The most prevalent medication abortion regimen in the United States involves taking mifepristone and misoprostol. Although the U.S. Food and Drug Administration (FDA) recently eliminated the in-person dispensing requirement for mifepristone, they still require it to be prescribed by or under the supervision of a certified health care professional who meets certain qualifications.1 Medication abortion is an ideal candidate for over-the-counter (OTC) use.2 The medications are extremely safe, effective, and not toxic; have low potential for abuse; people can self-screen for eligibility and contraindications; and laboratory testing and ultrasonography are not required.3–7 As more people access medication abortion without the in-person visit, evidence supporting an OTC switch grows.
Telemedicine, mail-order, and online models of care that reduce or eliminate the in-person visit are proving to be as safe, acceptable, and effective as in-person care.4,8–15 Given medication abortion's established safety record6,16 and the need to make abortion more accessible,17–19 research assessing whether it should be moved OTC is warranted. For an OTC switch, the FDA requires label comprehension, self-selection, and actual-use studies demonstrating safe use of the medications without clinical supervision.7 Label-comprehension studies assess whether people can understand key concepts in the drug facts label, such as indications for use, dosing regimen, eligibility, contraindications, risks, warning signs, and side effects.20 These drive the development of primary and secondary communication objectives, which, according to the FDA, should be stated a priori and achieve greater than 80% comprehension.20 The FDA also recommends that label-comprehension studies include a general population of consumers with varying literacy levels. To move levonorgestrel emergency contraception OTC, the FDA required a separate label-comprehension study among adolescents aged 17 years and younger.21,22 An exploratory pilot label-comprehension study for an OTC medication abortion product conducted in South Africa among 100 reproductive-aged women demonstrated moderate understanding of key concepts and identified areas for modifying the label, which informed our drug facts label design.23
The current study developed a drug facts label prototype for a combination mifepristone and misoprostol medication abortion product and conducted a label-comprehension study to evaluate understanding of key label concepts among people of varying literacy levels and ages in the United States.
This study included three phases: 1) development of an initial drug facts label prototype, 2) preliminary cognitive interviews to test and refine the drug facts label, and 3) implementation of a label-comprehension study assessing understanding among a large sample of people living across the United States.
Informed by previous studies and following FDA guidance, we developed a drug facts label prototype by converting the FDA-approved prescription label for mifepristone 200 mg to the OTC format, assuming the same eligibility criteria as the prescription label.20–23 However, unlike most drug facts labels, this one describes two medications (mifepristone and misoprostol) that are taken separately, on different days, and in different ways (orally and buccally). We engaged a multidisciplinary group of experts for input on the design of the study and drug facts label. Experts included people with experience developing drug facts labels and implementing label-comprehension studies, clinicians, researchers, an OTC switch expert, and an advisory board representing reproductive health and justice organizations, who, in their professional roles, represent the lived experiences of people with limited access to abortion.
We developed 11 primary and 13 secondary communication objectives to test label comprehension (Table 1). Key concepts included indications for use (one question), eligibility for use (two questions), dosing regimen (seven questions), contraindications (eight questions), warning signs (one question), side effects (one question), and recognizing the risk of treatment failure (four questions). For each primary communication objective, we set a target performance threshold ranging from 80% to 90% accuracy, depending on the clinical significance (Table 1). We developed secondary communication objectives that were less critical to the safe and appropriate use of the product and did not set performance thresholds, per FDA guidance.20 The questions included in the structured interview guide used in the preliminary cognitive interviews and main study mirrored the communication objectives (see below).
We first conducted cognitive interviews to test initial versions of the drug facts label with 42 female participants aged 12–49 years living in the United States. We recruited people from May through August 2020 through Craigslist ads and community outreach, which included posting on listservs. During video interviews, we shared the drug facts label onscreen, solicited feedback, assessed understanding of key label concepts, assessed literacy using the 66-item REALM (Rapid Estimate of Adult Literacy in Medicine) or the REALM-Teen (Rapid Estimate of Adolescent Literacy in Medicine) for people aged 17 years and younger, and collected demographic information.24,25 People who scored 60 or lower (equivalent to less than a 9th-grade literacy level) on the REALM were coded as having limited literacy. We reviewed responses throughout the interview process and revised the drug facts label language, formatting, and interview questions iteratively until reaching saturation in participant feedback. We describe the cognitive interview methods and participants in Appendix 1, available online at https://links.lww.com/AOG/C686.
