There is consensus within the medical and scientific communities that induced abortion is a safe and effective health care procedure. A 2018 report by the National Academies of Sciences, Engineering, and Medicine found that abortion—including medication abortion—is a safe and effective way to end a pregnancy.1 The report also demonstrates that various regulatory and legislative restrictions on abortion provision create unequal and inadequate access to abortion in areas of the United States, particularly in the South and Midwest.1–3 A recent study that replicated a person's experience searching for an abortion provider online found that there are 27 U.S. cities of more than 100,000 people that are “abortion deserts”—the city is more than 100 miles from the closest abortion provider.4 Outside of large cities, access to abortion is worse—90% of all U.S. counties lack an abortion provider.5 Limited access to abortion imposes medical, social, and economic burdens on people seeking abortion and often means they must wait or travel long distances to obtain abortion care.6–9
One strategy to address barriers and increase access to abortion is the provision of medication abortion through telemedicine. Telemedicine is a growing field in health care more broadly,10–13 and medication abortion is a natural application for this care delivery model because it is a simple treatment requiring minimal hands-on assessment. Telemedicine may be especially useful in states with physician-only laws, which require that a patient meet with a physician rather than an advanced practice clinician (eg, nurse practitioner or physician's assistant) to obtain an abortion (whether by medication or aspiration). In remote locations where it is difficult to maintain physician staffing, telemedicine may allow patients to receive care sooner and at a health center closer to home.
In 2008, Planned Parenthood of the Heartland launched telemedicine for medication abortion in Iowa. An evaluation of this program using clinic- and patient-reported data found that telemedicine for medication abortion was comparably effective with services provided through standard care.14 A subsequent analysis of adverse events related to medication abortion across Iowa over 7 years also found that telemedicine was comparable with standard services.15 Surveys and interviews conducted with patients demonstrated comparable and high satisfaction with both service delivery models.14 Health care providers also found telemedicine for medication abortion to be highly acceptable, reporting that it facilitated patient-centered care by allowing patients to be seen sooner or closer to home.16
Despite this strong evidence, published research about outcomes of telemedicine for medication abortion in the United States is currently limited to Iowa. The present study expands the evidence base to four additional states now offering telemedicine for medication abortion services (Alaska, Idaho, Nevada, and Washington). We sought to assess outcomes of telemedicine medication abortion—including ongoing pregnancy, receipt of or referral for aspiration, and adverse events—compared with outcomes of standard medication abortion at Planned Parenthood health centers in four U.S. states.
In 2016, Planned Parenthood Federation of America launched a program to expand provision of telemedicine for medication abortion to increase access to abortion in the United States. Medication abortion involves the provision of two medications: mifepristone and misoprostol. Currently, the U.S. Food and Drug Administration requires that mifepristone be provided only in clinics, hospitals, or medical offices under the supervision of a certified prescriber. At Planned Parenthood, patients seeking medication abortion visit a health center and receive in-person care from staff, including ultrasound scans, laboratory testing, education, and counseling, and provide informed consent before meeting with the clinician. At health centers with an available clinician able to provide medication abortion, patients meet with the clinician in person. We refer to this as standard medication abortion. At participating health centers without an available clinician able to provide medication abortion on-site, the patient meets with a clinician virtually using a secure videoconference platform and takes the mifepristone in view of the on-screen clinician. We refer to this as telemedicine medication abortion. For both groups, misoprostol, taken up to 48 hours later, is either dispensed at the health center or prescribed along with other auxiliary medications.
As with standard medication abortion, telemedicine patients are instructed to complete follow-up evaluation within 1–2 weeks by ultrasound scan or serial human chorionic gonadotropin (hCG) testing at the health center or an outside facility. In the event of an unsuccessful medication abortion, patients can receive an aspiration procedure within the provider network or at an outside facility, depending on service availability or patient preference. The present study is limited to four states where Planned Parenthood health centers offered telemedicine for medication abortion between 2016 and 2018. In this retrospective cohort study, we compared outcomes for patients receiving telemedicine medication abortion with those for patients receiving standard medication abortion at Planned Parenthood health centers in these same four states. Study outcomes included occurrence of ongoing pregnancy, receipt of or referral for an aspiration procedure, and clinically significant adverse events. This retrospective analysis of routinely collected administrative data was reviewed and approved by the Allendale Investigational Review Board.
