Finding Fault? Exploring Legal Duties to Return Incidental Findings in Genomic Research

Abstract

The use of whole-genome sequencing in biomedical research is expected to produce dramatic advances in human health. The increasing use of this powerful, data-rich new technology in research, however, will inevitably give rise to incidental findings (IFs)—findings with individual health or reproductive significance that are beyond the aims of the particular research—and the related questions of whether and to what extent researchers have an ethical obligation to return IFs. Many have concluded that researchers have an ethical obligation to return some findings in some circumstances but have provided vague or context-dependent approaches to determining which IFs must be returned and when. As a result, researchers have started returning IFs inconsistently, giving rise to concerns about legal liability in circumstances in which notification could have potentially prevented injury. Although it is clear that ethical guidance should not be automatically codified as law and that crafting ethical obligations around legal duties can be inappropriate, the ethical debate should not proceed unaware of the potential legal ramifications of advancing and implementing an ethical obligation to return IFs.

This Article assesses the legal claims that could be brought for a researcher’s failure to return IFs. The potential for researchers to be held liable in tort is still uncertain and turns largely on a number of factors—including customary practice and guidance documents—that are still in flux. Unlike medical care, which has a well-defined duty into which evolving scientific knowledge about genetics and genomics can readily be incorporated, a researcher’s duty to return IFs is less well defined, making it difficult to determine at the outset whether and when legal liability will attach.

This Article advocates for a clearer, ethically sound standard of requiring that researchers disclose in the informed consent document which approach to offering IFs will be taken. This approach enables participants to know at the outset which findings, if any, will be returned, allows researchers to ascertain when their failure to appropriately return incidental findings will give rise to liability, and enables courts to make determinations that will produce more consistent legal guidance.

Introduction

A little over a decade ago, researchers first sequenced a rough draft of the human genome, the complete set of human genetic information.¹ Sequencing the human genome took over a decade and cost approximately $2.5 billion.² The hope and expectation was that delving further into the human genome would advance scientific knowledge and human health through genomic research. Few could have imagined the progress that would be made in the decade that followed.

Ten short years later, the estimated cost of sequencing an entire genome has dropped below $10,000, and the $1,000 genome is within sight.³ Whole-genome sequencing has quickly been incorporated into research, with researchers in a wide variety of disciplines—from mental health to cancer research—incorporating whole-genome sequencing into their research in hopes of uncovering a genetic component to the diseases they study.⁴

With the incorporation of whole-genome sequencing into research comes concerns about incidental findings⁵ (IFs)—genetic or genomic discoveries about individual research participants that are otherwise beyond the scope of research⁶—and the extent to which researchers have an ethical obligation to offer to return these results.⁷ For example, does a researcher looking for the underlying genetic cause of heart disease, who discovers that a research participant’s genetic sequence indicates a predisposition to an untreated condition, such as breast cancer, have an obligation to disclose that information? The references in this Article to the return of IFs are, in most cases, shorthand for an offer to return IFs which a participant can autonomously reject.

Those who have considered whether and to what extent researchers have an ethical obligation to offer to return IFs arising from genomic research have generally found there to be an ethical obligation to return some IFs in some contexts, but they have offered varying or context-specific standards or frameworks that are not sufficient to solve this difficult and controversial problem.⁸ Perhaps because of this lack of unanimity, and in the absence of clear guidance, researchers have taken a variety of approaches to returning IFs in practice, and some researchers have chosen not to return IFs altogether.⁹ This gap between the view that researchers have an ethical obligation to return at least some IFs, and the reality that some, but not all, researchers choose to return IFs has led many to worry about the potential for legal liability arising from inconsistent approaches to returning IFs.¹⁰ Given that tort law duties are determined by the prevailing standard of care, recognition by scholars and the research community of an ethical obligation to return IFs could ultimately lead to a legal obligation.¹¹

We examine whether, and under what circumstances, there may be a legal obligation to offer to return IFs—an analysis that is currently lacking.¹² Although there is no law or case law directly on point,¹³ this Article concludes that there is a small possibility that a failure to appropriately return IFs could result in legal liability under law as it stands today. Furthermore, there is a greater likelihood of legal liability as scholars and researchers continue to advocate for an ethical obligation, particularly if returning IFs becomes widespread practice.¹⁴ We also recognize that fragmented and varying ethical guidance could give rise to uncertain and inconsistent legal duties—a result that would make it difficult for researchers, participants, and courts and that could cause gridlock in this emerging scientific arena.

For this reason, this Article argues that the research community should consider prospectively adopting a clearer approach to returning IFs that could give rise to a more definitive legal standard delineating when IFs should be returned. A clearer legal standard would tell researchers when they are expected to offer to return IFs, and when they will face liability for falling short of this standard; a clearer approach would also let participants know which IFs they can reasonably expect to have returned.¹⁵

Part I provides an overview of the issue of IFs in the context of whole genome sequencing research, including a discussion of the IFs that might arise in genomics research; the ethical principles that potentially underlie a researcher’s ethical obligation to return IFs; and the different approaches setting forth whether, when, and which IFs ought to be returned. Part II reviews the possible tort law claims that could be brought alleging a failure to adequately return IFs and recognizes that there may only be a narrow set of circumstances in which researchers could be held liable today. As the research community increasingly recognizes an ethical obligation to return IFs, as guidance documents set forth which IFs should be returned and when, and as research custom moves toward routine return of IFs, researchers will be increasingly likely to face legal liability. Part III recommends that researchers develop clearer policies for when IFs will be returned and communicate to research participants whether, when, and which IFs will be returned. Research participants can thereby develop reasonable expectations about the IFs they can expect to receive. This Part sets forth and analyzes several approaches that would provide clear and consistent guidance and selects the approach that best allows researchers to make informed decisions about appropriate policies for returning IFs. This approach also makes clearest to research participants which IFs they can reasonably expect to have returned. Part IV suggests that courts that are faced with lawsuits alleging a researcher’s failure to adequately return IFs adopt the proposed standard, bringing clarity to this uncertain terrain.

I. An Ethical Obligation to Return Incidental Findings from Genomic Research?

The high likelihood—or virtual certainty—that IFs will arise in genomic research has given rise to questions about a researcher’s ethical obligation to return IFs. This Part notes that certain ethical principles—beneficence, respect for persons, a duty to rescue, reciprocity, and partial entrustment—could give rise to an obligation to return IFs; other notions, however, including non-maleficence, concerns about the therapeutic misconception, and worries about burdening the research enterprise—caution against finding such an ethical obligation. This Part examines the guidance documents—documents that seek to establish best practices and that are often relied upon by practitioners—and frameworks that assess whether and when IFs should be returned.

A. INCIDENTAL FINDINGS IN GENOMIC RESEARCH

Over the past decade, biomedical research has increasingly incorporated whole-genome sequencing, whole-exome sequencing (a process of selectively sequencing the small portion of the genome that codes for proteins, which is believed to cover approximately eighty-five percent of the disease-causing regions in the genome), and other large-scale genomic methodologies into research in hopes of better understanding the genetic causes of disease.¹⁶ With these new technologies, researchers are no longer limited to interrogating certain targeted portions of the genome; instead they can conduct genome-wide analyses, casting a wider net in the hopes of finding answers to their research questions.¹⁷ Although a powerful research tool, this broader sequencing gives rise to a greater likelihood of finding IFs. Furthermore, as the cost of whole-genome sequencing falls, researchers who had heretofore used targeting sequencing (or who had not used genetic methodologies at all) will increasingly incorporate whole-genome sequencing into their research, increasing the number of studies in which IFs are likely to become a relevant issue.

The issue of IFs is not unique to genomic research; IFs can arise from a number of different types of tests.¹⁸ What distinguishes the issue of IFs in genetic research from IFs in imaging and other research procedures is the likelihood and scope of potential IFs. In imaging research—research conducted using imaging techniques that include MRI, CT scans, and PET scans to examine particular parts of the body—the question is whether one of a limited handful of types of IFs will occur.¹⁹ In the context of genetic research, the question is how to deal with the dozens or hundreds of disparate IFs that can potentially be uncovered in any individual research participant’s genetic material.²⁰

IFs can exhibit a wide range of usefulness.²¹ For example, an IF can indicate an urgent, clinically actionable condition such as a predisposition to non-polyp-forming colon cancer—a cancer that is fatal and otherwise undetectable through standard screening.²² An IF could predict a propensity to develop disease in the future, such as the possibility of developing breast cancer or Alzheimer’s.²³ Even when the variant’s meaning is known, “the probabilistic character of genetic information and the pleiotropic nature of genes make accurate interpretation … particularly challenging.”²⁴ Some IFs will reveal information that is not clinically significant, such as information about ethnic or geographic heritage.²⁵ Yet other variants, perhaps even most variants, may instead be something the meaning and import of which is unknown.²⁶