We aimed to recruit 800 participants for the main label-comprehension study, including a minimum of 300 young people aged 12–17 years, as requested by the FDA in the label-comprehension study for emergency contraception in adolescents.21 This sample size was set to assess whether expected comprehension met the target threshold of 90% (lower limit of the 95% CI) for the primary communication objective “indications for use,” using an exact binomial test and setting alpha at 0.05 and power at 80%. A subgroup sample size of 300 assures that the lower limit of the two-sided 95% CI for the comprehension rate is greater than 90% if comprehension is 94.5%; a subgroup sample of 150 is sufficient power to assess whether this objective was met if comprehension was 96%. We contracted with PEGUS Research, an OTC consumer behavior research organization, to recruit and interview participants. Participant eligibility criteria included having a uterus, aged 12–49 years, able to read and speak English, living in the United States, had not participated in a PEGUS-conducted market research study in the previous 3 months, and had a computer and internet access. From November 2020 to February 2021, PEGUS recruited through Facebook, Instagram, and community partner listservs.
Participants self-screened for eligibility by completing a brief online survey, and those eligible scheduled an interview online or selected to be contacted to schedule an interview. Before the interview, we emailed study details to the participants. At the start of the interview, a trained interviewer obtained verbal informed consent from participants aged 18 years and older or verbal assent for participants aged 12–17 years and verbal consent from their parent or guardian, who was present at the start of the interview. The interviewer then shared their video screen and assessed participant literacy using the REALM or the REALM-Teen for participants aged 17 years or younger. The interviewer showed the drug facts label (Fig. 1) on their screen for participants to review, then asked them a series of questions, most of which posed hypothetical situations designed to address each primary and secondary communication objective, as listed in the first column of Tables 1 and 2. The interviewer invited the participant to refer back to the drug facts label as they were answering questions. Participants were also asked open-ended questions regarding their thoughts about the blue and yellow shapes indicating each medication and the red and green table format. At the end of the interview, we asked participants to report on a series of sociodemographic characteristics including self-reported race (according to prespecified categories); Hispanic, Latina, or Latinx ethnicity; highest level of educational attainment; employment status; gender identity; and household and pregnancy characteristics (Table 3). We collected data on race, ethnicity, and other household characteristics to ensure that our sample was demographically diverse and representative of the U.S. reproductive-aged population as a whole.
PEGUS trained all interviewers to strictly follow a script and to give no indication as to whether the respondents' answers were correct or not. After 50 interviews, we paused the interview process to make minor modifications to the drug facts label and interview guide. In this iteration, we changed the drug facts label language from “Light or no bleeding” to “No bleeding or only light bleeding.” We digitally recorded and transcribed all interviews verbatim. Interviews lasted approximately 45 minutes on average. We remunerated participants $50 for their participation in the study. All procedures received ethical approval from the University of California, San Francisco, Institutional Review Board.
We reviewed all transcribed interview transcripts and coded responses for all primary and secondary communication objectives. We coded responses that were accurate according to the drug facts label as “correct.” Responses that demonstrated sufficient but not exact understanding as instructed on the label were coded as “acceptable” (eg, responded “see a doctor” instead of “do not use”). Clearly incorrect responses, responses of “don't know” and “not sure,” and skipped items were coded as “incorrect.” We estimated the proportion and exact binomial 95% CIs of correct responses by age group (12–17 years and 18–49 years) and literacy level (limited literacy and 9th-grade reading level or higher). We examined the proportion of “acceptable” responses if the lower bound of the CI for the primary communication objectives fell below the prespecified performance threshold. All analyses were conducted in Stata 15.
