An 18-item electronic survey was developed using Qualtrics (Qualtrics Inc). Privacy-related survey items published by Seltzer et al [23] and validated and published by Doherty et al [25], as well as demographic-related survey items validated and published by Zhu et al [26], were selected, adapted, and revalidated for bias, relevance, and cognition among a convenience sample of 5 individuals who identified as both patients and health consumers. Based on the pilot participants’ feedback, the survey questions were refined to improve item quality and clarity and overall instrument clarity, appropriateness, and relevance.

In April 2021, the electronic survey was administered to adults (aged ≥18 years) registered in the ResearchMatch database [27] who agreed to be contacted to engage in the survey after receiving an informational electronic invitation letter via the ResearchMatch platform. ResearchMatch is a registry and collaborative project that is maintained at Vanderbilt University and overseen by the Vanderbilt University Institutional Review Board. ResearchMatch participants live within the US and Puerto Rico, are of all ages, races, and ethnicities, and comprise healthy volunteers as well as those living with medical conditions. Access to the ResearchMatch platform was provided through Ohio University.

Individuals were invited to participate in the survey regardless of health status, race, gender, or any other mutable or immutable characteristics. Survey participants were informed that their participation was entirely voluntary. No survey questions were mandatory, and participants were informed that they could skip any question at any time. Survey participants were welcomed to contact the research team at any time with any questions or concerns about the study. Reminders were sent up to 3 times to participants who began but had yet to complete and submit the survey within the study timeframe.

Participants’ personal contact details were received only after participants agreed to participate in the survey. Those who completed and submitted the survey were offered a random chance to receive a US $250, $100, $50, or $25 gift card. A random selection tool was developed and deployed using Microsoft Excel (Microsoft Corp) to randomly select email addresses of survey participants and deliver the gift card incentives. Participant email addresses were deleted or destroyed to prevent reidentification at the conclusion of the study.

This study is part of a larger investigation to explore the privacy-related experiences and perspectives of US adults and their willingness to share digital data with researchers [24]. This analysis centers on a sample of ResearchMatch participants with a 100% response or completion rate (inclusive of completed surveys with items containing no responses). A Qualtrics tool was used to calculate an ideal survey sample size (n=384) based on the total ResearchMatch population (95% CI; 5% margin of error). Descriptive statistical analyses were conducted using Microsoft Excel. Manual queries (single layer, bottom-up topic modeling) were used to categorize illuminating quotes from free-text responses.

This study was reviewed and approved by the Ohio University Institutional Review Board (20-E-457).

This is the first and only study to explore the perspectives of a subpopulation of research participants of childbearing age about the health-relatedness of their digital data and accompanying privacy concerns. As mentioned, this assessment is part of a larger study to explore the privacy-related experiences and perspectives of US adults and their willingness to share real-world digital data with researchers [24]. Our findings indicate, in light of Dobbs v Jackson and other related events [28], a need to not only understand these phenomena, but also determine strategies and opportunities to educate research participants about how seemingly non–health-related data can become contextually relevant to abortion and miscarriage care.

For instance, a recent article published by the Brookings Institution [28] discussed how, recently, courts of law seeking to convict abortion seekers admitted data and evidence collected from individuals’ location data, text messages, and online activity. They highlighted 4 recent events:

Our survey findings show that ResearchMatch participants are concerned about these very types of privacy-related events occurring in their lives. Our findings also indicate that ResearchMatch participants are likely correct in their judgments that social media data, email data, text message data, Google search history data, and online purchase history data are health-related in this context. Ironically, however, these events and our findings indicate that ResearchMatch participants are likely incorrect in their assumptions that ride-sharing history and geolocation data are not health-related. Moving forward, it will be critical to identify and implement resources to educate ResearchMatch participants and others about cases that demonstrate the contextual health-relatedness of ride-sharing history and geolocation data [33-37]. It will also be critical for data companies, app developers, and their collaborators to provide and support research and education about privacy-invasive software development kits and application programming interfaces that collect geolocation data.

