Features to Build Consumer Trust in Patient Data Sharing

Patients are interested in sharing their own medical information, so long as that patient data sharing comes with features like informed consent, transparency, and consumer oversight, according to a new paper in JAMA Network Open.

These findings are important as more medical research hinges on patient opt-in for health data sharing. The bevy of data patients generate simply by interacting with their own health and well-being—clinical data, wearable data, information from social media, and data gleaned from mobile apps—can all be valuable to improving patient and population health, the researchers said.

But that movement toward using more consumer-generated data has also come with privacy concerns, as patients worry about how different entities will actually use that data and how they can keep their information private and secure.

“The growing collection of digital health information and blurred lines between health and nonhealth data also raise privacy and security concerns that are in tension with benefits,” the researchers pointed out.

For instance, the Dobbs v. Jackson Women’s Health Supreme Court decision has raised questions about the privacy and security of data aggregated from period tracker apps or even whether website purchases could compromise reproductive health information, the team said.

Given these fears, researchers are working to strike the right balance between patient autonomy and trust in health data sharing, ensuring that they might opt into sharing their information with actors who will use it to improve patient and population health.

This latest study, completed by experts at the Johns Hopkins School of Medicine, particularly looked at the issue of informed consent in patient data sharing, underscoring that consent is not well studied and can sometimes fall short of actually informing the consumer.

The researcher surveyed more than 3,500 about their willingness to share their personal digital information during 192 different scenarios. The different scenarios reflected four possible privacy protection elements, three uses for the information, two users of the information, and two sources of information.

Respondents became increasingly more willing to share their health information with each individual privacy protection—consent, transparency of collected data, consumer and regulatory oversight, and ability to delete data—that was added to a scenario. Consent was the most impactful privacy protection, followed by data deletion, oversight, and then transparency.

“The central role of consent may reflect the value placed by consumers on preserving autonomy and the ability to choose whether and how their personal data are used,” the researchers said. “These results affirm the importance of establishing consent as a baseline model of promoting digital data privacy.”

The purpose of the health data use was also extremely important, the researchers found. The purpose of use was more important than any privacy protection on its own, but having all four privacy protections in place ended up being more impactful than the purpose of use.

More specifically, patients were more likely to opt into data sharing if the information was being used for research purposes, compared to clinical purposes or marketing purposes.

There were some differences by demographic, the team added. Older adults with conservative views were less likely to opt into data sharing than younger, liberal-leaning individuals. Those with conservative ideology were more lenient if data was being used for national security purposes, while liberal-leaning people were more lenient when data was used to quell the spread of COVID-19.

Willingness to share health data was similar across racial groups, although White people were more interested in giving consent before sharing their information.

These findings do pose a key contradiction: patients want some sense of autonomy over their health information privacy, and particularly value consent, but often click through privacy disclosure without reading them.

“Several potential reasons explain this contradiction, including that the agreements are often cumbersome to read and privacy protections difficult to understand; the desired product or service is more appealing than the potentially lost privacy; and the responses in our study based on hypothetical scenarios may not be trustworthy,” the researchers posited.

Consumers may also feel a sense of resignation regarding how well their medical information may be protected.

Still, these findings give a key glimpse into the factors that could motivate patient data sharing, a practice that will be essential should more big data projects continue in light of population health research imperatives.

“Given the growing complexities of data sharing, unpredictable future uses of data, and the infeasibility of repeatedly acquiring consent for new uses, one approach to protecting consumer privacy is to implement a combination of individualized and early consent with collective and ongoing governance,” the researchers concluded. “Such a model would reduce individual burden while maintaining protections. Moreover, transparency and comprehensibility must apply both to the specific data being shared as well as how the data were collected and used.”