A total of 2,522 people completed the screening process; of those, 1,507 were eligible to participate and scheduled an interview, and 851 completed an interview (851/1,507, 56.5% response rate). Among those who completed the screening process (n=1,507), those participating were significantly more likely to self-identify as Black race than White race and did not differ significantly by age or Hispanic, Latina, or Latinx ethnicity. Reasons for ineligibility included being a health care professional (n=147), not having video capability (n=171), having participated in research in the previous 3 months (n=123), a minor without an available parent (n=87), born without a uterus (n=24), not interested (n=22), outside eligible age range (n=7), and unable to speak and understand English (n=5). We removed three people owing to poor audio or recording quality and four people because they were living outside the United States, leaving a final analytic sample of 844. By design, more than one third (35.7%) of participants were young people aged 12–17 years (Table 3). Across age groups and literacy levels, nearly half of participants self-identified as non-Hispanic White (46.4%); 15.2% as non-Hispanic Black; 15.0% as non-Hispanic Asian, Hawaiian, or other Pacific Islander; 10.8% as Hispanic, Latina, or Latinx; and nearly one in five (18.6%) had limited literacy scores (below a 9th-grade literacy level). Participants represented all U.S. regions. Demographic and household characteristics differed significantly by age group and literacy level (see Table 3).
For 10 of the 11 primary communication objectives, point estimates for the full sample exceeded 92%, and the lower bounds of the 95% CIs were well above prespecified performance thresholds (Table 1). However, the communication objective “seek medical help or talk to a provider if no or light bleeding within 2 days of taking misoprostol” was not met across age groups and literacy levels. Only 79% (95% CI 0.76–0.82) of the full sample, 78% (95% CI 0.71–0.85) of people with limited literacy, and 81% (95% CI 0.76–0.85) of young people aged 12–17 years understood that one should seek medical help if no bleeding occurs within 2 days of taking misoprostol. When we consider the 15 acceptable responses as correct for this objective, the point estimate for the full sample reaches 81% (95% CI 0.78–0.84), still below the performance threshold (not shown). Most incorrect responses erroneously indicated that the label says nothing or that one should do nothing if no bleeding occurs soon after taking misoprostol (n=89). Point estimates for understanding that the product is for abortion were 92% and above across age and literacy groups, although the lower limit of the 95% CI was below 90% for people with limited literacy (95% CI 0.86–96). The performance threshold for “take pregnancy test 4 weeks later” was met for the total sample (point estimate 0.93, 95% CI 0.91–0.95), adults (point estimate 0.93, 95% CI 0.91–0.95), and among people with normal literacy (point estimate 0.95, 95% CI 0.93–0.97); however young people (point estimate 0.93, 95% CI 0.89–0.95) and people with limited literacy (point estimate 0.83, 95% CI 0.77–0.89) did not meet this threshold. If we consider the 27 acceptable responses as correct, young people meet the threshold (point estimate 0.96, 95% CI 0.93–0.98) but people with limited literacy (point estimate 0.91, 95% CI 0.85–0.95) do not (not shown). On average, the full sample understood 95% (95% CI 0.95–0.96) of all primary communication objectives, people with limited literacy understood 92% (95% CI 0.90–0.94), adults understood 95% (95% CI 0.94–0.96), and young people understood 96% (95% CI 0.95–0.97).
For 11 of the 13 secondary communication objectives, point estimates were greater than 90% across groups. For the remaining two objectives (understanding side effects and when to take the second medication) point estimates were greater than 85% for the total sample. However, among people with limited literacy, only 78% (95% CI 0.70–0.84) understood that dizziness and cramping were expected side effects and 73% (95% CI 0.66–0.80) understood that the earliest you could take the next medication was the next day at the same time. On average, the full sample understood 95% (95% CI 0.95–0.96) of the secondary communication objectives, people with limited literacy understood 92% (95% CI 0.90–0.93), adults understood 95% (95% CI 0.94–0.96), and young people understood 96% (95% CI 0.95–0.97).
For the open-ended questions soliciting opinions about the blue and yellow shapes on the label (Fig. 1), most people found these symbols helpful (86.3%), primarily to differentiate the medications (75.2%) and to understand the dosing regimen (13.3%) (not shown). Most people also found the red and green table helpful (89.9%), which 78.1% said helped to differentiate between normal or expected symptoms and the “bad” symptoms. A few people (3.9%) suggested that the table could be improved by adding simpler and clearer headings to indicate what each red and green color or column means.
