Figure 2 compares the average RSI values over time for home abortion, self abortion, and buy abortion pill online. As shown, home abortion was the most popular of the 3 initial search terms explored. Home abortion had an average popularity search index relative to all other searches in the United States (RSI) of 30 in 2020 compared with almost 10 for self abortion and 8 for buy abortion pill online. Searches for home abortion remained higher throughout the year compared with those for the 2 other search terms we explored, and peaked in November. By contrast, searches for self abortion and buy abortion pill online peaked in January and February of 2020, prior to the official onset of the COVID-19 pandemic, although there was less variability in these searches over time and these peaks do not reflect significantly greater relative interest compared with other weeks in 2020.

There were many top ranked queries associated with home abortion, including abortion at home, at home abortion(s), abortion pill, abortion remedies, abortion clinic, home remedies for abortion, pregnancy symptoms, how to have an at home abortion, how to do an at home abortion, how to do an abortion at home, Planned Parenthood, and abortion clinic near me. Relative to all these top queries, Planned Parenthood had the highest RSV, followed by abortion pill and abortion clinic (Figure 3). In comparison, specific queries on abortion at home, home abortion remedies, and how to do an abortion at home or how to have an abortion at home were popular but had lower search volumes.

Self induced abortion was the only consistent top query for self abortion. Similarly, buy abortion pill online was associated with only 1 query, buy abortion kit online.

As shown in Table 1, among the 10 states with highest average RSI for home abortion, the majority were those where abortion was severely restricted (5/10) or restricted (3/10). Except for Pennsylvania, none of the top states with abortion restrictions were located in the Northeast. Conversely, New York and Connecticut, 2 states in the Northeast where abortion was accessible, also topped the list with highest searches for home abortion. Similarly, the majority of states with high average RSI for self abortion were severely restricted (2/10) or restricted (4/10) and some states where abortion is accessible, such as Massachusetts, New Jersey, Vermont, and Illinois, also topped the list for self abortion searches. There was no overlap in high-search states for these 2 initial search queries, suggesting differences in state-level interest in home abortion and self abortion searches. Google did not provide results for top states with the highest RSI for buy abortion pill online.

To better provide contextual evidence for our results, we took a snapshot of the top 10 webpages presented to users searching for home abortion, self abortion, and buy abortion pill online in the United States as of April 2021 (Multimedia Appendix 1). Regarding home abortion, only 2 of the top ranked webpages focused on health education, 1 from Healthline, a health blog that cautions against abortion home remedies while recommending that women consider physician-prescribed medication abortion at home; the other, an antiabortion blog from a Wisconsin clinic that focuses on the abortion process and on ways of reversing medication abortion. These 2 sites ranked first and ninth, respectively. A Wikipedia overview article on self-induced abortion ranked eighth. Several webpages focused on potential access expansions to home abortions under the Biden administration and to mail-order abortion pills approved in the United Kingdom during the pandemic. Webpages covering news and scholarly articles were more common than webpages from clinical settings.

For self abortion, webpages from the Guttmacher Institute (a proabortion advocacy organization) ranked first and third, followed by several academic publications.

As many as 4 out of 10 top webpages for buy abortion pill online were sponsored by prochoice groups such as Planned Parenthood, Vox, Ms. Magazine, and Plan C; these webpages openly discuss where and how to get abortion pills online. In contrast, 4/10 webpages presented antiabortion content, 2 managed by Crisis Pregnancy Centers that dissuade people from buying the abortion pill; 1 from a Florida county promoting a referendum to declare itself a “Pro-Life Sanctuary”; and 1 from a business blog cautioning potential online pill buyers from ending in jail because they are doing an illegal activity. Notably, the top webpage from the National Pharmaceutical Provider Association does not include any content on medication abortion. Images of top webpages for buy abortion pill online that were links to internal website search results providing no relevant content for medication abortion are presented in Multimedia Appendix 2 (webpages ranked 1, 7, and 10). The search query used for each of these internal searches included emojis or special characters or both.

The use of medication abortion has been increasing steadily in the United States since its introduction in 2000, accounting for just 6% of all induced abortions in 2001 and almost 40% in 2017 [12]. Such steady growth in uptake of medication abortion coupled with increased access barriers to abortion, including recent disruptions related to COVID-19, may have accelerated the use of out-of-clinic-abortion services in the United States in 2020 [23,24]. We aimed to describe, in near–real-time, population interest in out-of-clinic abortion information and services during the pandemic on Google.

