Dealing with ambivalence in the practice of advanced genetic healthcare: towards an ethical choreography

The implementation of next-generation sequencing (NGS) in diagnostic practice has stimulated ongoing debates on how to construct and perform “good” genomic care. Our multi-sited qualitative fieldwork at two large European centres for human genetics (CHGs) revealed tangible ambivalence in beliefs, norms, and actions in the enactment of NGS practices across sites stemming from differing expectations, interests, demands, and tensions. First, ambivalence was present around the boundaries of clinical diagnostic genetic care. The overlap between research and clinical work and diagnostics and screening led to ambivalence around “best” practices and norms concerning whom to offer NGS testing and how far to take testing. Secondly, the clinical value of NGS results, especially VUS and unsolicited findings, was ambivalently valued, resulting in an inconsistent approach towards these types of findings. Thirdly, ambivalence was recognized in applying guidelines in the reality of clinical practice. The ambivalence we encountered was often not made explicit or acknowledged, causing a failure to benefit from its possibility to encourage reflexivity and change. We propose to facilitate a more explicit ethical choreography [27], where ethics and science are developed iteratively whilst welcoming different perspectives and disciplines. Pulling experiences and practices of ambivalence into the light can help to understand the points of tension in the values and internal logic in care practices within the CHGs and facilitate a more informed, transparent, and consciously chosen direction for genetic care.


INTRODUCTION
Next-generation sequencing (NGS) has greatly impacted our ability to read and understand the human genome.In enabling a wider, faster, and more in-depth look at a genomic sequence, the implementation of NGS has led to substantial advances in our knowledge about, and ability to find, genetic causes for disorders [1].In diagnostic clinical practice, NGS has increased the chances of finding a genetic cause for patients with a wide range of complex developmental, physical, and neurological disorders to between 40 and 60% when looking at the whole genome [2].NGS results are, however, often more complex and fluid than those coming from traditional targeted testing [3].Diagnostic NGS tests lead to a vast amount of data to analyse, increasing the chances of encountering results with less explanatory power, such as variants of uncertain significance (VUS or class 3) and unsolicited or secondary findings (i.e., respectively accidentally or actively identified variants in disease-causing genes that were not related to the original testing purpose) [4].
Important developments in the life sciences, such as NGS testing, are often met with ambivalence: fuelling optimism with its promises and instilling fear with its potential to threaten core human values [5]."The genome had ethics in ways that other biological matter did not" from the start of the related research [6], sparking discussions on if and how NGS should be integrated into (future) care practice.Next to debates in public media, this has led to an extensive body of work in the social sciences (e.g., science and technology studies, ethical, legal, and social implication studies, (medical) sociology and anthropology, bioethics).The increased application of NGS in clinical practice has led to various perspectives on when and how to offer this type of testing and how to interpret, report, and communicate the results responsibly and understandably (e.g., [5][6][7][8]).It has also led to substantial policy efforts and the development of several professional guidelines and recommendations by (inter)national societies of human genetics, meant to guide both scientific as well as ethical aspects of NGS in healthcare (e.g., [9][10][11][12][13][14]).
These reflections on "best practices", however, must find their place in different local clinical settings, with differing interests, tensions, demands, and constraints.Conflicting expectations and interests in these healthcare practices can fuel tension and ambivalence on both an individual (in patients or professionals) and a structural (in, for instance, practices or guidelines) level.Literature on ambivalence in healthcare has pointed towards those experiences of patients and involved healthcare professionals in relation to ethical dilemmas as well as to specific health technologies and diagnoses [15][16][17][18][19][20].Work on ambivalence in research contexts, including genetic research, has shown how scientific "norms" can be oppositional, situated, and fluid in nature.This ambivalence is deemed pivotal in science-making and in understanding knowledge transformations [18,[21][22][23][24].Although often viewed as negative by healthcare professionals and lay people, ambivalence has shown to open space for personal and professional reflections and to allow for change by enforcing reflection and cooperation [18,20,24,25].
In this article, we look at ambivalence in healthcare practices around diagnostic NGS from a sociological perspective.We point out the ambivalence we encountered in domains that have an important role in defining the future of clinical NGS.In pointing out these foci of ambivalence in practice, we are able to trace "the contours and axes of pressure, opposition, and potentials for movement" in these practices and highlight where there is wiggle room to (re)define the direction of genetic care [24].The central aim of this article is to examine not only the experiences of ambivalence but also how it impacts current and future care practices.Based on the analysis of our multi-sited qualitative fieldwork in two European centres for human genetics (CHGs), we present the three most prominent domains of ambivalence across both sites [26].First, we look at ambivalence around the boundaries of diagnostic clinical genetic care.Second, we examine the ambivalence in the clinical value of NGS results, focusing on VUS and unsolicited findings.Lastly, we explore the ambivalence between ethical standards and clinical reality.
By acknowledging ambivalence as a common, and perhaps even necessary, component of genetic care, the considerations and choices behind the current on-the-ground practices can be made more transparent.This allows us to include them in a broader discussion on what "good" genetic care should look like and how to move the discipline forward.Although the different existing guidelines and recommendations have either a more scientific (e.g., how to classify variants, how to write reports) or ethical (e.g., what to do with unsolicited findings, which choices to offer a patient) focus, when putting these into practice, ethical and scientific considerations and their consequences become intertwined.As we propose, the achievement of "good" genetic care should involve an open ethical choreography in which good science and meaningful ethics "go hand in hand" [27].