We assessed abortion outcomes for a convenience sample of all patients undergoing medication abortion presenting at participating health centers in Alaska, Idaho, Nevada, and Washington for 1 full year: April 2017 through March 2018. We used administrative data from participating centers’ electronic health record (EHR) systems to assess patient characteristics and abortion outcomes related to effectiveness, including ongoing pregnancy (ie, presence of fetal cardiac activity or, in the case of hCG follow-up, any result indicating a drop of less than 80% within 2 weeks of mifepristone administration) and aspiration procedure. We identified patients undergoing medication abortion using the Current Procedural Terminology (CPT) code for mifepristone (S0190).17
We worked with participating health centers to extract encounter-level, partially de-identified patient data entered into the systems at the time of service as part of routine clinical charting. We obtained all encounters for patients on the day of their medication abortion when mifepristone was provided as well as any subsequent visit or communication with the health care network within 45 days of receiving mifepristone if recorded in the patient’s chart (including follow-up hCG results obtained elsewhere if documented in specific fields in the chart). We included data from the full 45-day follow-up window for all patients; thus, some follow-up encounters occurred after March 2018.
All participating health centers included in this analysis used similar EHR templates, though data are not shared across systems. Thus, we harmonized extracted reports to analyze comparable outcomes from each health center. We reviewed data extracts in consultation with participating centers; site-specific codes or terminology were discussed and resolved with staff before final coding. We then aggregated the harmonized reports, which included data for patients seen at 26 participating health centers across four states. We identified telemedicine for medication abortion visits based on specific EHR template fields indicating that the visit was completed through telemedicine; these structured data fields were tested before the launch of telemedicine services at participating health centers.
We assessed ongoing pregnancy and receipt of or referral for an aspiration procedure after medication abortion. We determined whether these outcomes occurred based on data reported in the EHR from patients’ follow-up visits to, or communications with, the health center network if documented in the patient’s chart. We therefore can assess abortion outcomes only for those patients who returned to, or otherwise had communication with, the health center within 45 days of their medication abortion.
We identified ongoing pregnancy based on results of all available follow-up testing, including ultrasound scans and serial hCG tests. We determined ongoing pregnancy based on the provider's documented interpretation of ultrasound images indicating an ongoing pregnancy—defined by the presence of fetal cardiac activity or, in the case of hCG follow-up, any result indicating a drop of less than 80% within 2 weeks of mifepristone administration, per protocol.
We identified a patient's receipt of an aspiration procedure based on the presence of a corresponding CPT code (59840, 59841, 59812) within 45 days of mifepristone administration. In addition, we examined an open-text encounter remarks field to identify any other documentation of an aspiration procedure within 45 days. For two sites where aspiration was not available at a local Planned Parenthood health center, we identified referral to an external provider for intervention using a specific structured field in the medication abortion follow-up EHR template. Finally, we reviewed adverse event reports (discussed in greater detail below) submitted by participating health centers to determine whether any ongoing pregnancies or aspiration procedures were reported for patients who did not return to the health center within 45 days or did not have other documentation of an ongoing pregnancy or aspiration procedure at follow-up encounters.
The unit of analysis was unique abortion episode rather than unique patient. Thus, patients who may have had multiple medication abortion episodes within the study period are included multiple times in the analysis if mifepristone CPT codes occur more than 45 days apart (ie, distinct abortion episodes). We compared abortion effectiveness outcomes using bivariate and multivariate analyses, with the exposure of interest being receipt of medication abortion through telemedicine compared with standard medication abortion. We included gestational age as a covariate given evidence from the literature that prevalence of ongoing pregnancy increases with advanced gestational age.18 To account for the nonindependence of observations within health centers, we adjusted the standard error using the Stata vce option.
We conducted sensitivity analyses to address missing outcomes for patients who did not follow up with the health center by applying published rates of ongoing pregnancy (0.8%, range 0–2.6%) and aspiration after medication abortion for any reason (3.3%, range 1.7–9.0%) from a systematic review18 to patients who were lost-to-follow-up. We then calculated estimated rates of ongoing pregnancy and aspiration procedures for the full population, including those without follow-up data, and compared these rates with those for the population of patients with documented medication abortion outcomes.
