OUP user menu

Ethics and neuropsychiatric genetics: a review of major issues

Steven K. Hoge, Paul S. Appelbaum
DOI: http://dx.doi.org/10.1017/S1461145711001982 1547-1557 First published online: 1 November 2012


Advances in neuropsychiatric genetics hold great hopes for improved prevention, diagnosis and treatment. However, the power of genetic testing to identify individuals at increased risk for disorders and to convey information about relatives creates a set of complex ethical issues. Public attitudes are inevitably affected by the shadow of eugenics, with its history of distorting scientific findings to serve socio-political ends. Nonetheless, the growing availability of genetic tests means that more patients will seek genetic information and physicians must manage the process of informed consent to allow meaningful decisions. Patients should be helped to understand the often-limited predictive power of current knowledge, potential psychological impact, risks of stigma and discrimination and possible implications for family members. Decisions for predictive testing of children raise additional concerns, including distortions of family dynamics and negative effects on children's self-image; testing is best deferred until adulthood unless preventive interventions exist. Pharmacogenomic testing, part of personalized medicine, may bring collateral susceptibility information for which patients should be prepared. The implications of genetic findings for families raise the question of whether physicians have duties to inform family members of implications for their health. Finally, participation in research in neuropsychiatric genetics evokes a broad range of ethical concerns, including the contentious issue of the extent to which results should be returned to individual subjects. As genetic science becomes more widely applied, the public will become more sophisticated and will be likely to demand a greater role in determining social policy on these issues.

Key words
  • Consent
  • ethics
  • genetics


In this article, we review the major ethics issues raised by advances in neuropsychiatric genetics. It should be recognized that this discussion of ethics reflects current thinking about how genetic information should be used. But we are still early in the evolution of this promising technology and it will be many years before its full impact is felt and understood by society. Only then will it be possible to determine how genetic science should best be used to the benefit of society and to fully understand its ethical implications. As experience is gained, ethical principles will emerge to guide the use of genetic information in ways that balance the interests of individuals and society.

Historical background: the spectre of eugenics

Clinicians and researchers generally have positive views of genetic research, anticipating dramatic improvements in understanding the causes of mental disorders, selecting appropriate treatments for patients and preventing or reducing the burden of illness (Lawrence & Appelbaum, 2011). But many lay people, particularly members of minority groups, react to genetic science with suspicion, the roots of which can be traced to the historical spectre of the eugenics movement (Bates et al. 2005). The term eugenics (literally, ‘well born’) was coined by Sir Francis Galton (1883), who was influenced by the publication of Origin of Species by his half-cousin Charles Darwin. Galton posited that progressive social policies had undermined natural selection by providing protection, support and sustenance to the ‘unfit’, which allowed them to survive and reproduce. Inevitably, he believed, the population or ‘race’ would degenerate unless countervailing social policies were adopted. The eugenics movement that developed from these ideas was based on the belief that scientific methods should be applied to improve the genetic stock of mankind. The movement quickly spread to the United States, then to nations around the world (Black, 2008).

Positive eugenics programmes promoted marriage and reproduction among those believed to have superior genes, while negative eugenics aimed to lower the birth rates of those with ‘undesirable’ characteristics. The US was in the vanguard of negative eugenics. Early evidence of the heritability of ‘insanity’ and ‘feeblemindedness’ led many states to ban marriage and compel sterilization of individuals in these groups (as well as criminals), beginning in 1907 (Black, 2008) through the early 1980s (Sullivan, 2002). Approximately 65 000 people in 28 US states underwent compulsory sterilization for eugenic purposes before these programmes were halted. During the eugenics era, numerous other nations enacted compulsory or coercive sterilization programmes that targeted, in various combinations, the mentally ill, mentally retarded, criminal offenders and those afflicted with inherited medical disorders, including Germany, Canada, Japan, Sweden and Switzerland (Bashford & Levine, 2010).

Under the Nazi regime's eugenics programme, Germany involuntarily sterilized >350 000 individuals, primarily persons with schizophrenia, mental retardation and alcoholism (Hassenfeld, 2002; Meyer, 1988). The Nazis also called on physicians to identify children and adults with a variety of heritable disorders (‘life unworthy of life’) for extermination. More than 70 000 psychiatric patients were put to death in this Aktion T4 programme. Later in the war, 200 000–300 000 additional psychiatric patients were killed by gunshot, starvation and other methods, often to provide hospital beds to accommodate war casualties (Meyer, 1988; Torrey & Yolken, 2010). Following World War II, eugenic policies were discredited and gradually abandoned in most countries. However, the eugenics movement is a cautionary example of how the power of genetic science and its predictive promise can be exaggerated, distorted and channelled by elites to justify unspeakable actions. Eugenic laws disproportionately affected the less powerful segments of society. Thus, it is not surprising that these groups remain the most leery of genetic research and its potential applications or that there is lingering concern about the motives of physicians who recommend genetic testing and how the information may be used.

