The moral prima facie right in context
Due to specific aspects of genomic raw data and the contexts of treatment and research, there might be reasons to question the data subject’s prima facie right to receive genomic raw data. Such reasons arise from concerns regarding the interests of the data subjects themselves and potentially conflicting rights and interests of other persons and parties involved: the researchers, physicians, the biomedical institutions and family members.
The data subject
Concerns and objections building upon the interests of the data subjects themselves claim that receiving raw data is not useful or that it is even potentially harmful for data subjects. Studies suggest that people have a great interest in accessing their raw data [1, 35, 36]. However, the largest relevant study also shows shortcomings in people’s understanding of genomic raw data: among the (non-professional) public who were keen to receive personal raw data, many said that, if they were given their genomic raw data, they would analyze the data themselves or ask their general practitioner or primary care physician for analyses. As both the lay public as well as general practitioners are unable to analyze and interpret genomic raw data, this raises the concern that people “may have perceived genome sequence data to be both simpler and more interpretable than it actually is” .
Data subjects’ potentially non-adequate understanding of the nature of genomic raw data plays a key role in concerns and objections that have been put forward to emphasize the uselessness or potential harm of raw data for the data subject [1, 2]. This way of reasoning can also be referred to as “Knowledge-deficit argument”. As to the utility, it is clear that in contrast to single findings with annotations concerning health relevance, raw data as such do not constitute information ready to be transformed into (medical) actions. Retrieving information of clear health relevance for the individual from genomic raw data requires sophisticated bioinformatic tools and highly specialized competencies in human genetics. Almost all people, even health professionals, lack such tools and competencies. Additionally, at least to date, the sequencing data produced in research or translational contexts are of limited reliability for they are not necessarily produced with clinically validated and approved methods. Beyond its questionable utility, the release of raw data might even be a source of psychological, social or economic burdens and harms for the individual data subject. Data subjects might be disappointed when they learn that the data are not what they erroneously took them to be, for instance that they are not easily interpretable genetic information ready for use. The data subject’s handling of the raw data might undermine the confidentiality and security of the data. This could enable other persons to (legally or illegally) gain sensitive personal information and result in ideal or material harms if the information is used against the data subject, for example by (future) employers, insurance companies, marketing and publicity companies or state agencies. Furthermore, genetic findings such as those that can be retrieved from raw data are often probabilistic, associated with uncertain clinical diagnosis and treatment options (actionability) and thus complicated to understand and communicate. These difficulties hold also for the communication of genetic information to relatives, which is another potential source of burdens for the data subject who, when passing genetic risk information to relatives, must also take into account the relatives’ right not to know about genetic dispositions.Footnote 5
As to the empirical basis of these concerns of lacking utility and potential harm, it is worth pointing out that there exist no representative studies on the understanding of genomic raw data and personal experiences of persons who really requested and received their raw data . References to the poor understanding of many lay persons also neglect that the understanding of people can change; indeed, we will argue that (poor) understanding of the nature of raw data should be addressed and improved by targeted information during the release process. Also, to deal with their lack of tools and competencies in interpreting their own raw data, data subjects can refer to professional bioinformaticians and geneticists from the public health and research sector, to patient driven organisations or to private companies offering analysis and interpretation of genomic data as a service. Public and commercial services for genome interpretation are not all and necessarily of poor quality, or, as Angrist  puts it: “to pretend that CLIA-certified testing and analysis are still the only game in town seems a bit naïve this late in day.”Footnote 6 Beyond medical utility (for health, clinical care or reproduction), genomic data may also serve personal and social interests of the data subject, for example as entertainment, learning, way to relate to others or as option for purely potential usage in the future [1, 36].