As part of the Planned Parenthood–affiliated network, health centers submit incident reports for adverse events related to medication abortions for the purposes of safety monitoring and quality improvement. The following types of events are reported through a centralized system: ongoing pregnancy, complications associated with ectopic pregnancy, hospitalization, hospital emergency department (ED) visit or intervention by an outside provider, blood transfusion, infection resulting in intravenous antibiotics, allergic reaction, and death. Aspiration procedures performed as a result of symptomatic bleeding, retained tissue, persistent (but not viable) gestational sac, or patient preference are not reported unless they are associated with other reportable events. If patients seek care at a hospital or other provider, this information is captured and reported if Planned Parenthood is notified by the patient or outside provider.
Before the study period, a structured data field was added to the incident report template to indicate whether a medication abortion was provided through telemedicine. The response options are “yes,” “no,” or “not applicable” (to indicate that the center does not offer telemedicine). We instituted active follow-up with affiliates beginning in January 2017 (ie, before the study period) to ensure completion of this field for every reported incident.
We analyzed incidents associated with medication abortions occurring from April 2017 through March 2018. To align with previous studies,14,15 we focused on prevalence of clinically significant adverse events, defined as those resulting in hospital admission, surgery (not including aspiration procedure), blood transfusion, or death. We also analyzed incidents of treatment given at a hospital ED, including intravenous fluids or oral medications, which were considered minor but clinically significant adverse events.
We did not include cases of ongoing intrauterine pregnancy as adverse events because this is a known possible outcome of medication abortion. Patients with an incomplete abortion were included as adverse events only if they were treated in the hospital or ED; outpatient treatment of incomplete abortion was not reported as an adverse event. Similarly, cases of ectopic pregnancy were captured only if the patient required hospitalization, surgery, blood transfusion, or treatment in the ED. It is worth noting that Planned Parenthood health centers have standard protocols for initiating and managing medication abortion for pregnancies of unknown location.
Individual patients may have experienced more than one clinically significant adverse event, all of which are reported here if known. For example, a patient who underwent surgery and also received a blood transfusion would be counted in each of these categories but would be counted as a single case of having experienced “any major adverse event.” We calculated the frequency of any event and event type overall and by abortion type (telemedicine compared with standard medication abortion). We calculated rates of clinically significant adverse events and performed Fisher's exact tests to assess differences between groups.
During the study period, a total of 5,952 patients underwent medication abortion across the 26 participating Planned Parenthood health centers included in this analysis. Of these patients, 738 (12.4%) underwent telemedicine medication abortion and 5,214 (87.6%) standard medication abortion (Table 1). There were differences between groups in terms of patient race or ethnicity, insurance, and parity. None of these characteristics was significantly associated with the study outcomes. Gestational age at time of abortion—the only patient characteristic that was prespecified for inclusion in the model based on extant literature—was slightly different between groups: 50.4 days for telemedicine patients vs 48.9 days for standard patients (prevalence ratio 1.02; 95% CI 1.00–1.03).
There was a difference in follow-up rates within 45 days of receiving mifepristone (including adverse event reports, if applicable) between the telemedicine and standard treatment groups. Overall, 4,456 of 5,952 (74.9%) patients followed up, including 4,011 of 5,214 (76.9%) standard care patients and 445 of 738 (60.3%) telemedicine patients (prevalence ratio 0.83; 95% CI 0.78–0.88). Rates of follow-up varied by state. Some states observed higher rates of follow-up among patients undergoing telemedicine medication abortion, others observed higher rates among patients undergoing standard medication abortion, and still others saw no difference in follow-up by treatment mode.
Overall, we identified 73 (1.6%) ongoing pregnancies across all 4,456 patients with follow-up data (Table 2). Of these, 2 of 445 (0.5%) were telemedicine patients and 71 of 4,011 (1.8%) were standard patients (adjusted odds ratio [OR] 0.23; 95% CI 0.14–0.39). A total of 188 of 4,456 (4.2%) patients undergoing medication abortion received or were referred for an aspiration procedure within 45 days of receiving mifepristone; of these, 6 of 445 (1.4%) were telemedicine patients and 182 of 4,011 (4.5%) were standard patients (adjusted OR 0.28; 95% CI 0.17–0.46). Ongoing pregnancy and aspiration are not mutually exclusive categories; patients can be counted in both categories to include aspirations performed to resolve ongoing pregnancies and those for other reasons such as uterine debris.
