The social shaping of a diagnosis in Next Generation Sequencing

Abstract

Although Next Generation Sequencing (NGS) has increased our ability to test and diagnose, its results are often not clear-cut and require a complex interpretation and negotiation process by both healthcare professionals and patients involved. In this paper, we explore how diagnoses identified through NGS are socially shaped under influence of the broader social context. Using an analytical framework stemming from the sociology of health and illness and science and technology studies, with a focus on the construction of diagnosis and the role of uncertainty in medicine, we analyze the existing corpus of literature on NGS diagnostic practices. We show how in this early implementation phase of NGS, in which standardization efforts are not yet solidified, the boundary struggles around diagnoses obtained through NGS are particularly visible. We discuss the different steps in the diagnosis shaping process and what it subsequently entails to give and a receive a diagnosis using NGS.

Keywords:

• diagnosis
• next generation sequencing
• uncertainty
• genomics

Introduction

While earlier DNA sequencing techniques could only examine one gene at a time, the introduction of Next Generation Sequencing (NGS) techniques in the mid 2000s has made it possible for many genes, or even the whole exome or genome, to be sequenced at once. NGS techniques allow for a much faster and more cost-effective way of determining the sequence of the approximately 3.2 billion DNA base pairs that make up someone's complete genetic code. Next to sequencing the entire code through whole genome sequencing (WGS), NGS includes targeted panel testing, where specific regions or genes that are linked to a particular genetic condition are examined, and whole exome sequencing (WES), which looks at the 1% of the genome that codes proteins and where most disease-causing mutations occur. The vast amount of data generated by this less targeted testing approach has greatly advanced the discovery and definition of genetic illnesses. The use of NGS techniques has led to the identification of many additional mutations and associations of genes with disorders. Simultaneously, NGS has contributed to an improved understanding of the complexity of genetic disorders and the mechanisms that underly their development (Shashi et al. 2016; Chakravarti 2011). By enabling both a more encompassing and more in-depth look at someone's DNA, NGS has heightened the chance of finding a genetic basis for illnesses that were previously difficult, or even impossible, to find diagnostic answers for. Although in this study we focus on its diagnostic use, NGS techniques have also allowed for the clinical introduction of non-invasive prenatal testing (NIPT), making prenatal screening available earlier on in the pregnancy, less risky and more accurate (Reinsch 2020).

In the clinical diagnostic setting, NGS is currently mainly used to find genetic causes for rare and not well-known types of illnesses that would not be found using more traditional DNA sequencing techniques. Often, patients undergoing NGS testing have already gone through long searches for a diagnosis (Sawyer et al. 2016; Stivers and Timmermans 20171998). These so-called “diagnostic odysseys” not only take up considerable time of all involved and produce significant healthcare costs, they are also emotionally exhausting for the patients and parents. As social science studies have demonstrated, diagnosis is a “powerful social tool” (Jutel 2009, 278). Diagnoses are pivotal in conferring social legitimacy to an illness: they give permission to be ill (Nettleton 2006). Diagnoses help to understand oneself and to explain oneself to others by providing a framework to create a coherent identity and illness narrative. Diagnosis is an important part of the framework within which medicine operates and helps organize illnesses and care: a diagnosis provides or eases access to care and other associated privileges, and often informs prognosis and treatment plans. Receiving a diagnosis can offer patients a collective identity to relate to and provides the potential of a network of peer-support (Gibbon and Novas 2008; Rabinow 1996). On the other hand, a lack of clarity on what causes the health problem(s) at hand, also referred to as “diagnostic uncertainty” (Bhise et al. 2018), is shown to have a significant negative social and psychological impact on patients and their support network (Stockl 2007).

For the patients undergoing NGS testing, it is often a last resort to find a diagnostic explanation for their symptoms (Rosell et al. 2016). NGS is often presented, and therefore perceived, as potentially having all the answers (Skinner, Raspberry, and King 2016; Stivers and Timmermans 2016, 2017; Walser et al. 2017). The press coverage in mainstream media on the diagnostic odysseys patients experience reveals a discourse of hype and hope, as formulated, for example, by this parent of a patient diagnosed using NGS: “Whole-genome sequencing can change lives and maybe save some (…) It changed ours.” (Yong 2013). Currently, the chance of finding a genetic diagnosis using WES is estimated to be between 25 and 35%, or as high as 40 to 60% when the more encompassing WGS is used (Mattick et al. 2018). In addition to a substantial group of patients not receiving the highly sought end to their diagnostic uncertainty following NGS, the results coming from NGS techniques come with their own set of complexities and uncertainties.

With the increased embeddedness of NGS techniques in clinical diagnostic practice, some of these complexities quickly surfaced and have been the focus of a substantial stream of work across research domains. Apart from more technical and practical issues, the main challenges coming with NGS techniques seem to evolve around their potential individual and societal impact, and the ethical and legal issues that emerge in their implementation. A great deal of insightful work on this has been put forward in the field of ethical, legal, and social implications (ELSI) studies, encompassing a wide range of sub-fields of the social sciences and beyond. The vastness and complexity of the data generated through NGS sequencing has led to concerns in the field about the interpretability and understandability of this data, both for the patients and different healthcare professionals involved. Due to the less targeted nature of NGS testing compared to more traditional forms of DNA testing, with the test not directed towards a specific disease or disease group, there is an increased chance of encountering more complex types of results. These include unsolicited findings (inadvertently found additional findings that were not part of the original aim for testing), secondary findings (additional findings that were actively sought for), and variants of uncertain significance (variants that could be involved in the illness, but on which there is not yet enough information to classify them further). In relation to these concerns, existing research has focused on the complexities around which results to return, how to return them, and how to achieve “true” informed consent and choices (Bertier, Hétu, and Joly 2016; Jamal et al. 2017; Yu et al. 2014; Tabor et al. 2012; Vears, Sénécal, and Borry 2019).

