As part of a prospective, observational study comparing the safety and effectiveness of telehealth versus in-person medication abortion care conducted from May 2021 to March 2023 [19], we recruited participants obtaining medication abortion from 4 abortion-providing organizations in 6 US states, including Colorado, Illinois, Maryland, Minnesota, Virginia, and Washington. These states were selected because they had no restrictions on the use of telehealth medication abortion.

To be eligible to participate in the study, partnering organizations must have been providing medication abortion both in-person and via telehealth at the time of participant recruitment. All sites provided synchronous (real-time) care [29], and 2 organizations offered telehealth services via phone or video, while the other 2 offered telehealth only via video.

Patients were eligible to participate if they were obtaining medication abortion, aged 15 years or older (≥18 y at 4 sites not offering telehealth care to minors), English or Spanish speakers, and ≤70 days of gestation on the date they intended to take mifepristone. Patients younger than the age of 15 years (or <18 y at 4 sites), those who were estimated to be over 70 days of gestation when taking mifepristone, those who were unable to speak English or Spanish, and/or those not seeking medication abortion care were ineligible to participate in the study.

Approximately 1 day, 2 weeks, and 1 month after receiving their medication abortion pills, we invited participants via email to complete electronic follow-up surveys. For each survey, we sent up to 5 reminders via email, text, or phone, depending on participants’ preferences. Clinic staff abstracted participants’ medical record data approximately 6 weeks after their enrollment.

We describe participants’ sociodemographic characteristics, the perceived importance of each abortion care feature by study group, and participants’ access to technology and experiences with telehealth appointments. To assess differences in participant characteristics by study group, we fit bivariable linear and multinomial regression analyses that accounted for clustering by recruitment site and state, followed by postestimation joint Wald tests assessing the overall association between study group and each variable. To examine differences in the perceived importance of abortion care features, we used bivariable ordinal logistic regression models with 4-point Likert-scale responses (ranging from 1=“not important” to 4=“extremely important”) and logistic regression models for dichotomized outcomes (“extremely important” vs all other responses), with study group as the primary independent variable. To assess whether the care received aligned with participant preferences, we examined 3 abortion care features associated with in-person care (getting an ultrasound; meeting in person with the clinician providing the abortion) or telehealth care (ability to get care at home without a clinic visit) and whether such care was received. To identify potential barriers to accessing telehealth services, we describe patients’ access to technologies and experiences with telehealth appointments. We conducted complete-case analyses for regression models as missing outcome data were low (<3%). All statistical tests accounted for clustering by site and state where abortion care was provided using robust SEs, and analyses were conducted in Stata 18 (StataCorp).

All study activities were approved by the University of California, San Francisco, Institutional Review Board (20‐32514). Participants provided electronic informed consent to participate in the study. The consent documentation described all measures to maintain patient privacy and confidentiality. We deidentified all survey data prior to analysis, and each participant was assigned a unique study identification number. Participants received a US $15 gift card for enrolling in the study, US $25 gift cards for completing the 1-day and 1-month surveys, and a US $15 gift card for completing the 2-week survey. Medical record data were abstracted, standardized, and entered into a secure REDCap database (Vanderbilt University) by clinic staff.

Among 583 participants enrolled, 487 initiated a survey, and 477 (242 telehealth and 235 in-person) answered at least one of our primary outcome questions and comprise our analytic sample. Among 242 participants in the telehealth group, 63 (26.0%) were dispensed medications in person. Among 235 participants in the in-person group, 35 (14.9%) did not undergo ultrasonography (not shown). Participants in the telehealth and in-person groups differed by age, race or ethnicity, education, and parity; however, both groups had a similar history of previous abortion, receipt of public assistance, and gestational duration (Table 1).

The features most commonly reported as “extremely important” by participants in both study groups included effectiveness (340/397, 85.6%), safety of the abortion (326/397, 82.1%), scheduling their appointment as soon as possible (307/397, 77.3%), and keeping their abortion private (299/397, 75.3%), which did not differ by study group (Table 2 and Figure 1).

According to logistic regression analyses, telehealth participants were more likely than in-person participants to consider it extremely important to get care at home without a clinic visit (65.9% vs 43.5%; odds ratio [OR] 2.48, 95% CI 1.57‐3.89; P<.001) and to have a medication abortion (as opposed to having an in-clinic procedure) (62.1% vs 51.1%; OR 1.58, 95% CI 1.11‐2.23; P=.01; Table 2 and Multimedia Appendix 3), and were less likely to consider it extremely important to meet with a clinician in person (16.1% vs 45.7%; OR 0.23, 95% CI 0.19‐0.26; P<.001) or get an ultrasound (15.2% vs 38.2%; OR 0.28, 95% CI 0.20‐0.40; P<.001).