When the consequence of the variant is known and the finding has clinical significance, research participants may benefit from learning about IFs discovered during the course of research. A research participant who learns that she has a BRCA mutation, an indication of a predisposition to developing breast cancer, could take measures aimed at preventing undesirable future medical outcomes, including participation in regular mammogram screening.²⁷ Even research participants who learn that they are predisposed to Alzheimer’s disease—a disease that is currently unpreventable and untreatable—may decide to monitor more vigilantly scientific advances to obtain the benefit of any potential future treatments.²⁸

Despite the potential benefits of learning about genetic predispositions, disclosing IFs—even IFs of known clinical significance—may nevertheless pose risks for individual research participants. Disclosing IFs that suggest genetic predisposition for disease can cause psychosocial harms and may raise concerns about loss of privacy or confidentiality.²⁹ Given the possibility that any IF could be a false positive, a research participant’s decision to undergo medical treatment in response to an IF “could lead research participants down inappropriate or dangerous clinical pathways.”³⁰ As noted in the article, The Future of Incidental Findings: Should They Be Viewed as Benefits?:

The positive or negative outcome depends partly on one’s circumstances (e.g., whether one has insurance to cover follow-up evaluation or must shoulder crippling costs and even job loss, whether discovery of the IF saves one’s life or leads only to expensive and burdensome tests yielding no useful medical information) and one’s values (e.g., whether a diagnosis of untreatable cancer prompts a clarification of one’s priorities, provides the opportunity for salvific suffering, or lowers a black cloud of depression over one’s final months).³¹

Disclosing IFs also has costs for the research enterprise. Before returning an IF, researchers must determine that an IF is analytically valid—that the IF was correctly identified by the sequencing process as a variant.³² Determining analytic validity may require confirmation from a laboratory approved under the Clinical Laboratory Improvement Amendments (CLIA).³³ Researchers must also ensure that results are clinically valid—that the genetic variant identified is predictive of a propensity towards developing a specific disease,³⁴ which may require a time-consuming analysis of the current state of scientific knowledge.³⁵ Some guidelines also require that IFs be clinically actionable before they can be returned—a requirement that research participants be able to act upon the finding to avoid unwanted medical outcomes.³⁶ A requirement of actionability means that researchers must go through an additional analytic process to assess whether a particular variant should be returned, a process that could require additional time and expense. The process of validating IFs imposes additional burdens on the research enterprise, both in time and money.³⁷

IFs in genomic research are therefore extremely common and likely to occur, with frequency increasing as one broadens the definition of what counts as an IF. IFs may indicate important, clinically actionable results or may instead be findings the meaning and importance of which are unknown. Returning IFs may benefit individual research participants, but it also risks harming individuals and burdening the research enterprise.

B. PRINCIPLES UNDERLYING AN ETHICAL OBLIGATION TO RETURN INCIDENTAL FINDINGS

Given the vast array of IFs that could be discovered in genomic research, the question arises as to whether and when researchers have an ethical obligation to return IFs to individual research participants. A number of principles support an ethical obligation to return IFs.³⁸ Beneficence, for example, requires that researchers take efforts to “maximize possible benefits and minimize possible harms.”³⁹ With regard to IFs, returning IFs could benefit participants if the information enabled participants to take additional precautions to prevent or minimize future medical harm.⁴⁰ To the extent particular IFs are likely to be valuable to research participants, beneficence suggests that the IF should be returned.⁴¹

Respect for persons—a recognition that individuals should be treated as autonomous agents, capable of making decisions and acting in accordance with their unique value systems—might also support an ethical obligation to return IFs.⁴² Respect for persons suggests that when researchers have information that participants would find useful in making their own medical decisions, such information should be disclosed.⁴³ Respect for persons may even support returning IFs for which disclosure may be harmful, recognizing that individuals should be free to make autonomous decisions based upon this information.⁴⁴

A third principle, and possible source of an ethical obligation to return IFs, is the duty to rescue, which obligates an individual to act when presented with an opportunity to alleviate the serious plight of another with minimal burden to oneself.⁴⁵ The extent of any duty to rescue turns on “the urgency of the situation, the severity of the outcome if nothing is done, the ability of the potential rescuer to prevent that outcome, and the fact that only modest effort or sacrifice on the rescuer’s part is required.”⁴⁶ The quintessential example of a general duty to rescue involves a passerby seeing a young child drowning in a shallow pool of water; with minimal inconvenience—perhaps a few minutes of time spent, or clothes that become wet—the passerby can save the life of a drowning child. The passerby is generally thought to have a moral obligation, even though not a legal duty, to rescue the drowning child.⁴⁷

Under a duty-to-rescue-based argument, researchers would have an ethical obligation to return certain IFs to research participants if the level of effort required to return them is minimal and the possible benefit to participants is great.⁴⁸ Arguably, the current process for returning IFs is burdensome,⁴⁹ and few known findings rise to the level of alleviating a serious plight of another.⁵⁰ A duty to rescue would therefore be unlikely to attach under these circumstances.

A fourth principle, reciprocity, suggests that those who contribute to the research enterprise—either with time or by providing biological material—are owed something in return for their participation.⁵¹ Though the exact contours of what the principle of reciprocity requires are subject to debate, a reasonable interpretation could include notifying those who contribute biological samples of relevant IFs.⁵²

Finally, the partial-entrustment model—developed to assess when researchers are ethically obligated to provide care beyond that which is necessary to protect participants from the risks of research or for the scientific conduct of the research, otherwise known as ancillary care—sets forth a framework to assess whether researchers are ethically obligated to return IFs.⁵³ The partial-entrustment model recognizes that participants grant researchers access to otherwise private health information, creating a relationship of limited trust, but that research participants do not generally rely exclusively on researchers for their medical needs.⁵⁴ Under the partial-entrustment/ancillary-care model, once a condition is determined to be within the scope of entrustment, the strength of any obligation turns on a variety of factors, including: (1) participants’ vulnerability; (2) participants’ uncompensated risks or burdens; (3) the depth, in intensity and duration, of the researcher–participant relationship; and, (4) participants’ dependence on the researchers.⁵⁵

Although several of the variables in the partial-entrustment model, including participant vulnerability and uncompensated risks, cannot be assessed in the abstract, analysis of the remaining variables suggest that researchers may not have a universal ethical obligation to return IFs.⁵⁶ For example, researchers conducting genomic research may have only a minimal, or nonexistent, relationship with research participants.⁵⁷ And participants in genomic research entrust their genetic information to researchers but do not generally entrust researchers with their medical care.⁵⁸

Several principles caution against finding that researchers have an ethical obligation to return IFs. The first principle is that of non-maleficence, often stated as “do no harm.”⁵⁹ As discussed in section I.A, returning IFs has the potential to cause harm, either by causing emotional unrest upon learning of upsetting information or by causing physical harm to the extent false positives result from unnecessary medical treatment.⁶⁰ Because IFs can cause harm, the principle of non-maleficence cautions against returning IFs in the absence of a plan to minimize the risks of harm.

Concerns about the therapeutic misconception may also counsel against an ethical obligation to return IFs.⁶¹ Many research participants suffer from what is called the therapeutic misconception—the mistaken conflation of medical research with clinical care.⁶² As a result, research participants often believe that researchers will provide treatment comparable to that provided by physicians and accordingly believe that researchers who come across information that can prevent future medical harm will disclose this information even absent a fiduciary relationship. As has been noted, however, though the primary goal of medicine is to treat individual patients, the primary goal of research is to produce generalizable knowledge.⁶³ Participants may rely on researchers in ways that were never intended. Researchers who choose to return IFs may therefore inappropriately contribute to the therapeutic misconception.⁶⁴

A final concern counseling against finding an ethical obligation to return IFs is the burden that this obligation would place on the research enterprise.⁶⁵ Returning IFs requires outlays of time and money that must be diverted from other research objectives. Investing this time and money into returning IFs—information that primarily benefits individual research participants—detracts from the goal of creating generalizable knowledge in ways that some may argue is unethical.⁶⁶

This section is merely intended to lay out the range of principled arguments that scholars have made for and against the idea that researchers have a duty to disclose IFs. Although it appears that these ethical principles may support some (possibly limited) duty to disclose incidental findings, the contours of any possible duty still remain unsettled. Appeal to ethical principles provides no clear answer as to whether researchers must, can, should, or should not return IFs.⁶⁷ As shown in the next section, this uncertainty carries into research practice.