Informing, among other things, these questions around result communication, a substantial concentration of research has been directed towards the clinical and social utility of NGS and its outcomes, i.e. what impact do the results have in terms of medical benefit and benefit for relatives and the broader society (Sawyer et al. 2016; Malek et al. 2017; Timmermans and Stivers 2018). That NGS results are not only deeply personal but also entail familial information has raised pertinent questions around privacy, data sharing, the right-to-know and not-know, and the ethics of family communication (Borry, Shabani, and Howard 2014; Dove et al. 2019; Shabani, Bezuidenhout, and Borry 2014; Wauters and Van Hoyweghen 2016). Although more limited in number, rich and in-depth work has been put forward touching upon some of the core interests of the sociology of health and illness, such as communication patterns around result disclosure in the clinic (Vears, Sénécal, and Borry 2019; Skinner et al. 2018), the impact of this new field on the medical profession (Rabeharisoa and Bourret 2009; Bourret 2005), and patient experiences (Whitmarsh et al. 2007; Malek et al. 2019; Quinlan-Jones et al. 2017).

This interdisciplinary body of work on NGS has shed light on some of the most pressing complexities and social implications. So far, however, the diagnoses coming from these techniques have received little scrutiny. In the discussions of the social implications of NGS results, little attention has been given to how these NGS results are defined and interpreted and how this is impacted by the broader social context. In this paper, we utilize our sociological understanding of the shaping of a diagnosis to have a closer look at the diagnoses coming out of NGS. We aim to identify how, at which moments, and by whom these diagnoses are delineated and defined. Subsequently, we examine the implications of the current demarcation on the experience of receiving and providing diagnoses that are identified through NGS. By focusing on the social aspects of NGS diagnoses with a different perspective, stemming from the sociology of health and illness and science and technology studies (STS), we seek to extend the discussions on the social implications of NGS. Simultaneously, we aim to advance the understanding of the construction and meaning of diagnosis through our exploration of the use of this relatively recent diagnostic tool.

In the following pages, we present the results of an analysis of the existing literature on NGS diagnostic practice, showing what these studies can teach us on the social shaping of diagnoses obtained using NGS. We will first take a closer look at literature from the sociology of health and illness and science and technology studies on the social construction of diagnosis and medical uncertainty, before we apply these theoretical arguments and conceptual repertoires to the specific literature on the diagnostic practices in NGS. We will consecutively examine (1) classification and standardization in NGS, (2) social shaping of result interpretation, (3) uncertainty in diagnosis (re)negotiation, (4) impact on giving diagnoses, and (5) impact on receiving diagnoses. In this paper, we demonstrate the flexibility of the boundaries surrounding diagnoses identified through NGS, and how the recent implementation and the nature of NGS results renders the boundary struggles particularly visible. By focusing on the role of uncertainty in the social shaping of NGS diagnoses, we put forward a nuanced understanding of the complexities and meaning of diagnoses obtained through NGS techniques.

Sociological understandings of diagnosis and medical uncertainty

To study the social shaping of diagnosis in NGS practices, we build on insights from the sociology of health and illness, more specifically the sociology of diagnosis, on the social construction of diagnosis. The interest in looking at diagnosis from a sociological perspective gained momentum after Annemarie Jutel (2009) echoed earlier calls made by Mildred Blaxter (1978) and Phil Brown (Brown 1990, 1995) to recognize its importance in understanding the practice of medicine and our experiences of health and illness (Jutel 2015). Studies in the sociology of diagnosis look at how diagnostic labels are established, how they operate, and what their consequences are but also ask why they are of such importance to begin with. Looking at diagnosis as the topic of study and treating it, despite its material reality, as a product of a social endeavor can provide valuable insight into the “broader social and structural constraints of the society” (Jutel 2015, 844).

Following this perspective, diagnoses are a way to divide up nature into categories that fit our understanding of the world (Jutel 2011; Aronowitz 2001; Bowker and Star 1999). They are classification tools that are used in both medicine and everyday life but that do not ontologically exist: diagnoses are concepts that bind together “the biological, the technological, the social, the political, and the lived” (Jutel 2009, 294). Biological and clinical factors set boundaries for which symptoms would fit a disease concept, but the social influences are key in determining which group of symptoms actually becomes a disease (Aronowitz 2001). For a diagnosis to come to be, the “disease” (i.e. the cluster of signs and symptoms) has to be not only undesirable, we must be technically able to differentiate it from the “normal,” and there must a collective will to turn it into a disease (Jutel 2011). Accordingly, a particular genetic mutation only becomes relevant in clinical practice and patients’ lives if it is combined with social mobilization of both (patient-)activists and experts with the right resources (Navon 2019).

