Martin Andersen was frustrated.
When location data company SafeGraph last week suddenly turned off access to abortion clinic data that Andersen had been using in his research, the associate professor in the department of economics at the University of North Carolina at Greensboro lost a valuable source of information. Andersen had been using the data to measure the real-world impact of laws limiting abortion rights.
“I think it’s incredibly valuable to help us understand how people are responding to laws like SB 8 or overturning of Roe v. Wade,” said Andersen. He was referring to SafeGraph’s data and its use in research gauging the effects of Texas’ Senate Bill 8, which bans most abortions after six weeks, and a potential decision by the Supreme Court to overturn Roe v. Wade.
“I am kind of frustrated they did it, but I understand why they did this as well. I think the decision to take down family clinic data is a very defensible decision,” Andersen said regarding SafeGraph’s data removal.
SafeGraph said on May 3 it would cut access to data associated with travel to and from family-planning center locations from its online self-serve data platform and from the API through which it distributes data to customers and researchers. The company sells data to business and government customers that shows where anonymized mobile devices are spotted to indicate which locations people traveled from, how long they stayed and where they traveled afterward. The company pointed to “potential federal changes in family-planning access” as the reason for its decision to remove access to the data.
However, while the move was designed to protect people’s privacy, it had broader effects on researchers studying the impact of laws on access to reproductive health services.
“Data limitation is a severe issue for research generally, but especially so for research in the area of reproductive health care,” said Jason Lindo, a professor of economics at Texas A&M University who studies the impact of abortion restrictions on reproductive health care access and the amount of travel necessary to obtain it.
“There are a lot of big questions in terms of who is served by abortion clinics, and how far do women come from in seeking care, and what factors affect the amount of care that they can provide. The SafeGraph data really opens new opportunities to get at this question,” said Lindo, who has not personally used SafeGraph data.
“If Roe falls this summer, as certainly seems a pretty distinct possibility, and policymakers and the public are trying to understand the impacts of state abortion bans, these data would allow us to have a near-immediate window into what begins to happen as women travel across state lines,” said Caitlin Knowles Myers, an economics professor at Middlebury College. She also has not used SafeGraph data in her research, but she said a student of hers has.
“There’s certainly a potential role for this type of location data to give really quick, high-frequency evidence on the effects of restrictions on people seeking abortions,” Knowles Myers said. She said location data can fill in data gaps, as some states do not keep track of the number of abortions performed in their locales — when they do, it can take years for this vital statistical data to be released.
Ethical gray areas
While some academic researchers believe that mobile location data showing travel to and from abortion clinics — such as what SafeGraph provided — is useful for their work, recent state laws limiting abortion services and the potential for the Supreme Court to overturn Roe v. Wade have exposed the privacy risks associated with the easy availability of location data.
The situation shines a spotlight on a complex debate that has plagued researchers, particularly those in need of sensitive data related to health care. Researchers want data, and the increasingly digitized world enabled through connected mobile communication creates new forms of information that can offer valuable new insights to researchers. However, researchers care about the privacy of research subjects, and must comply with academic review board data use and privacy rules.
“We see this issue constantly — the trade-off between access and privacy,” Andersen said.
For instance, in the past, privacy and intellectual property protections have limited data details and access to Facebook data for researchers; some have decided not to use it at all as a result.
Do I have an ethical obligation to think through those issues? Yes, and I am, and I don’t have an answer for you yet.
But the use of location data poses dilemmas beyond balancing privacy and access.
Most mobile location data providers rely on harvesting and selling people’s personal data obtained through covert data partnerships with unnamed ad tech companies and mobile app providers, which fuels what many perceive as surveillance capitalism. Tech and data providers are sometimes eager to open access to their information for research, not only because it helps show their products’ value and promote their companies, but because it gives them a “data-for-good” cover in an otherwise highly criticized industry.
“As I’m watching this conversation about SafeGraph play out, I have every obligation to really take a pause and think through these issues, knowing that more broadly there’s this question of whether people understand their data are being collected, whether they truly have provided consent and whether they are potentially being manipulated without even realizing it,” Knowles Myers said.
“Do I have an ethical obligation to think through those issues? Yes, and I am, and I don’t have an answer for you yet,” she said.
SafeGraph declined to comment for this story.
