Community Voices and Whole-Genome Sequencing for Tuberculosis: Storytelling and the Importance of Listening

In recent years, increasingly detailed genomic analysis on diagnosed TB cases (that is, on Mycobacterium tuberculosis (MTB) isolates, rather than human genetic studies) has been carried out by public health authorities with the goal of improving the inferences used to establish transmission chains [1]. Whole-genome sequencing (WGS) is now performed in many research and service laboratories worldwide, with a small but growing number of jurisdictions integrating results in real time into public health responses [2], particularly in high-income, low-incidence settings. Demonstrating that two or more strains of MTB are highly related may lead to determining that an outbreak of TB is occurring [3, 4]. Accordingly, an increasing number of programs globally perform routine WGS on MTB isolates as rapidly as possible for the purpose of improving real-time public health responses, rather than simply for identification of drug resistance mutations or for broader research purposes.

However, the use of WGS for TB surveillance continues to require grappling with the reality that the conclusions are driven by assumptions and an understanding of context and ultimately remain what is sometimes referred to as epistemically underdetermined; that is, the availability of evidence is insufficient to give us definitive conclusions [5]. As we argue below, the plurality of conclusions that one can draw with WGS for TB surveillance speaks to the need to take seriously the importance of narrative and power in interpretation of laboratory results. This leads us to the conclusion that the ethical conduct of WGS sequencing requires community engagement of two kinds: transparent and evolving consensus building among scientists and public health practitioners, as well as input from members of affected communities. Herein, we describe just how the stories we draw around inferences of transmission remain underdetermined. We then articulate what we believe are four key factors or values in using WGS for TB surveillance ethically, namely, transparency, diversity, reciprocity, and accessibility. Rather than conclude nihilistically that WGS for TB surveillance provides no real improvements for public health, we believe that acknowledging the natural limitations of WGS will hopefully make its use even stronger in practice.

While WGS analysis demonstrating that several strains of TB are highly related is often assumed to infer transmission between people, there are in fact a variety of possible explanations for such findings. For example, consider a hypothetical dendrogram showing five sequenced isolates, which are identical on WGS. One possible explanation for this observed pattern is that a single highly contagious individual has had direct contact with the other four people, leading to direct transmission of the remaining cases. However, other hypotheses could also be constructed to explain this finding. These five people may have never met but all migrated from the same country, where this TB strain is widely circulating. Alternatively, all five may have been infected locally by another person who remains undiagnosed in the community and has therefore not been identified in the analysis. It is also possible that a laboratory error in processing samples has led to one of these strains contaminating the test results for the other four and that these people do not have TB at all.

These alternative explanations are not exhaustive but are all possible stories that the same pattern of sequences could represent and in the absence of additional clinical or epidemiological information may be equally likely. It might seem that the truth of which is correct can be resolved with more information. For example, a public health service could interview people with TB to find out if there were opportunities for contact between them or investigate laboratory practices to seek evidence about whether a contamination event had occurred. However, further information may not actually resolve the issues. In the scenario above, what if all 5 came from the same country originally and did not know each other socially but had all been investigated for cough through the same hospital? Any of the three explanations would remain possible. In fact, it is most likely that further investigation, however detailed, might make some explanations seem more or less likely but would not answer the question of what had occurred definitively.

The different possible explanations given here are materially different and may lead a public health program to act in quite different ways. Whether local or overseas transmission is considered correct will change the extent and urgency of community testing for others at risk, while if laboratory error is felt to be true, some people otherwise diagnosed with TB may not be treated at all. These interpretive lenses would lead to different interventions or omissions, and perceived prior knowledge about factors such as the expected frequency of local transmission, or the quality of TB laboratory practices, will influence assessment about which interpretation should be considered correct. It must be highlighted, though, that interpretation is likely also to be stratified on the basis of logistic or political factors. If, for example, community screening in a particular group of people is considered by a public health program to be complex, expensive, or unwelcome, it may be much easier to conclude that transmission has not occurred locally and direct efforts elsewhere. These explanations of WGS truth are not merely academic or philosophical but directly impact health systems and communities.