A diagnosis is a simplification of reality that highlights some aspects of an illness and minimizes others (Aronowitz 2001; Bowker and Star 1999; Goldstein Jutel 2019). Choices are made on where the boundaries of the illness lay, and lists of possible symptoms and criteria that need to be met are created. To come to diagnostic definitions and classification criteria takes contestation and compromise, and this struggle takes place on different levels: between patients and healthcare professionals, between different healthcare disciplines, between policy makers and healthcare professionals, and so on. Tensions on how to define and demarcate are highly visible in contested illnesses, such as chronic fatigue syndrome and irritable bowel syndrome, and illnesses that are considered to be increasingly “medicalized”, including depression and hyperactivity, but can be recognized more broadly (Bowker and Star 1999; Danforth and Navarro 2001; Bröer and Besseling 2017; Conrad and Barker 2010; Nettleton 2006).

The categorizations of illnesses are not fixed: scientific but also moral, economic and aesthetic developments can cause labels, definitions and diagnostic criteria to change. The ongoing delineation work is, for instance, clearly recognizable in the field of psychiatry, with the illnesses taken up in the widely used Diagnostic Statistical Manual of Mental Disorders (DSM) varying greatly over time. Famous examples include the abolishment of homosexuality as a diagnostic category in 1974, and the more recent creation of the category of autism spectrum disorder (Jutel 2009; American Psychiatric Association 2013). Where the boundaries of mental illnesses might be easily perceived as rather subjective, the same type of “subjective” boundary setting takes place in illnesses that might seem more clear-cut at first. New illnesses rise, others are merged together or dissolve: “lumping” (i.e. the creation of broad categories that highlight similarities) and “splitting” (i.e. where categories are divided to be more specific according to their differences) of illnesses is a common practice in medicine (Joyce and Jeske 2020).

Where this delineation work might seem precarious, the culture of medicine, and its central task of diagnosing, characterizes itself by a quest for certainty. The often gray-scale narratives of patients are preferably transformed into a clear and informative black-and-white diagnosis (Simpkin and Schwartzstein 2016). Uncertainty, however, is also a key component of medicine (Hunter 2016; Fox 2000) and has become recognized as an important factor in patients’ illness experiences, clinical communication, and medical practice. While clear conceptualizations or delineations of medical uncertainty are rare, Han et al. (2017) have attempted to develop a taxonomy of the uncertainties around the use of NGS in clinical settings (Babrow et al. 1998; Bhise et al. 2018; Newson et al. 2016). They define uncertainty as “a conscious awareness of ignorance – a self-awareness of incomplete knowledge of some aspect of the world” that is more frequently experienced when the available knowledge is limited or involves the main sources of uncertainty: probability, ambiguity, and complexity (Han et al. 2017, 2). In their conceptualization, the concept of uncertainty is divided up into scientific uncertainty (on diagnosis, prognosis, causality, and treatments options), practical uncertainty (on the structure and process of receiving care), and personal uncertainty (on psychosocial and existential concerns). An even broader range of uncertainties, varying from epistemological to vocational uncertainty, can be found in the literature in STS and the sociology of health and illness (Fox 2000).

The recognition of the importance of the concept of “uncertainty” in healthcare began with Talcott Parsons (1951), who pointed out the many “unknowns” and limitations that healthcare professionals continuously encountered in medical practice, despite the rapid advancement of the field. Renée C. Fox further advanced the concept, by showing the crucial role that uncertainty plays in the socialization of medical students in her essay on “training for uncertainty” (1957). According to Fox, a critical part of becoming a doctor is learning to deal with uncertainties rising from the limits of what is known, the limit of what you as a doctor can know, and the difficulty in distinguishing between these two. By the 1980s, Fox has already pointed towards the impact of the development of “scientific medicine”, referring to the scientific and technological advancement of medicine from the 1940s until then, that “uncovered and created uncertainties and risks that were not previously known or experienced” (Fox 1980, 19). More recent work of her peers has reinforced this observation: the increasing use of complex technologies seems to lead to an increased need to deal with uncertainty in medical practice. Like in NGS, technological possibilities sometimes outdo interpretative knowledge, and medical results can ask for a complex negotiation process on what they mean (Fox 2000; Kerr et al. 2019; Horton and Lucassen 2019). Uncertainty, in all its facets, is an unavoidable part of the current medical practice (Pilnick and Zayts 2014; Newson et al. 2016).

The framing of uncertainty can easily be perceived as rather negative: as something that one needs to learn to deal with, reduce, avoid, or minimize (Newson et al. 2016; Timmermans, Tietbohl, and Skaperdas 2017). Diagnostic uncertainty can, for instance, lead to significant psychosocial discomfort for patients. Likewise, uncertainty has been found to be difficult to deal with for healthcare professionals (Kerr et al. 2019; Vears, Sénécal, and Borry 2019). Nevertheless, uncertainty can also be functional and even productive. For patients, uncertainties can also be experienced as positive, leaving room for hope and personal interpretations of the illness and the patient (Newson et al. 2016; Whitmarsh et al. 2007). Classic works by Glaser and Strauss (1965) and Davis (1960) show how uncertainty was used functionally by doctors to keep a certain distance to the patient and their experiences with illness and death. More recent work provides further insight into how everyday uncertainties in medicine and science can be productive in two opposing ways. On the one hand, the “uncertainty work” that is undertaken in practice can make research “doable in the face of uncertainty” (Hollin 2017, 227). By “disarming, displacing and deflating” (Hollin 2017, 211) uncertainty, expectations can be maintained and ongoing investment in the field ensured (Kerr et al. 2019). On the other hand, emphasizing uncertainty has proven to be productive by displaying the need for attention, discussion, and further funding to come to a solution (McGoey 2009; Moreira, May, and Bond 2009). In the pages which follow, we seek to use these insights on the social construction of a diagnosis and medical uncertainty as a lens to look at the existing literature on diagnostic NGS practices.