We conducted sensitivity analyses to address loss to follow-up, particularly given rare study outcomes. Findings were consistent in terms of directionality and significance when we applied published rates of ongoing pregnancy (0.8%) and aspiration post–medication abortion (3.3%) from a systematic review18 to the 1,496 patients for whom we do not have outcome data (Table 3). When we applied these rates to all those without follow-up data, we estimated an ongoing pregnancy rate of 1.6% for standard patients compared with 0.6% for telemedicine patients (OR 0.34; 95% CI 0.13–0.90). For aspiration post–medication abortion, we estimated a rate of 4.3% for standard patients and 2.1% for telemedicine patients (OR 0.50; 95% CI 0.30–0.83). The estimated differences in rates of ongoing pregnancy and aspiration procedure between groups were therefore attenuated in sensitivity analyses but remained statistically significant. We also ran additional multivariable models adjusting for covariates found to be associated with loss to follow-up in bivariate models (eg, insurance, race or ethnicity, and parity); none, excepting gestational age, was significantly associated with the study outcomes (results not shown).
To analyze adverse events, we reviewed a total of 77 medication abortion–related incident reports submitted by the same health centers included in the EHR data analysis for medication abortions that occurred during the same time period. Of these, 17 were clinically significant adverse events as defined above; 16 adverse events occurred after a standard medication abortion, and one occurred after a telemedicine medication abortion (Table 4). There were no reported deaths.
Patients could have more than one type of event. Eight patients (0.13%) had at least one major adverse event. Four were admitted to the hospital, two had surgery (for ectopic pregnancy), and five received blood transfusions; two of these patients were coded as having received transfusions based on hemoglobin levels and symptoms at time of referral or admission to a hospital. Nine patients (0.15%) had minor adverse events but were treated in hospital EDs and are therefore included here; the most common treatments in the ED were dilation and curettage and aspiration procedures. There were no statistically significant differences between the telemedicine and standard medication abortion groups.
As in previous research conducted in Iowa,14,15 this large study of patients visiting Planned Parenthood health centers in four states found that outcomes for telemedicine medication abortion were within ranges reported in the literature for standard medication abortion.18 Rates of adverse events were low in both groups and aligned with the published literature.17
We found that rates of both ongoing pregnancy and aspiration procedures were higher in the standard group than in the telemedicine group; this was also found in a previous Iowa study.15 These differences could be due to a lower proportion of telemedicine patients with follow-up data in the present study. It is possible, however, that meeting with a clinician virtually rather than in person may reduce one's likelihood to return for a follow-up visit. Nevertheless, there is no biologically plausible reason that medication abortion would be appreciably different with respect to safety or effectiveness given an identical medication regimen.
This study has several limitations. Patients undergoing telemedicine and standard medication abortion were not evenly distributed across sites; most telemedicine patients were in one state (Nevada), and most standard patients were in another (Washington). Therefore, higher rates of outcomes of interest in the standard group compared with the telemedicine group could be influenced by differing availability of aspiration abortion across sites (eg, two sites in Nevada had to refer patients with failed medication abortion to outside providers). We did not have adequate sample sizes to permit state comparisons. In addition, in the two sites where aspiration procedures were not offered by the Planned Parenthood affiliate, we used documented referral rather than receipt of aspiration procedure. We therefore cannot report definitively on the proportion of patients who ultimately received an aspiration procedure.
We did not have outcome data for the full patient sample because approximately one quarter of patients did not follow up with the health center. It is possible that patients may have sought care elsewhere and did not inform Planned Parenthood of doing so; however, patients may have been more likely to return to Planned Parenthood given that they would not have to pay additional fees for follow-up care, which they might incur at other providers. Although staff encourage patients to follow up and report concerns, practices may vary by provider in terms of how actively patients are encouraged to follow up. At least one contact attempt for patients undergoing medication abortion is required of all Planned Parenthood health centers for patients who do not follow up. We did not attempt to obtain outcome data for patients for whom we did not have follow-up data either through documented EHR encounters or adverse event reports. It is worth noting that, in a study conducted in Iowa, researchers surveyed local EDs and did not identify any additional adverse events related to medication abortion that were not already captured in health center data.15 Although complete data were not available for all patients, sensitivity analyses including patients for whom we have no follow-up data found similar differences between groups.
