Expanding the boundaries of previously obtained informed consent in research: Views from participants in the Personalised Risk‐based Mammascreening study

Abstract Introduction Understanding participants' concerns and information needs regarding broadened consent is crucial to ensure transparency and participant autonomy. Our study qualitatively examined these issues in women participating in the Personalized RISk‐based MAmmascreening study (PRISMA). The original PRISMA informed consent was project‐specific (i.e., breast cancer research), limiting the scope of secondary research. We explored participants' needs for broadened consent to preserve informed decision‐making while maximising the potential re‐use of data. Methods Focus groups (FGs) were performed following a semistructured discussion guide. Two independent researchers analysed the data thematically using an inductive approach. Findings Twenty‐three asymptomatic women and 13 women diagnosed with breast cancer were randomly divided into six FGs. Four superordinate themes were identified: (1) Normalization, (2) Attitude towards the pharmaceutical industry, (3) Privacy and (4) Knowledge. Our participants viewed data sharing as an important conduit for advancing medical science. Perceived integrity was more often attributed to noncommercial than commercial parties, with a marked mistrust towards the pharmaceutical industry. Most requested information needs related to data protection. Participants' ideal consent process would confer a range of options; for example, they would be able to choose with whom data can be shared, whether data will be de‐identified or anonymous, the expiration date of their consent and how, if requested, general and personal study results would be disclosed. Conclusion Our participants expressed clear information needs and a strong desire to be actively engaged in future data sharing decisions. Given that many researchers collaborate with commercial parties, building public confidence in these institutions would be beneficial. Illustrative examples addressing privacy concerns and clarifying difficult terms would aid consent decision‐making. Although our participants displayed great altruism in sharing their data and accepted that broad consent would ultimately facilitate future research, broad consent did not reflect their ideal situation. Dynamic consent may be an option but warrants further feasibility research. Patient and Public Contribution Women were recruited from the general breast cancer screening population. Their perceptions and information needs, as reported in this study, will not only inform broadened consent for PRISMA but ideally guide other consent templates and decisions regarding consent processes.

would be beneficial. Illustrative examples addressing privacy concerns and clarifying difficult terms would aid consent decision-making. Although our participants displayed great altruism in sharing their data and accepted that broad consent would ultimately facilitate future research, broad consent did not reflect their ideal situation. Dynamic consent may be an option but warrants further feasibility research.
Patient and Public Contribution: Women were recruited from the general breast cancer screening population. Their perceptions and information needs, as reported in this study, will not only inform broadened consent for PRISMA but ideally guide other consent templates and decisions regarding consent processes. consent has been deemed ethically acceptableprovided that participants are adequately informed. 3 Critics contend, however, that broad consent is too vague and therefore precludes informed decisionmaking. They purport that it is not truly informed consent but rather a generic permission that prioritises research interests over participant autonomy. 4 This position is further bolstered by evidence suggesting that participants' comprehension of broad consent is suboptimal. 5 Yet, given that many participants are supportive of research initiatives, 5,6 stakeholders in the field counter that trust in the institution may compensate for this inherent vagueness. 7,8 These conflicting viewpoints have consequently ignited debate over thresholds for informed decision-making. 1,6,9 Disconcertingly, there is no globally accepted consensus on the definition of broad consent. Following guidance from the World Medical Association's 2016 'Declaration on Ethical Considerations regarding Health Databases and Biobanks', 10 several stakeholders have drafted templates for broad consent. 4,9 Furthermore, there have been massive legal efforts undertaken in this sphere. To harmonize EU data protection laws, the General Data Protection Regulation (GDPR) was drafted and implemented in all EU Member States on May 25, 2018. 11 Although this wide-encompassing legislation has had profound effects on the sharing and use of data in medical research, it did not fully resolve the issue regarding the interpretation of broad consent. 12 Thus, EU member states have adopted varied safeguards for the use of data, 13 butoftentimes in the absence of clear national guidelines, local researchers and their institutions are left to define policies and requirements. 14 Undoubtedly research participants should serve an integral role in this discussion, underscoring the need for qualitative research to guide the review and adaptation of broad consent templates and policies. Their concerns and information needs are key elements in accepting broad consent and data sharing in research. 15 Since information needs may differ across research fields, exploring these dynamics within different subgroups of participants is essential. Thus, given this backdrop, we sought feedback from women who participated in the Personalised RISk-based MAmmascreening (PRISMA) study.
The primary aim of PRISMA was to investigate the potential value of risk-tailored versus traditional breast cancer screening protocols in the Netherlands. Mammograms (∼67,000), questionnaire data (∼38,000 surveys), blood (∼10,100 samples) and/or salvia (∼600 samples) were collected from participants, resulting in a unique cohort and a large population-based biobank. The PRISMA study is a rich source of health research data that can likely serve as a biomarker pipeline for more than breast cancer alone. However, the original informed consent obtained from participants was project-specific (i.e., research into breast cancer), hindering its use within the wider research community.
Reapproaching women for broadened consent for other research would expand and encourage using these data. Yet, to ensure transparency and participant autonomy, it is first critical to understand PRISMA participants' concerns and needs regarding broadened consent. Thus, the primary aim of our study was to qualitatively explore PRISMA participants' perceptions of data sharing and their information needs for broadened consent to optimize informed decision-making and garner confidence in the PRISMA data sharing initiative.