Classification and standardization in NGS

Although next generation sequencing has increased our ability to test and measure, the type of data that is generated by these techniques and the interpretation of this data leaves considerable room for uncertainty (Bertier, Hétu, and Joly 2016; Horton and Lucassen 2019). Dealing with NGS test results requires making sometimes demanding decisions at the different analytical and interpretation stages it goes through, by the different involved professions. These include decisions about which clinical indicators are to be used to direct the testing, which list of genes is then chosen to be the focus of the analysis, and how the variants are subsequently classified and reported. Variants found through NGS can be classified to be pathogenic or not pathogenic, but also as likely pathogenic, variant of uncertain significance (VUS), and likely not pathogenic. When there is limited information on a variant, or a high variability in how the variant or disease expresses phenotypically, the delineation of a disease (or variant) can be experienced as more ambiguous. In these cases, taxonomic uncertainty (Star 1985) can emerge, raising questions on what constitutes a certain disease and what delineates it from other diseases.

As already pointed out by Susan Leigh Star in the mid-eighties, taxonomic uncertainty in medicine is often countered by standardization efforts (Star 1985; Timmermans and Angell 2001). This standardization often aims to clarify the procedures, materials, and protocols in place to ensure “clear” data that fits the classification. On a more macro level, these type of standardization efforts can, for instance, be recognized in the increasing standardization and codification of the clinical gaze over the years, in which the diagnosis and its diagnostic code (a translation of the description of a disease or injury into codes that fit a classification system) play a crucial role (Green, Carusi, and Hoeyer 2019; Atkinson 1981). Standardization efforts related to NGS have mainly taken the shape of guidelines, set up by professional bodies such as the American College of Medical Genetics and Genomics, the Association for Molecular Pathology, and the European Society for Human Genetics.

The guidelines for NGS classification and reporting aim to provide a framework to lighten the interpretative dilemmas, but they still leave room for considerable uncertainty and subjectivity in interpreting and applying them in everyday clinical practice. The different guidelines issued by these professional bodies, for instance, lack uniformity on how to interpret and report VUS and unsolicited findings (Vears, Sénécal, and Borry 2017). Different usages of the guidelines, different interpretations of the boundaries of the variant classes, and different availability of the lines of evidence to assess the pathogenicity can lead to different classifications across sites and ultimately result in differences in the medical care that patients receive (Hanquet, Vinck, and Thiry 2018; Horton and Lucassen 2019; Kim, Ki, and Jang 2019; Bland et al. 2018). As standardization studies have shown, guidelines can point healthcare professionals to what is considered the “right” way, but standards are often taken up pragmatically and only if they fit in the existing practices (Bowker and Star 1999; Timmermans and Berg 2010). And although standards can provide shortcuts for classification, they are not by definition beneficial: in the clinical context they might even run counter to the practice of medicine, in which clinicians traditionally have great autonomy in describing and labeling phenotypes (Mol 2002).

In cases of taxonomic uncertainty, applying the classification can lead to diagnostic uncertainty on the healthcare professionals’ side: a lack of clarity on the classification itself can complicate the decision making on who fits this classification and can therefore be rightfully diagnosed. The difficulties associated with encountering this type of diagnostic uncertainty can contribute to a reassessment of the classification (Star 1985; Navon 2019; Latimer et al. 2006). As mentioned already, classification criteria can change over time and diagnostic categories can become more or less inclusive. Genetic technologies are found to contribute to both: leading to the refinement of existing illnesses and the development of sub-categories. They are even found to be at the base of the creation of entirely new disease categories: so-called “genetically designated conditions” (Navon 2019).

The limited research on the use of standards in the classification of NGS results in practice shows that the boundaries of what is considered enough “evidence” for the different classes in variant interpretation are on the move and are ongoingly reshaped in everyday clinical practice. Laboratory personnel has been found to use their professional autonomy to make decisions on variant interpretation and reporting: utilizing the freedom of interpretation that is left by the current guidelines. Local practices are developed that best fit the local experience of how to deal with uncertainties of interpretation (Vears, Sénécal, and Borry 2017; Kerr et al. 2019). Healthcare professionals have been found to use standards on NGS in a reflexive way: ongoingly assessing the fit of the standards at hand, and if deemed necessary working around them (Timmermans 2015). Local solutions to deal with the complexities of variant interpretation and reporting are generated in everyday practice, such as the creation of subcategories to the VUS category (Vears, Sénécal, and Borry 2017).