C. INCONSISTENCY IN PRACTICE

Given the inconclusive guidance provided by an analysis of the relevant ethical principles, scholars and practitioners have sought to resolve the issues of when, whether, and which IFs ought to be returned through practical guidance. Researchers can choose to take a range of actions with regard to returning IFs, from returning nothing to returning everything (such as a comprehensively annotated analysis of the subject’s entire genome). Most of those who have studied this issue conclude that neither of these extremes is practicable or appropriate.⁶⁸ Rather, they generally conclude that there is an ethical obligation to return at least some IFs in some circumstances⁶⁹ but offer a variety of sometimes inconsistent approaches about which IFs must be returned and when.⁷⁰

A majority view is beginning to emerge that IFs should be returned only when they are analytically valid,⁷¹ have significant health implications, and are clinically actionable.⁷² A number of institutional efforts to promulgate guidance on this issue, and on the related issue of when individual research results should be returned, have reiterated these factors.⁷³ The consistency of these criteria belies the vagueness in how these terms are defined.⁷⁴ For example, what constitutes a “significant” health implication? Must the consequences be life or death? Or is an IF that, once revealed, leads to actions that can lower one’s chance of contracting an unpleasant disease sufficiently significant? Must the IF have individual significance, or is reproductive significance—significance to offspring—sufficient?⁷⁵

The same definitional uncertainty applies to the requirement that findings be clinically actionable. One group, the 2010 National Heart, Lung, and Blood Institute (NHLBI) Working Group, defined “clinically actionable” as follows:

Actionable means that disclosure has the potential to lead to an improved health outcome; there must be established therapeutic or preventive interventions available or other available actions that may change the course of disease. Actionable may include surveillance and interventions to improve clinical course, such as by delaying onset, leading to earlier diagnosis, increasing likelihood of less burdensome disease, or expanding treatment options.⁷⁶

This definition seems to foreclose returning IFs that are not clinically actionable now but may become clinically actionable in the future.⁷⁷ And still other approaches have been suggested for defining clinical actionability.⁷⁸

In addition to vagueness in the definitions, there is inconsistency in how these factors are to be applied. For example, the National Bioethics Advisory Commission concluded that IFs should only be returned if they satisfy the criteria, whereas in 2004, the NHLBI Working Group concluded that IFs should not be withheld if they satisfied the criteria.⁷⁹

The uncertainty in the way the factors are defined and applied has led some to conclude that determinations about whether to return results should be made on a case-by-case basis.⁸⁰ The uncertainty has also led to ongoing efforts by a number of advisory bodies to promulgate lists of which IFs ought to be returned.⁸¹ Although preliminary data suggested that there may be a great deal of consensus about the limited set of particular results that should be returned,⁸² as guidance documents have begun to take more concrete positions about what should constitute an appropriate standard of care, the pushback has been intense. For example, in 2013, the American College of Medical Genetics and Genomics recommended that for clinical whole-genome sequencing, a limited set of IFs should always be returned, without deference to individual patient or clinician preferences or to patient age.⁸³ This guidance document generated a flurry of positive and negative responses, with critiques focusing on the move away from a “right not to know” and the abandonment of seemingly settled thoughts about the inappropriateness of testing minors for adult-onset conditions.⁸⁴ A contemporaneous publication from the European Society of Human Genetics covered similar ground, although it offered a different set of recommendations.⁸⁵

By and large, scholars, practitioners, and advisory bodies agree that researchers have an ethical obligation to offer to return some IFs. The various frameworks that have been offered to assess IFs are starting to converge around three variables: validity, a significant health impact, and clinical actionability. But this seeming consensus belies the vagueness and inconsistency in how these variables are defined and applied. The current thinking about when IFs ought to be returned provides little practicable guidance to researchers.

D. CONCERNS ABOUT LEGAL LIABILITY

Despite the emerging majority view that researchers have an ethical obligation to return at least some IFs arising from genomic research, the practice of returning IFs has developed more slowly and in a highly variable manner. Rather than adopting comprehensive policies to determine whether, when, or which IFs should be returned, some researchers have chosen to avoid the problem, stating in informed consent documents that no individual results, incidental or otherwise, will be returned.⁸⁶ Others have not precluded the possibility of return but have set forth policies indicating that IFs will be returned under extraordinarily narrow circumstances. Many others have remained silent in their consent forms about the entire concept of IFs. Even researchers who conscientiously choose to return IFs often do not clearly define the circumstances under which they will do so.⁸⁷

This gap between the emerging majority view that some IFs ought to be returned and the reality that IFs often are not returned—or are returned under a variety of narrowly or ill-defined plans—has led to growing concern about potential legal liability arising from a researcher’s failure to return IFs.⁸⁸ Consider, for example, a heart disease researcher conducting whole-genome sequencing who accidentally stumbles upon a finding that a participant has a propensity to develop otherwise difficult-to-detect non-polyp-forming colon cancer. Because the consent form clearly states that only results related to heart disease will be returned (or perhaps because the consent form states that no individual results will be returned whatsoever), the researcher does not disclose the individual’s heightened risk of developing colon cancer. When that individual later goes on to develop the disease and learns that she could have had a better treatment prognosis with earlier notification, she could bring a lawsuit against the researcher for failing to disclose the potentially lifesaving information.

Although there is, as yet, no law on point,⁸⁹ there is a growing sense that we may be standing at the precipice of legal liability for failing to adequately return IFs.⁹⁰ Liability could arise from three situations: (1) failing to return IFs at all; (2) returning some but not the ones that could prove critical; or (3) returning IFs that turn out to be false positives that give rise to harms as a result of attempting treatment. To the extent there continues to be a push towards an ethical obligation to return at least some IFs, and to the extent offering to return IFs develops into a research standard or custom, the ethical obligation to return IFs could give rise to a legal duty.⁹¹ Although it is clear that ethics is not the law⁹² and that deriving the parameters of ethical obligations from concerns about potential legal obligations can be inappropriate, the ethical debate should not proceed unaware of the potential legal ramifications of advancing and implementing an ethical obligation to return IFs.

II. Potential Tort Law Claims for Failure to Return Incidental Findings in Genomic Research

The discrepancy between the emerging ethical view that at least some IFs should be returned and the reality that IFs are being returned inconsistently has given rise to concerns about legal liability.⁹³ There is growing concern that research participants who later develop a disease that could have been predicted may bring suit against the researcher who knew about the potential harm but did not disclose it. Injured research participants could allege that, had they been notified about their propensity to develop a disease, they would have taken additional precautions to prevent or minimize the consequences.⁹⁴

Because there is no case law directly addressing researchers’ legal obligations to return IFs,⁹⁵ plaintiffs bringing suit—and the courts hearing these cases—will have to reason analogically to areas in which courts have considered liability for failure to disclose. The discussion below examines possible tort law claims for failing to return IFs.⁹⁶ As a general rule, injured research participants will almost always have difficulty recovering in tort. This is so, in part, because informed consent documents have been viewed by courts as means by which participants assume the risks of research, because the duties that researchers owe are uncertain and have been held to be limited to compliance with federal regulations, and because research interventions can result in injury even if no one was at fault.⁹⁷ Nevertheless, as additional voices call for an ethical obligation to return IFs, and as this emerging ethical obligation increasingly becomes standard or customary practice in research, the emerging ethical obligation could give rise to a legal obligation to return IFs, the failure of which could result in legal liability.

A few illustrative examples of possible scenarios are set forth here and analyzed throughout the section that follows. These examples set forth a number of factors that courts would likely consider under the current tort law regime. Most notably for our analysis, we will discuss: (1) the extent of any researcher–participant relationship; (2) the foreseeability and proximity of any harm; and (3) the extent to which the harm uncovered is preventable.

Example 1 (Anesthesia Death): Having suffered a heart attack once before, Arnold Adams is altruistically motivated to enroll in research to contribute to a deeper understanding of the genetic causes of heart disease. Arnold is referred by his heart specialist to Dr. Ames, a physician specializing in heart disease research, for enrollment in Dr. Ames’s non-therapeutic research protocol. Dr. Ames meets with Arnold and sequences Arnold’s entire genome. Dr. Ames learns that Arnold has a rare variant that causes a potentially fatal outcome should Arnold be exposed to a particular anesthesia that can be used in heart surgery. Dr. Ames does not tell Arnold of this genetic variant. One month later, Arnold goes into surgery for his heart condition, is given the anesthesia, and dies.

Example 2 (De-Identified Data): Betsy Banks donates blood and agrees that her blood can be used in research so long as no other identifying information is provided. Dr. Bell, a Ph.D. scientist, is conducting research using de-identified samples and notes that the blood marked Specimen X, Betsy’s blood, contains a genetic marker for a predisposition to breast cancer. Unaware of her predisposition to breast cancer, and in accordance with clinical guidelines for those with no predisposition to breast cancer, Betsy gets regular mammograms every two years. The guidelines for those with a predisposition to breast cancer suggest getting mammograms every year. In the second year following her mammogram, Betsy develops breast cancer and ultimately succumbs to it.

Example 3 (Large-Scale Research): Cal Crawford decides to enroll in a research study as a healthy research participant. Dr. Cruz, an oncologist, is sequencing the whole genome of both healthy and sick research participants to ascertain the underlying genetic causes of a particular cancer. The study is a large-scale study with thousands of participants, none of whom Dr. Cruz meets personally. In sequencing Cal’s whole genome, Dr. Cruz discovers that Cal’s genome contains the ApoE gene, indicating a predisposition to Alzheimer’s disease. Dr. Cruz does not return this result to Cal. Fifteen years later, Cal developes Alzheimer’s disease. Had Cal been told about his predisposition, he would have monitored scientific developments about emerging ways to stave off Alzheimer’s disease and might have made a different set of life choices.