Assessing the concerns and objections referring to data subjects themselves on a more theoretical and normative level, we first want to emphasize that most references to potential lack of utility and to risks of burdens and harms for data subjects are not based upon the perceptions and values of the data subjects, but on what others presume to be good or bad for the data subjects. Determining what is good or bad for others without knowing their opinions, values and life plans (and taking decisions for others on this basis) is sometimes necessary, for instance with respect to very young children. Yet, from a liberal perspective, every person is considered to know best what is good or bad for her and should have the right and responsibility to decide for herself what kind of conception of a good life and plans to follow . To repudiate or restrict a person’s liberties on the basis of a determination of what is good or bad for her without relying on the person’s opinions and will is ethically problematic and challenging to justify. To restrict or deny a person’s self-determination in order to protect her from (self-induced) harm qualifies as what is usually referred to as paternalism. In some cases, paternalistic interventions may be justified. However, to be justified, paternalistic interventions must meet several conditions to a sufficient degree: the person to be treated paternalistically must be badly informed; the potential harm to be avoided through the paternalistic intervention must be relatively likely to occur and of substantial magnitude, i.e. threatening the person’s life or important resources and capacities to plainly use her liberties in the future; there must be no other reasonable alternative option to avoid the harm or protect the person without violating her self-determination and liberties. This criticism (lack of respect for persons’ individuality and liberties, and unjustified paternalism) also holds for the knowledge-deficit argument which – without information about the data subject’s personal values, plans and goals – starts with the presumption that the data subject’s understanding of the nature of genomic raw data is not adequate. Finally, implementing a general rule that denies the liberty of all in order to protect just some persons is additionally problematic and challenging to justify. Not recognising the right of data subjects to access their raw data would prevent even those subjects who could handle that data without (unconsciously) inducing risks or harms to themselves. Overall, the conditions to justify paternalistically motivated violations of data subjects’ prima facie right to raw data are not met.
One might also argue that, still, instead of releasing genomic raw data, biomedical institutions should offer the return of individual genetic information based upon a list of serious and actionable genetic conditions (such as the list elaborated by the American College of Medical Geneticists . However, as long as offering the return of individual genetic findings is understood as an exclusive alternative which justifies not releasing genomic raw data to subjects, we dismiss it for the following reasons: First, as other before mentioned objections it relies on an “objective” determination of the “real” relevance and “real” utility of genomic raw data for people (defined as individual information based upon a determined list of clinically actionable genetic variations). For many persons, release of raw data and return of results will not be equivalent alternatives, or, as Lunshof et al.  put it:” There is a crucial difference between providing access to data and returning results”. Second, we take the release of raw data as a moral right of persons that cannot be cancelled or circumvented by the bound party through “offering” some other (and remarkably different) thing – not to mention the specific challenges, responsibilities and costs in terms of genetic counseling and other resources that biomedical institutions would incur in case of a return of results policy. Third, at least in EU member states, persons do have, in terms of principle, a legal right to receive their genomic raw data.
We conclude this section by stressing that, rather than addressing the concerns regarding data subjects’ lack of understanding, limited utility and potential harms from raw data release by means of prohibitions, restrictions or forced “alternatives”, they should be addressed through offering general elementary information to help data subjects make an informed and autonomous use of their right which suits their personal interests and way of life. That is why we advocate that the release of raw data be accompanied by (non-obligatory) information offers – which must not be confused with individual genetic counseling.Footnote 7
Researchers, research institutions and research participants
We will now scrutinize concerns and objections to data subjects’ prima facie right to receive raw data which rely on (potentially) conflicting rights of other persons and parties involved, starting with researchers and research institutions. So far, we have argued that data subjects have a prima facie right to receive their raw data in a way which allows them to make an informed use of the data. Accordingly, this means that researchers conducting genomic research have the correlative duty to release the raw data to participants in a way to help them to understand the general elementary nature and implications of raw data and their request. Additional reasons in favor of research participants’ right to raw data can be retrieved from a closer look at the ideal of the researcher-participant-relation. The researcher-participant-relation should be understood as a cooperative relation between two citizens and partners who interact in the spirit of equality and reciprocal emancipation and respect.Footnote 8 By this we do not mean that there are no de facto differences, asymmetries and inequalities between the researcher and the participant, for instance with regard to competencies to understand genomic data. We rather claim that the two should interact with due respect for each one’s rights and equality and feel bound by the ideal of partnership and civil fraternity.Footnote 9
Against their duty to release raw data in an appropriate manner, researchers might claim that the pertaining costs (in terms of financial resources and work load etc.) keep them from doing their job and duty, i.e. carrying out research , and that it jeopardizes the success of their research projects. Potential costs are the main reason for Wright et al.  when they “conclude that unless the return of genomic sequence data to individual participants is specifically and appropriately resourced, potentially be research funders themselves, to do so could do more harm than good”. Addressing the costs concern, Lunshof et al.  argue that the costs for raw data release are limited. However, they do not take into consideration that the release should meet certain criteria and thus requires more resources. We understand the process of adequate handling of requests and release of raw data to roughly encompass the following work load: 1. technical part: identification of the data subject’s raw data, production of a copy of the data, quality control, annotation of the copy in terms of data format and quality, making the raw data available to the data subject; 2. Coordination: authentication of the requesting data subject , maybe de-pseudonymization of the genomic data set; record of communication with the data subject; 3. Communication and information work: passing general basic information to the data subject, personal discussion with the data subject on demand. Overall costs (including around 50€ for the storage medium for the copy) might amount to around 170€.Footnote 10 Needed resources in terms of time might be higher for the first requests but are likely to decrease with increasing routine and increasing availability of tools , models and recommendations for best practice. As NGS biomedical research projects are cost intensive, it appears possible but highly unlikely that such relatively low costs could jeopardize the success of the research project – even if a relatively high number of participants request their raw data.
