Australian public perspectives on genomic data governance: responsibility, regulation, and logistical considerations

Abstract

Genomic sequencing generates huge volumes of data, which may be collected or donated to form large genomic databases. Such information can be stored for future use, either for the data donor themselves or by researchers to help improve our understanding of the genetic basis of disease. Creating datasets of this magnitude and diversity is only possible if patients, their families, and members of the public worldwide share their data. However, there is no consensus on the best technical approach to data sharing that also minimises risks to individuals and exploration of stakeholders’ views on aspects of genomic data governance models—the ways genomic data is stored, managed, shared and used—has been minimal. To address this need, we conducted focus groups with 39 members of the Australian public exploring their views and preferences for different aspects of genomic data governance models. We found that consent and control were essential to participants, as they wanted the option to choose who had access to their data and for what purposes. Critically, participants wanted a trustworthy body to enforce regulation of data storage, sharing and usage. While participants recognised the importance of data accessibility, they also expressed a strong desire for data security. Finally, financial responsibility for data storage raised concerns for inequity as well as organisations and individuals using data in ethically contentious ways to generate profit. Our findings highlight some of the trade-offs that need to be considered in the development of genomic data governance systems.

Introduction

Genomic sequencing generates unprecedented volumes of genomic data, which has the potential to be used for a wide range of purposes [1]. Genomic data (which we use to mean any data generated from genomic sequencing) may be donated or collected following clinical testing or as part of a research study [2, 3]. Such information can then be used to improve our understanding of the genetic basis of both common and rare diseases, bettering the clinical care of patients. Furthermore, stored genomic data could be used for the future benefit of the data donor; it could be reanalysed at specific time points throughout an individual’s life to identify variations in genes that cause disease unrelated to the original reason for testing (secondary findings) [4].

To realise the huge potential of genomic medicine, the storage and sharing of large and ethnically diverse genomic datasets with researchers and clinicians is required [1, 2, 5–8]. Creating datasets of this magnitude and diversity is only possible if patients, their families, and members of the public worldwide share their data [8–10]. However, exactly what data is stored and shared (for example, the raw genomic data, a list of genetic variants, or accompanying clinical information) can vary. Furthermore, how this information is managed and used has important implications for these data donors and the healthcare system in general [11].

The collection and use of personal (including genetic) information in Australia is governed by Commonwealth, State and Territory Legislation. Relevant Commonwealth legislation includes the federal Privacy Act [12] and the accompanying Australian Privacy Principles [13]. Similar, though not identical, privacy principles are contained in relevant State and Territory legislation [14]. However, despite (or perhaps partly due to) this complex regulatory landscape providing for several aspects of genomic data storage and sharing (including basic requirements for consent and privacy), there exists no nationally accepted way to collect, store and share genomic information [14]. Models of genomic data governance vary regarding the method of data storage (e.g. cloud-based versus physical); storage/sharing location (e.g. centralised versus decentralised); type of data stored (e.g. sequence data alone versus including clinical data); storage length; responsibility for data management; and data access permissions. Which model of data governance we ought to adopt depends on the weight we assign to different values, such as protecting privacy, autonomy, or collective wellbeing, and how these link with practical considerations, such as technological capability and cost. Currently, there is no consensus on the best approach to data sharing that also minimises risks to individuals [11] and exploration of stakeholders’ views on aspects of genomic data governance models—the ways genomic data is stored, managed, shared and used—has been minimal. Understanding how members of the public weigh the different aspects of genomic data storage, sharing, and use is important to help guide policy decisions regarding data management in genomic medicine.

We aimed to address this need by exploring the Australian public’s views and preferences for different aspects of genomic data governance models and what values drive these preferences.

Subjects and methods

Methods for this study are described in detail elsewhere [15].

Advertisements were distributed via Facebook to English-speaking members of the Australian public over 18 years of age. Prior to the focus groups, participants watched a 10-minute video to provide background information about issues relating to genomic data storage and sharing [16]. Focus groups explored participants’ preferences and values regarding secondary use of genomic sequencing data obtained in the paediatric diagnostic setting. The focus group guide is included as Supplementary Material.