Social shaping of NGS result interpretation

Uncertainties not only arise in the shaping of classifications and standards but also come into play when these are applied in clinical practice. They become especially clear with the majority of the cases where there is no (complete) causal explanation found after NGS testing (Belkadi et al. 2015; Mattick et al. 2018). In these cases, there may be likely pathogenic variants, pathogenic variants that only partly explain the symptoms, or variants of uncertain significance. There could be several potentially causative variants at play, several variants found of which none fit the phenotype of the patient, and even unexpected unsolicited findings. Having to deal with these types of results can create additional uncertainty for everyone involved (Bertier et al. 2017; Howard and Iwarsson 2018; Vears, Sénécal, and Borry 2017). The involved healthcare professionals have to make decisions on the certainty of the results, what is relevant enough to report, and how to report these results in a way that is both comprehensible for the patient and close to the truth (as then known). Having to deal with the complexities and uncertainties associated with genomic results has been found to increase the workload of healthcare professionals, requiring more management of expectations and greater navigation of the level and amount of information to provide (Kerr et al. 2019; Vears, Sénécal, and Borry 2019). A study by Timmermans, Tietbohl, and Skaperdas (2017) on decision-making in genetic board meetings shows the interpretative and narrative work that comes with making sense of VUS in practice.

The complexity in interpretation is also visible in the interactions between patients and healthcare professionals around the results. In discussing the results, doctors and patients together try to situate the results and fill possible gaps of knowledge using personal experiences, phenotype information, and medical and personal history. Discussing and questioning the results is often actively encouraged by the healthcare professionals, taking into account the personal theories and additional concerns of the patients (Stivers and Timmermans 2016; Skinner, Raspberry, and King 2016). Results can even become up- or downgraded: over the course of the discussion, healthcare professionals have been found to present the qualification of the identified variants either as more or less certain than originally reported in the laboratory report. In discussing parents’ questions and assertions, the doctor and parents collectively define the boundaries of the diagnosis and the level of uncertainty attached to it (Stivers and Timmermans 2016). Similar types of negotiation and modifications of risk were previously found in studies focusing on pre-NGS genetic testing. A study by Brookes-Howell (2006) shows how genetic counsellors were found to negotiate the boundaries of non-diagnosis to an acceptable level for the parents involved by attaching subjective levels of certainty to uncertain results and providing so-called “everyday life” diagnoses, which are not to be seen as a “proper” clinical explanation but as a way to explain the illness to others. Geneticists have also been found to modulate genetic results according to the anxiety levels of the patient in cases of increased risk of cancer (Wood, Prior, and Gray 2003). The boundaries of what constitutes a diagnosis seems to be, at least in some cases, malleable to the circumstances.

Uncertainty in NGS (re)negotiation

The flexibility in interpretation of NGS results that is present during the interpretation and communication also remains after the patient has left the clinic. Whether results do or do not provide more clarity about the condition at hand, patients and doctors alike are often still confronted with diagnostic and prognostic uncertainty in NGS (Howard and Iwarsson 2018). Additionally, questions might remain about whether the current classification will change over time. With ongoing developments in the field, new variants are continuously discovered, and existing variants are frequently up- or downgraded due to new information about their (non-) pathogenicity (Jamal et al. 2017; Navon 2019). To date, there have been numerous false attributions of disease to variants, with many patients being diagnosed incorrectly with a genetic condition as a result (Biesecker and Green 2014; Horton and Lucassen 2019).

Reclassification of genetic variants can have strong medical and personal consequences. A study by Turner et al. (2019) demonstrates that in patients seen for hereditary cancer evaluation, 7.3% of the patients’ variants were reclassified between 2012 and 2017. Of these reclassified cases, 11.3% had a high potential for clinical impact, with 94% of these cases leading to changes in clinical management. In many cases where a variant was downgraded, there had been years of unnecessary surveillance or, for some, even surgical interventions. A study on changes in classification of the well-known and well-characterized BRCA1/2 variants in the Advanced Molecular Diagnostics Laboratory database in Toronto, showed a reclassification of 12.4% of the variants between 2012 and 2017 (Mighton et al. 2019). Questions on responsibilities and organization of re-analysis and reinterpretation have subsequently become a prominent concern in the field (Appelbaum et al. 2020; Chisholm et al. 2018; Knoppers, Thorogood, and Zawati 2019).

Changing diagnostic criteria and reclassification of diagnoses are encountered in all facets of medicine, as with the previously mentioned ongoing processes of definition through lumping and splitting. In NGS, however, the boundary struggles are more at the forefront due to the relatively recent implementation of these techniques in clinical diagnostic practice. Categorizations, over time, have the ability to become increasingly less visible as categorizations (that are based on choices, made by people) and appear to us more as a natural order of things. Categorizations can become very resilient to change, just because they have become so “real”: they are no longer questioned or put to the test (Porter 1995). The categorization surrounding NGS techniques has not yet been reified in such a way that they appear to us as “natural” or “real” and therefore cannot hide in plain sight (Latimer et al. 2006; Joyce and Jeske 2020). The visibility of the boundary struggles in NGS is proliferated by the current situation in which the technological possibilities to measure outdo the human interpretation of these results (Horton and Lucassen 2019). In the following section, we will look at what it means for patients and healthcare professionals to receive and provide the different types of diagnostic outcomes coming from NGS.

