Page 76 - DP-2-2

Design+ Meaningful digital consent: A VSD study

a plan for when they are in crisis. At present, all data in frequently asked questions (FAQs; Strava), and key point
the Hope app are stored on the phone. However, there is summary (Canva and Coursera).
30
growing interest in exploring how the data might be used Interviews were recorded and transcribed verbatim.
to improve suicide risk detection, prevention, and care Transcripts were analyzed using rapid qualitative analysis, a
services. As such, this study will inform the development pragmatic implementation science method suited for time-
30
of a meaningful consent process as the app and other AI sensitive research. 34,35 A template of key domains from the
initiatives at CAMH evolve. This study received ethics interview guide was used to summarize transcripts, with
1
approval at CAMH (CAMH REB#82_2020). notable quotations integrated where relevant. The research
3.1. Subject recruitment team independently summarized three transcripts to
ensure consistency, and discrepancies were resolved by the
In this study, we conducted semi-structured interviews lead author (NS). Transcripts were divided among the team
with Hope app users (past, present, and potential users). for paired summarization, with all summaries reviewed
Participants were recruited through electronic posting for accuracy. The summaries were exported into Microsoft
on the CAMH online research registry and the EENet Excel for matrix analysis, identifying common themes
Community of Practice, and flyers posted in waiting across key domains. Themes were reviewed and finalized
areas at CAMH. Interested participants contacted the through team discussions. Participant characteristics from
research team. To be eligible, participants had to be pre-interview surveys were summarized using descriptive
English-speaking, at least 18 years old, Hope app users, statistics.
and provide informed consent to participate in the
study. Recruitment occurred between October 2023 and 4. Results
August 2024.
4.1. Participant characteristics
A maximum variation strategy was used to recruit
a diverse range of perspectives so as to identify cross- Nineteen participants were included in the diverse sample,
cutting themes. Maximum variation was sought based on capturing maximum variation across gender, age, race/
user type, digital health use, history with CAMH, trusting ethnicity, history with CAMH, trust, privacy knowledge,
disposition, privacy knowledge, and privacy protection and confidence in protecting privacy (Table A1). Three-
behaviors. These characteristics were selected based on the quarters of the participants identified as men (42.1%)
eHealth Trust Model – a theory-grounded, evidence-based and women (31.5%), with the remaining individuals
patient privacy research framework. The model constructs identifying as either non-binary (21.1%) or transgender
31
were also validated in follow-up studies on privacy and (5.3%). More than half of the participants were less than
5,32
consent. Race, age, and gender were also monitored for 29 years old (52.6%). There was a diversity of ethnic
socio-demographic diversity, as there are equity and digital identities, with 36.8% of participants identifying as White
divide implications across different population groups. individuals. While most participants use digital tools and
5
Variation was determined through responses to an 11-item services routinely (68.4%), approximately one-quarter of
pre-interview survey through REDCap. 33 the participants have used the Hope app at some point
(26.3%). Less than half of the participants have received
3.2. Data collection care at CAMH (47.3%). Most of the participants self-
One-on-one interviews were conducted virtually reported as being trust (47.3%) and knowledgeable about
through Webex video conferencing software, each lasting privacy rights (47.3%). Their trust levels varied, with 47.3%
approximately 90 min. A semi-structured interview format of participants leaning toward trust (rating of 5+ out of 7),
was selected to allow the interviewer (NS) to pursue ideas 31.5% leaning toward distrust (rating of 5+ out of 7), and
in greater depth, when necessary. Key aspects related 21.1% being neutral. There were mixed levels of knowledge
to the use of the research data, the voluntary nature of privacy rights and risks, with more participants reporting
of participation, and assurance of confidentiality were good or very good (47.3% and 42.1%, respectively). There
emphasized before starting the interview. Participants was also a mix in levels of confidence in protecting privacy
were given the opportunity to ask questions before rights; however, 52.6% of participants reported being
electronically signing the consent form (via REDCap). The confident (42.1%) or completely confident (10.5%).
research team developed and piloted the interview guide Overall, most participants described their experiences
to explore participant values (empirical investigation) and with privacy notices as negative due to the length and
thoughts on innovative privacy notice designs, such as the complexity of the notices. Engagement with privacy notices
nutrition label approach (iOS/Android app store), table varied, with most skim reading the full text or skim reading
of contents (Slack), emphasized key points (Twitter/X), for key points on the apps’ data practices and looking for

Volume 2 Issue 2 (2025) 5 doi: 10.36922/dp.8158