Edinburgh Explorer Using a knowledge exchange event to assess study participants’ attitudes to research in a rapidly evolving research context

The UK hosts some of the world’s longest-running Background: longitudinal cohort studies, who make repeated observations of their participants and use these data to explore health outcomes. An alternative method for data collection is record linkage; the linking together of electronic health and administrative records. Applied nationally, this could provide unrivalled opportunities to follow a large number of people in perpetuity. However, public attitudes to the use of data in research are currently unclear. Here we report on an event where we collected attitudes towards recent opportunities and controversies within health data science. The event was attended by ~250 individuals (cohort members Methods: and their guests), who had been invited through the offices of their participating cohort studies. There were a series of presentations describing key research results and the audience participated in 15 multiple-choice questions using interactive voting pads. or 5/5) pregnant mothers (97% agreed) or extending invitations to their own children and grandchildren (98%). Our participants were broadly supportive of research access Conclusions: to data, albeit less supportive when commercial interests were involved. Public engagement events that facilitate two-way interactions can influence and support future research and public engagement efforts. The study describes a real-time survey of cohort participants’, and their guests’, attitudes about research. The limitations of the work are driven mainly by selection bias and missing information. All of the participants are closely connected with longstanding cohort studies. Unfortunately, the authors did not collect how long each participant was in the cohort, only which cohort. That said, because of the nature of the cohorts, some participants have been involved for over half a century. Keeping research subjects engaged is a huge problem. Any research subject who continues for over 50 years is not typical of the population. A valuable reference point would have been the overall percentage retention of the various cohorts. It is possible that these individuals are a highly selected fraction of the starting cohort, and that most participants already dropped out. Since each individual identified their cohort and guest could be separated from cohort participants, it seems surprising that among-cohort comparisons were not done. This would have strengthened the paper. The results were further biased by the fact that this was a gala celebrating the successes of cohort-based research. Therefore, the participants were presented with success stories, and then asked their opinions. The positive nature of the evening is likely to bias results to the positive. The negative reaction of pharmaceutical research was striking. However, one is left to wonder whether results would be different had the gala been celebrating the many breakthroughs and lives saved by pharmaceutical companies. Overall, the work is sound in terms of the conclusions drawn from the data. However, the selected population is so narrow and atypical that few, if any, generalizations can be made from the results to the general population. We are left with the unsurprising conclusion that people who have chosen to participate in research for decades are positively disposed toward research. Stylistically, the

pregnant mothers (97% agreed) or extending invitations to their own children and grandchildren (98%). Our participants were broadly supportive of research access Conclusions: to data, albeit less supportive when commercial interests were involved. Public engagement events that facilitate two-way interactions can influence and support future research and public engagement efforts. AMM has received research support funding from the Sackler Trust, Eli Lilly and Janssen and has received speaker's fees Competing interests: from Illumina. DJP is PI for Generation Scotland and a member of ACONF and of UK Biobank. He reports no other conflicts of interest.