In 2020, home abortion was the most popular search term among those we explored. As shown by the multiple associated queries, this search term reflects interest from searchers in out-of-clinic abortion care. We cannot know the precise type of out-of-clinic abortion information or services being sought by consumers searching home abortion. However, we do know that there were more searches for home abortion—which could encompass any abortion happening at home regardless of clinical support—than for self abortion, a term that we believe implies an interest in self-management of abortion, and buy abortion pill online, a term that implies an interest in self-procurement of medication abortion.

Of the multiple top queries for home abortion, Planned Parenthood received the highest search traffic relative to other search queries for this keyword (based on RSV), followed by abortion pill. The higher search traffic may be associated with the high ranking of the Planned Parenthood website on Google. A previous study indicated that Planned Parenthood was the top webpage for medication abortion searches on Google and the site that provides the most accurate information on medication abortion [30]. Additionally, the higher RSV for Planned Parenthood in relation to home abortion searches may reflect an interest in out-of-clinic abortion involving a provider that can give oversight, medications, and support for an abortion at home and the role of Planned Parenthood as a recognized abortion provider offering telemedicine abortion services both prior to and during the pandemic [31]. In this model, clinicians remotely prescribe medication abortions by collaborating with Planned Parenthood centers that do not have on-site abortion providers. The patient visits their laboratories, undergoes ultrasound, and receives medications from their local Planned Parenthood Clinic.

Until 2020, the US Food and Drug Administration (US FDA) required certified providers to dispense mifepristone, 1 of the 2 drugs in the abortion regimen most commonly prescribed, in clinics or hospitals [9]. However, a federal district court ruled in July 2020 that the US FDA was required to lift this restriction and allow remote distribution of mifepristone via telemedicine during the pandemic. We note that searches for home abortion peaked in November of 2020, during the third wave of the COVID-19 pandemic and national allowance of telemedicine provision of medication abortion [8,9].

Novel platforms such as AidAccess and the Plan C Campaign are facilitating the online provision of abortion pills by offering information, support for, and access to medications for self-managed abortion [9,24]. Our results for self abortion and buy abortion pill online, searches that denote minimal provider support or fully online self-managed abortion care, showed that searches for self abortion were slightly more popular, but data for both were sparse. Our findings indicate relatively low population interest in these search queries for Google searchers in general, and compared with home abortion. Minimally supported or totally self-managed medication abortion with the 2-drug regimen of mifepristone and misoprostol is difficult to implement in the United States, even in states considered supportive of abortion rights, because ordering the drugs online through direct order without a prescription is considered illegal and many foreign clinics or pharmacies do not ship to the United States. By contrast, clinically supported home abortions may overcome multiple legal, economic, and cultural barriers because they may be more private, convenient, affordable, and less stigmatizing than in-clinic abortions. Aiken et al [24], using AidAccess data spanning January 2019 through April 2020, found an increase in the rate of requests for self-managed medication abortion in the United States [24]. Nonetheless, past research estimated that 7% of women in the United States would attempt to self-manage an abortion during their lifetime [14]. Notably, we found that searches for these 2 queries in 2020 peaked before the pandemic, in contrast to searches for home abortion that peaked during a third wave of COVID-19, during a time when telemedicine provision of medication abortion was allowed.

Previous studies have shown that legal restrictions to abortion do not reduce desire or intention to seek abortion care, but may push abortion seekers to virtual sources of information and services [15,24]. In light of such findings, our study provides important contextual evidence about the differences in relative queries across states with varying social attitudes and legal positions on abortion. Home abortion searches were predominantly most popular in states with restricted abortion access such as Arizona, Missouri, Arkansas, Indiana, and Kentucky. Arizona, Missouri, and Arkansas are 3 of the 5 states that prohibit the use of telemedicine for abortion while Indiana prohibits prescription of medication abortions without a prior in-person patient examination [9,31]. In 2020, Kentucky and Arizona enacted laws requiring physical presence of prescribing clinicians, hence effectively blocking the use of telemedicine [9]. Furthermore, several of the states with highest search traffic for home abortion attempted to limit abortion access during the COVID-19 outbreak by deeming abortion “non-essential” [32].