| Original PRISMA consent protocol
Participants in the PRISMA study were asked to consent to use their data and biomaterial for research into early detection, prognosis and treatment of breast cancer. They were further asked to consent to use their data and any residual biomaterial for ancillary studies into breast cancer; this included the possibility of genetic research.
Participants could consent to (1) store their unprocessed screening mammogram and/or (2) complete an online questionnaire on breast cancer risk factors and/or (3) venous blood collection to obtain a DNA sample, blood serum and blood plasma and/or (4) collection of a saliva sample to obtain DNA. Informed consent was obtained before participation in the PRISMA study. A signed paper consent form was submitted at the screening site. The online questionnaire also began with a short consent form.

| PRISMA Biobank
Biomaterials are stored at the Radboud Biobank, a hospitalintegrated central biobanking facility located within the Radboud University Medical Center (Radboudumc), Nijmegen, the Netherlands. 16 Information regarding request procedures and material transfer agreements can be found at www.radboudbiobank.nl. In line with the facility's local policy, the collaboration between commercial partners and Radboudumc researchers is permitted for biomarker and genetic medical research and innovation. The use of biomaterial to produce commercial products is strictly prohibited.

| Participants
Women were randomly selected from the PRISMA database using a digital random number calculator, stratified across asymptomatic participants and incident breast cancer cases. This stratification aimed to explore whether breast cancer status affects women's opinions of data sharing. Inclusion criteria included: 1. Consented to be approached for follow-up studies, 2. Valid email address and, 3. In possession of a computer, laptop, telephone or tablet with a functional camera and microphone.
Personalised study invitation letters with an information leaflet were sent via email. In October 2020, invitations were sent to 200 asymptomatic women (50 per screening region) and 50 breast cancer cases (divided per region as equally as possible, with consideration given to regional follow-up time). Due to the initial low response (16/ 250 women; 6.4%), a second batch of invitations was sent to an additional 200 asymptomatic women and 50 cases in November 2020. All women received one general reminder one week after the initial email invitation. Interested women contacted our research team, after which an FG date was scheduled and an informed consent form was mailed to the potential participant. The FGs were carried out between November and December 2020.

| Procedures
Due to the Covid-19 pandemic and subsequent government regulations, the FGs were performed digitally. We used Zoom, 17 which serves as a highly suitable platform for qualitative data collection when compared with other video and conference technologies. 18 All participants were provided with a short manual via email and offered help from a researcher (L. R.) in setting up Zoom on their laptop, tablet or telephone the week preceding the FG. All FGs followed a semistructured discussion guide and were led by one moderator with extensive experience in qualitative interviewing (L. R.).
The discussion guide, available in Supporting Information: 1, provides potential prompts to ensure topics of interest were addressed in the FGs. In brief, this guide includes open-ended questions about the perceived benefits of data sharing, concerns regarding data sharing, parties with whom data can be shared, potential consequences of data sharing, information needs and procedural preferences. The notion of broadened consent was presented in its most general form to better understand participants' receptiveness to using their data by academic research institutions, pharmaceutical/medical companies and nonmedical commercial parties (such as department stores).
Whereas the discussion guide served as a reminder of topics to address for the interviewer, the aim was to have the conversation between participants flow as naturally as possible. To avoid premeditated responses, the discussion guide was not provided to the participants before the FG discussions.
FGs were recorded and transcribed verbatim. FGs were organised until data saturation was achieved, that is, until no new themes were raised. Participants also completed a short demographic questionnaire.

