perform and interpret the test, and a questionnaire to use at home. Tests were conducted by dipping the test panel into the urine sample cup and stirring it for a minimum of 10 seconds to achieve strip saturation. Either first morning or random urine could be used. The tests were then placed face up on a flat surface and read approximately 15 minutes later. Each woman's baseline hCG as determined by the semiquantitative test was noted on her at-home symptom checklist. Participants were instructed to complete the pregnancy test and the symptom checklist before their scheduled telephone follow-up weeks later. On the assigned date, participants were contacted by phone by the clinic staff. During the call, a study nurse reviewed the pregnancy test result and checklist responses with the woman. The pregnancy test result at follow-up was compared with the baseline result and a drop of at least one bracket was considered a success. In the event that she had not performed the pregnancy test at the time of the call, she was asked to perform the test immediately and the nurse would call back in approximately 30 minutes.
The symptom checklist used was derived from an earlier study12 and included the following questions: 1) Did you experience less than 2 days of heavy bleeding during treatment? 2) Are you experiencing any of the following symptoms now: breast tenderness, nausea or morning sickness, frequent urination, or exhaustion or tiredness or both? 3) Do you still feel pregnant? If the woman answered “yes” to any of the questions on the checklist or her pregnancy test was invalid or indicated an increase or no change in the hCG level, she was asked to return to the clinic for follow-up. If a woman screened negative for referral, she was released from the study after the call. Participants who screened positive for clinic referral were asked by the health care provider to return for any reason or requested a clinic visit were scheduled for additional clinic follow-up. During the follow-up phone call, participants were also asked about the satisfaction and acceptability of the phone follow-up procedure.
Participants were randomized to a study group by opening the next sequentially numbered sealed opaque envelope. The envelopes were prepared by staff at Gynuity Health Projects in New York. Randomization was stratified by study site using random blocks of eight to ensure equal distribution in the two study groups. Regardless of study group, any participant failing to return for the follow-up visit or make the scheduled telephone follow-up call was contacted by telephone up to three times by a study provider.
The sample size was based on the least common outcome of interest: ongoing pregnancy after medical abortion. Based on previous studies conducted in Vietnam, we estimated that ongoing pregnancy would occur in approximately 2.0% of medical abortion cases.12,13 Assuming that routine clinic-based follow-up will diagnose 100% of ongoing pregnancies, we designed the study to have 80% power to demonstrate that phone follow-up (using a semiquantitative pregnancy test in combination with a self-administered checklist) would be as effective as standard clinic-based follow-up in detecting ongoing pregnancies with an α of 0.05. To achieve this, we estimated that 28 ongoing pregnancies (14 in each study arm) would be needed, with a total sample size of 1,400 participants (700 in each study arm).
Data entry and analysis were performed using SPSS 15.0. Study data were entered and cleaned in Vietnam for analysis jointly by the investigators in Vietnam and New York. The two study groups were compared using t tests or the Mann-Whitney U test for continuous variables and χ2 or Fisher’s exact tests for categorical variables. The main study outcomes were compared using relative risks (RRs), 95% confidence intervals (CIs), and P values. A P value <.05 was considered statistically significant. Sensitivity and specificity of the pregnancy test and checklist in detecting ongoing pregnancy was also calculated. Ethical approvals were obtained at Hung Vuong Hospital, Hoc Mon General District Hospital in Ho Chi Minh City, and National Hospital in OBGYN and Hanoi OBGYN Hospital in Hanoi, Vietnam, and the trial is registered at clinical trials.gov as NCT01150422.
One thousand four hundred thirty-three women were enrolled at the study sites between May 26, 2010, and April 14, 2011. Seven hundred thirteen women were randomized to the phone follow-up group and 720 women assigned to clinic follow-up (Fig. 1). There was no statistically significant difference between the two groups (Table 1). The medical abortion outcomes for the two groups are presented in Table 2. There were no statistically significant differences in rates of complete abortion without surgical evacuation (clinic: 94.6% [626/662], phone: 94.8% [672/709], RR 1.002, 95% CI 0.977–1.028) or ongoing pregnancy between the two groups (clinic: 2.7% [18/720], phone: 2.5% [18/713], RR 0.934, 95% CI 0.985–1.019).