Just prior to Dobbs (May 2022), in an attempt to educate the public about their privacy practices, SafeGraph released a public statement and formal letter in light of requests from legislators to explain how SafeGraph handled physical location and geolocation data [29]. The company explained their practice of aggregating data, implementing rigorous “differential privacy” [38], and removing device-level information and identifiers from data products prior to sharing them with third parties. They also stated that for their Patterns data product, which includes mobility data, they remove “aggregated, anonymized visit statistics for all businesses categorized by the North American Industry Classification System (NAICS) Code for family planning centers.” In the absence of a comprehensive federal data privacy law and amid an emerging patchwork of US state-level comprehensive privacy laws, companies like SafeGraph and others who track ride-sharing history and geolocation data should also collaborate to develop best practices for privacy and publicly endorse human rights protections for managing, processing, using, and sharing data that implicates businesses categorized by the North American Industry Classification System code as family planning centers (code 621410) [39]. By doing so, companies would not only provide transparency regarding their current practices to educate policy makers, regulators, and the public, but also cultivate strategies to improve their privacy practices in collaboration or consultation with neutral or third-party privacy experts.

In light of Dobbs v Jackson and related events [40], trust as well as data sharing and use transparency are necessary to recruit and retain participants in health research involving digital data. For instance, although a patchwork of comprehensive US state privacy laws [41] and a recent US executive order [42] exist to help preserve digital data privacy and protect access to reproductive health care services, respectively, further assurances are needed or should be made explicit to affirm the following: (1) the right to consent to the use, processing, storage, and deletion of data most perceived as health-related (ie, social media data, email data, text message data, online search history data, online purchase history data, electronic medical record data, fitness tracker or wearable data, credit card statement data, and genetic data); (2) protection against risk of fraud or abuse based on personal information by governments, corporations, health care providers, and insurers; (3) protection against data uses beyond the scope of consent or explicit statements in privacy policies; and (4) data confidentiality, security, and discretion.

Limitations should be noted when interpreting our results and directing future work. First, while our results provide important insights, they are merely descriptive due to our small sample size. Second, our analysis is not based on a representative sample of ResearchMatch participants or US adults of childbearing age. Most of our study participants reported high levels of education and most had at least a bachelor’s degree. Future work should endeavor to explore this topic in a more representative and larger sample of US adults of childbearing age.

Although income levels were not captured in this survey, educational level may serve as a proxy for income. As mentioned, most participants in this study had at least some college-level education or above, rendering it possible for them to have affordable and safe access to family planning services. According to the US Centers for Disease Control and Prevention [43], in 2019 women in their 20s accounted for the majority (56.9%) of those receiving abortion care nationally, though abortion ratios were highest among adolescents aged ≤19 years. Given the relatively low proportion of participants aged ≤20 years who completed our survey and given that the largest proportion of individuals engaged in the survey were aged 21 to 40 years, it is possible that the concerns and perspectives herein do not reflect those of adolescent women, who are most vulnerable to a lack of safe and affordable access to family planning services.

Our survey was developed on the basis of a recent systematic review and a survey study that both determined that age, income level, and education level were the strongest predictors of online or digital-footprint activity [44,45]. Thus, we collected these demographic variables in the survey but not sex, gender, or gender identity, as these variables were not suggested to be predictors of online or digital-footprint activity. Yet it should be noted that since the start of this survey, a more recent assessment of ResearchMatch participants has shown a large and growing number of participants (n=154,200) within the registry, with nearly 70% reporting as female and nearly 42% reporting as female parents. Thus, there is opportunity to explore the effect of gender on ResearchMatch participants’ views on the health-relatedness of their data.

Although our survey was conducted in April 2021, prior to the June 2022 US Supreme Court ruling in Dobbs v Jackson, ResearchMatch participants’ data privacy concerns may have been affected by COVID-19 and other major events that occurred during 2020 [23]. This is especially relevant given that recent studies show that individuals’ data privacy concerns were likely heightened during the COVID-19 pandemic, as privacy laws and regulations were relaxed to accommodate digital public health surveillance and remote health care [19,46]. Therefore, further research should explore how this recent US Supreme Court decision might impact female and female parent perspectives on the health-relatedness of their data and their concerns about donating data to researchers today or in the future.

The US Supreme Court decision in Dobbs v Jackson has sharpened the focus on the ways in which populations perceive the health-relatedness of their data, on discretion around how data are used and processed, and on the organizations with whom data can be shared or disclosed. Recent uses of digital-footprint data to convict women in courts of law for engaging in abortion or undergoing certain types of miscarriage are both controversial and alarming. Therefore, it has become drastically important to understand the perspectives of health research participants of childbearing age regarding the health-relatedness of their digital-footprint data and their privacy concerns about sharing their digital-footprint data with researchers. After Dobbs, opportunities to educate research participants about the health-relatedness of their digital data are paramount. Determining strategies and privacy best practices for using digital-footprint data related to family planning with discretion should be a priority for companies, researchers, families, and other important stakeholders today and in the future.