There are several notable strengths to our study. We had a large overall sample size of patients, which allowed us to evaluate outcomes across four states, including in both urban and rural settings. Although use of EHR data has some limitations, these data are systematically recorded at the time of abortion and follow-up. Across all states included in the analyses, common systems were used to collect similar data elements, allowing for harmonization of data across the network of participating providers. Finally, the research was conducted in real-world settings, increasing the external validity and potential relevance of the findings.
In sum, this study provides additional evidence that telemedicine for medication abortion results in comparable outcomes with standard provision of medication abortion when implemented across different geographic settings. To the extent that state bans on telemedicine for abortion rest on arguments of improved patient safety, the findings of this and previous studies do not support such contentions. As more health care services, including abortion, become dependent on or facilitated by technology, future research should continue to assess the effects of telemedicine on patient experience and access to care.
1. National Academies of Sciences, Engineering and Medicine. The safety and quality of abortion care in the United States. Washington, DC: The National Academies Press; 2018.
2. Baum SE, White K, Hopkins K, Potter JE, Grossman D. Women's experience obtaining abortion care in Texas after implementation of restrictive abortion laws: a qualitative study. PLoS One 2016;11:e0165048.
3. Baum S, White K, Hopkins K, Potter J, Grossman D. Rapid response to evaluate policy: assessing changes in medical abortion using real-time data collection from Texas abortion providers. Contraception 2018;98:339–40.
4. 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.
5. Jones RK, Jerman J. Abortion incidence and service availability in the United States, 2014. Perspect Sex Reprod Health 2017;49:17–27.
6. Bearak JM, Burke KL, Jones RK. Disparities and change over time in distance women would need to travel to have an abortion in the USA: a spatial analysis. Lancet Public Health 2017;2:e493–500.
7. Jerman J, Frohwirth L, Kavanaugh ML, Blades N. Barriers to abortion care and their consequences for patients traveling for services: qualitative findings from two states. Perspect Sex Reprod Health 2017;49:95–102.
8. Ostrach B, Cheyney M. Navigating social and institutional obstacles: low-income women seeking abortion. Qual Health Res 2014;24:1006–17.
9. Jerman J, Jones RK. Secondary measures of access to abortion services in the United States, 2011 and 2012: gestational age limits, cost, and harassment. Womens Health Issues 2014;24:e419–24.
10. American Telemedicine Association. About telemedicine. Available at: https://www.americantelemed.org/resource/why-telemedicine/
. Retrieved November 20, 2018.
11. Mehrotra A, Huskamp HA, Souza J, Uscher-Pines L, Rose S, Landon BE, et al. Rapid growth in mental health telemedicine use among rural Medicare beneficiaries, wide variation across states. Health Aff (Millwood) 2017;36:909–17.
12. Kruse CS, Bouffard S, Dougherty M, Parro JS. Telemedicine use in rural Native American communities in the era of the ACA: a systematic literature review. J Med Syst 2016;40:145.
13. Kahn JM, Cicero BD, Wallace DJ, Iwashyna TJ. Adoption of intensive care unit telemedicine in the United States. Crit Care Med 2014;42:362–8.
14. Grossman D, Grindlay K, Buchacker T, Lane K, Blanchard K. Effectiveness and acceptability of medical abortion provided through telemedicine. Obstet Gynecol 2011;118:296–303.
15. Grossman D, Grindlay K. Safety of medical abortion provided through telemedicine compared with in person. Obstet Gynecol 2017;130:778–82.
16. Grindlay K, Grossman D. Telemedicine provision of medical abortion in Alaska: through the provider's lens. J Telemed Telecare 2017;23:680–5.
17. HCPCS Codes. HCPCS code S0190: mifepristone, oral, 200 mg. Available at: https://hcpcs.codes/s-codes/S0190/
. Retrieved November 21, 2018.
18. Chen MJ, Creinin MD. Mifepristone with buccal misoprostol for medical abortion: a systematic review. Obstet Gynecol 2015;126:12–21.