A recent Comment, published in The Lancet Digital Health, has provided key insights into public perceptions on data sharing.
Data science and artificial intelligence (AI) have the ability to transform how healthcare is delivered. The ongoing COVID-19 pandemic has been a key example of how global data sharing improves data-driven responses. Researchers have been able to rapidly collect, analyse, model and report data timely.
Nevertheless, data sharing in some cases is still a controversial subject. The most common concern is in relation to privacy. Despite efforts to ensure anonymity, malicious hackers remain a risk in re-identifying individuals. For many, the risk of data sharing outweighs the benefits.
Previous surveys have explored the range of views that the public hold about data sharing. It is clear that the public question the safety of their data. In particular, they are concerned about sharing their health data with third parties.
In this study, researchers surveyed individuals across the UK and USA to examine public attitudes towards data sharing, data access and the use of AI in healthcare. The team selected these countries as comparators, as both are high-income countries with significant investment in advanced health information technology and data-driven care. They also wanted to observe whether the contrasting healthcare delivery systems would influence public attitudes.
The researchers developed an online YouGov survey. From this, they obtained data from 3,194 respondents aged >18 years across the UK (n=2,080) and USA (n=1,114).
Perceptions on data sharing
Whist more respondents were aware of electronic health records, their use was not widespread, particularly in the UK (14.3% vs 44.5% in the USA). Access to medical records was also low (59.0% in the UK and 70.1% in the USA). Nonetheless, many respondents still regarded access to their medical records as useful.
Willingness to share anonymised health information was different depending on who was receiving it. Respondents were less willing to share their anonymised data to institutions using it for a more commercial purpose, e.g. insurance or tech companies. Older respondents were less trustful on sharing their personal information with any organisation compared to younger respondents.
Most respondents, from both countries, felt happier to receive advice from a healthcare professional without assistance from AI (58.2% vs 54.8% in the UK and 51.8% vs 38.2% in the USA). However, a small majority of UK respondents, in contrast to the USA, felt happy to receive care from AI software if it was there just as a support for healthcare professionals. Younger respondents were more open to receiving advice from independent AI than older respondents.
These results highlight that access and use of EHRs is not widespread. There are several barriers to accessing health records, including poor usability and interoperability, language barriers and overall mistrust in healthcare systems. Previous studies have shown that in the USA, ethnic minorities and lower income patients adopt EHRs less often than white and higher income patients. This emphasises the importance of tackling inequity in digital health. Patient involvement in the design of digital applications will likely be key in responding to the rapidly evolving expectations of digital services.
Another concern is the lack of trust. In the USA, over a third of patients were not willing to share their data with their doctor. It is clear that individuals are concerned about people profiting off their data. This was reinforced by the fact respondents were less willing to share their data with commercial end-users.
The team highlighted that these results represent a sharp shift in public opinion compared to the 2016 Wellcome Trust survey. In this survey, respondents were willing to share data with commercial organisations. This shift represents the increased public awareness of risks and consequences of data sharing with commercial industries.
The researchers emphasise the need for public standards that secure regulation and transparency of data use and sharing. They also suggest the importance of communicating this to patients clearly so that they understand how and why their data is being used.