Focus groups were facilitated by DV and FL via Zoom, audio-recorded, transcribed, and analysed using inductive content analysis [17]. Data analysis was managed using NVivo (released March 2020) [18].

Results

Thirty-nine members of the Australian public aged 18–67 participated in one of seven focus groups (range 4–7 per group). Participant characteristics are summarised in Table 1. Six (15%) participants were medical or allied health professionals, fourteen (36%) reported experience with genetic testing for themselves or a close family member, and six (15%) reported that they or their child had a diagnosed medical condition.

Table 1.

Participant characteristics.

	n (%)
Age (years)
Mean	37
Range	18–67
State/territory
Victoria	22 (56%)
New South Wales	14 (36%)
South Australia	2 (5%)
Western Australia	1 (3%)
Metro/rural
Metro	29 (74%)
Regional/rural	10 (26%)
Children
Y	29 (74%)
N	10 (26%)
Country of birth
Australia	27 (69%)
Other	12 (31%)
Language spoken at home
English only	12 (31%)
English plus another language	27 (69%)
Gender
Female	32 (82%)
Male	7 (18%)
TOTAL	39

Here, we report two categories: (1) logistics of data storage and sharing; and (2) responsibility, control and regulation.

Quotes are used to illustrate categories and subcategories. An ellipsis (…) reflects where a significant portion of speech has been removed, and square brackets represent where a word has been replaced for clarity or to protect participant anonymity. Quotes are deidentified and labelled based on participants’ focus group number (e.g. FG1 P1 refers to focus group 1, participant 1).

Logistics of data storage and sharing

Participants described their preferences and perspectives on many aspects of the logistics of genomic data storage and sharing. Illustrative quotes are displayed in Table 2.

Table 2.

Participants’ perspectives about the logistics of genomic data storage and sharing.

