JMIR Publications

ABSTRACT

Background:

Growing interest surrounds how internet search behaviours might provide digital signals of disease prior to diagnosis, for example when people search symptoms or potential remedies online. Internet browsing data offers novel opportunities for understanding response to symptoms, public health surveillance and early intervention in conditions such as cancer. However, the acceptability of using such sensitive data in medical research remains unclear, particularly among individuals at higher risk of health and digital exclusion, such as older adults and those from minority ethnic groups or with a lower socio-economic status.

Objective:

To explore the feasibility and acceptability of using internet browsing history data for health research.

Methods:

Twenty participants (ten with a history of cancer and ten without) were purposively sampled to ensure representation from groups at risk of digital and health inequalities. We conducted semi-structured and think-aloud interviews allowing participants to reflect on hypothetical research involving sharing their internet browsing data. The Adapted Theoretical Framework of Acceptability guided the interview structure and coding. Interviews were transcribed, coded in NVivo and thematically analysed. Patient and public involvement informed the study approach, participant-facing documents and interpretation of the findings.

Results:

Trust and transparency were fundamental to participants’ willingness to share data. Trust in researchers was key and would have to be earned through clear communication, ethical data handling and familiarity with a named research team. Privacy concerns were prominent, particularly regarding non-health-related information, with participants wanting control over what was shared. Potential (mis)use of data beyond the original research purpose caused more concern than the nature of the shared data itself. Participants valued potential advancements in early detection but doubted that their internet browsing history would have sufficient value. Digital literacy varied; many expressed concerns over the technical aspects of sharing data.

Conclusions:

Participant recommendations balanced privacy concerns against the potential of internet history data for early diagnosis and health research. The study highlights ethical and inclusive approaches to health research using internet browsing history. Future researchers should consider defining the scope of health specific data filters; providing user-friendly information and guidance for study participants; and ensuring that participants are able to contact research team members to build trust and facilitate data sharing. Clinical Trial: Test