Introduction
The UK has hosted some of the world's longest running longitudinal cohort studies of health and wellbeing. These long-term projects make repeated observations of their participants and use these data to explore how factors such as health, wealth, family, and education influence health outcomes and mortality. Together, these studies have led to several thousand publications (e.g. Generation Scotland, 2019;Lothian Birth Cohort, 2019;UK Biobank, 2019), and to policy changes that have impacted national and global health (Pearson, 2017;Power & Elliott, 2006).
Cohort studies are, nevertheless, highly resource intensive and subject to participant attrition. It is also difficult to make them future-proof; for example, it is not possible to anticipate every measure that may become of interest to researchers in the future. Furthermore, inevitable changes to lifestyle and technology can make cohort data collected decades ago less relevant to current circumstances.
An alternative method to longitudinal 'face-to-face' follow-up of individual participants is record linkage: the linking together of electronic health and administrative records, which are routinely collected (e.g. as part of a hospital visit or census). Although not collected with research in mind, these records can nevertheless be combined to produce a comprehensive and longitudinal dataset. If applied nationally, this type of dataset could provide unrivalled opportunities for researchers to follow a large number of people in perpetuity. Data linkage also has several advantages over face-to-face follow up, not least the fact that it is participant-passive and of negligible burden to the participant. This in turn reduces study attrition and increases the representativeness of study findings. Record linkage is also more flexible than face-to-face assessments, as it can be updated to capture new events, exposures and outcomes.
A recent extension to record linkage studies, particularly in Denmark and Sweden, has been the identification and analysis of dried neonatal blood spots. These were originally obtained as heel-prick neonatal blood samples and used to detect inborn errors of metabolism. Nick-named 'Guthrie Spots' after the physician who devised them, these dried blood spots have been collected and archived by NHS Scotland since 1965; and now number around 3 million in total. In Denmark, there is a longestablished biobank of newborn blood spots that is available for anonymised research (Norgaard-Pedersen & Hougaard, 2007). However, such research access has not yet been granted in Scotland. Nevertheless, Generation Scotland have demonstrated the feasibility of using adults dried blood spots for DNA methylation studies, and have shown that they can accurately replicate the findings made with fresh peripheral blood (Walker et al., 2019). Thus, the NHS Scotland blood spot archive has a high potential research value.
Nevertheless, record linkage studies, including those that use archived blood spots, also have several drawbacks. These include their dependence upon administrative recording processes, which may not be standardised within large organisations like the NHS. It is important to consider that administrative records are not collected with research in-mind, and data may be of lower quality or need substantial pre-processing before it can be used. Furthermore, the systems and legal basis for the use of archived data and samples may vary depending on which organisation is responsible for their retention. Even in situations where the data and samples are available for research, it would be impossible to obtain informed consent from all of the individuals to whom the data and samples relate.
Public attitudes to the use of such data and samples for research are currently unclear. It is not known what proportion of the public are aware of their retention, their value and whether they would approve of their diversion for approved forms of research. It is also unclear whether the public would approve the use of samples such as blood spots for all research, and if so with what sort of regulatory oversight and approval mechanism?
Similarly, researchers are interested in public opinion on other tricky issues such as: Should children be allowed to consent to their own participation in research? Who would you trust with your data? Should predictive brain scans be offered for later mental illness?
Here we report on an event at which we sought to engage with individuals and their families from across diverse Scottish research cohorts. We aimed to both share our recent research findings and to assess attitudes towards recent opportunities and controversies on topics such as: electronic health record linkage; the repurposing of biological samples for research use; and the involvement of commercial interests (amongst other topics). By collecting these opinions from our cohort members, we sought to better understand their views and to provide a basis for further public engagement on these issues. In particular, by asking individuals who had taken part in research to bring along a guest, we also sought to test whether the individuals who had participated in research differed in their attitudes towards data linkage and analysis when compared to those who had not.