	Illustrative quotes
Length of data storage	“…if genomic data is going to be stored for a lot of people, potentially on cloud servers, that would have huge environmental impacts because imagine your data is lying on a server for some decades, and electricity is going to be used to keep your data on those servers for a long time…so with current conditions it might have a high carbon footprint because we are still relying on fossil fuels to make electricity…” [FG1 P3]
	“Look, the researcher side of me says collect it all and store it all and use it for as long as you possibly can because, I mean, it’s a bit of a no-brainer, the more of that you have and the longer it’s available for, the more powerful and more useful it can be.” [FG3 P1]
Location of data storage	“…the centralisation of storage of these things brings with it a whole lot of potential benefits as well for that sort of advancement of science in terms of having it all in one place and enabling the linkage of different datasets…” [FG2 P6]
	“I just think that if they weren’t stored centrally and they were stored in each research facility, it would be, something like a breach of privacy may more likely be occurring in that sort of area because…each organisation is different, maybe the way they store the data may be different as well…” [FG5 P4]
	“…if you’re talking about something that is a physical storage device then you’ve got less of a chance of hacking or that information being accidentally shared, but when you’re talking about something that’s stored in some kind of internet space, that does open up a can of worms for things like viruses and other technological issues.” [FG2 P3]
	“…from a storage and cost perspective, it would be interesting to consider individuals being given the option to, you know, ‘we can’t afford to store your data currently but do you want to take your data and we can then put you on a list of people where if someone wants to access that, they can contact you?’ I mean, obviously, yes, it’s a big file, but if everyone’s only got their own file, then that significantly eases the burden of storing it. Makes it hard to access, but when the alternative’s potentially deleting it entirely, it would be an option.” [FG3 P4]
Payment and cost of storage	“I’m assuming from a financial point of view, [institutes are] keeping it because it’s worth something. Are they going to sell it? Is this going to be a business?…I’m thinking of it as a commodity value.” [FG5 P3]
	“…once you give your consent for [your data] to be stored, can people buy it? Are you kind of taken out of the equation of who you consent to have and share the information with? Is money part of that? Can people buy it from the company that’s storing it?” [FG4 P1]
	“I think realistically, the financial side of things, at the end of the day is probably going to fall back to a government level…if you’re talking about medical research institutes or universities as being the custodians of the data, in order to do that they would need to get funding, and realistically they’re going to be funded by the government, so the flow on effect is essentially, that’s where it’s going to come from.” [FG3 P1]
	“I think a lot of people would be open to the option of even paying a fee to have their data stored, like a once-off fee…I think if you personally want your data stored then you should pay something towards that because obviously it is an expensive process…but at the same time, a lot of people can’t afford this sort of thing so there should be some kind of concession available or according to the circumstances too.” [FG2 P3]
	“…you’d get buy-in from your big pharma type places…It might be a really practical way of getting the money to make it happen, but then that’s got to be offset with the trust and what the end goals of storing it all really are.” [FG3 P1]
Type of data stored	“…if there were clinical information attached to [the genomic data], that would be more relevant for the research that they could do. It expands the field and if people would give consent for that, I mean, there would be more progress.” [FG4 P2]
	“I think with big pharma and research, I think the information needs to be deidentified and then it’s safe, it’s not connected to someone, it’s just used as, ‘this is the information, we don’t know who it is’. When it’s for health reasons, then it should be identified.” [FG4 P4]
	“If you’re going to have the full clinical dataset where it’s got absolutely everything, then I think you’re going to need to have tighter controls about who has access and how access is approved and whatever else, whereas if it’s more limited, sure, you can’t do as much with it, but then it could, realistically, be made more available to more people.” [FG3 P1]
International data sharing	“In our situation it’s such a rare disease that it would have to be [an international] database, only because there’s like, for example, there’s only three genetic cases that we know of in Australia of anyone that’s got her disease so you’re not going to get much research done with three cases.” [FG1 P5]
	“I do think [sharing data internationally] would need to be done very carefully because not everywhere is as careful with data as our system might be…” [FG3 P4]
	“…it doesn’t necessarily matter what country the information is being shared with, it’s more who in that country is the information being shared with. And yeah, what regulations they have in place, what, how they store it, how they use it…” [FG4 P1]

Length of data storage

When asked how long they thought genomic data should be stored, some participants suggested data should be discarded after a set time period to reduce costs and environmental impacts. Others recognised that the longer data was stored, the more useful it became for research.

Location of data storage

Participants generally favoured a centralised over a decentralised model for genomic data storage, citing accessibility and linkage potential as its greatest assets. Centralised storage was also viewed as more secure than decentralised due to individual organisations having different policies. Despite this, participants raised security and technology concerns about cloud-based storage compared to physical storage.

Some participants suggested individuals may want to store their own data, rather than have an organisation store it on their behalf. They reflected that while this may ease the storage burden for organisations, it makes accessing the data for future health uses and research more difficult.

Payment and cost of storage

Participants recognised the value of genomic data as both a health and financial asset; they queried whether, once stored, genomic data could be sold to other organisations for unapproved uses.

Some participants suggested data storage should be government-funded. While some suggested individuals could take responsibility for paying for their own data storage, they also recognised the health inequities this could create. Others felt that if the main purpose of the data was for research or commercial use, then organisations using the data should pay for storage. Some participants suggested private companies would probably be willing to pay for data storage, provided they had access. Ultimately, many groups reached the consensus that those who want to use the data should either directly or indirectly (by paying for access) fund data storage.

Type of data stored

Participants felt storing both genomic and clinical data was most useful, particularly for research purposes. Participants specified that whether the information being stored and shared was deidentified or identifiable mattered; while deidentified data was seen as more secure, participants recognised the benefits of making identifiable data available for personal health and research uses. Mostly, participants agreed that there should be stricter regulations around identifiable data than deidentified data.