The claim most likely to be brought by a participant in one of the above scenarios is a tort law claim alleging negligence.⁹⁸ To prove negligence, a research participant must show: (1) that a researcher owed a duty to participants to return IFs arising from research; (2) that the researcher breached the duty; (3) that the breached duty was the “but for” and proximate cause of the participant’s injury; and (4) that the participant suffered measurable harm.⁹⁹ An analysis of each of these elements is set forth below.

A. A LEGAL DUTY TO RETURN INCIDENTAL FINDINGS?

As a general rule, individuals owe a duty of reasonable care under the circumstances.¹⁰⁰ Absent particular circumstances, tort law imposes no affirmative duties to act for another’s benefit,¹⁰¹ and individuals are not required to warn others of impending harm.¹⁰² A number of factors can overcome this general tort law notion that individuals do not owe others affirmative duties, including: (1) special relationships;¹⁰³ (2) contractual agreements to act in ways that differ from common law presumptions;¹⁰⁴ (3) particular undertakings by either party;¹⁰⁵ and (4) particular skills, training, and ability.¹⁰⁶ Analysis of each of these factors, and the extent to which they give rise to tort law duties to warn, is context specific.

1. The Fiduciary Duty of Physicians to Return Findings

The medical profession with the most clearly articulated and legally recognized duties is that of physicians providing medical care to patients.¹⁰⁷ The heightened duty owed by physicians is a duty to comply with medical customs or standards.¹⁰⁸ In other words, physicians have an obligation to act using the same skill, knowledge, and care possessed and exercised by comparable physicians.¹⁰⁹ Physicians providing medical care owe a fiduciary duty to their patients, which entails a duty to act in the patient’s best interest.¹¹⁰ As articulated by Professors Thomas L. Hafemeister and Selina Spinos:

Because physicians have superior medical knowledge and skill and are the gatekeepers to medical services, patients are dependent on them. Patients lack the knowledge or skill to assess their own health conditions. Instead, they must depend on their physicians to provide critical information about their medical well-being. … Because patients are so vulnerable and dependent on their physicians, the law imposes a “trust” on doctors—a fiduciary responsibility stemming from the dependence and vulnerability of the patient, and from the disparity between a patient’s and a physician’s knowledge and ability to act.¹¹¹

A fiduciary duty to act in the patient’s best interest means that physicians must disclose unexpected findings that indicate potential future harm, even if these findings fall outside of the physician’s particular area of expertise.¹¹² Failure to properly disclose unexpected findings can lead to tort liability and damage awards to patients.¹¹³

Were a researcher to be treated like a physician, a researcher—like a physician—would have a duty to return IFs. But researchers qua researchers are not generally thought to owe fiduciary duties, despite often having similar knowledge, skills, and abilities as physicians. In critically important ways, a researcher is not given the flexibility to act in a participant’s best interest¹¹⁴ and thus cannot be a fiduciary.¹¹⁵ Researchers, therefore, owe participants something less than a fiduciary duty. Courts assessing claims that researchers failed to appropriately return IFs should not find a treating physician’s duty to disclose dispositive unless a researcher was actually or effectively acting as a treating physician in addition to acting as a researcher.

2. Third-Party Physicians’ Duty to Return Findings

The area of established case law that more closely resembles the researcher-participant relationship is the body of case law dealing with third-party physicians. Third-party physicians are physicians hired by employers or insurance companies to conduct physical examinations to determine whether an individual has any disqualifying or costly health problems.¹¹⁶ Because, like researchers, third-party physicians do not provide clinical care to patients, they are similarly thought not to owe fiduciary duties to examinees.

Nevertheless, third-party physicians have the knowledge, skills, and ability to make medical assessments; are privy to private health information about the examinee; are uniquely positioned to discover preventable harms; and may develop a special relationship with the examinee.¹¹⁷ A number of courts have therefore held that third-party physicians have a duty to return unexpected findings to the individuals being examined, even absent a fiduciary relationship.¹¹⁸

For example, in Stanley v. McCarver, a third-party physician evaluated a chest x-ray of Ms. Stanley as part of a pre-employment tuberculosis screening.¹¹⁹ Dr. McCarver recognized that the x-ray showed particular abnormalities, yet these abnormalities were not disclosed to Ms. Stanley by her prospective employer, in violation of their company policy.¹²⁰ Approximately ten months later, Ms. Stanley was diagnosed with lung cancer and brought suit against Dr. McCarver, his employer, and her prospective employer alleging a negligent failure to “timely and adequately diagnose and/or communicate to [her] the abnormality evident on her chest x-ray.”¹²¹ The court noted that “Dr. McCarver did agree, for consideration, to interpret Ms. Stanley’s confidential medical record” and that “[b]y doing so, he undertook a professional obligation with respect to Ms. Stanley’s physical well being.”¹²² The Court concluded that “Dr. McCarver placed himself in a unique position to prevent future harm to Ms. Stanley” and that Ms. Stanley could have “reasonably expect[ed] the physician to sound the alarm if any serious abnormality [was] discovered.”¹²³

In the case of Reed v. Bojarski, the New Jersey Supreme Court also decided that third-party physicians can owe duties to their examinees even outside a traditional physician–patient relationship.¹²⁴ As part of a pre-employment physical conducted by Dr. Bojarski, Arnold Reed underwent a chest x-ray that revealed a widened mediastinum, known to be a potential indicator of Hodgkin’s Lymphoma.¹²⁵ In violation of the terms of the contract, Dr. Bojarski never notified Reed’s potential employer about his potentially dangerous condition, and Reed’s potential employer responded to Reed that he was in good health.¹²⁶ About six months later, Reed returned for another examination, having already lost twenty-five pounds and suffering from flu-like symptoms.¹²⁷ A month following that, Reed was admitted to the hospital and diagnosed with Stage IIB Hodgkin’s disease.¹²⁸ Reed died eight months after his initial examination at the age of twenty-eight.¹²⁹ The New Jersey Supreme Court concluded that Dr. Bojarski owed Reed a duty of reasonable care and that in light of the evaluative purpose of the exam, Reed “had an absolute right to expect that he would be told if something was wrong.”¹³⁰

Like the third-party physicians in Stanley and Reed, researchers will also have access to a patient’s confidential medical record and may also be in a unique position to prevent harm. Although a researcher does not accept monetary consideration from research participants, a court could conclude that researchers have a similar duty to “sound the alarm” if any serious abnormality is discovered.

3. A Researcher’s Duty to Return Findings?

As a general rule, researchers are not fiduciaries of participants, so they do not owe fiduciary duties to act in a participant’s best interest.¹³¹ Courts that have considered the duties that researchers owe to participants have generally found that researchers have a legal duty to comply with the federal regulations governing research.¹³² The regulations are silent about how researchers must handle IFs.¹³³

The question then is what level of responsiveness to an IF would satisfy a researcher’s duty of care. Is a researcher’s duty discharged merely by returning results that the researcher stumbles upon? Or is a researcher obligated to look for IFs?¹³⁴ Is a researcher’s duty time limited? That is, must a researcher report only those IFs that are known or validated by today’s science? Or is a researcher obligated to continually reexamine the whole-genome-sequence data in light of new, emerging science—an approach that would undoubtedly be burdensome? Existing legal doctrine provides no clear answers.

Ultimately, a researcher’s duty to return IFs will turn on whether researchers are treated more like physicians, who have a duty to comply with medical customs and standards,¹³⁵ or more like third-party physicians, who have a duty of reasonable care under the circumstances taking into account a number of factors.

If a researcher owes a duty comparable to that owed by a physician—a duty to act in accordance with the customs and standards of comparable researchers—a court would be unlikely to conclude that researchers owe a duty of care to return IFs.¹³⁶ At present, there is too much inconsistency among researchers concerning the return of IFs for any custom to develop.¹³⁷ Similarly, although courts can use guidance documents as evidence of a duty owed,¹³⁸ the current guidance as to which IFs should be returned in which circumstances is too varied and has yet to reach a level of consensus. Accordingly, it is too soon for these guidance documents to give rise to a researcher’s duty.¹³⁹ Once a consensus does emerge, courts in some states would look to these guidance documents as evidence of a standard of care.¹⁴⁰

To the extent that a researcher’s duty of care is determined like that of a third-party physician, a court would look to a number of factors to determine whether a researcher acted reasonably under the circumstances. For example, a court may consider: (1) whether the researcher was in a unique position to prevent harm, (2) the burden of preventing harm, (3) whether the plaintiff relied upon the researcher’s diagnosis or interpretation, (4) the closeness of the connection between the defendant’s conduct and the injury suffered, (5) the degree of certainty that the plaintiff has suffered or will suffer harm, (6) the skill or special reputation of the actors, (7) the scope of any agreement, and (8) public policy.¹⁴¹

Our Example 1 (Anesthesia Death) presents a strong case for a researcher having a duty to return the particular IF. In Example 1, Dr. Ames was one of the few people who knew or could have known about Arnold’s vulnerability to a particular anesthesia and was therefore in a unique position to prevent harm; the burden of preventing the harm through notification was likely to be low given that the Dr. Ames and Arnold had already met, and there was a high degree of certainty that Arnold would suffer harm. Without more information, however, we cannot determine the extent to which Arnold relied on Dr. Ames’s offer to return IFs.