Participants
This study reports on the purpose and findings for an 'all cohorts' meeting under the banner "A Celebration of Scottish Health Research: Participatory Research in Cohort Studies of Mental and Physical Health" held in Edinburgh on 10th June, 2018.
The event format and venue was based upon the successful 'reunion' model developed by Professor Ian Deary and his team at the Disconnected Mind project, who regularly update their Lothian Birth Cohorts 1921 and 1936 members about their study findings.
The event was held at The Assembly Hall, Mound Place, Edinburgh. This is the meeting place of the General Assembly of the Church of Scotland and was previously home to The Scottish Parliament between 1999 and 2004. This meant that the venue had experienced technicians who could provide and install high quality presentation equipment, filming equipment and up to 600 interactive voting pads, allowing the collection of participant responses in real time.
Participants from a number of Scottish cohort studies (Aberdeen Children of the 1950s (Leon et al., 2006), Generation Scotland (Smith et al., 2013, Lothian Birth Cohort (Deary et al., 2012), and Theirworld Edinburgh Birth Cohort (TEBC, 2016)) were invited to attend an event at which they would hear key results from the studies in which they had participated. The event was also used as an opportunity to measure attitudes towards future research, including routine health record/sample linkage and its subsequent analysis.
Participants were personally invited using paper invitations which were posted out via their cohort managers or, in the case of Generation Scotland, via The Health Informatics Centre at The University of Dundee. (Data protection and GDPR laws meant it was not possible for us to obtain cohort members' names and addresses, so invitations could not be posted out directly.) The invitation is available as Extended data (Beange et al., 2019). Selection for invitation was done by the cohort managers, based upon factors such as permission to re-contact, postcode, etc. All participants of the Lothian Birth Cohort were invited; for practical reasons, a randomly selected subsection of the other cohorts were invited. They also received an interactive voting pad on a lanyard (see Figure 1). The number of voting attendees from each study, and their accompanying guests, are shown in Table 1. The majority of voting participants were cohort members, but 31% were guests (e.g. the partner, child or friend of a cohort member). NB: Some participants who attended are not included in Table 1, either because they arrived late or because they did not use the voting pads.

Event logistics
No other demographic information was collected from participants on the day, although we know that Lothian Birth Cohort members were approximately 82 years old and Aberdeen Children of the 1950s members were between 62 and 68 years old. Participants came from across all regions of Scotland.
Ethical permission was granted by The Psychology Research Ethics Committee (PREC) at the University of Edinburgh (Ref No: 327-1718/3). As no identifying data were collected from participating individuals, it was deemed that written consent to participate was not necessary. The information sheet given to participants is included in Extended data (Beange et al., 2019). Participants had the option to take part (press clicker button) or not for every question as it arose.
Written photography and filming consent was obtained from all speakers, so the talks could be filmed and uploaded to the ccacevideo YouTube channel. (The videos and slides can also be found in the Extended data, Beange et al., 2019).
For the public, photography notices were displayed prominently on the walls and on seats which had the potential to be captured by photography or video recording. A more detailed photography notice, which indicated potential uses for the photographs/video was also included in the delegate pack. To comply with data protection regulations (GDPR), these notices included contact details to allow people to withdraw their consent after the event, should they wish to do so. Alternative seating was available for those who preferred not to be captured in this way.

Event programme
The meeting began with a short introductory talk by Prof Andrew Morris, Vice Principal of Data Science at The University of Edinburgh, who outlined the importance of medical research and extended his thanks to the cohort members.
He was followed by Professor Andrew McIntosh, Professor of Biological Psychiatry at The University of Edinburgh, who introduced the concept of a health cohort study, set out how healthcare data was used in research and instructed the audience in the use of the voting system.
These introductory talks were then followed by 6 topic-specific presentations, each of approximately 20 minutes in duration (see Table 2), and which included 2-3 voting pad questions (see Table 3).

Voting procedure
Multiple choice questions were posed by each speaker, during or at the end of their talk (see Table 3. Participants were given a list of the questions in their delegate packs (see also cohort meeting slide deck, Extended data (Beange et al., 2019)). Participants had 10 seconds to respond via an interactive voting pad ( Figure 1). If they pressed more than one button, only their last result was recorded. When the vote closed, the results were immediately displayed on the screens (Figure 1b-d).
Participants could choose not to vote at any point. Videos of all talks given, as well as each of the slides used during these talks, are available as Extended data (Beange et al., 2019).

Statistical analysis
The majority of the results are reported as a percentage of total respondents. Where people answered both question 1 (cohort membership/guest) and one of the other questions, Mann-Whitney U-tests were used to compare groups (see Table 5).