Overall comprehension for this drug facts label prototype was excellent, meeting the prespecified performance criteria for all but one primary communication objective—recognizing what to do if there is little or no bleeding. Despite the complexity of describing two different medications and dosing regimens, comprehension was markedly higher than typically reported in other label-comprehension studies.26,27 More than 95% of the full sample understood that the product is used for an abortion and understood the appropriate pregnancy duration for use, and more than 90% correctly identified contraindications. The one primary communication objective that did not meet its target threshold was related to understanding that lack of bleeding soon after taking misoprostol could indicate that the medication is not working and requires contacting a health professional. Lack of bleeding may be an indication that the pregnancy is continuing or, in very rare cases, of an ectopic pregnancy. People may have had difficulty distinguishing among the many bleeding-related symptoms included on the drug facts label. Nonetheless, the point estimate of 79% achieved a moderately high level of comprehension. Changes to the label design, for example describing this concept in bold font or grouping the information on bleeding together, might improve understanding.
Although people across age and literacy groups demonstrated clear understanding that this product is not intended for people more than 10 weeks pregnant or people unsure of how far along they are in pregnancy, some people interested in using this product may have difficulty accurately assessing the duration of their pregnancies. Studies suggest that, although most people can self-determine pregnancy duration based on the date of their last menstrual period, this exact date can be difficult for some to recall.28 A recent study of patients seeking abortion across the United States found that a combination of three non–last menstrual period questions achieved high accuracy in self-assessment of pregnancy duration; only 2.3% incorrectly self-screened as less than 10 weeks pregnant when using their responses to whether they were 1) more than 10 weeks pregnant, 2) more than 2 months pregnant, or 3) had missed two or more periods. Integration of these three statements into the label instructions and as part of an interactive online screening platform could help ensure that people have the best tools to self-screen for pregnancy duration with high accuracy and sensitivity.29
Young people aged 17 years and younger demonstrated excellent comprehension of the drug facts label, achieving comparable levels of understanding with adults across communication objectives, suggesting that young people can understand drug facts label instructions without the supervision of a licensed practitioner. Similarly, a label-comprehension study for emergency contraception also found high levels of comprehension among young people.21
An important innovation in the drug facts label development was that we engaged a multidisciplinary group of stakeholders, including an advisory board representing reproductive justice organizations, researchers, and health care professionals, to provide input into the overall research design and drug facts label wording and format. During quarterly meetings with this diverse group of stakeholders, we considered how the intersections of race, ethnicity, language, gender, age, and immigration status create unique challenges and health care needs among the individuals and communities who are likely to benefit from an OTC product. Although this process took more than 2 years, it likely contributed to a drug facts label that was more clearly understood by a diverse range of people. In particular, the color formatting and shapes that were added to the drug facts label, based on group discussions and early feedback, were well-received by participants. These improved people's ability to distinguish the medications and to differentiate normal side effects from warning signs. We recommend including such colored formatting in future drug facts labels while also pairing these with clear labels, as suggested by participants.
Although this study captured a diverse range of perspectives across age, income, race, ethnicity, literacy, and geography, our response rate of 56.5% raises the possibility that there are unobserved differences between our sample and the general population. Furthermore, our sample is limited to people with access to a computer and the internet. Studies suggest that people with limited or no internet access disproportionately live on low incomes, live in rural areas, and are less likely to be confident in their ability to obtain health information.30,31 People without access to digital technology may also have limited access to facility-based abortion care and benefit from an OTC product, yet their perspectives are not captured as part of this study.32 Although people with limited literacy did not meet performance criteria for three of the primary learning objectives (lower limit of the 95% CIs were below the threshold), only one point estimate was below the prespecified performance threshold, suggesting that we may lack statistical power given the small sample size (n=157) of this group and also that they may have more difficulty understanding label instructions. Further testing of these three label concepts among people with limited literacy is warranted.