Example 2 (De-Identified Data) or Example 3 (Large-Scale Research) present much weaker cases. In neither of the cases was the researcher in a position to prevent harm. In Example 2, this is because Betsy’s data was shared without identifying information, and in Example 3, this is because Alzheimer’s disease is currently unpreventable and untreatable. In both examples, the burden of preventing harm through notification could be quite high given that these examples both involve research with large numbers of participants.

Nevertheless, it is not entirely clear how a court would rule given the context-specific nature of this analysis. Although this context-specific approach is a fundamental aspect of tort law, it provides very limited guidance to researchers about when they can expect tort law duties to attach. And a fact- and context-specific approach is particularly problematic in cases in which the ethical and scientific analysis is evolving and uncertain. Though it is not clear that a court today would find a researcher’s duty of reasonable care to include a duty to notify participants about IFs, a continuing push for an ethical obligation to return IFs, guidance documents that advocate returning IFs in particular circumstances,¹⁴² and movement towards returning more IFs in practice could foreseeably give rise to a legal duty to return IFs in the future.¹⁴³

B. WAS THE DUTY BREACHED?

Assuming courts find that researchers owe participants a duty to return some IFs, assessing whether the duty has been breached is no simple matter and may, in fact, raise more questions than answers.¹⁴⁴ Merely concluding that individuals have a duty to return particular findings does not, however, mean that the duty has been breached.¹⁴⁵ The issue of whether a duty was breached is in most cases a question for a jury.¹⁴⁶

The Court of Appeals for the Tenth Circuit has addressed the issue of when a third-party physician’s failure to notify examinees should be considered a breach of a duty owed. In Pehle v. Farm Bureau Life Insurance Co., a husband and wife applied to Farm Bureau for life insurance.¹⁴⁷ Blood tests conducted in 1999 revealed that both husband and wife were HIV positive.¹⁴⁸ The laboratory testing company notified both Farm Bureau and the state department of health, as required by law.¹⁴⁹ Farm Bureau notified the couple that their application for insurance was being denied based on the results of their blood test and offered to send the blood-test results to their physician upon written authorization.¹⁵⁰ The couple made no such request until the wife developed symptoms of AIDS in 2001.¹⁵¹

The Court of Appeals for the Tenth Circuit held that “the trust and confidence that the [couple] put in Farm Bureau was sufficient to create a relationship giving rise to a limited duty”¹⁵² and that “if an insurance company, through independent investigation by it or a third party for purposes of determining policy eligibility, discovers that an applicant is infected with HIV, the company has a duty to disclose to the applicant information sufficient to cause a reasonable applicant to inquire further.”¹⁵³ Despite this heightened duty, the Tenth Circuit affirmed summary judgment in favor of the laboratory, while concluding that whether Farm Bureau satisfied its duty was a genuine issue of material fact.¹⁵⁴

In the context of researchers who conduct whole-genome sequencing, adopting a standard that researchers have a duty to disclose to the participant “information sufficient to cause a reasonable [participant] to inquire further” would not settle the issue.¹⁵⁵ Unlike researchers, third-party physicians have a duty to thoroughly investigate the examinee and are tasked with discovering health problems that are well-characterized and understood. Researchers, by contrast, can potentially find thousands of IFs of which the meaning and import could be unknown. Even those IFs that are well-characterized and understood often only indicate a probability of developing disease and are much less definitive than a clear diagnosis of HIV.

C. DID THE BREACHED DUTY CAUSE INJURY?

A successful tort lawsuit generally requires that the defendant’s conduct be both the “but for” and proximate cause of the plaintiff’s harm.¹⁵⁶ Under a conventional causation analysis, participants alleging a researcher’s failure to appropriately return IFs will have difficulty showing that the researcher’s failure to return IFs was the direct cause of their injury; the but-for cause will always be the disease resulting from the underlying variant.¹⁵⁷ Courts have nevertheless recognized the legal doctrine “loss of a chance,” which acknowledges that the possibility of preventing a medical misfortune is itself a valuable thing¹⁵⁸ and that the diminution of a chance to prevent or minimize future harm should be compensated if it results from a medical professional’s failure.¹⁵⁹

Loss-of-a-chance doctrine speaks to both the causation and injury elements of a tort law claim. Although causation and injury are generally considered separately,¹⁶⁰ loss-of-a-chance redefines the interest that was injured (the chance of recovering or preventing injury rather than the ultimate outcome)¹⁶¹ and then considers whether the medical professional’s action was the but-for cause of the lost chance.¹⁶²

Although loss-of-a-chance is a relatively recent doctrine, it has started to make inroads in state courts and has been adopted in nearly half of all states.¹⁶³ The loss-of-a-chance doctrine is viewed as being more in accord with the ideals of tort law generally. Without loss of a chance, any situation in which an individual had less than a fifty percent chance of recovering would be wholly immunized from tort liability.¹⁶⁴

Thus far, loss-of-a-chance has been applied primarily in the context of medical malpractice,¹⁶⁵ but the doctrine could be applied more broadly.¹⁶⁶ One reason that the loss-of-a-chance doctrine has been used in medical malpractice is that “reliable expert evidence establishing loss of chance is more likely to be available in a medical malpractice case than in some other domains of tort law.”¹⁶⁷ Second, “it is not uncommon for patients to have a less than even chance of survival or of achieving a better outcome when they present themselves for diagnosis, so the shortcomings of the all or nothing rule are particularly widespread.”¹⁶⁸ Third, the “failure to recognize loss of chance in medical malpractice actions forces the party who is the least capable of preventing the harm to bear the consequences of the more capable party’s negligence.”¹⁶⁹

Some of the factors that justify application of loss-of-a-chance to medical malpractice also support applying the doctrine to research. First, the same reliable expert evidence about the chances of succumbing to, for example, undisclosed breast cancer in a medical malpractice case could also be provided in the research context. Second, the same diseases for which individuals may have a less-than-even chance of survival could occur in both the medical and research contexts. Third, as with medical malpractice, the failure to recognize loss-of-a-chance in the research context forces the party who may be the least capable of preventing the harm to bear its consequences.

Some reasons used to justify loss-of-a-chance in the context of medical malpractice are not similarly applicable in research. Courts have looked to the specific physician–patient relationship—particularly the expectation that “the physician will take every reasonable measure to obtain an optimal outcome for the patient”—as warranting these additional protections.¹⁷⁰ Such expectations are generally not justifiable in research, and courts should be cautious before imposing additional burdens on researchers that were specifically disclaimed during the informed consent process. Moreover, courts have justified loss-of-a-chance in the context of medical malpractice because “reasonably good empirical evidence is often available about the general statistical probability of the lost opportunity.”¹⁷¹ In research, where so much is unknown, this simply may not be true. Certain IFs, such as those unlikely to be uncovered by other means, would be more likely to lead to recovery under loss-of-a-chance principles than would others. Example 1, introduced at the outset of this section, presents just such a case. Other IFs, such as the possibility of developing breast cancer or Alzheimer’s disease in the future (Examples 2 and 3), may be less likely to satisfy this criterion.¹⁷²

A research participant must also show that the researcher’s conduct was the proximate cause of the participant’s harm—“that the harm was the general kind that was unreasonably risked by the defendant, the kind of harm the defendant should have been more careful to avoid.”¹⁷³ Whether harm is foreseeable turns “on what facts the defendant actually knew or those he should have known.”¹⁷⁴ Regardless of whether the other elements, including but-for causation, have been satisfied, one could argue that in each of the examples set forth at the outset, the harm that materialized was precisely the harm predicted by the IF: Arnold’s reaction to the particular anesthesia, Betsy developing breast cancer, and Cal developing Alzheimer’s disease were all probabilistically foretold by the particular variants discovered. But these cases may be among the easiest examples. In other cases, the proximate cause analysis is far less straightforward.

In many, if not most, cases, the IF will be of unknown significance so that any harm that occurs would not be foreseeable. In other cases, the risk could foreseeably be discovered through routine medical tests, and thus the harm would not be foreseeably likely to occur. In all cases, an IF arising from genomic research only gives rise to a probabilistic chance of harm occurring that may or may not be sufficiently proximate.