| FAIR data statement
The authors are highly committed to FAIR data to ensure that data are Findable, Accessible, Interoperable and Reusable. 20

| FINDINGS
Twenty-three asymptomatic women (response: 5.8%) and 13 women who had been diagnosed with breast cancer (response: 13.0%) participated in our study. We organized six FGs ranging in size from four to seven women; on average, there were six women per group.
All participants were actively engaged in the FGs and many exhibited strong opinions on certain issues. Numerous topics from the discussion guide naturally emerged during the FG conversations without a prompt. In some instances, discussions of a specific topic reached a natural closure without organically progressing into a new topic. However, even when prompts were necessary, we found that the general flow of discussion was free and spontaneous with minimal silent pauses. The digital platform did not impede the FG discussionsand the average duration of the FGs was 58 min (range: | 1311 (medical) science had not wavered and they wanted their data to provide more relevant scientific insights. In the words of the participants, 'More data mean more insights', 'More data mean more women will be helped in the future using more reliable statistics', and 'I think it's important that things are researched so solutions can be found and lives can be saved',. They also felt a keen desire to help others, for example, 'I am happy if I can help somebody' and 'But even if you don't have children, like me, I think it's important to pass information on to other girls, women, to assess whether you have cancer'. Women added that sharing personal information is ubiquitous in contemporary society, from supermarket discount cards to online cookies. Data sharing, therefore, feels habitual and relatively unremarkable, as illustrated by the following quotes, 'We spend all day on the internet and I think Google knows more about us than we care to think about' and 'I think sharing scientific data is safer than being on Whatsapp or Facebook'.

| Theme 2: Attitude towards the pharmaceutical industry
Attitude towards the pharmaceutical industry was a major theme in the FGs and it was evident that topical events, namely the COVID-19 pandemic, influenced this discussion. Participants expressed a great aversion towards sharing with pharmaceutical companies, particularly against their aim to make a profit, for example, 'I don't want my data

| Theme 3: Privacy
Concerns about privacy were raised frequently. In general, participants agreed that data sharing is acceptable as long as their data are de-identified and the key to their identifiable information is kept at the university, for example, 'If anonymity can be guaranteed, I'll allow data sharing' and 'I assume that the link to my name remains at the university medical centre'. Participants also emphasised the integrity of the research team, distinguishing between university researchers and commercial parties. As one participant stated, 'I assume that researchers deal with your data in a decent and thorough fashion.
I think that yes, researchers, yes are a different type of person to people who work for a commercial party'. They clearly voiced that their data should only be used for medical research, not commercial purposes, such as marketing. This was in part due to a lack of trust regarding privacy when sharing data with commercial and international parties, for example, 'I don't want to be bombarded with commercial offers', 'What if you release your data commercially, where does it go?! I wouldn't know', and 'I believe that data security has been established in the Netherlands, however, I wouldn't want my data to be shared with the USA because that country leaks like a sieve'.

| Theme 4: Knowledge
Choosing whether or not to be informed about future findings, particularly from genetic research, was critical for participants. Some