In cases in which costs are unsustainable for the research project, charging participants with a fair part of the costs is an acceptable solution. Yet, when charging participants, the costs should not be prohibitive, not exclude economically disadvantaged populations  and be calculated solely on the basis of the additional costs associated with the release of the raw data, not on basis of the costs caused by the production of the raw data initially carried out for the sole sake of research. It has been argued in favor of the release of raw data that it serves the interests of researchers and the scientific community by means of fostering trust and public involvement  and through producing a “tangible return of investment in the form of participants goodwill and support for research” . However, these points appear to be rather speculations on possible consequences and might hold, if holding at all, rather for the system of research as a whole than for single researchers or research projects.
Researchers could also object that the release of “their” data violates their academic freedom. However, first, as mentioned above, the data are not theirs; second, it is not plausible that the release of raw data could - through costs or other consequences - prevent or discourage scientists from engaging in what is at the core of academic freedom: to freely choose research questions and methods, carry out research and publish results without being hindered, intimidated or discriminated against. Academic freedom does not ground a monopoly and property right on collected personal data from data subjects . Third, in research on human subjects, academic freedom needs to be balanced against the rights of those subjects. When compared to other and well-recognized elements of the right to informational self-determination within biomedical research such as the right to withdraw consent to research usage (including the right that the data be deleted), the right to receive a copy of one’s raw data represents a minor disturbance for research.
Another concern refers to unfair competition and market distortion. If a publicly sponsored biomedical institution releases raw data to data subjects free of charge (or just charging the costs of the release of the genomic raw data as described above), this might negatively affect private for-profit companies which offer the sequencing and releasing of one’s genome as a commercial service to customers. However, to date, this does not seem very realistic, one reason being that private companies such as 23andMe do not confine their service on solely sequencing the genome, but offer genetic analysis and interpretations that are explicitly not part of the release of raw data as understood in this article. However, the worry of market distortion and unfair competition might be a reason why publicly sponsored genomic research projects should renounce to recruit research participants through advertising the possibility of release of raw data free of charge. Likewise, charging research participants some of the costs associated with the release of raw data could be a way to partially respond to the concern.
Contrary to Middleton et al.  who suggest that the release of raw data might be an alternative option to reporting back individual health relevant findings, including clinically actionable results, we strongly advocate that releasing raw data shall not abrogate researchers’ ethical duty to report back incidental or additional findings of significant health relevance to the individual participant.Footnote 11Understanding it as alternative option appears to assume that every finding that a highly specialized researcher might make during a cutting edge research project (and with maybe additional information about the participant) could be equally discovered and interpreted by the participant herself through referring to available services or tools. The position seems to assume also that every participant requiring and obtaining raw data has the economic and intellectual capacities to refer to genetic testing services, which warrant high validity and quality of the data and interpretation, and actually refers to them (within time). These assumptions are hardly realistic. Respect to the right to raw data is no matter of good will or generosity, but a moral duty which cannot be offset against respect to other duties of different content.
Researchers and research institutions have a vital interest to protect themselves in advance against potential and unjustified allegations by the participant or third parties in case of negative consequences from the participants’ handling of their raw data. To prevent harms to reputation and clarify ethical and legal liabilities and responsibilities we suggest that releasing research institutions may and should require participants to sign a very short acknowledgment (“receipt”; see section III). To conclude this section, researchers’ rights and interests do not justify restrictions or derogations of participants’ right to raw data except for the requirement to sign the short acknowledgment (“receipt”).
Physicians and patients
Looking at the context of medical care we need to examine patients’ right to raw data within the physician-patient relationship and how patients’ requests might concern physicians’ role and rights. Patients (with exception of extraordinary, mainly psychiatric, cases; see the legal section) have a widely recognized ethical and, as seen in the legal section, legal right to access all data and information generated and collected during their treatment – to which we refer here as “medical data”. From an ethical viewpoint, patients disclose medical data within the patient-physician relation, which is a relation of asymmetric power, competencies and dependencies. The logic and goal of the physician-patient relation is that of medical care and does not include nor ground physicians’ property or monopoly rights regarding the patients’ medical data. Treating patients as subject and partnerFootnote 12 and respecting patients’ autonomy entails the duty to meet patients’ request of access to their data.