We saw the highest search traffic for self abortion among different states than for home abortion. Although self abortion traffic was mostly concentrated in Southern states where abortion is more restricted, traffic also came from 4 states, predominantly located in the Northeast, that make abortion accessible via telemedicine and by allowing nonphysician-certified clinicians to authorize the medication abortion [33]. Previous research suggested that barriers to clinic access are present even in states with more supportive abortion policies [15]. In these states, barriers to abortion include cost, difficulties taking time away from work or arranging childcare, abortion stigma, and the need to keep an abortion secret for fear of negative consequences [15].

As people turn to the internet for information and resources on out-of-clinic abortion, they can face challenges to informed reproductive choice and abortion access. Consistent with previous research on abortion webpages [30], we found that the top webpages listed in our snapshots provided limited evidence-based clinical content on self-management of abortions, particularly abortions without clinical provider supervision. The webpages linked to buy abortion pills online were neither relevant nor helpful. In fact, some pages offered no content related to abortion (self-managed or otherwise) and their appearance as top search results could be related to efforts to leverage search algorithm optimizations to appear higher on search results either through spam or erroneous linking. Furthermore, several sites presented disinformation about abortion, the pill, and other aspects of sexual and reproductive health, a finding that aligns with past research on the contents of abortion search results on Google [30,34,35]. Multiple legal, financial, cultural, and logistical barriers to abortion care underscore the need to support consumer access to accurate webpages that provide high-quality information and resources.

Our study faced several limitations. Although we researched 3 keywords for out-of-clinic abortions and their associated top queries, it is not an exhaustive list of every term searched. For instance, we considered telemedicine abortion and telehealth abortion, but Google did not give results for these search queries. Moreover, we were not able to identify the number of unique users or their individual characteristics, nor the reasons that prompt individuals to search for out-of-clinic abortion. Additionally, this research did not compare searches in 2020 for out-of-clinic abortion terms with previous years. Rather, we chose to scope our study to pre- and post-pandemic US searches within 2020. Nonetheless, additional sensitivity analyses comparing searches in 2020 with those in 2018 and 2019 showed similar search trends for home abortion and self abortion. Further research should be done to exhaustively explore differences in searches over time, in consideration of the impacts of changes in the volume of all Google searches over time on the frequency of searches for abortion-related terms. For example, time-series heatmaps and other visual representations of key terms by geographic region could be useful in future research and as general tools for understanding these trends.

We also cannot assume that online searches for out-of-clinic abortion reflect intention to use or current use of this type of abortion care. Google data allow us to assess relative population-level search interest and concerns about key topics and search queries in near or real-time. However, Google Trends only shows data for popular queries, so search queries with low volume appear as “0.” The Google Health Trends API does not give RSV below a certain threshold (unknown to us). Following prior research, we addressed this limitation by implementing a resampling approach over several months. We treated each of 30 data extractions from the Google Trends and Health Trends API as an independent sample and calculated average measures (RSI score and RSV score with nonmissing averages). We believe this is a valid way to account for inherent sampling fluctuations created by Google’s own mechanisms that intentionally obfuscate precise search activity at any one point in time [36]. We urge other researchers to consider the resampling approach in future analyses of infrequent Google search data.

Additionally, although we chose our initial search queries carefully, with consideration of relevant literature and search interest, further research should explore other search queries related to self-management of abortion to establish user interest for other relevant queries. As for top webpages that searchers of out-of-clinic abortion are shown on Google, we took a snapshot of these based on a 1-day retrieval of data; these listings and rankings are likely to fluctuate over time. Future research should examine top webpages and their rankings by frequent resampling over time and do a thorough content analysis to gain further insights into the content and quality of webpages providing information on abortion self-management to people searching on Google.

Our analysis provided meaningful insights into population-level interest in out-of-clinic medication abortions in the United States during the first year of the pandemic. Our findings demonstrate greater interest in home abortions, which presumably have varying degrees of clinical support than in minimally or nonclinically supported self-induced abortions. While our study was mainly descriptive, showing ways in which infrequent abortion-related search data can be analyzed, future studies should explore correlations between the keywords denoting interest in out-of-clinic abortion and abortion care measures and test models that allow for improved monitoring and surveillance of abortion concerns in our rapidly evolving policy context.