| Procedural preferences and information needs
Participants drew on both their professional and personal experiences when stating their procedural preferences and information needs. Although broadly posed, this question elicited quite specific responses. One participant who worked in the area of knowledge transfer, for example, mentioned the availability of a 'socially responsible licensing toolkit' to help translate academic knowledge into practice.
From a practical standpoint, our participants acknowledged the utility of an informed opt-out consent procedure to enable optimal use of their data and biomaterial for medical research. This type of consent procedure implies that participants' data and biomaterial may be used for multiple studies in a similar research domain unless they explicitly object. It is often referred to as informed opt-out since, in comparison to historical opt-out procedures, greater measures are typically undertaken to communicate the scope of the consent to the participants. However, our participants emphasised the ability to partake in selective sharing would positively influence their willingness to participate in research studies. Participants would prefer to be able to choose: 1. The type of data that can be shared (i.e., mammogram, blood/ saliva and/or questionnaire data in the case of the PRISMA study), 2. With whom the data can be shared (commercial vs. noncommercial), 3. Whether data are shared (inter)nationally, 4. The level of de-identification/anonymity, 5. Whether to receive general study outcomes and/or personal (genetic) test results, 6. How feedback is reported and 7. How data sharing is handled after death.
They also stated a preference for illustrative examples to aid their understanding. PRISMA participants would like to update their consent after a previously agreed time period, for example after ten years, when new data are added to their source data or when conditions for data sharing or third parties change. Table 1 provides an overview of information needs that PRISMA participants would like to be included in the information letter accompanying the request to expand the boundaries of the previously obtained consent.

| DISCUSSION
In our study, participants could reflect on their previous projectspecific consent when discussing under what conditions they would be willing to broaden their consent. Since this was not solely a hypothetical exercise, it was evident that their receptiveness to broadened consent was largely rooted in their underlying altruism for participating in the original PRISMA study, to improve screening practices and outcomes for women diagnosed with breast cancer. Our participants viewed sharing of their data and biomaterial as a conduit for advancing breast cancer research as well as medical research in general. Given the widespread use of digital monitoring technologies in their everyday lives, they experienced a normalization of data sharing, which they felt extended into scientific research. Yet, participants countered that data sharing should not come at the expense of privacy, as evidenced by most of their requested information needs relating to data protection. Perceived integrity was critical for our participants; integrity was more often attributed to noncommercial than commercial parties. They displayed a marked mistrust towards the pharmaceutical industry. Our participants emphasised that their ideal consent process would confer a range of options; for example, they would be able to choose with whom data can be shared, whether data will be de-identified or anonymous, the expiration date of their consent and how, if requested, general and personal study results would be reported to them. Notably, participants diagnosed with breast cancer voiced similar opinions regarding data sharing to those who had not.
In general, our participants took a broad stance on re-using their data and biomaterial. Whereas their original PRISMA consent referred to breast cancer research, our participants were not only open to the reuse of their data for other types of cancer research but also other (non-cancer) medical-related research, given that their procedural preferences and information needs were met. However, broad consent is not typically viewed through this lens. Ethicists and legal experts concur that broad consent does not imply complete unrestricted use of data and should, at minimum, be confined to research domains (for example, broadening breast cancer research to cancer research in general). The European Data Protection Board recently clarified this point in the context of the GDPR, stating that although Recital 33 GDPR may permit broad consent, the requirements stated in Article 5 GDPR ('Principles relating to processing of personal data') must still be observed. 12 In brief, this means that even broad consent should have reasonable boundaries resembling the conditions under which the original data were collected. 12 Furthermore, these boundaries should be in line with the expectations of the participants. 12 In contrast to their openness regarding research outside the scope of breast cancer, our participants exhibited reservations T A B L E 1 PRISMA participants' information needs regarding expanding the boundaries of previously obtained informed consent.