METHOD Study context
This article draws from extensive multi-sited qualitative research undertaken by the first author in two large academic European CHGs.The fieldwork in CHG-1, in Belgium, covered ten months starting in June 2019.The fieldwork at CHG-2, in the Netherlands, encompassed 11 months from September 2021 onwards.The first case was chosen based on pragmatic considerations regarding language proficiency, familiarity with the healthcare system, and ease of access.These considerations also influenced the second case selection, but the choice was mainly based on its complementarity on theoretical and contextual dimensions that were deemed salient after the preliminary analysis of the fieldwork at CHG-1 [28].Conducting two case studies allowed for comparison and contrast, facilitating a more in-depth and richer look at each because of this.The most relevant characteristics of the genetic care offered at the CHGs we visited are summarised in Table 1.

Data collection and analysis
Data collection involved three components: 1. Multidisciplinary team meetings (MDTs) were observed at both centres.The MDTs at CHG-2 included not only the discussion of results but also presentations and discussions of research opportunities, syndromes/mutations, and publications.At CHG-1, the meetings were usually attended by 15-25 participants, whereas at CHG-2 an average of 30-45 participants joined.At both sites, attendees consisted of laboratory specialists, clinical geneticists with different expertise, trainees, and, occasionally, other physicians or medical professionals.2. Semi-structured interviews with genetic healthcare professionals (GHCPs) were conducted.Interviews focused on their experiences with, and views on, the different steps of the NGS process: the testing offer, interpretation, reporting, and result communication.All interviewees (except for the genetic counselors) also frequented the MDTs and were included through a combination of purposive and snowball sampling.3. Consultations between patients and their family members and a clinical geneticist were followed.In these consultations the possibility of an NGS test was discussed or the test results were received and potentially discussed.Clinical geneticists from the CHGs selected potential participants and notified us prior to their appointment.Potential participants (patients and family members above 18 years old) were excluded from the study if they were deemed unable to provide informed consent by the geneticist or were not proficient in English or Dutch.
The study was approved by the local research ethics committee of the involved Belgian hospital and exempted by the Dutch hospital involved 1 .Participant consent was obtained prior to participation; MDT participants gave verbal consent and participants from interviews and consultations provided written consent.Due to the COVID-19 pandemic and the subsequent switch from face-to-face meetings to teleconferencing and teleconsultations, a portion of the observations and interviews took place online, in which case verbal consent was obtained.
The consultations and interviews were audio-recorded, transcribed verbatim, pseudonymised, and complemented with field notes made during the MDTs and consultations.The data was analysed by JK in NVivo using inductive coding based on the principles of grounded theory.A constant comparison approach ensured a close reading of the data.The two cases are used complementarily to analyse ambivalence: only differences between the centres that were deemed striking are noted.
The selection of the participants in our observations in consultations by the participating clinical geneticists may have biased our case selection towards interesting but, in their assessment, unproblematic cases.Our presence may have also led to more "correct" behaviour in MDTs and consultations.Moreover, our findings are situated in their local contexts, with unique constellations of practical circumstances, practices, interests, and values.Our fieldwork took place at two large European CHGs that are influential in their respective countries and in the shaping of NGS practices and genomic care more broadly.With a strong enmeshing of the local, national, and international levels in the field of genomic medicine (through, e.g., international conferences, representative bodies, research, and guidelines), Table 1.Main characteristics of practices conducted at the research sites.