We also assessed whether the care that participants received was concordant with abortion care features they identified as “extremely important.” Among participants who reported that it was extremely important to get an ultrasound, approximately one-third (33/103, 32.0%) did not receive one (Tables 3-5). Among those who thought it was extremely important to meet in person with a clinician, 28.6% (34/119) did not. Among those who felt it was extremely important to get care at home without a clinic visit, 54.1% (119/220) had a clinic visit to meet with a clinician (n=81) or to pick up their medications (n=38). Among those preferring to get care at home but who ultimately had an in-person visit, 56.3% (not shown) lived in a state with a telehealth or total abortion ban in effect.

Almost all participants in both the telehealth and in-person groups owned a smartphone (99.5% and 96.8%, respectively; OR 6.96, 95% CI 4.24‐11.43; P<.001), had access to Wi-Fi (95.6% and 91.7%; OR 1.99, 95% CI 1.12‐3.52; P=.02), and most had an unlimited data plan (82.9% and 80.0%; OR 1.14, 95% CI 0.89‐1.47; P=.20). Smaller proportions of both groups owned a computer (72.2% vs 73.9%; OR 0.92, 95% CI 0.60‐1.42; P=.71) and/or a tablet (55.5% vs 48.4%; OR 1.33, 95% CI 0.82‐2.16; P=.25; not shown).

Participants in the telehealth group were significantly more likely than those in the in-person group to report that they were given the option to have a phone appointment (59.9% vs 27.2%; OR 4.08, 95% CI 2.43‐6.85; P<.001) and/or a video appointment with a clinician (96.7% vs 23.4%; OR 95.73, 95% CI 50.72‐180.69; P<.001) and to have had a phone (50.0% vs 8.5%; OR 10.93, 95% CI 6.24‐19.14; P<.001) or video (92.6% vs 1.7%; OR 718.67, 95% CI 95.74‐5394.38; P<.001) appointment (Table 6). Among the 34 telehealth participants who reported being offered a phone appointment but declined one, the majority (n=28, 82.4%) preferred a video appointment. Among the 48 participants in the in-person group who reported being offered a telehealth appointment but declined, a majority preferred to be seen in person rather than having a phone (n=37, 77.1%) or video (n=41, 85.4%) appointment.

Almost all telehealth participants had a video appointment (224/242, 92.6%), and half had a phone appointment (121/242, 50.0%); a small proportion had only a phone appointment (7/242, 2.9%). Telehealth participants largely noted positive experiences with their video- or phone-based care: most participants (213/242, 88.0%) reported feeling very or extremely comfortable asking questions during their appointment, had no difficulty finding a time or place to have their appointment (200/242, 82.6%), and experienced difficulty hearing or seeing staff during their appointment (181/242, 74.8%; Table 7).

Among the 224 participants who had a video appointment, almost all (n=217, 96.9%) reported no difficulty accessing a device to participate in their video appointment. Most participants used a smartphone for their video appointment (n=170, 75.9%), followed by a computer (n=48, 21.4%), a tablet (n=8, 3.6%), or another form of technology (n=3, 1.3%). Over half (52.9%) of those who had a phone or video appointment had it at home, while approximately one-quarter (28.5%) had it in a car or vehicle, among other locations (Table 7).

In this study of people obtaining telehealth and in-person medication abortion care, participants receiving both models of care reported similar features as extremely important to them: safety, effectiveness, privacy, cost, and scheduling their abortion as soon as possible. These findings are consistent with previous research [23,32]. However, some differences in preferences emerged between the 2 study groups. Compared to the in-person group, telehealth participants were more likely to report that it was extremely important to get care at home without a clinic visit and were less likely to report that getting ultrasonography or meeting with their clinician in person was extremely important. Notably, we found that both telehealth and in-person abortion models of care offer features that are uniquely preferred by many, highlighting that the ability to access both telehealth and facility-based in-person care is critical to addressing patients’ needs and preferences, including those choosing medication abortion.