Stepping back from the specific context of WGS and public health, the construction of an explanation for our observations of the world is not straightforward or self-evident. Although millennia of arguments exist as to the nature of reality and humans’ ability to access it and shape it, most philosophers agree that a simplistic model of reality or truth as immutable and being “out there” for us to discover is false and potentially dangerous [6]. As famously highlighted by Jacques Derrida, an individual’s view of which narrative most accurately reflects reality is shaped by their “previous thoughts, beliefs and values” [7]. This insight does not mean we ought to resort to skepticism about scientific knowledge; on the contrary, one common – perhaps the most common – view is that medical and public health research and as such medicine and public health are community endeavors where epistemic values, for example, accuracy, can only be understood partially and dynamically and where such values must consistently be contested and defended within community [8]. Returning to Derrida, expressions of phenomenon as “real” or “true” do not represent neutral positions but are often attempts to hold and exert power as the arbiters of truth, even if this is not recognized or acknowledged [9]. Bringing together these ideas then, reality is at least partially constructed through values set by community whereby members of that community stand in different relationships and often in power asymmetries.

When considering scientific hypotheses such as the interpretation of a given pattern of WGS results, it is easy to suggest that additional information should at least lead us to prefer some explanations over others being as more likely to reflect the true situation and to assume that a single underlying truth is available for reference [10]. However, the variety of possible explanations for any given set of findings and plurality of viewpoints about the most plausible interpretative hypothesis persists, and it is unclear that increasing information necessarily leads to consensus about optimal explanations [11]. This may be in part due to conflict about what information can be legitimately considered as evidence, as well as bias regarding whose perspectives or positions are eligible for inclusion [12]. Fricker’s [13] description of epistemic injustice is relevant here as individual experience and legitimate perspective may be inappropriately dismissed or rejected due to preexisting bias. This is especially important in health care contexts, where the viewpoints of patients in general, and of marginalized patient groups more specifically, are widely recognized to have been historically undervalued or excluded [14, 15].

While not exhaustive, these approaches to critiquing how scientific truth is constructed are salutary and should certainly lead to caution in assuming “the” true phenomena underlying given observations. We consider it helpful in broad terms to describe explanations of scientific phenomena as “narratives,” in the sense of being a constructed explanation, which imposes order on observations. The use of the language of narratives serves as a reminder that, at best, such explanations can only approximate the underlying reality observed. Equally important, however, is that a narrative necessarily arises from a perspective rather than an absolute reflection of reality. In a health care context, the accepted narrative will be determined from a position of authority, which typically also connotes differences in power. Power differentials in health care may involve invoking consequences such as diagnostic labels, treatment, or even mandatory isolation on others according to accepted narratives [16]. Returning to descriptions about transmission of TB, then, if there are competing possible narratives to explain any given pattern of WGS results, what are the implications of this information for public health and for actions which might be taken in response?

Despite the contested nature of descriptions of community transmission based on WGS information, we would argue that programs do not need to be nihilistic with regard to capacity for response [17]. Indeed, if taking seriously the idea of perspective or positionality meant that no legitimate action could be taken on possible public health implications of WGS, it would be paralyzing and limit the possibility of offering genuine benefit to individuals and communities. Accepting both that there is genuine uncertainty and a real potential for value from action, we would offer four suggestions for shaping the integration of WGS into a public health response – relating to concepts or values of transparency, diversity, reciprocity, and accessibility.

First, acknowledging that an interpretation of WGS tells a story should lead us all (clinicians, researchers, and communities) to want more transparency about these decisions, and the basis on which they are made, in order to facilitate debate and discussion about the meaning of WGS results. Health care actors can support greater transparency through publication and dissemination of protocols, consensus statements, and worked case examples, which should be explicit about the approach taken to considering WGS and what is considered reasonable from a clinical and public health perspective. While publication is not the only avenue to transparency, written materials allow for reflection and external challenge as others are invited to reflect and respond to public health responses [18]. Developing these materials is also about the work of articulation as practitioners force themselves to be clear and consistent in their approach and open about the facts and reasoning on which it is based.

Second, understanding that stories are told by storytellers, there is an opportunity to ensure that a diversity of voices are part of that process. In existing programs, there may be laboratory scientists, epidemiologists, public health workers, and clinicians routinely considering WGS results. Those different disciplinary perspectives are often helpful [19]. However, we suggest that public health programs should look for ways to increasingly involve people outside of medicine and public health, including ethicists and affected community members. In part, this recognizes that the wider community can be impacted by public health interventions and that representation allows for greater agency and justice. Diversity of perspective, though, is not simply about representation, but it also allows opportunities for public health programs to improve understanding about what this information means in, and for, affected communities and how to use it most effectively for public benefit. These perspectives may be invaluable for understanding relevant community activities. From our experience, for example, the active participation of community members in explaining likely contact patterns within periods of mourning (“sorry business”) for Aboriginal communities has been essential for allowing respectful and effective follow-up after TB exposure [20].