Given the high levels of comprehension of this prototype label among people of all ages and literacy levels, we recommend only minor modifications to the label in a future OTC medication abortion product. To support an OTC switch, in addition to label-comprehension studies, there is a need for self-selection and actual-use studies demonstrating that people can take medication abortion appropriately without clinical supervision. As barriers to abortion care mount,17–19 OTC access has the potential to reduce patient burden, ensure access to abortion care earlier in pregnancy, and offer a more person-centered model of care.33,34
2. Kaye J, Reeves R, Chaiten L. The mifepristone REMS: a needless and unlawful barrier to care. Contraception 2021;104:12–5. doi: 10.1016/j.contraception.2021.04.025
3. American College of Obstetricians and Gynecologists. Joint statement on abortion access during the COVID-19 outbreak. Accessed September 17, 2021. https://www.acog.org/news/news-releases/2020/03/joint-statement-on-abortion-access-during-the-covid-19-outbreak
4. Raymond EG, Grossman D, Mark A, Upadhyay UD, Dean G, Creinin MD, et al. Commentary: no-test medication abortion: a sample protocol for increasing access during a pandemic and beyond. Contraception 2020;101:361–6. doi: 10.1016/j.contraception.2020.04.005
5. Mark A, Foster AM, Grossman D, Prager SW, Reeves M, Velásquez CV, et al. Foregoing Rh testing and anti-D immunoglobulin for women presenting for early abortion: a recommendation from the National Abortion Federation's Clinical Policies Committee. Contraception 2019;99:265–6. doi: 10.1016/j.contraception.2019.02.008
6. Upadhyay UD, Desai S, Zlidar V, Weitz TA, Grossman D, Anderson P, et al. Incidence of emergency department visits and complications after abortion. Obstet Gynecol 2015;125:175–83. doi: 10.1097/AOG.0000000000000603
7. Kapp N, Grossman D, Jackson E, Castleman L, Brahmi D. A research agenda for moving early medical pregnancy termination over the counter. BJOG Int J Obstet Gynaecol 2017;124:1646–52. doi: 10.1111/1471-0528.14646
8. Raymond EG, Tan YL, Comendant R, Comendant R, Sagaidac I, Hodorogea S, et al. Simplified medical abortion screening: a demonstration project. Contraception 2018;97:292–6. doi: 10.1016/j.contraception.2017.11.005
9. Gambir K, Garnsey C, Necastro KA, Ngo TD. Effectiveness, safety and acceptability of medical abortion at home versus in the clinic: a systematic review and meta-analysis in response to COVID-19. BMJ Glob Health 2020;5:e003934. doi: 10.1136/bmjgh-2020-003934
10. Raymond E, Chong E, Winikoff B, Platais I, Mary M, Lotarevich T, et al. TelAbortion: evaluation of a direct to patient telemedicine abortion service in the United States. Contraception 2019;100:173–7. doi: 10.1016/j.contraception.2019.05.013
11. Chong E, Shochet T, Raymond E, Platais I, Anger HA, Raidoo S, et al. Expansion of a direct-to-patient telemedicine abortion service in the United States and experience during the COVID-19 pandemic. Contraception 2021;104:43–8. doi: 10.1016/j.contraception.2021.03.019
12. Upadhyay UD, Koenig LR, Meckstroth KR. Safety and efficacy of telehealth medication abortions in the US during the COVID-19 pandemic. JAMA Netw Open 2021;4:e2122320. doi: 10.1001/jamanetworkopen.2021.22320
13. Kaller S, Muñoz MGI, Sharma S, Tayel S, Ahlbach C, Cook C, et al. Abortion service availability during the COVID-19 pandemic: results from a national census of abortion facilities in the U.S. Contracept X 2021;3:100067. doi: 10.1016/j.conx.2021.100067
14. Godfrey EM, Fiastro AE, Jacob-Files EA, Coeytaux FM, Wells ES, Ruben MR, et al. Factors associated with successful implementation of telehealth abortion in 4 United States clinical practice settings. Contraception 2021;104:82–91. doi: 10.1016/j.contraception.2021.04.021
15. Upadhyay UD, Schroeder R, Roberts SCM. Adoption of no-test and telehealth medication abortion care among independent abortion providers in response to COVID-19. Contracept X 2020;2:100049. doi: 10.1016/j.conx.2020.100049