Impact on receiving an NGS diagnosis

In the general public, a genetic explanation is often assigned greater value and found easier to accept than other types of diagnostic explanations, since it is seen as more fundamental, unambiguous, and natural (Dar-Nimrod and Heine 2011; Navon 2019). Likewise, parents have been shown to give genetic information a privileged position, mainly due to the perceived unambiguity (Whitmarsh et al. 2007). However, genetic results often entail risk information or information on possible future developments with a particular level of uncertainty attached to it. Additionally, as we have seen, a substantial portion (between 40 and 75%, depending on the used technique) of NGS tests do not lead to a diagnosis and a considerable part of the provided diagnoses have an unclear significance (Mattick et al. 2018; Biesecker and Green 2014). In these cases, the test results might not reduce the amount of uncertainty but merely shift the location of the uncertainty or even amplify it (Kerr et al. 2019; Robinson et al. 2019). Where a diagnosis is often expected to bring about information on the causality and prognosis of the illness, this is often not the case for NGS results. Even if a diagnosis is identified through NGS, the functionality of these results is often limited due to the many unknowns surrounding these diagnoses (Rosell et al. 2016; Walser et al. 2017).

Although this “lack of actionability” of genetic diagnoses is not specific to NGS, the current application might contribute to its prevalence (Vears et al. 2019). Currently, NGS sequencing in clinical practice is mainly used to detect rare and not well-known types of illnesses since the more well-researched and well-characterized illnesses would likely have been discovered earlier on in more targeted testing (Biesecker and Green 2014). The understanding of the different expressions and patterns of these rare illnesses is often still limited. The newly formulated diagnoses are frequently based on information of a small group of patients, which makes it more difficult to provide fitting information to the individual patient and their family. The patients found to have these types of genetic conditions may also not be very far down the illness pathway, which means information on the long-term course of these conditions is sparse, making it challenging to inform future perspectives (Horton and Lucassen 2019). Hence, in the majority of the cases where a diagnosis is found, this does not lead to any change in clinical management, treatment or prognosis (Biesecker and Green 2014; Weiner et al. 2017). The impact seems larger in neonatal intensive care settings, with a recent study in Texas showing a change in medical management in 52% of the 102 diagnosed infants. A substantial impact on treatment has also been shown in cancer genetics, where NGS testing can guide treatment choices by identifying variants in tumors that might respond to specific therapies (Horton and Lucassen 2019).

For patients, NGS testing is often perceived as the end to their diagnostic odyssey, regardless of whether a cause is identified or not. After this, patients can be sure to have done everything they can (Basel and McCarrier 2017). Ethnographic work by Skinner, Raspberry, and King (2016) on the interpretation of negative test results has shown that after this final test the active search can cease. Both those who deem a genetic cause unlikely afterwards and those who see it as a “not yet” diagnosis experience a relief of the responsibility to search. This responsibility is now transferred to the hands of the clinicians. The so-called “not yet” diagnoses keep patients engaged in the genetic imaginary: fostering hope for a diagnosis in the future and ensuring an ongoing care relationship (Jamal et al. 2017; Skinner, Raspberry, and King 2016). Research on how patients experience and give meaning to the different types of results and diagnoses coming from NGS testing is still limited, especially considering the less explanatory and stable forms of diagnoses.

For those who receive a genetic diagnosis, the opportunity arises to relate to a collective identity and to feel part of a group. While this may be available on different scales depending on the visibility and familiarity of the diagnosis, there are many examples of patient organizations for genetic diseases. Although active participation might not be for everyone, these patient organizations can provide online and local support and information (Clarke et al. 2009). Having a label can ease the communication about the illness with others, and make the illness seem more “real”. And although in many cases where an NGS diagnosis is received the access to care has already been arranged, having a clear diagnosis can increase the ease of access and provide opportunities for different services and coverage of costs (Stivers and Timmermans 2017).

The search for a genetic diagnosis often also leads to relief about the refutation of personal theories. Parents frequently come into the genetic consultation with various hypotheses regarding environmental causes for their child's illness, such as problems during pregnancy, delivery, or early childhood. Although these alternative explanations are clearly disproved by a genetic diagnosis, Stivers and Timmermans (2017) have shown that these types of concerns are raised even when a genetic diagnosis is already provided. These personal theories sometimes have worried the family for years, and can come with feelings of guilt and blame. Being able to address these personal theories and concerns in diagnostic consultations can provide substantial emotional relief (Malek et al. 2019). In several studies on clinical interactions, geneticists are found to firmly dismiss these theories in an attempt to lift the burden of potential feelings of guilt (Featherstone et al. 2005; Stivers and Timmermans 2017).

Impact on providing an NGS diagnosis

The search for a diagnosis plays a big role in the lives of patients, but it is also a crucial part of the medical profession. With rising “administrative, commercial, professional, economic and patient forces”, today's healthcare professional is increasingly pushed to diagnose in ways that suit these forces (Jutel and Nettleton 2011, 798). The diagnosis and the diagnostic code have become increasingly codified and have taken a central role in medicine. The introduction of genetic testing has given rise to so-called “bioclinical collectives” that encompass the network of different disciplines and specialties involved in the clinic, the laboratory, and research. This new configuration in the clinic intertwines laboratory and clinical work, research and clinical practice, and, in this, transforms everyday clinical and diagnostic work. The medical decision-making in genetics has become more collectivized and often includes multiple professions, interpretive tools, and feedback moments (Rabeharisoa and Bourret 2009; Bourret 2005; Bertier et al. 2017). Although more collectivized, diagnosing seems to still be at the core of “doctoring”: the authority to label what is an illness and what is not provides status and power (Jutel 2011; Goldstein Jutel 2019). As can be seen in the fact that a lot of genetic syndromes are named after their “discoverer”, there is status to be gained by identifying rare cases and subsequently publishing on them (Featherstone et al. 2005). Despite the central role of diagnosing in the medical profession, there is a lack of empirical research on the impact diagnosing has on those providing the diagnosis: the focus is predominantly placed on the impact it has on patients (Jutel and Nettleton 2011).