Offered tests
Mainly large panel, clinical exome and whole exome: started whole genome.
Large panels and whole exome.

Majority of patients seen a
Broad range of complex phenotypes (incl.developmental delay, physical disabilities, and neurological disorders).
Broad range of complex phenotypes (incl.developmental delay, physical disabilities, and neurological disorders) and cardiovascular disorders.

Place of analysis
In-house laboratory, only on occasion outsourced.Mostly in-house laboratory, certain specific panels/gene testing outsourced.
a Both CHGs also offer prenatal, preconception, and cancer-related NGS testing.However, these fall outside of the scope of this study as these are much more focused in nature and therefore do not bring up the same questions and concerns around, for instance, VUS or unsolicited findings.
we feel our findings may, at least partially, reflect broader shared practices and experiences of ambivalence.However, more research in other institutions and regions is needed to understand the impact of structural contextual differences in the organization of healthcare (e.g., the involvement of genetic counsellors who can offer NGS testing, the levels of standardization of care, and the access to care).Similarly, further research on non-diagnostic NGS testing practices (e.g., preconception, prenatal, and cancer-related genetics) may reveal the impact of the differing patient populations, power dynamics, and ethical concerns in shaping the care in these subfields.

RESULTS
The results of the analysis are reported under the following categories: 1) the boundaries of diagnostic genetic care; (2) the clinical value of NGS results; and (3) ethical guidelines versus clinical reality.Representative and illustrative quotes from this analysis are provided.In Table 2 we provide an overview of the conducted fieldwork.
At both the CHGs, the variant classification was loosely based on, mainly, the criteria of the American College for Medical Genetics (ACMG) [13].Variants were placed on a scale of suspected pathogenicity from 1 to 5, with the final classification involving manual curation.At CHG-1, the variants deemed most relevant by the laboratory were subsequently discussed at the weekly multidisciplinary team meeting (MDT), where the relevance of the variant(s) for the patient's phenotype was assessed and the team aimed to decide what to report and how.At CHG-2, although to a lesser extent, difficult results were also discussed in a weekly MDT.At both CHGs, the laboratory drafted the reports, after which the results were discussed with the patient (or their legal representative) in a consultation.At CHG-2 some of these were conducted over the phone or via video consultation.
In the Dutch context, national recommendations issued by the Vereniging Klinische Genetica Nederland [Association Clinical Genetics Netherlands] and supported by all Dutch CHGs were in place that take a position on, most prominently, sharing unsolicited findings with patients (letting an internal committee decide whether to report), informed consent (and the included opt-in and out options for unsolicited findings), and what information on the different types of possible test results to provide to patients.While the Belgian Society for Human Genetics refers to the European Society for Human Genetics (ESHG) recommendations on whole genome sequencing (WGS) testing and some more illness-specific guidelines, there were no underwritten guidelines for the aforementioned topics.The ESHG recommendations leave room for the individual centres and GHCPs to develop fitting practices by recommending that "stakeholders from relevant fields in research and the clinic should set up structures for sharing experiences and establish testing guidelines at local, national and international levels."[11].Some elements were lacking guidelines across the two CHGs, such as the reporting of class 3 results (which the ESHG refers to as "debatable" [11]), the offering of testing, and how to deal with the reclassification of variants and results.As we will show, ambivalence in NGS practice was, at times, related to the difficulties of dealing with these guidelines, or the lack thereof, in practice.