For the reasons discussed above, even a research participant able to prove negligence will have a difficult time proving causation. An injured participant could only succeed in a state that has adopted the loss-of-a-chance doctrine, and in a court that is willing to extend the doctrine from medical malpractice to research. Even then, an injured research participant could recover only if the researcher’s failure to disclose was the but-for and proximate cause of the participant’s injury—a claim that, in most circumstances, will be quite difficult to prove.

D. IS THE PARTICIPANT’S INJURY COMPENSABLE?

A final requirement for a successful tort lawsuit is that a plaintiff suffer actual, compensable harm.¹⁷⁵ In the context of loss-of-a-chance, the plaintiff is entitled to recover only the value of the chance that was destroyed.¹⁷⁶ An injured participant’s damages are determined “by subtracting the decedent’s postnegligence chance of survival from the prenegligence chance of survival.”¹⁷⁷ Thus, if an injured research participant had only a forty percent chance of surviving a potentially disclosed IF and the researcher’s negligence deprived the participant of that chance entirely, the researcher would be liable for forty percent of the total damages suffered.¹⁷⁸

In Example 3, Cal’s recovery for loss-of-a-chance would approach zero percent because given today’s treatment options, Cal would have had no chance of preventing his Alzheimer’s disease regardless of what Dr. Cruz disclosed. In Example 1, Arnold’s recovery would be close to one hundred percent because Dr. Ames’s failure to notify Arnold wholly removed his chance to use a different anesthesia during surgery. Betsy’s recovery in Example 2 would fall somewhere between these two extremes because she would have had a better chance had she obtained more frequent mammograms, but she still would have had a risk of contracting, and succumbing to, breast cancer.

The foregoing analysis suggests that a researcher will have a duty to act with reasonable care under the circumstances, although it is unclear, at present, the parameters of the duty given the highly contextual analysis. A continued push for an ethical obligation, guidance documents that set forth the IFs that should be returned in particular circumstances, and an emerging research custom of returning IFs could lead a court to conclude that researchers have a legal duty to return IFs. Whether any such duty has been breached, however, raises questions about which IFs must be returned, how hard a researcher must look for IFs, how researchers should handle probabilistic information, and whether a researcher’s duty is time limited or requires continued reexamination. Once negligence is established—a difficult proposition in and of itself—to succeed with a legal claim, participants would have to rely on the loss-of-a-chance doctrine to show that researchers were the but-for cause of their inability to avoid succumbing to a subsequent adverse medical outcome and that the researcher’s failure to disclose was the proximate cause of the participant’s injury—claims that would be difficult to prove. In the unlikely circumstance that a participant satisfies these criteria, a participant would then have to demonstrate a compensable injury, the value of which would be reduced proportionately under the loss-of-a-chance calculations.

Research participants would also have to show that had they known about the particular information, they would have acted differently in courts adopting a subjective standard, or that a reasonable person would have done so if the objective standard is used.¹⁷⁹ Arnold would have to show that, had he or a reasonable person known about his potential reaction to anesthesia, he or a reasonable person would have used a different one during his surgery. Betsy would have to show that she or a reasonable person would have gotten more frequent mammograms. Cal would be unable to show any actions that would have prevented his Alzheimer’s.

Although this analysis describes the complexities and uncertainties inherent in describing whether tort liability will attach, it does not answer the broader question of whether tort liability should attach. The broader goals of tort law include compensating those who are injured;¹⁸⁰ deterring undesirable behaviors;¹⁸¹ notions of corrective justice, or righting wrongs;¹⁸² distributive justice, or addressing equitable distribution of goods in society;¹⁸³ and policy and utility, or assessing what is good for society as a whole.¹⁸⁴ Tort law is also fundamentally concerned with fault.¹⁸⁵

In the context of a researcher’s liability for failure to return IFs, these considerations lead to different conclusions. Considerations of policy and utility counsel against tort liability because potential liability could hamper socially beneficial research. Notions of distributive justice similarly counsel against liability because it is not clear that particular research participants have a stronger claim to funds that would otherwise be used for socially beneficial research. Considerations of compensating those injured and corrective justice weakly support liability to ensure that those who are injured are adequately protected. Ultimately, however, the analysis comes down to fault and the idea that researchers should not be held accountable or found to be at fault for a duty so broadly conceived. As set forth in Part III, a duty more narrowly conceived could more reasonably give rise to fault-based liability.

III. Potential Standards for Disclosing Incidental Findings

Under current law, the questions of when and whether a researcher can be held liable for failure to return IFs are still uncertain. As a result of this confusion, researchers are left with little guidance about when liability for failure to return IFs will attach, participants lack meaningful guidance about which IFs they can reasonably expect to have returned and when, and courts are left to make complicated determinations about researchers’ legal duties in light of uncertain and evolving scientific knowledge and customary practice, potentially giving rise to an inconsistent body of case law. A clearer legal standard would address each of these concerns.

One challenge in adopting an appropriate standard is that many of the ethical, scientific, and legal issues regarding a researcher’s duty to return IFs are still in flux.¹⁸⁶ Waiting until these areas settle, however, is an untenable solution. Courts will be asked to weigh in on this issue and, absent clearer legal standards, could decide cases inconsistently—an approach that could bring confusion and gridlock to an emerging scientific arena.¹⁸⁷

To avert confusion and gridlock, this Article proposes adopting a clearer standard that provides ethically sound, consistent guidance about when IFs ought to be returned and when the failure to return IFs can and should give rise to legal liability.¹⁸⁸ This Part argues that biomedical researchers should adopt one of several possible standards for disclosing which IFs will be offered that are appropriately transparent. We then advocate that courts recognize the proposed standard and grant researchers that select one of these options the benefit of accepting the terms of that selection, and then in Part IV, explore some of the nuanced interpretations courts will have to make when adopting our proposal.

A. DISCLOSING THE PROCESS OF OFFERING INCIDENTAL FINDINGS IN THE INFORMED CONSENT DOCUMENT

A clear and appropriate rule would require that researchers disclose in the informed consent document the extent to which IFs will be returned to participants, with a corresponding legal duty to act in a manner consistent with the terms set forth in the informed consent document.¹⁸⁹ The approach to returning IFs must be understandable to research participants and must set forth clear guidelines that allow participants to ascertain which IFs will be returned to them and the extent to which a researcher’s silence can be interpreted to indicate good health. Negligence could be found when researchers act in ways that are inconsistent with the policy set forth in the informed consent document.¹⁹⁰

This approach recognizes the value of both informing research participants about what to expect throughout the informed consent process and meeting those expectations.¹⁹¹ This approach is also consistent with established case law holding that clinicians¹⁹² and third-party physicians¹⁹³ can establish the parameters of which results will be returned and can be held liable in tort for failing to comply with these terms—terms upon which a recipient could reasonably be expected to rely. This approach also recognizes, as many scholars have, that there are many ethically permissible approaches to returning IFs.¹⁹⁴

The next three sections outline three possible approaches that can be articulated clearly enough to satisfy our goal of adequately informing research participants. These approaches are: no return of IFs, returning all genomic data without interpretation, and returning results consistent with a compendium. Although they each have strengths and weaknesses, and they might not be universally embraced as ethically optimal, we believe that they are at least prima facie ethically defensible and provide the clearest articulation of the IFs that will be offered to research participants. After outlining these three options, we then articulate a framework that explores why researchers should generally be on safer legal footing if they select from one of these options and explain why other common, possible return paradigms are inappropriate.

1. No Return of Incidental Findings

The first proposed standard is that researchers not return IFs to participants regardless of the seriousness of their findings. This approach acknowledges the difficulties inherent in tracking down and locating large numbers of participants and recognizes that a large amount of resources—both time and money—may be needed to analyze, return, and provide counseling for the thousands of variants that could arise in any given research participant’s whole-genome-sequence data.¹⁹⁵ Under this approach, no liability would attach for failing to return IFs, but liability could attach once a researcher chooses to return IFs but does so improperly.

An advantage of this approach is that it appropriately recognizes the distinctions between research and clinical care and does not contribute to the therapeutic misconception.¹⁹⁶ This approach also respects an arguably proper allocation of research resources. Research is funded to create generalizable knowledge that advances the public good.¹⁹⁷ Returning IFs can be costly, diverting funds from research that benefits the general public to the process of returning IFs, which generally benefits particular individuals.¹⁹⁸

This approach does, however, have a number of disadvantages. First, given the emerging ethical guidance encouraging the return of certain results in certain circumstances, and given that some researchers have already started returning IFs, a court may be unwilling to conclude that researchers have no duty to return IFs, particularly in a case with a serious IF, a substantial researcher–participant relationship, and a sympathetic plaintiff.¹⁹⁹ Second, the reason that some researchers started returning IFs is because they felt morally bound to return IFs with significant health implications. Creating a standard whereby researchers are not expected to return IFs, and in which researchers may be penalized for doing so, may lead researchers to compromise their moral conviction. Third, this approach does not address research participants’ reasonable expectations that someone with particularized knowledge and skills who is granted privileged access to otherwise private health information would inform them of any significant health findings.