International data sharing

Regarding sharing with researchers outside of Australia, participants recognised this kind of sharing was often necessary for research into rare and ultra-rare diseases, as it allowed the generation of larger and more diverse datasets. However, they also expressed concerns with data crossing international borders, as other countries might not have the same legal and regulatory protections. They stressed that sharing data outside Australia should require strict regulation and oversight.

Other participants thought where the data was shared was irrelevant; rather, who the data was shared with, and for what purpose, were more important. As such, they felt regulation for international data usage should reflect local regulation: that is, the rules for how data can be used in other countries should be the same as for data use in Australia.

Responsibility, control and regulation

Comparisons with other forms of data storage

Participants noted similarities and differences between storing genomic data and other types of data, which informed their understanding and perspectives of genomic data governance. These included cord blood storage, COVID vaccination data, Facebook data, health data, egg or embryo storage, and banking details. Participants reflected that data sharing in general was becoming more widely accepted and that people were often willing to share other forms of data with little consideration.

“…that’s the direction that we’re going in, anyway, with other parts of our lives and our data being stored…I think that it’s something that people are slowly getting their heads around and getting used to, so it’s part of the way that society’s moving.” [FG6 P3]

Who should manage data storage and access

While some participants suggested that government was the most obvious entity to be responsible for managing genomic data, others were wary of a government body being in charge based on their track record.

“I think there’s been numerous occasions where the government has failed to properly protect our data and I think I’d be more wary about genomic data storage if I knew the government was in charge.” [FG2 P5]

As an alternative, some suggested not-for-profit research institutes could oversee storing the data as they were perceived as having the right motivations.

“I think that it should be held in control of something like the Murdoch Research Institute…rather than a third-party body who would be more interested in the money side of it. So I think it needs to be held by a responsible body who would literally be there for the genetics and not as a business.” [FG2 P3]

Participants identified organisational trust as an important factor influencing their decisions about data management.

“I would be less trusting of, say, a government health system versus a private organisation because I would be questioning whether the private organisation is for profit and whether that is going to influence what they do with the data.” [FG1 P4]

Who owns the data

Participants questioned who owns stored genomic data. Although most thought that the data donor (or their legal guardian) should continue to have ownership over it, some recognised that, in reality, this is not always the case.

“…who owns that data? Can I sell it? Is somebody going to financially profit off that data at some point? Does it belong to me? Who does it belong to?…I think it should belong to the person who has it, but I think the way our society works is whoever paid for it owns it, really…I think in reality they own that, but I think that the person should.” [FG1 P6]

Regulation of data storage and sharing

Participants felt it was important to have regulations and processes in place for accessing and sharing stored data. They agreed independent regulation of data storage and sharing was important to ensure data is stored securely and used ethically.

“…I would think something this massive and this, potentially, sensitive, would almost need its own separate entity established for this specific purpose…it almost would need to be something completely independent from anything that already exists.” [FG3 P1]

They stressed the importance of having well-constructed regulatory frameworks, or even legislation, in place to protect individuals’ data.

“I think if you have a framework set up, like a regulatory framework…I think it’s actually a good thing and can actually stop people doing the wrong thing, and I think if you have a framework in place…then it should go some way to actually stop its misuse…” [FG6 P6]

Having some form of accountability and repercussions for data misuse was seen as important.

“…if [your data] was misused and…if it was to cause an individual some sort of harm…do the people who store the data, keep the data or use it, what sort of liability they might accept to compensate or make amends or rectify any unintended consequences or harm that might come about as a result of misuse…” [FG1 P7]

Consent

Participants agreed that consent for data storage and sharing should be ‘opt-in’, requiring explicit consent, and that donors should be given as much choice as possible about what their data is used for as people will have different levels of comfort with data sharing.