The boundaries of diagnostic genetic care
In the two CHG, where the focus is on identifying genetic conditions in (a)symptomatic individuals and increased risks for asymptomatic The average length of our fieldwork is solely indicative, as a precise measurement was not always possible. c The presentations and discussions of research opportunities, syndromes/mutations, and publications at CHG-2 took up ~30 min and have been included in this calculation. d The break-up below represents the focus of the consultation.
individuals, the boundaries between research and clinical work and diagnostic testing and screening, were not always clearly defined or understood.Screening entails both going beyond the test or analysis needed to answer a diagnostic question and offering a test or analysis without a clear indication for this in those to whom it is offered.Different ethical frameworks may apply in diagnostic versus screening contexts, with a more precarious risk-benefit balance at play in screening [29].The CHGs we followed are both located within academic hospitals and thus, by default, involved in education and research alongside clinical care.The necessity to work across disciplines in NGS testing led to an intertwinement of research and care practices and of screening and diagnostic work.Many GHCPs were both involved in screening (prenatal or other) and diagnostics or worked, either currently or previously, both in research and clinical care.The close coexistence of different practices of research versus clinic care and screening versus diagnostic testing in the genetics clinic led to a cross-pollination of certain "best practices" and norms.This, at times, led to ambivalent practices and experiences of these practices.The overlapping of these different cultures of care [30], building on differing norms and practices of care, contributed to ambivalence about which patients to offer NGS testing and how far to take testing to better understand a certain variant.Whom to offer NGS testing (as opposed to more targeted testing) or when cascade testing or screening was warranted remained vague, leading to observable ambivalence in the MDTs on "how far you should go" (MDT_180122_CG-2-08).At times, the search for a diagnosis turned into testing of family members that was more resemblant of screening than testing related to a diagnostic question: That in the VUSsome say like "Ok we ask information from the doctor to see if we can do anything else with it" And others [in clinical care] have, after so many years, decided that doesn't actually yield much.Unless in very exceptional casesright.If you have a family with ten affected and ten non-affected with a disease that is strongly penetrantok.But if you come into questions of penetrance.Yesthen you have a non-affected family member with a mutationdoes that then mean that your mutation does not segregate with the disease?Or does that then mean that the disease does not always express?So, there we are often stuck.And the problem there is that we started working with an 'à la tête du client'.(INT_CG-1-07)This case-by-case "à la tête du client" approach meant the GHCPs made decisions on test offering based on an untransparent balancing act of interests and concerns, related to both themselves and the patient, which ultimately made the offering to patients ambivalent.This practice also seemed to widen the initial clinical focus of whom to offer NGS testing, where a diagnostic question and relevance would lead, without a clear reflection on this on a national, CHG, or even GHCP level.