Results
In total, 234 people voted at least once during the event and the number of responses to each question ranged from 176 to 220. Data are presented below in terms of frequency counts, and we examined in each case the difference in opinion between cohort participants and other event guests. Summary frequencies for the participants' responses are shown in Table 4, organised according to the topic of the presentation that immediately

Presenter (order) Presentation title Brief description
Prof J Boardman,

University of Edinburgh
Growing up following premature birth Theirworld Edinburgh Birth Cohort. Purpose: to investigate the causes and consequences of being born too soon or too small on brain development and long term outcomes on children and their families.

University of Aberdeen
Whatever happened to the Aberdeen Children of the 1950s?
The Aberdeen Children of the 1950s Purpose: to study the determinants of health and ill health in a group of individuals born in Aberdeen in the 1950s

University of Edinburgh
Generation Scotland -Next Generation Generation Scotland Purpose: to conduct a family and population based study of genetic and environmental determinants of physical and mental health.

Youth Mental Health in Families at High Risk
The Edinburgh High Risk Study and Bipolar Family Study Purpose: to follow a group of unaffected young people at high genetic risk of schizophrenia or bipolar disorder and identify the baseline predictors and trajectories of those who would later become unwell.

Ten Lothian Birth Cohort Commandments
The Lothian Birth Cohort The Lothian Birth Cohort study aims to examine non-pathological cognitive ageing and its determinants. Individuals born in 1921 and 1936 and living in the Lothians were first invited to participate in 1999. The cognitive ability and health of participants has been monitored as they have aged.

Living Longer in Scotland Combining Scottish and English Cohort Studies
For the last 4 decades the people of Scotland have experienced markedly shorter life expectancy than their English counterparts. We report on our attempts to understand the reasons for these differentials. preceded the questions being asked. Raw and summary results are available as Underlying data (Beange et al., 2019).

Summary of results
Overall, our respondents were very positive about health data research (See Figure 2).
When asked if all pregnant women should be given the opportunity to take part in research, 97% of our respondents replied 'yes'. But the response was more mixed when they were asked at what age a child participating in such a birth cohort should consent to continued participation, with a fairly even spread of results across 12, 14 and 16 years of age (25%, 29% and  There is no indication that failure to answer one question had any impact on the next, so each question has been analysed individually. There are many different reasons why people may not have answered a particular question including practical factors like arriving late or leaving early, or failing to press a button within the allocated 10 seconds. 29% respectively). A total of 12% of respondents suggested the age of consent should be as low as 10 years old.

Number of guests
Similarly, our respondents were very positive about researchers accessing data held by the NHS such as neonatal 'Guthrie Spots' (97% yes) and routinely collected health care records (95% yes).
On the issue of trust, our participants showed a high degree of trust in university health researchers (87% of participants scored them 4/5 or 5/5) and doctors (81% scored them 4/5 or 5/5). However, less trust was expressed for companies with commercial interests (only 35% scored them 4/5 or 5/5).
For research data collection purposes 78% of our respondents were willing to wear a 'Fit-bit' style activity monitor and 73% were willing to change an aspect of their lifestyle (e.g. attend a social club or change their diet). However, only 45% were willing to take a new drug as part of an intervention study. Nevertheless, 68% of our respondents were prepared to undergo a brain scan to help researchers predict later mental illness and 73% would like access to such a test more generally, if it were safe and accurate.
Post-mortem brain donation is an option within one of our older age cohorts and all members have already been approached once about this decision. Of our respondents, 45% said that cohort members should be approached again and a further 18% said 'possibly yes'. We also offered a more nuanced option of 'yes, if they didn't reply before' which 25% of our respondents selected. Of our respondents, 7% said they were 'not sure' and 4% said cohort members should not be contacted again about this option.
Encouragingly, 98% of our participants would encourage their children and grandchildren to take part in a research cohort.
Finally, we asked if participants would be willing to repeat the testing that they have already done, but on a more frequent basis (i.e. every 2 years). In total, 87% of respondents said yes, highlighting again our respondents' high level of enthusiasm for health data research.
Differences between guests and cohort participants Significant differences between cohort participants and their invited guests are described below. All test results (including non-significant results) are reported in Table 5.
Trust. A significant difference was observed between groups for trust in researchers (Mann-Whitney U = 3349, p=0.008) and doctors (Mann-Whitney U = 3110, p=0.001), such that cohort participants showed higher trust in researchers and doctors than guest participants (Figure 3). This is perhaps to be expected, as cohort members have self-selected to participate in health research studies. Responses to trust in companies was not significantly different between groups (p=0.095).