“…whether the consent is opt-in or implied, I would err on the side of opt-in, ‘cause it seems to me to lend itself most naturally to the most informed consent rather than the sort of, the assumption that this is just part of the process of the treatment of the child or whatever…” [FG4 P1]

“I think we kind of need to take it as a case by case. Some people might be fine for their data to be accessed and then other people that are a bit more wary…” [FG7 P4]

Overall, participants felt that donors should maintain some level of control over their own data. As such, specific consent was viewed as more attractive to some individuals because they wanted to know and have a say in what their data was being used for, although there was acknowledgement of the barrier this could present for researchers.

“…I think there’s a tension there, isn’t there, between making these processes as efficient as possible and not having too many hurdles in terms of going back to patients each time or having broader consent and having opt-out processes can help to remove some of those really inefficient steps and costly steps. Some of these things can be such big barriers to research that the research doesn’t happen…” [FG2 P6]

“…I think that regaining consent every time the data’s shared is really important…for me personally, I’d like to be contacted every single time because you just would want to know where your data was going…” [FG6 P4]

This desire for control led participants to favour a dynamic consent model as this would allow donors to be able to change their preferences over time.

“I think that it would be good to have it being able to change it any time you like…just because you might have different opinions in the future…because people’s thoughts keep changing.” [FG7 P3]

They expressed that this should be easy, suggesting an online portal as one possible method of facilitating this.

“If it is one of these online systems where yeah, I think that would be great for convenience that you can log in and just say ‘don’t wish to share this’ or ‘these are the specific people I want to’ or even if you do want the data to then be destroyed.” [FG7 P5]

Yet participants recognised that broad consent would benefit researchers by facilitating greater volumes of data stored.

“…having broader consent and having opt-out processes can help to remove some of those really inefficient steps and costly steps. Some of these things can be such big barriers to research that the research doesn’t happen…” [FG2 P6]

As such, many participants suggested consent should only be re-sought for new uses of the data, or when things like legislation or policies changed. However, others suggested that, provided it was not too frequent, being recontacted for specific consent would be acceptable.

“I think for convenience, it’s probably better to have a pre-set consent for a list of trusted uses. However, if there are uses outside of that list or outside of the scope or the purpose, then I don’t mind to get a, to consent separately, but I wouldn’t think it’s convenient to consent too frequently.” [FG6 P1]

“…to have to provide consent every single time would be a hassle if it’s frequent access, but if it’s like once a year or once every couple of years, then it’s not such a concern.” [FG2 P1]

Participants were adamant that people’s decisions about data storage and sharing should be well-informed. To achieve this, they explained the information provided needs to be detailed, easily accessible, and describe both the risks and benefits associated with the decision.

“I think when you’re trying to store data, you should be telling them about the benefits and the risks and allowing them to have that information before they make that decision to allow you to access all of that information…” [FG2 P5]

In the event of an individual’s death, participants raised several suggestions, including following the directions of an advanced care directive or will, automatic data donation, data destruction, increasing data availability, decision-making responsibility passing to next of kin, or adhering to the initial consent.

“…what happens if the person whose data it is passes away? ‘Cause I think that is a really important thing that needs to be taken into consideration, whether it be included in the initial consent that the person says either ‘when I pass away, I want my data deleted’ or ‘I want responsibility passed onto next of kin’ or ‘donated to a university’.” [FG3 P4]

Lastly, participants identified additional complexities related to parents providing consent on behalf of their children, particularly because of the long-term nature of data storage. They felt that children should be provided with the opportunity to either re-consent or change their preferences when they reach 18 years of age.

“…as that child reaches an age where they can consent, they should be asked again whether they continue to use it or whether they don’t want their data used anymore, I do think that that’s an important consideration when you’re talking about children’s data, that they have a say once they’re older.” [FG3 P4]

Discussion

To our knowledge, this is the first study to provide an in-depth exploration of the Australian public’s views and preferences for storing and sharing genomic data acquired in a clinical setting. Understanding these perspectives is critical in order for the huge potential of genomic medicine to be realised [11] and to ensure that genomic data governance models are developed that align with the public’s values. While previous research has explored global perspectives on genomic data donation and sharing through the ‘Your DNA, Your Say’ project [2, 8–10, 19], our work provides an Australian-specific perspective, with a particular focus on the practical aspects of genomic data governance models.