In the phone follow-up group, 2.2% of women (16/713) did not complete the scheduled follow-up call because they came to the clinic early (Table 2; Fig. 1). Seven women had a complete abortion without surgical evacuation. Four women underwent a uterine evacuation for a continuing pregnancy. Five participants had aspirations for other reasons including profuse bleeding (n=1) and woman's request (n=4). One woman made an interim visit to the clinic 10 days after mifepristone administration and was hospitalized for treatment of a suspected infection. She received intravenous antibiotics and her fever resolved after 4 days.
In the clinic follow-up group, 1.4% of women (10/720) made an interim visit to the study clinic before their scheduled appointment (Table 2; Fig. 1). Half of women who returned early (5/10) underwent a uterine evacuation for an ongoing pregnancy confirmed by ultrasonography. Two women underwent an evacuation for retained products of conception. The remaining women (n=3) had a successful abortion confirmed by ultrasonography.
Most women in both study arms completed follow-up. In the phone follow-up group, almost all women who did not make an interim visit to the clinic had their telephone follow-up as scheduled (99.4% [693/697]). Four women (0.6% [4/697]) could not be contacted and were classified as lost to follow-up. In the clinic follow-up group, 8.1% (58/710) did not return for their scheduled 2-week clinic visit. The difference between the two study groups in the proportion of women lost to follow-up (0.6% compared with 8.2%) was statistically significant (RR 0.070, 95% CI 0.025–0.191).
In the clinic follow-up group, most women who returned for a scheduled appointment had a successful medical abortion confirmed by ultrasonography (95.5% [623/652]) (Fig. 1). At this visit, 13 women received a uterine aspiration for an ongoing pregnancy (1.9% [13/652]). Fifteen women underwent an aspiration for other medical reasons and one woman requested a surgical evacuation.
Most women in the phone follow-up group who were contacted were not referred to the clinic for additional medical abortion related care (84.7% [587/693]) (Fig. 1). In the phone follow-up group, only 15% of women (106/693) were referred to the clinic for additional follow-up care after their phone assessment (Fig. 1). More than one third of these women referred reported either no change or an increase in hCG from the time of mifepristone administration (39.6% [42/106]). More than half of these women reported a positive response to one of the checklist questions (56.6% [60/106]). Almost one fourth of women referred to the clinic for follow-up (25.4% [27/106]) screened negative on both tests, ie, reported a decrease in hCG from the time of mifepristone administration and answered “no” to all of the symptom checklist questions. Among the women referred without a positive screen test (n=27), the reported reasons for referral included ongoing bleeding (25.9% [7/27]), suspected ongoing pregnancy based on other symptoms (25.9% [7/27]), inconclusive pregnancy test (18.5% [5/27]), woman felt anxious (11.1% [3/27]), or other reasons (18.5% [5/27]) (data not shown).
The outcomes of the 106 women in the phone follow-up group who were referred to the clinic for further follow-up are presented in Figure 1. All women who were referred returned to the clinic. Most women who were referred (76.4% [81/106]) had successful medical abortions confirmed by ultrasonography. Thirteen women underwent a uterine evacuation for an ongoing pregnancy confirmed by ultrasonography. Twelve women underwent a surgical evacuation for other medical reasons. All nine cases of missed abortion were identified by no decrease or change in hCG from initial baseline at the time of the scheduled phone follow-up call.
Three women who screened negative and were not asked to return to the clinic for follow-up subsequently received additional medical abortion-related care. In one case, the woman reported in her phone interview that she had experienced more than 2 days of bleeding, had no pregnancy symptoms, and that her pregnancy test result showed a decrease in hCG level. Still, she returned to the clinic 5 days later and was diagnosed with an ongoing pregnancy. Another woman underwent an aspiration for retained products of conception when she returned to the clinic 9 days after her phone interview (study day 23). A third woman returned to the clinic 10 days after her phone call (study day 24) and requested a surgical evacuation for management of ongoing bleeding.
Most women (98.8% [685/693]) who did not make an interim visit to the clinic and were successfully contacted by phone had a functional pregnancy test result recorded at the phone follow-up. In eight cases, the test line did not appear and the result was not readable. Among women with a functional pregnancy test (n=685), the sensitivity and specificity of the pregnancy test alone for the identification of ongoing pregnancy was 92.8% and 95.7%, respectively (Table 3). The specificity was slightly lower with the use of the pregnancy test and symptom checklist (90.6%) and the screen-positive rate slightly higher (11.1% for the combined symptom checklist and pregnancy test compared with 6.1% for the pregnancy test alone).