For these reasons, scholars have argued that not returning IFs regardless of their severity is unethical,²⁰⁰ and very few have advocated for this approach.²⁰¹ Given that some researchers feel a moral obligation to return certain IFs—and that many researchers have begun returning IFs, thereby fostering a reasonable expectation among participants that particular IFs will be returned—setting a threshold legal standard at no return is not an appropriate bright-line rule in all cases, though it may be appropriate in certain circumstances (for example, secondary research using stored specimens, or research using de-identified samples).²⁰²

2. Return All Genomic Data Without Interpretation

An approach that has been suggested but rarely seriously considered is an obligation to return all genetic sequence data—the nucleotide base pairs of As, Ts, Cs, and Gs—without interpretation or annotation.²⁰³ This approach is generally only considered before settling on a more “reasonable” position somewhere between the extremes of not returning anything and returning everything.²⁰⁴ Nevertheless, a bright-line rule suggesting the return of all whole-genome-sequence data without interpretation does have some advantages.

First, this approach allows researchers to fulfill their moral convictions about returning information to participants while minimizing the burden of doing so. Returning all genomic information without interpretation allows researchers to bypass the requirements of additional confirmatory testing in a CLIA-approved laboratory, additional searching, interpreting, seeking out genetic counselors, and making decisions about which results to return on the basis of shifting criteria of clinical utility or actionability. And as more websites and services allow individuals to input their whole-genome-sequence data to obtain reasonably priced interpretations, participants can take control over when and whether they learn more about their genetic information.²⁰⁵ This approach is responsive to data that suggest that participants prefer to receive significantly more information than has generally been returned, even if the information is not clinically actionable.²⁰⁶ This approach also does away with a potential obligation to reexamine and re-notify participants with updated analyses because participants have access to their own genomic sequence data and can seek out their own reexaminations in light of evolving science.

This approach, like the others before it, also has some downsides. First, this approach may not satisfy researchers’ moral conviction about alerting research participants to IFs that may have significant health consequences. The actual genetic information returned—the As, Ts, Cs, and Gs—does not itself alert participants to any actionable anomalies. Second, handing genetic information to research participants without any interpretation may overwhelm those who do not know where to turn for interpretation or who may get their genomic data interpreted without appropriate counseling or guidance about what the information means.²⁰⁷ Finally, this approach may be technologically and logistically difficult and expensive, particularly with regard to large genomic studies that involve thousands of participants.

3. Return Results Consistent with a Compendium

A third proposed standard would require that researchers return IFs that are listed in a compendium of consensus returnable IFs, as determined by a body of expert medical geneticists. Legal liability would attach if researchers fail to return such results. There are currently several ongoing efforts to determine which results rise to the level of a returnable genomic findings.²⁰⁸ These groups would be expected to reexamine and modify these compendia over time, consistent with evolving science.²⁰⁹ Though the exact number of variants will depend on the criteria used, one notably (and arguably conservative) early effort resulted in a list containing fifty-six possible variants.²¹⁰

One advantage of this approach is that results will be returned consistently across all types of research and in accordance with the most rigorous determination of what ought to be returned.²¹¹ Although this approach establishes a limited obligation on the part of the researchers to hunt for findings beyond those that are stumbled upon, the obligation is narrow and confined to those specific variants that are well-characterized and understood.²¹² This approach minimizes the more costly elements of returning IFs. Because there would be a known universe of IFs that must be returned, deciding which results to return and how to return them could become standardized and less costly.²¹³ This approach also likely satisfies a researcher’s moral convictions about returning IFs that may have important health implications.

This approach also has some disadvantages. First, as mentioned above, there is not yet one standardized compendium, but several emerging ones.²¹⁴ And although there is some common ground among geneticists about the IFs that merit routine return, there is not absolute agreement, which could result in differences among the compendia.²¹⁵ Second, as mentioned above, these compendia evolve over time, raising questions about when a researcher’s duty to examine a participant’s data attaches and whether researchers have an ongoing obligation to reexamine participant data in light of changes to the compendium.²¹⁶ Third, some compendia may be geared to the IFs that should be returned in the clinic; it is unclear whether the same guidelines should apply in research.²¹⁷ Finally, an argument could be made that even this level of return contributes to the therapeutic misconception and causes research participants to rely on researchers and the research process in a way that was not intended.²¹⁸

B. A PROPOSED FRAMEWORK FOR RETURNING INCIDENTAL FINDINGS

As applied, researchers would be able to choose their approach to returning IFs from the options set forth above—no return of IFs, offering to return consistent with a specified compendium, or offering full return of data without interpretation or annotation. Researchers who select and comply with one of these approaches should be found to have satisfied their legal duty under tort law. Allowing researchers to choose a justifiable approach to returning IFs from among these options and subjecting that chosen approach to Institutional Review Board (IRB)²¹⁹ review permits researchers to choose an option that is consistent with their moral convictions about returning IFs, sets reasonable expectations among research participants about the results that they should expect to receive, and is consistent with the oft-made argument that the appropriateness of returning IFs turns on a variety of factors.²²⁰ Although this approach potentially protects researchers who choose not to return any IFs from liability, it only does so in cases for which a researcher’s determination not to return IFs has been reviewed and approved by an IRB. An IRB presumably would only accept a researcher’s proposal to not return any IFs when accompanied by well-articulated and compelling scientific or practical justifications.

Researchers who take a different approach—including a more subjective promise to return all findings that, for example, have “significant health implications” or that are “clinically actionable” or a promise to comply with participant preferences, regardless of the clarity with which participants express those preferences—are setting themselves up to comply with a heightened legal duty. Because these researchers will have begun an undertaking that they now have a duty to satisfy responsibly,²²¹ and because these researchers have made a promise that can induce reliance,²²² these researchers have taken upon themselves a duty to comply with the heightened fiduciary duty generally applied to physicians.²²³

Similarly, some researchers might want to return IFs consistent with a complex solicitation of participant preferences.²²⁴ To the extent that return consistent with participant preferences hinges on terms that are difficult to define, like “actionability” and “clinical significance,” this approach would not bring clarity to participants about the IFs that they can reasonably expect to have returned and should not be one of the options for researchers. Although this approach may lead to an ethically sound policy for returning results, it would remove researchers from the safety of a clearly stated and articulable approach to returning IFs.

We acknowledge that our proposal could be controversial in that it would allow researchers to escape legal liability in some cases where people will have a strong moral intuition that they should have returned patently life-saving medical information. For instance, in Example 1, in a circumstance where an approach of not returning IFs was approved by an IRB, Dr. Ames would be insulated from suit for failing to disclose his participant’s risk of being exposed to certain anesthetic agents. As is often noted, however, law and ethics can frequently diverge, and this seems to be a case in point. We believe that the value of a clearly articulated legal rule in this complicated and important research field outweighs the moral distress that some people will feel regarding very difficult marginal cases.

IV. Guidance for Courts

Courts that are faced with legal challenges brought by research participants who allege inadequate or improper disclosure of IFs will almost certainly find themselves faced with a variety of confounding fact patterns. This Part provides guidance to courts to help navigate this difficult terrain. We address (1) the implications of inadequate or silent consent-form language, (2) factors relevant to determine the scope of a researcher’s duty, and (3) how to assess whether adequate disclosure has been made.

A. INTERPRETING CONSENT-FORM LANGUAGE

In Part III, we proposed a paradigm in which researchers could choose from a menu of ethically justifiable options (subject to IRB review) and would have to clearly articulate their plan for managing IFs during the informed consent process. This does not, however, account for cases in which the informed consent is silent about—or inadequately addresses—the issue of returning IFs. The question for courts, then, is how to assess informed consent documents that do not explicitly or sufficiently address IFs.

As set forth above, because genomic sequencing generates so much data, researchers are expected to explicitly disclose a number of pieces of information, including: (1) that the subject’s entire genome will be sequenced; (2) that there is a high likelihood that IFs will arise; and (3) the plan for managing any IFs that do arise.²²⁵ To the extent researchers set forth an IRB-approved approach to returning IFs, that should be the standard by which researchers are judged for purposes of assessing legal liability. For informed consent that is silent, or that inadequately addresses the return of IFs, the general assumption should not be that researchers are therefore protected from legal liability. Rather, courts should look carefully and skeptically at inadequate consent documents.

There are, however, two circumstances in which a silent or inadequate informed consent document might be justified. The first justification for inadequate consent language relates to the evolving nature and adoption of new sequencing technologies. As genomic-sequencing technologies became available, they were first adopted by a handful of researchers with significant genetic-research expertise.²²⁶ During this period of early adoption when there was no clear ethical guidance, researchers and IRBs were understandably unsure about how to address the IFs arising from genomic sequencing during the consent process. As a result, informed consent disclosure about the possibility of IFs was varied, less evolved, and perhaps—if judged by today’s standards—inadequate. Courts should therefore be sympathetic to a researcher with an inadequate consent document if the relevant document originated when there was no established practice regarding the informed consent disclosure of IFs. Similarly, a researcher might have a justifiably insufficient informed consent form if she can demonstrate that the specific language was consistent with the IRB’s practices or institutional policies during this early adoption period.²²⁷

This early stage has given way, however, to much more widespread adoption as the price of sequencing has dropped and a wider range of researchers have become intrigued by the power of whole-genome sequencing. The increased uptake has been accompanied by a proliferation of debate and discussion in the research ethics literature,²²⁸ such that no researcher today could reasonably claim ignorance about the need for clear language about IFs in the informed consent document. Reasonable researchers should be expected to have a well-articulated plan as soon as they should have been made aware of the contours of this ethical controversy.