Location data-based research used in court
Andersen told Protocol he has two research papers in progress that rely on data from SafeGraph, both related to travel associated with abortion clinics: one that analyzes the effects of the COVID-19 pandemic and another that focuses on the effect SB 8 has on travel to reproductive health facilities.
Andersen said the SafeGraph data he had already downloaded showed the number of unique devices that had traveled from a specific census block to a reproductive health clinic on a weekly and monthly basis. SafeGraph’s decision means he will not be able to gather updated information showing future implications of legal abortion restrictions to compare against that historical data.
Andersen also co-authored a 2020 paper for the National Bureau of Economic Research that used SafeGraph foot traffic data to estimate the impact of early COVID-19-related state lockdowns on the number of abortions that may have been performed.
Research quantifying the travel burdens created by abortion restrictions has been considered in abortion-related legal cases since the 2016 Supreme Court case Whole Woman's Health v. Hellerstedt, in which the court ruled that Texas cannot place restrictions on abortion services that create an undue burden on people seeking an abortion.
“In the case of reproductive health care and abortion access, the research has mattered. Showing how travel distance affects abortion rates has been used by courts to determine that laws have constituted an undue burden,” Lindo said.
While Lindo has not used SafeGraph data in his work, his research quantifying the degree to which abortion rates decrease when people are forced to travel farther to reach an abortion provider than they would have without restrictions in place was used to inform a court decision blocking enforcement of an Arkansas law.
Research by Lindo and Knowles Myers predicting the effects on travel distances to access abortion services if Roe v. Wade was overturned or substantially weakened is cited in a 2021 amicus brief submitted to the Supreme Court.
The nuances of location data privacy
Mobile location data has been the subject of great scrutiny by privacy, data security and human rights advocates for many years. Amid the threat of more abortion bans, there is more concern than ever that location data could be used by law enforcement officials, government agencies or everyday people to detect clinics providing interstate abortions, or people obtaining abortion services outside their home states.
When Motherboard reported last week that another location data company, Placer.ai, offered easy access to data showing the approximate home locations of visitors to specific Planned Parenthood facilities, Sen. Ron Wyden linked to the article on Twitter, noting, “Researching birth control online, updating a period-tracking app or bringing a phone to the doctor's office could be used to track and prosecute women across the U.S.” Wyden pointed to his proposed Fourth Amendment is Not For Sale Act, which he said “would make it harder for Republican states to persecute anyone seeking an abortion by weaponizing their personal information.”
But some researchers say they appreciate SafeGraph’s data because, unlike other location data providers they could purchase from, SafeGraph reports foot traffic data according to census block groups rather than providing device-level information or precise GPS coordinates.
“I could go out and buy a device-level data set, but there’s no way I’m going to touch that,” said Andersen — he is more comfortable using SafeGraph’s data for that reason.
While Andersen and others argue SafeGraph’s approach provides a greater degree of anonymity than what other location data suppliers provide, in cases where a small number of devices appear in a given location, even census block-level information could be used to identify someone.
A chance for middle ground?
Already, academic researchers must satisfy the criteria of review boards to get approval for their research plans, including data sources. “Researchers are super, super accustomed to dealing with these circumstances in using these sorts of sensitive data all the time,” Lindo said. “The balance is in finding out how to make sure that people are protected.”
Lindo and others suggest that rather than shutting off the spigot to abortion clinic data to everyone, SafeGraph could consider ways to provide it safely to vetted researchers. “It is very reasonable for them to require [Independent Review Board] approval to buy data,” Lindo said. “It doesn’t have to be an all or nothing thing.”
Knowles Myers agreed. “I would like to hope that there’s that middle ground,” she said. “All these data vendors should be thinking about protecting individuals anyway.”
In a recent interview with Protocol, SafeGraph CEO Auren Hoffman said the company might consider altering its approach to data access. “We could say, ‘Only vetted researchers can get access to this data, whereas the broader public can get less access to the data,’ and that's something we might do. So we are evaluating those types of things,” he said.
“In general, we would support carve-outs for research, especially if they’re focusing on institutions that are well-vetted,” said Caitlin Seeley George, campaign director at Fight for the Future, a digital and privacy rights group. Ultimately, she said, the federal government should step in to define national rules related to data use.
“We need data privacy legislation, because in the end, we can’t rely on companies to make the decision to just share with researchers,” Seeley George said. “They’re going to do what makes the most money, which is likely selling that data to anyone who wants to buy it.”