One might nonetheless argue that patients’ general right to access all medical data does not apply to genomic raw data. For instance, one could argue that to date genomic raw data are not stored in the electronic patient record or in the clinical information system. However, this argument is too technical and misses the sense of the patient’s right to access all personal data and information collected or produced during the course of care. One might also object that within the medical context genomic raw data should be handled in an exceptional way - as individual genetic findings should be. In Germany for instance the Genetic Diagnostics Act (§ 7) requires that genetic findings are passed to patients only by genetically trained physicians through genetic counseling. In line with this reasoning one might argue that genomic raw data must not be released to patients unless through a process of individual genetic counseling. To this we respond that, as shown above, there are essential differences between genomic raw data on the one hand and single findings of individual health relevance on the other. Single findings might be retrieved from genomic raw data, but this requires further analyses and working steps (by third persons) and does not occur necessarily. Middleton et al.  refer to the physicians’ duty to give patients the information they want or need in a way they can understand it. From this duty they conclude that physicians and clinician researchers are not allowed to issue raw data to participants of clinical studies for this would entail giving information to patients that they cannot understand (due to the quantity and complexity of the data). Yet, this way of interpreting and referring to physicians’ duty to appropriately inform patients is misleading. The duty’s overall goal is to strengthen patients’ autonomy and rights, not to serve as a justification to restrict them. The duty applies to information deemed relevant for the diagnosis and treatment and to be discussed with patients, not to all medical data collected and generated during the course of care.
The data subject’s relatives
A data subject’s genomic data also concern the data subject’s relatives. Close relatives share roughly 50% of the genome, so that genomic raw data can be used to identify relatives or to learn (with a 50% likelihood) about their genetic dispositions. A data subject’s genomic data thus concern the relatives’ right to informational self-determination. Furthermore, if through further analysis and interpretation of the raw data the data subject learns about a hereditary genetic disposition and passes this information inappropriately over to relatives, this can lead to a violation of the relatives’ right not to know about genetic risks and to social and psychological burdens and harms. However, relatives might also benefit from appropriately being given genetic information by the data subject about serious but actionable genetic dispositions. Even more than in the case of the concerns regarding the data subject herself, potential negative consequences for relatives’ rights are rather hypothetical for they do not stem from the release itself but from potential actions (or omissions) by the data subject and other persons subsequent to the release of raw data. To conclude, the implications for relatives do not justify not releasing raw data to data subjects; however, the implications bear sufficient ethical weight to require awareness building efforts by means of being addressed within the information and individual discussion offered to data subjects requiring their raw data.
The main application oriented ethical consequences from the above analysis are:
data subjects (research participants and patients) have a right to receive their genomic raw data;
the right must be respected in a substantial way that helps data subjects to make an informed use of their right and released data;
concerns relating to the data subjects themselves, researchers, physicians and relatives should be addressed through an information process and do not justify a refusal to release genomic raw data.
Hence, we propose the following recommendations for the practice of release (with most recommendations intended to legally and ethically shape the upcoming practice of handling requests for genomic raw data right from the beginning, and some recommendations urging for immediate changes of the current practice of planning and funding NGS research projects):
basic information on general and elementary features and implications of genomic raw data – to be clearly distinct from individual genetic counseling – should be offered to the data subject as follows: first, when data subjects request genomic raw data, they should be briefly informed about the data’s sensitivity for themselves and their relatives and the complicated usability and limited reliability; second, if data subjects continue to wish to access the data, they should be offered written information (addressing again these points and further aspects) as well as personal discussion with a professional.Footnote 13 Accepting the information and discussion offer is not obligatory for data subjects in order to receive their raw data;
the receiving data subjects should be required to sign a very brief document (“receipt”) in which they declare that they were given the data on their request and after being offered basic information about the general nature and implications of raw data and that they assume responsibility for the consequences that might come from their handling of the raw data for themselves and third parties;
the raw data files to be released should have a common format that allows for further processing and be accompanied by meta-information about quality and methods used in acquiring it .
To prepare for raw data requests in the future:
future NGS research projects should consider in advance the possibility of participants requiring their raw data when planning grant submissions and projects, and set up a work flow;
the right to receive genomic raw data should be mentioned in the informed consent and be clearly distinguished from plans for the return of individual findings ;
funders should provide funding for potential costs of appropriate release of raw data.