Fourteen women in the phone follow-up group who did not make an interim visit to the clinic were diagnosed with an ongoing pregnancy at or after the follow-up visit. The results of the two follow-up screening methods are shown in Table 4. The semiquantitative pregnancy test identified 92.8% (13/14) of ongoing pregnancies. The symptom checklist alone identified 71.4% (10/14) of ongoing pregnancies. One woman with an ongoing pregnancy was not identified by either the checklist or the pregnancy test.
We also examined the estimated hCG levels reported by women on the day of their follow-up call. More than one third of women (37.8% [259/685]) reported a result of less than 25 milli-international units/mL. Most women (94.7% [649/685]) reported an hCG level of less than 2,000 milli-international units/mL when they administered the second urine pregnancy test. Most women who were diagnosed with an ongoing pregnancy on or after the follow-up appointment (n=14) reported an hCG level of 10,000 milli-international units/mL or above (85.7% [12/14]). However, two women reported lower levels of at least 500 milli-international units/mL and at least 25 milli-international units/mL, respectively.
Women in the phone follow-up group were asked about the feasibility and acceptability of the pregnancy test and symptom checklist during their follow-up phone call (Table 5). Most participants completed the checklist (98.3% [681/693]) and the semiquantitative pregnancy test (99.1% [687/693]). Almost all women reported that completion of the checklist and pregnancy test was easy or very easy (checklist: 97.3% [673/691]; semiquantitative pregnancy test: 97.5% [674/691]).
Women in both groups were asked about their preference for follow-up should they require another medical abortion in the future (Table 5). In the phone follow-up group, most women (88.3% [606/686]) indicated that they preferred to complete their follow-up at home with a phone call from the health care provider. In the clinic follow-up group, fewer women (39.9% [256/242]) reported that they would opt for phone follow-up at home in the future.
Phone follow-up after medical abortion is a safe and effective alternative to clinic-based follow-up after medical abortion. The data from this study show that phone follow-up with a semiquantitative urine pregnancy test and symptom checklist could enable 85% of women to avoid a routine follow-up visit without any decrease in safety. The semiquantitative urine pregnancy test and symptom checklist accurately screened for ongoing pregnancy and resulted in fewer women being lost to follow-up.
Women reported the hCG levels measured with the semiquantitative pregnancy test at 2 weeks after mifepristone administration. A follow-up system using only a high sensitivity pregnancy test at follow-up could successfully identify all ongoing pregnancies but would still require nearly two thirds of women to return to the clinic for follow-up. In contrast, only 5.3% of women reported a value greater than 2,000 milli-international units/mL at the follow-up call, but two women diagnosed with an ongoing pregnancy reported levels of at least 500 milli-international units/mL and at least 25 milli-international units/mL. A follow-up scheme using a single low-sensitivity pregnancy test would permit fewer women to return to the clinic but could still fail to identify cases of ongoing pregnancy after medical abortion.
The semiquantitative urine pregnancy test was more successful at screening for ongoing pregnancy earlier than high- and low-sensitivity urine pregnancy tests used in previous research.9 In the present study, women were able to confirm their abortion outcome 2 weeks after mifepristone administration with repeat semiquantitative pregnancy tests with a higher sensitivity and specificity (92.8% and 95.7%, respectively). Two previous studies have also documented that the semiquantitative pregnancy test effectively detects continuing pregnancy when performed 1–2 weeks mifepristone administration.11,12
Although the phone follow-up failed to identify one woman who was later identified with an ongoing pregnancy, she still returned to the clinic for follow-up but was not clear about why she had come. Counseling, together with the semiquantitative pregnancy test, may help to ensure that women can safely and effectively assess their postabortion status at home and identify when they need to return.