Post-mortem brain donation.
Guests were significantly less positive than cohort members about re-approaching someone to ask for their consent to donate post-mortem brain tissue (Mann-Whitney U = 2291, p=0.005; Figure 4).

Frequency of research testing.
Guests were also significantly less positive about the possibility of more frequent research visits, compared with cohort members (Mann Whitney U = 1886, p=0.015; Figure 5).

Discussion
The current results demonstrate the feasibility of interactively measuring public attitudes to research, including data linkage, through the use of live voting pads. The substantial majority of our audience were very positive about health data linkage and sharing. Most were prepared to consider new and more frequent   forms of data collection such as the use of 'Fit-bits' and brain scans and were keen to encourage others to take part in research. Responses were less uniformly positive when the question of trust in commercial companies was posed, or when they were asked to consider taking a new drug or changing an aspect of their lifestyle.
Although our results are from a relatively small pool of people, larger scale public surveys corroborate several of our results. For example, The Wellcome Trust found that 77% of the UK public would be willing to share their anonymised medical records for the purposes of medical research (Ipsos MORI, 2016).

Trust
Unsurprisingly, given the high rates of audience participation in research, our participants showed a high level of trust in researchers and doctors. This result was echoed again in a recent Generation Scotland email survey (Edwards et al., 2019). Nonetheless, our findings are in line with previous research which reported that 92% of the UK public trust doctors to tell the truth and 85% trust scientists (Ipsos MORI, 2018). Our more mixed response to trust in 'commercial companies' replicated previous reports of a lower level of trust in 'business leaders' (34% of those surveyed trusted them to tell the truth, Ipsos MORI, 2018).
In a similar 2016 survey, people trusted doctors and nurses to provide accurate and reliable information about medical research (64% of those surveyed trusted them 'completely' or 'a great deal'). University researchers came a close second (59%). By contrast, a much lower degree of trust was expressed for 'Pharma Scientists' (32%) and 'Industry Scientists' (29%) (Ipsos MORI, 2016), who were perceived to 'exaggerate information' and 'only show positives'.
Using NHS data for 'big-data' research leads to significant ethical questions around access, privacy, confidentiality, trust and rights (Adibuzzaman et al., 2018). As researchers are increasingly encouraged to develop collaborations with industry, it is important to consider what steps could be taken to maintain trust and transparency, especially when working with public or donated data. These might include public consultations, or campaigns and collaborative knowledge exchange efforts which include a wide range of stakeholders.

Policy implications -Guthrie Spot
Our data have implications for policy, by demonstrating strong support for research access to 'Guthrie Cards' (neonatal blood spots). This is in line with the results of a recent email survey by Generation Scotland (Edwards et al., 2019) and a more in-depth Citizens' Jury, which was unanimous in its conclusion that research access to Guthrie cards was in the public interest, subject to appropriate ethical considerations, governance and oversight (June 2017, Porteous et al, in preparation, see also Edwards et al., 2019). Nevertheless, there continues to be an embargo on the use of Guthrie spots for research in Scotland (and the rest of the UK), pending the conclusion of an ongoing stakeholder and public consultation.