Historically, an important component of the diagnostic practice of a doctor is based on observation (Featherstone et al. 2005; Latimer et al. 2006). Recent ethnographic research in a genetics clinic focusing on dysmorphology has shown that this also holds true for genetic diagnostic practices. Even though one might expect genetic testing to trump other forms of clinical knowledge, this does not seem to occur: not even for those conditions that are seen as inherently genetic. The clinical judgement still remains central and the authority of doctors lies in the art of immediately “seeing” a look that matches a diagnosis and subsequently directing further testing (genetic or non-genetic). When genetic testing is undertaken, expertise lies in correctly choosing and using the test, including a correct interpretation of the results (Featherstone et al. 2005; Shaw et al. 2003; Latimer 2013). Although the focus on “seeing” might be particularly strong in dysmorphology, where there is a significant abnormality in the physical appearance, it is yet to be studied if this authority coming from diagnosing with minimal use of genetic technology is also found in other genetic specialties and if it remains central with the increasing availability of more expansive techniques like NGS.

Although medicine is loaded with uncertainty and, as we have shown, this may be even more the case in genetic medicine, it seems to struggle to accept uncertainty in its practice. Even if medical students develop an accepting attitude to uncertainty, as Fox (1980) proposes, difficulties often remain in addressing these uncertainties in contact with the patients: doctors often associate expressing uncertainties with showing ignorance and therefore try to avoid doing so (Simpkin and Schwartzstein 2016). Some have argued that having to deal with many uncertainties can even lead to an attitude of over-certainty or dogmatism. Paradoxically, in trying to gain control over the uncertainties, clinicians can become insensitive to them (Atkinson 1984; Light 1979; Buetow 2011). More recent work in the field of psychiatry has shown that, at least in this field, clinicians commonly express uncertainty about the individual diagnostic outcome but refrain from voicing ontological uncertainty (Buetow 2011; Gordon, Joos, and Byme 2000). In this way, the source of uncertainty is placed within the patient rather than locating it within the diagnosis, the classification system, or medicine itself. Providing patients insight into these more structural ontological uncertainties can, however, lessen the burden for the diagnostically uncertain and reduce the pressure to find a (completely) fitting diagnosis (Lane 2020).

In the field of genetics, managing uncertainty seems to entail not so much dealing with the uncertainty but mostly “conveying the grounds for the uncertainty” by focusing on what is known and certain (Sarangi and Clarke 2002). A focus on presenting numerical risk values has been found to be one of the practical ways of dealing with the uncertainty and maintaining authority (Sarangi 2002). Recent research in a pediatric practice using NGS shows a similar tendency to focus on explaining (too) difficult genetic concepts and details (Walser et al. 2017). Conveying the more complex information that comes with NGS results is proven to be challenging for healthcare professionals, with concerns centered around the importance of tempering the expectations and the level of understanding (Walser et al. 2017). The unease in acknowledging (diagnostic) uncertainty in NGS can also be recognized in the frequently used “not yet” disclaimer when no genetic explanation is found using NGS techniques (Skinner, Raspberry, and King 2016; Timmermans, Tietbohl, and Skaperdas 2017). A study by Skinner, Raspberry, and King (2016) showed that in the case of a negative result following WES, the result was presented as provisional in about 75 percent of the cases: emphasizing that a genetic explanation was not ruled out. By coupling uncertain findings with a commitment to continue the search and care relationship, patients and their families are offered “what they were certain they could deliver” (Walser et al. 2017, 417).

Furthermore, the high provision of “not yet” diagnoses seems exemplary of the strong belief in the genomic promise, in which all unknowns will become known over time. Even though the claims of the geneticization thesis might have proven excessive, the strong focus on genetics and genomics in scientific agendas, policy strategies, and the media has led to a powerful genetic imaginary (Arribas-Ayllon 2016). Timmermans, Tietbohl, and Skaperdas' (2017) study on dealing with VUS in board meetings, shows how a more flexible, rather than standardized, interpretation of uncertain results creates the opportunity to hang on to a positive genomic imaginary even when the attached promises are not (yet) fulfilled. Yet, there are also voices rising that paint another picture. A study by Kerr et al. (2019) of 25 cancer practitioners involved with molecular cancer research, diagnosis and treatment in the UK, reveals a more tempered enthusiasm on the promise of genetics. Those interviewed recognized the danger of overpromising, and saw the possibility of genomics increasing rather than decreasing uncertainty, due to the infinite layers of complexity that it reveals. Nonetheless, even if the perspective on the role of genomics was found to be more nuanced in the UK cancer clinic, the uncertainty work (Star 1985) undertaken in the foci of both these studies results in a functional and productive genetic optimism: keeping up a positive perspective to ensure ongoing investment and advancement of both clinical work and the associated research.