Information needs
• What conditions do a third party need to meet before my data will be shared?
• With which parties will my data be shared?
• What type of studies can be performed with my data?
• What type of genetic research can be performed with my biomaterials?
• Will my data be shared anonymously?
• Why are commercial parties interested in my data?
• Will I be informed of the general study results?
• Will I be informed of personal test results, for example, genetic results? If not, please provide a reason.
• Will my data be shared with integrity?
• Will my data only be shared with parties whose primary goal is to progress medical science?
• Will third parties be forbidden to pass my data on to other parties?
• Who has the key to my identifying information?
• Will all study results be published, including negative results?
• Will the 'Socially Responsible Licensing Toolkit' be referenced?
• How long will my data be stored?
• How long will my consent be valid?
• academic institutions since they viewed the latter as more trustworthy. Thus, our participants did not entirely reject sharing their data and biomaterial with external parties but instead took a conditional stance (e.g., medical research only, cooperation with an academic partner).
Our participants' wariness towards medical commercial partners is not an isolated observation. Mistrust of commercial parties, 21,22 particularly amongst women 23 is a widely recognised issue with potentially serious ramifications for medical research. In the United States alone, the industry supports more than half of all funded clinical trials and their sponsorship has led to numerous clinical advancements. 24 Yet, restoring the general population's confidence in commercial parties requires a multifaceted approach. Numerous suggestions have previously been proposed, 24,25 and our research supports that serious consideration should be given to them. In brief, we should not expect participants to inherently trust commercial parties, but rather we should ensure the system prioritises safeguards against data misuse to gain participants' confidence. 25 It was clear that the COVID-19 pandemic, a highly publicised showcase of industry in action, patently influenced our FG discussions regarding data sharing with pharmaceutical companies. For some, the pandemic provided a concrete example of industry profiteering, thus underscoring the need for restricted data sharing.
Whereas for others, it exemplified the pivotal role industry serves in global health, which is better achieved through data sharing. In overarching terms, the experience of the pandemic appeared to modestly improve the industry's favorability amongst our partici- pants, yet not enough to abolish their pervasive underlying scepticism.
Optimal data security and privacy were marked concerns for our participants. Similar to their views on pharmaceutical research, our participants did not outrightly reject data sharing due to privacy concerns. They understood that data sharing could drive scientific research and innovation, but they placed a strong emphasis on the integrity of research teams and institutions. Data should only be shared with reputable parties prioritising privacy and data security and ideally linked with academic research. Although our participants displayed a high level of trust in their own national infrastructure for preserving privacy, there were clearly doubts regarding how international researchers would maintain their data. Mechanisms for data sharing and maintenance vary, with some more vulnerable to security risks than others. Cloud computing and virtualisation, however, are rapidly becoming the norm in the health sector. [26][27][28] These advances have given rise to digital research environments, which permit secure, auditable data sharing and analysis within a virtual workspace, 29 eliminating the need to transfer data between local hard drives. In the Netherlands, the AnDREa project is one of the numerous working examples of such endeavours, 29 and PRISMA data will be accessible within this environment.
Nonetheless, it is questionable whether this form of data sharing is enough to assuage the fears expressed by our participants.
Whereas cloud providers may be located locally, their data centres and server farms may be located worldwide. These information and  37 Dynamic consent is envisioned to increase participant engagement and autonomy, 38,39 and arguably, the reliance on broad consent may be viewed as archaic in light of such ICT advancements. 37 Previous qualitative research exploring participants' views of a dynamic consent interface have been generally positive, 40 which supports our interpretation that participants seek this type of consent process.
Yet, despite these findings, many challenges with dynamic consent persist. Consent fatigue is a noted concern that may lead to an increased number of study withdrawals. 1 43 no strong support was found for adding a dynamic consent interface to an existing longitudinal cohort study with healthy volunteers. 38 Participants were highly satisfied with their broad consent arrangement and had a high level of trust in the research team and their ability to securely share data. 38 Interestingly, despite their high levels of trust in the PRISMA team, our participants generally wished to be more involved in data sharing decisions. This may be an artefact of cultural differences or selection bias between the source populations. Our FG participants represented a highly select population who chose to participate in breast cancer screening, the original PRISMA study and this subsequent qualitative research.
Speculatively, these women are highly engaged in their own healthrelated decision-making and feel strongly about the importance of consent. Given these aforementioned issues in implementing dynamic consent, it would be prudent to conduct further feasibility and implementation research into its sustainability and practicality.
There were two main limitations in our study. Our response rate was low; however, our final sample size is comparable with other qualitative researchand data saturation was achieved within the FGs.
Second, although participants were recruited from several screening centres across the Netherlands, our results might not represent all women eligible for breast cancer screening. Reassuringly, however, the information needs requested by our participants bared strong semblance to needs reported from other qualitative studies, thus strengthening our observations.

| CONCLUSION
Our participants expressed very clear information needs and a strong desire to be actively engaged in future data sharing decisions. Given