This study used both the semiquantitative pregnancy test and a symptom checklist to help women assess their postabortion status. The semiquantitative pregnancy test was highly effective at identifying ongoing pregnancies. When used with the semiquantitative pregnancy test, the symptom checklist appears to offer no additional value for identifying cases of ongoing pregnancy. Indeed, the one case of ongoing pregnancy missed in the phone follow-up group was not detected by either the participant's self-reported pregnancy test or symptom checklist (Table 4). The results from this study confirm previous reports that symptomatic assessment of ongoing pregnancy is not highly specific.14
The study has several limitations. The results obtained in a single country, Vietnam, may not be generalizable to other settings. The large majority of participants reported having completed at least secondary school education; results may differ among women with lower levels of educational attainment. It is possible, although highly unlikely, that some of the participants in the phone follow-up group may have sought additional abortion care elsewhere. Given that all outcomes, including the number of ongoing pregnancies, were similar in the two randomized study groups, we feel that ascertainment of ongoing pregnancy was equally good in both groups.
Additional research should seek to further simplify service delivery, perhaps with pictorial phone-based assessment of the pregnancy test results for various settings. It might also be possible to document the downward trend in hCG earlier than 2 weeks postmifepristone. If so, this would allow for more expeditious integration of other aspects of service delivery such as provision of postabortion contraception, which also deserves greater attention. More documentation about novel service delivery modalities such as the one described in this article should contribute to international efforts to encourage task shifting in medical care. In the case of medical abortion follow-up, this could mean giving women more of the responsibility to manage their follow-up at home.
The introduction of semiquantitative tests in lieu of universal clinic-based follow-up would simplify care and potentially reduce costs for women and the health care system by providing an alternative follow-up tool to confirm the absence of an ongoing pregnancy. Such efforts may be especially beneficial in resource-poor countries, where clinics may be overcrowded and understaffed with limited access to serum hCG testing and transvaginal ultrasonography.
1. Clark W, Gold M, Grossman D, Winikoff B. Can mifepristone medical abortion be simplified? A review of the evidence and questions for future research. Contraception 2007;75:245–50.
2. Grossman D, Ellertson C, Grimes D, Walker D. Routine follow-up visits after first-trimester induced abortion. Obstet Gynecol 2004;103:738–45.
3. Paul M, Schaff E, Nichols M. The roles of clinical assessment, human chorionic gonadotropin assays and ultrasonography in medical abortion practice. Am Obstet Gynecol 2000;183(suppl 2):S34–43.
4. Schaff E, Walker K, Fielding S, Fuller L. Monitoring serum chorionic gonadotropin levels after mifepristone abortion. Contraception 2001;64:271–3.
5. Creinin MD. Change in serum beta-human chorionic gonadotropin after abortion with methotrexate and misoprostol. Am Obstet Gynecol 1996;174:776–8.
6. Royal College of Obstetrics and Gynaecologists. The Care of Women Requesting Induced Abortion. London: RCOG Press, 2011.
7. Safe abortion: technical and policy guidance for health systems. Geneva (Switzerland): World Health Organization; 2012.
8. Clark W, Bracken H, Tanenhaus J, Schweikert S, Lichtenberg S, Winikoff B. Alternatives to a routine follow-up visit for early medical abortion. Obstet Gynecol 2010;115:264–72.
9. Cameron S, Glasier A, Dewart H, Johnson A, Burnside A. Telephone follow-up and self-performed urine pregnancy testing after early medical abortion: a service evaluation. Contraception 2012;86:67–73.
10. Grossman D, Grindlay K. Alternatives to ultrasound for follow-up after medication abortion: a systematic review. Contraception 2011;83:504–10.
11. Blum J, Shochet T, Lynd K, Lichtenberg ES, Fischer D, Arnesen M, et al.. Can at-home semi-quantitative pregnancy tests serve as a replacement for clinical follow-up of medical abortion: a United States study. Contraception 2012;86:757–62–.
12. Lynd K, Blum J, Ngoc NTN, Shochet T, Blumenthal P, Winikoff B. Simplified medical abortion using a semi-quantitative pregnancy test for home-based follow-up.Int J Gynaecol Obstet 2013;121:144–8.
13. Blum J, Raghavan S, Dabash R, Ngoc NTN, Chelli H, Hajri S, et al.. Comparison of misoprostol-only and combined mifepristone–misoprostol regimens for home-based early medical abortion in Tunisia and Vietnam. Int J Gynaecol Obstet 2012;118:166–71.
14. Jackson A, Dayananda I, Fortin J, Fitzmaurice G, Goldberg A. Can women accurately assess the outcome of medical abortion based on symptoms alone? Contraception 2012;85:192–7.