Age of consent
A key ethical debate for our birth cohorts concerns age of competence. What is the appropriate age at which to seek informed consent from children who were enrolled in a cohort study by their parents at birth? Our respondents were fairly equally split between 12, 14 and 16 years old.
Scottish law has specific rules which govern a child's participation in clinical trials (i.e. the testing of a medicinal product). In these circumstances, consent must be given by a parent or legal representative for all children under 16 years of age (NHS Health Research Authority, n.d). However, there is no such legal provision for other types of research. Instead, guidelines are offered by The Health Research Authority, which equate research consent with treatment consent (Griffith, 2016); i.e. children under 16 have a right to consent to treatment (and by extrapolation, research) if they are deemed, by a medical practitioner, to be competent to do so (HRA, n.d.). This competence depends heavily on the child's capacity to understand the research being proposed and any risks that it entails. Furthermore, attempting to achieve consent from a child puts the onus on the researcher to present the information in an age-appropriate way that fosters true voluntary decision making (HRA, n.d.). Conversely, if a child objects to participation, this is assumed to be their legal right, with case law suggesting that parents will not be able to overrule this.
Yet, even when an age is decided, this is just the beginning of the debate; what should happen to the data collected so far if a teenager withdraws their consent? Should they be offered an opportunity to 'rejoin' the study aged 16, 18 or 21? These and other questions will continue to be discussed and debated in the coming years.
Public engagement with research Finally, in line with all major UK funding councils, we believe that public engagement should be a priority for all research studies, especially cohort studies. For medical research, altruism is a key motivator, (Jones et al., 2016; National Institute for Health Research, 2019), but our participants also expect to benefit personally (McCann et al., 2010). This event brought together researchers and cohort members with a spirit of openness and community; to learn new things and engage with one-another. However, other methods such as circulating newsletters and sending birthday cards have also been used by these cohort studies to share results and build social bonds. Madsen et al. (1999) found that amongst former clinical trial participants who now held a negative attitude towards future participation, a common reason was the lack of information about results. Thus, feeding back research results in a clear and useful way was important not only for our own cohorts, but to the future of research participation in general.

Strengths and limitations
A significant limitation is that the majority of our audience already participate in research and had chosen to attend a university-run knowledge exchange event. Therefore, our results cannot be assumed to be representative of public opinion, though the contrasts between cohort members and guests give some indication of the likely direction and scale of differences between cohort members and the general population. Another methodological limitation is that not all participants answered every question. This is particularly pertinent for those who did not answer the first question (e.g. arrived late) and therefore could not be allocated to a cohort or labelled as a guest for further analysis. That said, the voting pad data collection method successfully served the dual purpose of improving engagement during the event and providing useful data for researchers to use in the future (e.g. on ethics forms and grant applications).

Conclusions
Public engagement events that allow participants to express their opinions have value to both researchers and the general public. Using a simple voting pad system, we were able to collect data which will likely influence and facilitate our future research and public engagement efforts. We would encourage other researchers to consider how they might facilitate such two-way interactions during their own public engagement events. Our findings reveal that both research participants and their guests are broadly supportive of research access to the data and samples, albeit they are less supportive when commercial interests are involved.

Justin Starren
Feinberg School of Medicine, Northwestern University, Chicago, IL, USA The study describes a real-time survey of cohort participants', and their guests', attitudes about research. The limitations of the work are driven mainly by selection bias and missing information.
All of the participants are closely connected with longstanding cohort studies. Unfortunately, the authors did not collect how long each participant was in the cohort, only which cohort. That said, because of the nature of the cohorts, some participants have been involved for over half a century. Keeping research subjects engaged is a huge problem. Any research subject who continues for over 50 years is not typical of the population. A valuable reference point would have been the overall percentage retention of the various cohorts. It is possible that these individuals are a highly selected fraction of the starting cohort, and that most participants already dropped out. Since each individual identified their cohort and guest could be separated from cohort participants, it seems surprising that among-cohort comparisons were not done. This would have strengthened the paper.
The results were further biased by the fact that this was a gala celebrating the successes of cohort-based research. Therefore, the participants were presented with success stories, and then asked their opinions. The positive nature of the evening is likely to bias results to the positive.
The negative reaction of pharmaceutical research was striking. However, one is left to wonder whether results would be different had the gala been celebrating the many breakthroughs and lives saved by pharmaceutical companies.
Overall, the work is sound in terms of the conclusions drawn from the data. However, the selected population is so narrow and atypical that few, if any, generalizations can be made from the results to the general population. We are left with the unsurprising conclusion that people who have chosen to participate in research for decades are positively disposed toward research.
Stylistically, the paper could combine figures and tables more efficiently. For examples, tables 3 and 4 could be combined so that questions and answers are adjacent. Similarly, figure 2 could have been incorporated into the table as "sparkline" bar charts to provide graphical and numerical data side by side.