Still, some participants received a model of care that did not necessarily align with the features they rated as extremely important. While a growing body of evidence supports the provision of medication abortion without ultrasonography [33,34], one-third of participants who noted that having an ultrasound was extremely important to them ultimately did not receive one, highlighting that people’s care preferences are highly individual in nature. Laws that restrict access to abortion may limit abortion patients’ ability to access their desired care preferences and reduce their reproductive autonomy [35]. Additionally, over half of participants who reported that it was extremely important to get care at home without a clinic visit had a clinic visit and/or picked up their medication abortion pills in-person. Regardless of their preferences, some people may have no choice but to visit a clinic if they are ineligible for no-test telehealth abortion care (ie, presenting for care later in pregnancy and/or having a contraindication for medication abortion [8]) or if they live in a state with a telehealth or abortion ban in effect. For example, qualitative interviews with people seeking medication abortion in Mississippi found that many were interested in but unable to access telehealth models of abortion care due to state-level abortion restrictions [36]. However, in the evolving landscape of shield law provision, continued research in this area is needed. It is also important to note that despite some discordance between participants’ abortion care preferences and the care they received, most study participants reported being largely satisfied with the abortion care they received, irrespective of study group [37].

We also found that nearly all participants, including those who received care in person, had access to the technologies required for telehealth abortion. These findings address an important gap in the literature, as many studies have examined the effectiveness and acceptability of telehealth medication abortion care but have not captured abortion patients’ access to the technologies required for this care [38]. Consistent with research showing that people accessing telehealth contraceptive and abortion care are very satisfied with telehealth services [24,39,40], we found that participants who had a phone or video appointment largely noted that it was not difficult to find a device or location, were able to hear or see staff, and were comfortable asking questions during their appointment. These findings suggest that telehealth can provide a feasible alternative to in-person abortion care for those who prefer it, as few people experienced technological barriers to telehealth abortion care. Lastly, while telehealth abortion is often described as facilitating “at-home” abortion care, we found that nearly half of respondents had their phone or video appointment at a location other than their home, which may be an important consideration for designing telehealth abortion care protocols that prioritize patient privacy and autonomy.

This study had some important limitations. Given the observational nature of our study design, participants were not randomized into study groups, and thus their “choice” of model of care received may have been restricted by abortion-related laws in their state of residence, by clinic policies, or by other personal factors that remained unmeasured. While our eligibility criteria included people aged 15 years and older, no youth participants enrolled in our study, limiting the generalizability of our findings to adults. As our recruitment occurred at 4 partner organizations within 6 states, the generalizability of our findings to other contexts also remains limited.

Additionally, participants were asked about their abortion care preferences approximately 1 month after their abortion, and thus their responses may have been influenced by the model of care they received and may not necessarily reflect their preferences prior to obtaining an abortion. Survey items did not undergo rigorous pretesting, although they were informed by prior literature and input from a community advisory board. The survey did not allow participants to rank their care preferences, making it difficult to determine the relative importance of participants’ reported preferences, particularly where features may conflict with one another (eg, where someone reported both a preference for having an ultrasound and for being able to get care at home without a clinic visit).

While we found that almost all survey participants had access to technologies that would allow them to participate in telehealth care, this finding may have been influenced by those who self-selected to participate in the study and/or those who opted to complete subsequent surveys. Additionally, our study was limited to people eligible for medication abortion, thereby excluding the perspectives of people ineligible for medication abortion and/or underrepresenting the views of people who may have preferred procedural abortion. Our sample also excludes those who may have faced insurmountable barriers to accessing abortion care. Furthermore, we conducted the study before states implemented telehealth shield laws, so our findings do not reflect the perspectives of people accessing telehealth abortion under shield laws or through asynchronous care. Future research should explore abortion care preferences among a broader sample of people, including minors, to gain further insight into the factors that matter most to abortion patients.

This paper offers additional evidence on the care features that patients seeking medication abortion prefer. We found that participants prioritized safety, effectiveness, privacy, and timeliness regardless of the care model they received. Participants receiving telehealth care were more likely to prioritize features that limited in-person interaction, whereas participants receiving in-person care placed greater importance on ultrasonography and in-person clinician interaction. While the abortion care features perceived as extremely important generally aligned with the model of care received, not all participants received the care features they deemed important. This may be, in part, due to a lack of choice, overly restrictive eligibility criteria associated with no-test screening [19,40,41], and/or policy barriers, including telehealth and total abortion bans. Thus, additional research is needed to better understand people’s reasoning for their reported preferences. The variability in the perceived importance of abortion care features across participants also emphasizes the importance of offering a range of medication abortion options. Together, these findings support policies and clinical protocols that allow for autonomous, person-centered approaches where patients can choose their preferred model of abortion care, which would require maintaining access to both in-person care, including optional ultrasonography, and telehealth models of care. We urge health care providers and policymakers to promote evidence-based policies that enable patients to access abortion services that meet their preferences and needs.