A second justification for inadequate informed consent language is that for certain kinds of research, the relationship between the researcher and the individual participant might be sufficiently distant so as to obviate the need for language about IFs. Under this line of reasoning, certain kinds of research would effectively exempt researchers from a duty of having to offer IFs, and from a corresponding duty to include informed consent language about IFs. This category might include researchers whose valid methodological choices make it extraordinarily difficult (or even impossible) to recontact participants (such as biobanks or studies involving de-identified samples) or where the inherent characteristics of the research mean that offering to return IFs would be prohibitively resource intensive (such as large-scale public-health studies).

If an informed consent document is silent or inadequate and neither of the justifications have been met, the scope of a researcher’s legal duty will be defined not by the silent or inadequate informed consent language but instead will turn on a number of factors set forth below. Allowing researchers to establish the parameters of potential legal liability through appropriate informed consent disclosure is part of recognizing the importance of adequate disclosure. A researcher’s failure to adequately set forth an IF disclosure policy results in legal liability that hinges on the multifactored analysis set forth below, thereby introducing an element of uncertainty to a researcher’s defense against liability.

B. FACTORS IMPORTANT FOR DETERMINING THE SCOPE OF A RESEARCHER’S LEGAL DUTY TO RETURN INCIDENTAL FINDINGS

Under our proposed framework, if plaintiffs can demonstrate that the informed consent language was inadequate and there is no justification for that inadequacy, the analysis then shifts to the difficult task of determining the extent of a researcher’s legal duty concerning the return of IFs. As discussed in Part II, because of differences in clinical care and research, we disagree with the notion that courts should use the stricter fiduciary standard typically applied to physicians. Rather, we suggest that courts should use a more general, multifactored test often associated with third-party physicians. Under this test, researchers have a reasonable duty of care, taking into account a range of contextual factors and circumstances, including the factors described in section II.A.3.

Not all of the factors to be considered are equal, however, and some might carry more weight than others. One critical factor is whether and to what extent a research participant reasonably relied on a researcher to return IFs. This factor could be satisfied by showing that the researcher (by word or action) explicitly took on a duty of disclosure on which it was reasonable for the participant to rely. Evidence demonstrating that a participant reasonably relied on the researcher to disclose important IFs would strongly weigh in favor of a researcher’s legal duty to return IFs. Absent evidence of an explicit promise, such an argument would, in large part, be a function of the depth of relationship between a researcher and participant and the extent to which the relationship resembled a typical doctor–patient relationship.

Courts might look to a number of other factors to evaluate a claim that the researcher–participant relationship gave rise to justifiable reliance. One potentially important factor is the nature or design of the research. Late-stage clinical trials would be most likely to give rise to justifiable reliance, given that participants generally receive clinical interventions with a prospect of direct individual benefit. Studies in which there is a long-term relationship between the researcher and participant (such as natural-history protocols) might similarly be convincing. This would be particularly true when that relationship has an explicitly clinical component, like if the study team provides standard-of-care treatment as a component of the research. Studies on samples from biobanks, particularly when there is minimal or no contact between the research team and the participant, are less likely to give rise to circumstances that justify reliance.

Other study characteristics might also provide evidence of justifiable reliance. The existence of study resources (such as genetic counselors) for returning certain genetic information might support a plaintiff’s claim. Conversely, the complete absence of any infrastructure for returning results would suggest the opposite. Researcher expertise in genomic science (or lack thereof) would follow a similar line of reasoning: clinicians—and medical geneticists in particular—would be more likely to lead participants to justifiable reliance; basic scientists and non-clinicians are presumably much less likely to do so.

C. ASSESSING WHETHER ADEQUATE DISCLOSURE HAS BEEN MADE

Even if a court concludes that researchers have a duty to return IFs, questions will remain about whether and when researchers have adequately satisfied their disclosure obligations. This question will be complicated by the fact that, as sequencing technology improves and is increasingly incorporated into research protocols, a growing number of professional societies and governmental agencies will weigh in with their thoughts about the contours of researchers’ obligations to return IFs. To the extent that these guidance documents point towards returning IFs, they have the potential to give rise to a corresponding legal duty to return IFs in genomic research, thus increasing the chances that participants will successfully be able to win liability suits. This analysis will likely be far from straightforward; much of this guidance is divergent and evolving, and many of these documents make self-limiting disclaimers about how they should be used by the courts.

As with all questions of fact about medical practice, courts will have to weigh a variety of kinds of evidence. In this case, professional association and government guidance documents are just one part of the picture. It is arguably even more important to determine the extent to which certain categories of researchers (such as clinical researchers) as a group have begun implementing this guidance to determine whether any standard of care has developed. The challenge will be to determine how to draw appropriate boundaries around each relevant professional group. Courts must consider whether they should establish a prevailing standard of care by looking to the entire population of researchers, to the set of clinical researchers, or to some more limited subset. We endorse the view that courts should draw the line broadly as an initial starting point but suggest they carefully consider contextual factors that suggest excluding certain groups for whom imposing a duty would be unduly burdensome—such as biobanks or basic science researchers working with de-identified samples.

Another question remains: to what lengths will researchers reasonably be required to go to inform participants of IFs? One circumstance in which this question could arise is when researchers no longer have a participant’s updated contact information. In most cases, researchers ask participants to provide contact information during initial intake, but participants retain the responsibility of updating their information over time. If a participant fails to do so and later sues, a defendant researcher will presumably argue that the failure to provide updated contact information releases the researcher from an obligation to disclose. As a complete defense, this seems unconvincing because it places all of the responsibility for maintaining contact on the participants and because conducting some basic inquiries about an individual’s whereabouts is not terribly burdensome in the modern digital age. With that said, courts must recognize that there are limits on a researcher’s ability and obligation to actively seek out lost participants. As discussed above, researchers are not clinicians and do not have the same clear obligation to act in the best interest of their participants.²²⁹ Research is undertaken with the goal of producing socially beneficial, generalizable knowledge,²³⁰ so it is unreasonable to expect researchers to devote resources to tracking down participants to the extent that doing so undermines their ability to answer important research questions.

The manner of disclosure might also be subject to dispute. For example, can a researcher discharge his or her duty by returning the information in a letter or phone conversation? Or are researchers instead required to provide robust genetic counseling to ensure that participants fully understand the clinical meaning of the incidental finding? We argue that although researchers are not necessarily obligated to provide genetic counseling, they must responsibly convey information in a way that is readily understandable. The manner of disclosure should be tied to the importance and urgency of the information being disclosed or the complexity or nuance of a particular finding.

Conclusion

The increasing use of whole-genome sequencing in research is likely to give rise to a wide range of potential IFs—from indications of a propensity to develop breast cancer to findings about which the meaning and importance are still unknown. Though the extent of any ethical obligation is still unsettled, a number of scholars, researchers, and institutional groups have promulgated guidance documents that attempt to set forth the IFs that should be returned and when they should be returned, or that have advanced criteria to be used in making these determinations. Researchers have responded by implementing their own piecemeal approaches to returning IFs. This push toward an ethical obligation to return particular IFs, coupled with a move among researchers towards returning IFs, has given rise to concerns about tort liability for failure to offer to return these results.

The extent to which a researcher could be held liable for inadequately returning IFs is uncertain. The extent of a researcher’s legal duty to return IFs turns on a number of factors, including the extent of any researcher–participant relationship, research customs or standards with regard to returning IFs, and guidance documents addressing when IFs should be returned. The last two factors are still very much in flux. A researcher’s liability will also turn on a court’s willingness to import the loss-of-a-chance causation analysis into research. But the researcher would be unable to ascertain at the outset when the failure to return IFs will give rise to liability.

Accordingly, this Article advocates adoption of a transparent standard that makes clear to participants which IFs, if any, will be returned to them. A researcher should decide at the outset which of several explainable approaches to returning IFs will be taken. The approach chosen, reviewable by an IRB, must be set forth in the informed consent document and will form the basis of a tort law duty—a researcher’s failure to comply with the terms in the informed consent document could lead to a finding of negligence.

This framework gives deference to a researcher’s considered judgment about the best approach to returning IFs given the particular circumstances of a research protocol, while allowing participants to know from the outset about which IFs will be returned and when, and it provides guidance to courts about whether liability should attach, given the evolving ethical, legal, and scientific considerations.