Discussion and conclusion

As Next Generation Sequencing (NGS) is increasingly making its way into the clinical diagnostic practice, it is often perceived as the next step in resolving the diagnostic odysseys of patients with rare genetic diseases. A substantial portion of patients undergoing NGS, however, are still faced with a not (completely) explanatory result or a result that has an unclear significance. The various types of NGS results, with their various levels of certainty and explanatory power, can require quite some interpretational effort of healthcare professionals and patients. Although more traditional DNA technologies and NGS screening techniques (such as NIPT) pose similar questions on how to deal with these more complex and uncertain results, these difficulties are even more at the forefront in the case of NGS. The questions that are asked when using more traditional DNA technologies or screening tests are more targeted and therefore so too are the answers. While results from these tests also often ask for a personal interpretation of the meaning and implications of the results and future perspectives can be uncertain, the diagnostic answers are often quite clear: there is or is not a variant found. As we have shown, the vast amount of data coming from NGS, the higher chances to encounter more complex results, and the fluid boundaries of the results add to the complexity of the social implications of NGS techniques. Where our knowledge of the genome and its interpretation is still very much in development, NGS currently leads to less stable results and diagnoses.

In this paper, we have scrutinized how these less stable NGS diagnoses are shaped under influence of the broader social context and what providing these diagnoses subsequently entails. By examining the literature on diagnostic NGS practice with a different gaze, stemming from the sociology of health and illness and science and technology studies (STS), we sought to extend the discussions on the social implications of NGS techniques to include the diagnoses themselves. By looking at the social construction of NGS diagnoses, we have provided insight into the broader social and structural processes that play a role in shaping this recent type of diagnosis. Focusing on the role of uncertainty helped us understand the complexities and nuances in the value and meaning of NGS diagnoses.

We have shown how the boundary struggles are particularly visible in the case of NGS techniques. Their recent implementation and the difficulties in interpretation still ask for the active setting of boundaries and leave room to question the existing ones. These boundary struggles take place on several levels and take shape in several moments: in the creation of taxonomies and standards, in the application of such standards in everyday clinical practice, in the communication and interpretation of results with the patient, and in the renegotiation of these boundaries based on developments in the field. We showed how the existing standards and classifications take and change shape in their application in everyday clinical practice. Ongoing developments in the field sustain the fluidity of the boundaries of variant classification by continuously providing new “evidence” and keeping the boundaries of what is enough evidence on the move. Next to the identification of additional variants, this has also led to several instances of downgrading of a variant over time, with sometimes significant personal and medical impact for the involved patients. The interpretation and meaning of NGS results take further shape in the consultation rooms and have proven malleable to the circumstances of the patient.

In the last two sections of this paper, we explored what it means to receive and provide diagnoses identified through NGS techniques. We showed that although genetic results are often highly valued and perceived as unambiguous, a considerable portion of NGS results are not only difficult to interpret but also have an unclear significance for those on the receiving end. The results can fall in more ambiguous categories, such as “likely pathogenic” or “variants of unknown significance”, or can just partly explain the symptoms at hand. The not entirely conclusive results frequently provide little information to help resolve the experienced uncertainties, they merely shift the location of uncertainty or even amplify it. Even the clear NGS diagnoses commonly bring little to inform prognosis, treatment, or clinical management. Nonetheless, even with this limited information or no diagnostic explanation at all, NGS is usually perceived as an end to the diagnostic odyssey and relief is felt.

Providing diagnoses has become increasingly codified and standardized. With high levels of uncertainty and demanding choices, the decision-making around NGS results has become more collectivized and often includes different professions, interpretive tools, and feedback moments. Expertise is shown through a correct choice, usage, and interpretation of the (genetic) diagnostic test(s). Although uncertainty is a key component of genetic medicine, healthcare professionals seem to struggle to accept and show it in practice. Their unease in acknowledging uncertainty can, for instance, be recognized in the high provision of “not yet” diagnoses, where the believe in the genomic promise is fostered. Although with different levels of enthusiasm, healthcare professionals put forward a functionally positive outlook on the future of genomics. In doing so, they ensure ongoing investment in and advancement of the field.

The insights gained from looking at the social shaping of diagnoses obtained using NGS help to unravel the complexities in the shaping of diagnoses in general. Understanding how, by whom, and why NGS diagnoses are constructed the way that they are helps us recognize the agency in shaping them and in defining their meaning. The way that we shape and give meaning to diagnoses, and deal with the uncertainties arising in this process, ultimately impacts how we experience health and illness. More empirical work on diagnostic boundary setting and re-setting in practice would help to clarify what choices are being made in our definitions of genetic illnesses and which wider social processes are guiding these choices. We often focus on how scientific developments guide the boundary setting, but we could be more attentive to, for instance, the moral, practical, and economic influences that are at play here.

Simultaneously, there is a need for more research on how people experience the more fluid and less explanatory forms of diagnoses coming forth out of NGS testing, and what these results mean to them in everyday life. It is clear that the value and meaning of (NGS) diagnoses are not captured solely by their clinical utility, and that a narrow focus on resolving diagnostic uncertainty does not cater to the complex experiences of patients. Exploring the uncertainties that arise in dealing with these complex diagnoses in both everyday life and diagnostic clinical practice can contribute to finding more fulfilling ways of dealing with the uncertainties and needs that both social worlds experience.

Acknowledgement

Danya Vears acknowledges the infrastructure funding received from the Victorian State Government through the Operational Infrastructure Support (OIS) Program.

Disclosure statement

No potential conflict of interest was reported by the author(s).