We found that consent and control were essential to participants, as they wanted the option to decide who had access to their data and for what purposes. Critically, participants wanted a trustworthy body to enforce regulation of data storage, sharing and usage. While participants recognised the importance of data accessibility, they also expressed a strong desire for data security. Finally, financial responsibility for data storage raised concerns for inequity as well as organisations and individuals using data in ethically contentious ways to generate profit.

Comparisons with other forms of data storage

Previous research has shown that attitudes to sharing of genomic data are affected by individual health experiences [20]. Similarly, our participants drew on their personal health experiences and existing understanding of storage, sharing and use of health and other personal data to make sense of data storage and sharing in the context of genomics. The way people viewed these other types of data appeared to impact their perceptions of the risks and benefits of genomic data usage; for example, those who highlighted the widespread collection and sharing of personal data via social media were less concerned about similar uses of genomic sequencing data.

Accessibility and security are equally important

While accessibility of genomic data to trusted entities was important to participants, so too was security. There has been repeated suggestion of the need for increased connectivity of genomic databases to enable both research and clinical uses of genomic data [2, 21]. One obvious way of facilitating this is to store and access genomic data using cloud-based systems. Our participants assumed that cloud-based storage was less secure and therefore expressed a preference for physical storage. However, cloud-based systems can enable high levels of data security, for example, through the use of encryption algorithms [22]. Physical systems of data storage are also not immune to data breaches. This is clearly an area for public education to facilitate acceptance of storage systems based on evidence of their security.

Advancing technologies may also present opportunities for enhancing the open sharing of genomic data whilst maintaining privacy and security. These technologies will likely impact the perspectives of the public, particularly by addressing the gap in trust. These significant opportunities will need constant consideration as technology advances.

Along with increased accessibility, many have recognised a need for accompanying ethical and legal frameworks to enable national and international data sharing [2, 23, 24]. The Australian Privacy Act places specific responsibilities upon organisations disclosing personal information to overseas recipients. In summary, these include that organisations with responsibilities under the Act must take reasonable steps to ensure that the overseas entity does not breach the Australian Privacy Principles and the disclosing organisation is accountable for any breach of the principles by the overseas entity [13]. While some participants were concerned about sharing data outside of Australia, others felt that the specific organisation or individual with whom the data was being shared (and for what purpose) was more important. Previous research shows that patients with rare diseases are positively disposed towards sharing their genomic data for research, including internationally [20], suggesting that international data sharing may be acceptable if adequate restrictions are implemented.

Regulation of data storage and sharing: trust is essential

As per decisions regarding whether to share genomic data [5, 10, 15], organisational trust was an essential factor in participants’ opinions about who should be responsible for storing and regulating genomic data sharing. While participants were divided on exactly who should be responsible for making decisions about data access, strict regulation, processes, and even legislation were important to participants, particularly when data was seen as more identifiable. It was also important to participants that there be consequences for data misuse. Previous research has also shown that trustworthy data governance is important in genomic data sharing [19], and that medical research participants rely on regulation of such research to ensure its safety [25]. Our results support the view of others that improving public trust in genomic data sharing will be essential to facilitate its widespread adoption [11].

Who should be legally and financially responsible?

While some participants trusted governmental organisations to store genomic data, others raised concerns for governments’ capability to safely and effectively manage such large and sensitive datasets. Focus groups with current and potential genomic research participants also demonstrated widespread concern for governmental regulation of genomic research data [5]. Despite their reservations about governments managing genomic data, our participants’ concerns about anyone profiting from the data meant they also did not want private companies to be responsible for data management.

Participants also discussed who should be financially responsible for data storage. When considering a ‘donor-pays’ model, some recognised that health disparities were likely to increase if only the wealthy could store their data. Conversely, a ‘user-pays’ model was raised as an alternative solution, whereby those who wanted to access the data would fund the storage and regulation costs. While a donor-pays models may provide a temporary compromise at this stage of the technology, it does raise issues of social justice and inequity. Yet a user pays system may open the door for private entities, such as pharmaceutical and insurance companies, to generate profit from the data [15].