The clinical value of NGS results
In both genetic practices, the value of two much-debated categories of results, VUS and unsolicited findings, was surrounded by ambivalence.In the case of VUS, many differing opinions were voiced across the clinics.While some GHCPs felt VUS should not receive (too) much attention, investment of effort, or tinkering in diagnostic practice, other GHCPs actively engaged in keeping VUS somewhere between a result and a non-result.Categorizing VUS as "class 3 plus" (INT_L-1-01) or "suspicious" (MDT_161121_CG-2-09) and creating tables for class 3 results below the "actual results" (at CHG-1) helped to differentiate these variants from negative results.With a lack of clear guidelines on VUS reporting, both centres had settled on a practice in which they report and communicate VUS if "it has a clear link with the phenotype" (INT_L-1-04).However, the whole point of a VUS is that, based on the current knowledge, this link cannot (yet) be confirmed.By reporting VUS in the report sent to other involved healthcare professionals, these results became reified and, therefore, potentially attained a value that they may not have previously had.As described by a clinical geneticist in training: But then if you have a class 3 like that then it's really hard to write a negative report -because you will always in head go "yes" and otherwise you could include it in your report -but so with a class 3 we say "We don't know what it means" but then you risk, of course, very much that it will start leading its own life.Those people, of course, will be seen again by the doctors that requested the testbut-And then it is up to them to frame that but stillit is already a label that you put on it and those people will maybe always be very interested in that gene.That eventually may not explain anything.So yeah.(INT_GT-1-02) The ambivalence between GHCPs on the value of VUS variants left space for ambivalence to arise in dealing with individual cases in which they were found.There was variation in how VUS were communicated to patients: You have of course class 3s, yes, that are very uncertain.And class 3s of which you think, well that is very close to a four.So, you try to convey that also in that way.(INT_CG-2-02) Similarly, the extent to which a VUS was researched to see if it could get promoted to class 4 varied.This ambivalence was present between GHCPs and within GCHPs individual practices, underlining their internal ambivalence.Choices on cases were informed not only by the "scientific" potential of a variant, based on their clinical insight, but also by the relevance for patients' current and future lives and whether it potentially contained interesting information for themselves and the field.
A similar ambivalence could be found in the stances and actions towards unsolicited findings.Where at CHG-1 "some say "Ah, we are definitely going to pass those on [to patients]" and others say "No" " (INT_CG-1-08), at CHG-2 this decision was put into the hands of their internal unsolicited findings committee.At both centres GHCPs often mentioned unsolicited findings as a reason not to conduct broad testing because of the higher chances of running into them.At CHG-2, these findings were a focus of the offered informed consent form and related discussion with the patient.However, at both sites they also often emphasized that unsolicited findings occur infrequently.Especially in contact with patients, they minimized their relevance: "We have had it happen only once in all those thousands of samples right, but eh…" (CONS_CG-1-02).Furthermore, GHCPs at CHG-2 were reluctant to extensively discuss the opt-ins for certain types of unsolicited results included in the informed consent forms, which will be discussed in the next section.

Ethical guidelines versus clinical reality
Ambivalence also became visible in the tension in applying the guidelines on ethics-laden topics in the clinical reality.The morals and ethics in or behind the guidelines were sometimes met with ambivalence.For instance, we observed a tension between the norm of patient autonomy, the desired starting point for the actions of caregivers according to the Dutch national guidelines, and a reluctance to burden patients, and to a certain extent themselves, with written informed consent forms and an abundance of choices in their everyday practices.
Of course, it's good that those options are there, because people need to also be able to make their own choices.The very paternalistic-it's good that we got rid of that.But we have thought about it a lot also together and our department has decided that this is the best way, so why would we then spend a whole lot of time on something that we don't feel is such a good idea?(INT_GT-2-02) In practice, the earlier mentioned informed consent form in place at CHG-2 was not often offered in a manner that facilitated a process of information exchange and conscious decision-making about (certain aspects of) the test: But that is also agreed upon, right?That we don't discuss that unless people really come with that themselves that you then give the option to do it differently.Uhm.So it is then written on the form in a kind of light grey.(INT_CG-2-02) In discussing the informed consent form with patients, the opt-in and out options 2 were often not discussed or touched upon solely in a brief manner.GHCPs often noted that they would hint to patients that the default options were the "standard way" and that patients would mostly "go along with that" (INT_CG-2-05).The following exemplifies how the reporting options would be discussed in consultations: And then there is also written, for instance, what you do and do not get reported.So read that through properly.And then you can also potentially talk to us more about that if you do not want to know things or have questions.(CONS_GT-2-01) The GHCPs seemed negligibly concerned about patients making an "informed" choice, and often felt, as GHCPs, they were best placed to know which options to choose.
Regarding ethical and scientific standardization, the GHCPs were not always so certain that there can be a one-size-fits-all approach.Concerning scientific standardization of variant classification, many referenced the usefulness of "all using and reporting these in the same manner" and "everybody speaking the same language".However, in their practices of dealing with the ACMG criteria also a desire for "flexible" (INT_CG-1-03) guidelines became apparent; "you shouldn't drive yourself crazy with these rules" (INT_L-2-02).In relation to both scientific and ethical concerns, which commonly overlap, there seemed to be a tension between standardization to enhance the ability to make morally and legally "correct" and reflected choices on the one hand and, on the other, wanting a level of flexibility and freedom to tinker and make decisions on a case-by-case basis that fit their expert views and practices.
Without clear ethical guidelines or underpinnings, GHCPs were left to guide themselves towards "good" genetic care in relation to aspects such as deciding whom to test, when and how to report VUS, unsolicited findings, and non-paternity outcomes, and dealing with and sharing the reclassification of variants.The choices in practice seemed to be led more by what is currently feasible (for instance, concerning infrastructure for reclassification and follow-up) and by letting care find its route in practice.