Figures 3 and 4 present comparisons between cohort members and guests differently. Figure 3 presents
Figures 3 and 4 present comparisons between cohort members and guests differently. Figure 3 presents separate graphs for each group while 4 mixes the groups and uses color coding.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Partly

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate? Yes understanding of process and for those wishing to replicate the methods. As such, I appreciated the level of detail provided about the questions asked; these were appropriate and clear. The event clearly generated high levels of audience engagement (indicated by numbers answering questions). Questions were sufficiently sensitive to detect differences between groups and questions, such as the apparent negative influence of commercial interests on audience trust. The use of Likert scales with neutral points and clear options follows good assessment practice.
I felt the authors dealt well with the balance between collecting nuanced data, placing a burden on participants (it was important that the event was interesting and beneficial to them) and research ethics. For example, it would have been fascinating to explore paired data (cohort participants and their guests) or understand more about demographic influences, but this would have required a more intrusive approach. Nevertheless, the unpaired data from this event generated some interesting findings and confirm the utility of this general approach such that future studies could delve deeper as appropriate to their audience and research questions. There is also scope for these methods to be used to elicit more detailed information or ideas from participants. This could contribute to greater control or ownership of research by patients and members of the public, a stated goal for several funders.
The authors acknowledge that participation in research probably contributed to the audience's high trust and interest in health research. I agree with the first reviewer that this bias is likely to extend to the guest group. It is still valid to compare their responses and interesting that attitudes were consistently more positive in the cohort group. The authors are careful not to attribute causality to the differences found. However, I feel that the difference in attitude towards more frequent testing (q14) could be explained by guests (who have not participated in testing) being unsure how to answer. A number selected "not sure", which they may have used in the absence of a "Not Applicable" option.

Minor comments on question set, analysis and results
The authors are explicit that corrections for multiple comparisons were not included. It would have been good to flag this in the discussion when assessing the potential significance of the differences between the cohort and guest groups.
They could also have acknowledged that the Lothian Birth cohort contributed by far the highest percentage of participants. Presumably this group has a much older age range than most cohorts and may therefore be less generalizable.
For Figs 2, 4 and 5, it would be helpful to include an n= in corner of graph for ease of reference (this info can be obtained from tables).
For the brain donation question, it would have been interesting to understand the context further. Were people being asked about a scenario several years later, or soon after the research?

Other minor comments
To contribute to the purposes of learning from the outputs of the event, it would be helpful to include in the supplementary information a copy of the feedback form used.
My assessment of accuracy is subject to the discrepancy already noted by the first reviewer.
Is the work clearly and accurately presented and does it cite the current literature?
Yes and relatives of the cohort participants, this might in fact still be less representative of a general population audience.
In all, this is a useful paper on the methodology with some insights the value of dynamic two-way communication channels in the delivery of events of this type.
Suggestions: Table 5. I feel it is worth reiterating here why the total number in the tests do not add up to the total numbers of respondents in Table 4 as it was confusing and required some backtracking to understand.
Errors: Page 12 contains what seem to be typographical errors in number reporting: Para 4 line 1. 'For research data collection purposes 78% of our respondents were willing…' From Table 4 this figure should be 88%. Note this error is reflected in the abstract.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: Knowledge Exchange. Public Engagement. Institutional Culture Change.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.