Consent and control are key

Control over their data was clearly important to participants in our study, and most favoured a dynamic consent model for this reason. Previous work has shown that the option to withdraw one’s genomic data is essential in fostering trust in genomic research [9]. Focus groups with patients from 16 countries with rare diseases also demonstrate a necessity for re-contact and re-consent, and that “ongoing, long-term interactions can build trusting relationships between researchers, clinicians and patients that can ultimately have a positive impact on participation” [20].

While most participants favoured a specific form of consent, they also recognised the benefits to researchers of broader types of consent. In response to this, some suggested that data donors should only be asked for additional consent for new or different uses of their data. This mirrors perspectives of patients with rare diseases who want to retain autonomy through opportunities to re-consent to new or changing uses of their stored genomic data [20].

Participants were in agreement that paediatric data donors should be able to decide what happens to their data when they reach the age of maturity. This is particularly pertinent due to the high proportion of paediatric samples likely to be shared to genomic databases. This will also be an important consideration for the implementation of genomic newborn screening programs in which genomic data for large proportions of the paediatric population will be stored.

Study limitations

We chose to use a concrete scenario based in the paediatric diagnostic setting to orient participants to the ethical issues arising from genomic data storage and sharing. While this scenario allowed us to explore subsequent uses of genomic data throughout the lifespan, we recognise that initially focussing on one context may have shaped subsequent discussions. Therefore, the transferability of these findings to other relevant settings (for example, in population screening or direct-to-consumer testing) should be further explored.

While this study included the perspectives of a diverse group of members of the Australian public, representativeness was not its aim. Some subgroups of the Australian population may have been overrepresented, including those with prior experience of genetic testing. Although we were not able to capture the views of those who did not speak English at all, our sample included a high proportion of participants who spoke a language other than English at home. While recruitment was restricted to those who were comfortable using a computer (including Zoom), thereby missing those who could not or did not have access to such technology, this model made it possible to include rural and interstate participants. Inclusion of these and other under-represented groups in the development of genomic data governance models will be essential as they are established and implemented.

Conclusion

Our findings highlight some of the trade-offs that need to be considered in the development of genomic data governance systems. Data should be stored so that it is safe, while simultaneously allowing fast and efficient access for appropriate uses. Importantly, data donors want a high level of control over what happens to their data. While some form of dynamic consent may address this desire, there needs to be recognition of the barriers this may create for sharing. Further work is needed to illuminate the relative value members of the public place on each aspect of genomic data governance.

Author contributions

All authors contributed to designing the study and revising the paper. FL and DV conducted the focus groups, analysed the data, and drafted and revised the paper.

Funding

The authors acknowledge the infrastructure funding received from the Victorian State Government through the Operational Infrastructure Support (OIS) Program. This work was supported by the Australian Government through the Medical Research Future Fund, as part of the Genomics Health Futures Mission (Grant number 76749).

Data availability

The datasets generated during the current study are available from the corresponding author on reasonable request.

Competing interests

YM is an associate at Analysis Group, Ltd. Research for this paper was undertaken when she was working at the University of Melbourne. JS is a Partner Investigator on an Australian Research Council Linkage award (LP190100841, Oct 2020‐2023) which involves industry partnership from Illumina. He does not personally receive any funds from Illumina. Julian Savulescu presented at a Genomic Prediction‐organised webinar (2021), but received no payment or other benefits from Genomic Prediction Ltd. JS is a Bioethics Committee consultant for Bayer. The remaining authors declare no competing interests.

Ethics approval

This study was reviewed and approved by The Royal Children’s Hospital Human Research Ethics Committee (HREC/79393/RCHM-2021). Participants provided voluntary, informed consent.

Supplementary information

The online version contains supplementary material available at 10.1038/s41431-023-01381-1.