DISCUSSION AND CONCLUSION
Our observations in the everyday practices around diagnostic NGS testing in two large European CHGs revealed tangible ambivalence on how to perform and construct "good" genomic care.Our fieldwork made clear that what genetic healthcare should look like is not only heavily debated in the literature and representative bodies but is also shaped by what happens on the floor of the CHGs, even if this presents itself as ambivalent or unexplicit.In the care practices of both centres, GHCPs openly but mostly performatively questioned the existing guidelines and "best" or common practices and, in doing so, contributed to the negotiation on what genetic care should look like.Our work shows how different and fluid configurations of, at times conflicting, norms, ethical reflections, and contextual factors can lead to ambivalence on and in certain practices of care.
The experiences and practices of ambivalence are important in encouraging reflexivity; a much-needed element in coming to a consciously chosen direction for genetic care.In our study, however, ambivalence was often not made explicit or acknowledged and therefore the reflection on it was lacking.We encountered predominantly implicit forms of ambivalence, where the practices go against the "structure" but are not openly contested, causing a failure to benefit from the discussion and change that acknowledged ambivalence can bring about.Open discussions on what good genetic care should look like were lacking on significant elements such as whom, when, and how to offer NGS testing.Even if there were guidelines, applying them in practice seemed more complex.At times, guidelines conflicted with the views of the GHCPs, such as in the case of the informed consent form at the Dutch centre, which could lead to confusing or mixed messages for patients.
A more explicit "ethical choreography" [27], where ethics and science are developed iteratively with attention to different perspectives, would be beneficial to understand the tensions in and different stances towards shaping NGS care.In this approach, GHCPs should be encouraged to be open about experiences of ambivalence.Furthermore, additional voices should be invited to help assess the internal logic of the current local practices and the values, interests, and constraints these are built on.More information is, for instance, needed on the patients' side: e.g., how do they actually value and deal with different results and informed consent.Likewise, social science insights on, for instance, the societal effects of the care offering should be incorporated into the decision-making around who to offer NGS.Including additional voices will make the decision-making more informed and transparent to all and allow for a more reflective approach to the future of genomic healthcare.
Facilitating an open ethical choreography, however, does not mean diminishing ambivalence.On the contrary, it means ambivalence should be shared and taken seriously to ongoingly create the best genetic care.It means acknowledging that knowledge, as reflected in for instance guidelines, context, and practice, are mutually constitutive and that sharing local practices and issues is thus pivotal to building fitting genetic care practices.Choices on, for instance, whom to include in NGS testing, how to deal with VUS and unsolicited findings, or how to deal with standardization in practice do not need to be black and white or unanimously shared.They should be able to be ongoingly questioned, not solely by displaying passive ambivalence but also by openly and actively engaging in challenging the status quo, whilst acknowledging that the concept of "good care" is not objective nor static.

The Medical Ethics Review Committee of the Dutch
Academic Medical Center (CHG-2) exempted this study from an official approval by their committee, as they concluded the Medical Research Involving Human Subjects Act (WMO) did not apply to our study.2. The current informed consent form offers the following options to be ticked off by the requester of the test, in case he/ she has come to the agreement with the patient to deviate from the default options listed above: 1) The predisposition to a condition for which medical treatment or checks are possible is not reported ( = opt out); 2) The predisposition to a condition for which, based on current knowledge, no medical treatment or checks are possible will be reported ( = opt-in); 3) If there is a high chance of a condition for the offspring, this will not be reported ( = opt-out).

Table 2 .
Overview of conducted fieldwork a .
a Labels for fieldwork are constructed as follows: (1) an indicator of the type of fieldwork: MDT, INT or CONS; (2) a date in case of multiple observations with the same participants such as MDTs; (3) an indicator of the role of the participant: CG, L, GT, or GC; (4) the number of the CHG; and (5) the participant number.b