16. Medication abortion up to 70 days of gestation. ACOG Practice Bulletin No. 225. American College of Obstetricians and Gynecologists. Obstet Gynecol 2020;136:e31–47. doi: 10.1097/AOG.0000000000004082
17. Thompson KMJ, Sturrock HJW, Foster DG, Upadhyay UD. Association of travel distance to nearest abortion facility with rates of abortion. JAMA Netw Open 2021;4:e2115530. doi: 10.1001/jamanetworkopen.2021.15530
18. Gerdts C, Fuentes L, Grossman D, White K, Keefe-Oates B, Baum SE, et al. Impact of clinic closures on women obtaining abortion services after implementation of a restrictive law in Texas. Am J Public Health 2016;106:857–64. doi: 10.2105/AJPH.2016.303134
19. Myers C, Jones R, Upadhyay U. Predicted changes in abortion access and incidence in a post-Roe world. Contraception 2019;100:367–73. doi: 10.1016/j.contraception.2019.07.139
21. Raymond EG, L'Engle KL, Tolley EE, Ricciotti N, Arnold MV, Park S. Comprehension of a prototype emergency contraception package label by female adolescents. Contraception 2009;79:199–205. doi: 10.1016/j.contraception.2008.09.004
22. Raymond EG, Dalebout SM, Camp SI. Comprehension of a prototype over-the-counter label for an emergency contraceptive pill product. Obstet Gynecol 2002;100:342–9. doi: 10.1016/s0029-7844(02)02086-0
23. Kapp N, Methazia J, Eckersberger E, Griffin R, Bessenaar T. Label comprehension of a combined mifepristone and misoprostol product for medical abortion: a pilot study in South Africa. Contraception 2020;101:46–52. doi: 10.1016/j.contraception.2019.09.009
24. Davis TC, Crouch MA, Long SW, Jackson RH, Bates P, George RB, et al. Rapid assessment of literacy levels of adult primary care patients. Fam Med 1991;23:433–5.
25. Davis TC, Wolf MS, Arnold CL, Byrd RS, Long SW, Springer T, et al. Development and validation of the Rapid Estimate of Adolescent Literacy in Medicine (REALM-Teen): a tool to screen adolescents for below-grade reading in health care settings. Pediatrics 2006;118:e1707–14. doi: 10.1542/peds.2006-1139
26. Cohen BR, Mahoney KM, Baro E, Squire C, Beck M, Travis S, et al. FDA initiative for drug facts label for over-the-counter naloxone. N Engl J Med 2020;382:2129–36. doi: 10.1056/NEJMsa1912403
27. Tong V, Raynor DK, Aslani P. Design and comprehensibility of over-the-counter product labels and leaflets: a narrative review. Int J Clin Pharm 2014;36:865–72. doi: 10.1007/s11096-014-9975-0
28. Schonberg D, Wang LF, Bennett AH, Gold M, Jackson E. The accuracy of using last menstrual period to determine gestational age for first trimester medication abortion: a systematic review. Contraception 2014;90:480–7. doi: 10.1016/j.contraception.2014.07.004
29. Ralph LJ, Ehrenreich K, Barar R, Biggs MA, Morris N, Blanchard K, et al. Accuracy of self-assessment of gestational duration among people seeking abortion. Am J Obstet Gynecol 2021 Dec 17 [Epub ahead of print]. doi: 10.1016/j.ajog.2021.11.1373
30. Oshima SM, Tait SD, Thomas SM, Fayanju OM, Ingraham K, Barrett NJ, et al. Association of smartphone ownership and internet use with markers of health literacy and access: cross-sectional survey study of perspectives from project PLACE (Population Level Approaches to Cancer Elimination). J Med Internet Res 2021;23:e24947. doi: 10.2196/24947
31. Saeed SA, Masters RM. Disparities in health care and the digital divide. Curr Psychiatry Rep 2021;23:61. doi: 10.1007/s11920-021-01274-4
32. Cartwright AF, Karunaratne M, Barr-Walker J, Johns NE, Upadhyay UD. Identifying national availability of abortion care and distance from major US cities: systematic online search. J Med Internet Res 2018;20:e186. doi: 10.2196/jmir.9717
33. Fay K, Kasier J, Turok D. The no-test abortion is a patient-centered abortion. Contraception 2020;102:142. doi: 10.1016/j.contraception.2020.05.018
34. Ramaswamy A, Weigel G, Sobel L, Salganicoff A. Medication abortion and telemedicine: innovations and barriers during the COVID-19 emergency. Accessed June 18, 2020. https://www.kff.org/coronavirus-policy-watch/medication-abortion-telemedicine-innovations-and-barriers-during-the-covid-19-emergency/