Third, a duty of reciprocity is owed to those with TB that should be reflected in how information from WGS is approached [21]. While it is generally considered that individuals do not “own” the organism with which they are infected, they do regularly provide information about themselves and their activities and interactions to public health officials. This sensitive and personal information contextualizes and informs our understanding of genomic data and is used to enhance public health responses to interrupt transmission. Individuals affected by TB thus contribute both directly and indirectly to providing benefit for the broader community through reducing risk of further transmission and do so at the potential cost of allowing access to their own personal stories and risks including stigmatization. The information arising from contextualized WGS is imperative and beneficial to the wider community, and public health practitioners who steward it have an obligation to honor that lived experience and contribution, which includes listening to the voices of affected individuals and communities in shaping a response [22, 23].

Fourth, we recognize that stories have power and should be used by public health program in ways which are accessible to challenge and correction; such power can deeply affect people’s lives and relationships. This has been recently displayed in the emerging of an international monkeypox outbreak, with inferences about transmission based on genomic data used in public health messaging and concerns regarding unreasonable stigma in affected communities [24]. In our system of public health and politics, TB programs have the authority to use that information in ways that help people greatly but also in ways that can restrict, impede, or embarrass them [25]. Where sequencing information is considered to reflect transmission and ongoing community risk, this may include application of public health powers for mandatory detection in many jurisdictions [16]. When coupled with the knowledge that our understanding is uncertain, it becomes essential that we build in checks and balances and that these be accessible to those most affected. That is particularly the case when it comes to interpreting and acting on the information for a public health intervention, such as through contact tracing, disclosure of information to a workplace or school, or in the most extreme cases potentially even informing public health orders where ongoing transmission risk is considered likely. In many jurisdictions, formal mechanisms for challenging public health orders exist, but the degree to which they are accessible and used is uncertain and controversial [26]. And the degree to which legal challenges might include interpretation of laboratory results is unknown, though likely little to nil. In addition to such structured formal processes, it may also be useful to provide informal and independent pathways, such as community advocates or ombuds, to improve accessibility [27].

Such cautions in interpretation of the inferences we can draw from WGS does not mean it cannot, as a tool, provide important information for understanding TB epidemiology or considering public health action. As we have suggested, however, an appropriate degree of epistemological humility regarding interpretation and implementation is necessary, drawing on the full scope of information and perspectives available, and should lead public health program toward a broader inclusion of voices from affected people and communities.

Some models for community engagement in the public health use of genomic data have previously been developed and evaluated, particularly in relation to HIV and sexually transmitted infections [28, 29]. These experiences, and more general approaches to effective community engagement, are informative in considering what structures may best support future efforts to include more diverse voices in TB public health and WGS [30, 31]. It is likely, however, that TB’s unusual characteristics may require some bespoke approaches to community involvement in this area. For example, TB is transmitted via aerosolization (and so may be acquired after brief or unrecognized contact), has a long latency period (and so infection may not become apparent until years later), and has a lengthy treatment course (meaning that individuals may still remain on treatment when WGS results are available so results may affect management or contact follow-up directly in real time). This contrasts with characteristics of other infectious diseases such as HIV where work on use of genomic data for inferring transmission is extant [32]. Recognition of potential harms from HIV genomics has led already to the development of models of community engagement for public health application [33]. While careful consideration of the structures best placed for community engagement in TB WGS and the public health response is required, the characteristics of TB also mean that there are important potential opportunities to improve treatment, prevent disease, and support communities in real time [34]. A strong and consistent community presence in all aspects of interpretation, implementation, and evaluation of TB WGS information is critical to support and pursue as such programs are established and should be incorporated routinely in public health programs.

The authors have no conflicts of interest to declare.

This manuscript was not supported by any sponsor or funder.

J.T.D. and D.S.S. jointly conceived this work and developed this manuscript for publication, with J.T.D. preparing initial drafts and both reviewing and approving the final manuscript.