Clinical applications of genetic tests in adults

Genetic testing can be used in clinical practice to establish more definitive diagnoses (‘diagnostic testing’), assist in management decisions (‘pharmacogenomic testing’), provide guidance for reproductive choices (‘reproductive testing’) and estimate the likelihood of disease onset in the future (‘predictive testing’). At present, the frequency of genetic testing in neuropsychiatric practice varies considerably, based on the disorder involved and the purpose for which testing is sought. Diagnostic testing is most commonly used by clinicians who care for patients with neurodevelopmental disorders, where it may also aid reproductive planning (Miller et al. 2010); to confirm diagnoses of neurodegenerative disorders, including Huntington's disease (US Huntington's Disease Genetic Testing Group, 2003), frontotemporal dementia (Goldman et al. 2004) and autosomal dominant forms of Alzheimer's disease (AD; Goldman et al. 2011) and for several syndromes associated with psychosis (Bassett et al. 2003) and epilepsy (Ottman et al. 2010). Pharmacogenomic testing for individualized treatment decisions is still at an early stage of development, without a robust evidence base, and hence tends to be infrequent (Arranz & Kapur, 2008). Reproductive testing is most useful for autosomal dominant disorders, such as Huntington's disease, and for X-linked disorders, including Fragile X syndrome (Garber et al. 2008). Given the low predictive power of most genes associated to date with common psychiatric and neurodegenerative disorders, predictive testing, although much discussed, plays a relatively small role in most neuropsychiatric disorders, with the notable exception of some autosomal dominant neuropsychiatric disorders such as Huntington's disease.

However, clinical genetic testing is likely to become increasingly common in many neuropsychiatric settings for three reasons. First, there may be significant latent demand for testing. For example, surveys of relatives of affected individuals indicate that many of them would seek genetic testing to determine their risk – or that of their children or other relatives – for developing neuropsychiatric disorders or to aid in reproductive decisions (Jones et al. 2002; Meiser et al. 2008; Milner et al. 1998; Smith et al. 1996; Trippitelli et al. 1998; Wilde et al. 2011). DeLisi & Bertisch (2006) found that 83% of family members of schizophrenia patients who had participated in genetic research would seek predictive testing and 80% of spouses of bipolar disorder patients indicated that they would want to be tested, if genetic tests for the conditions were available (Trippitelli et al. 1998).

Second, genetic tests are becoming more widely available. Advances in research have resulted in the identification of a growing number of genetic variants associated with neuropsychiatric disorders and the potential demand for such tests has encouraged their commercialization. In addition to traditional clinical laboratories, for-profit genetic testing companies have initiated aggressive direct-to-consumer (DTC) marketing campaigns for a broad range of tests (Hogarth et al. 2008). Gollust et al. (2002) found that DTC advertisements overstated the value of testing, provided misinformation about genetics, endorsed a deterministic relationship between genes and disease and tended to exploit consumers' fears to drive demand. In addition, a study by the US Government Accountability Office (2010) called into question the accuracy of DTC test results. Whatever its faults, DTC marketing is likely to increase public awareness of and demand for testing.

Third, as genetic testing migrates from research protocols to ordinary clinical settings, many physicians may be likely to comply with patients' requests to know their genetic status, even if it is not highly predictive of future disease. Moreover, the widely anticipated introduction of whole genome sequencing into clinical medicine will make available to physicians and patients data on their complete genetic make-up, including variants associated with neuropsychiatric disorders, magnifying the presence of genetic information in the clinical setting (Ashley et al. 2010).

As genetic testing moves into the clinic, the most salient ethical concerns will centre on promoting informed decision making by patients, a process that will be complicated by several factors. Few neuropsychiatric conditions follow classical Mendelian inheritance (Huntington's disease and autosomal dominant forms of dementia and epilepsy are among the exceptions). In general, the picture is much more complex. Single nucleotide polymorphisms (SNPs) associated with disorders such as bipolar disorder, schizophrenia and depression contribute only very modestly to individual susceptibility (Plomin & Davis, 2009). For example, the associations between single genes and common psychiatric disorders have odds ratios of 1.1–1.6 (Kendler, 2006). With regard to schizophrenia, replicated genetic loci have relative risks of about 1.1 (Kim et al. 2011). At present, testing for SNPs has no clinical value for these common disorders. On the other hand, recent data suggest that up to 2.5% of people with schizophrenia may have copy number variants (CNVs) associated with much higher odds ratios (Tam et al. 2009; Walsh et al. 2008), although penetrance may still be incomplete and variable (Vassos et al. 2010). Also, testing for CNVs among children with neurodevelopmental disorders may rapidly be becoming the standard of care (Miller et al. 2010). One of the more common and predictive variants is the microdeletion on chromosome 22q11.2 associated with velocardiofacial syndrome (DiGeorge syndrome). Approximately 25% of individuals with the deletion develop schizophrenia (Bassett et al. 2003 ), although the estimated prevalence is only about 2% in persons with schizophrenia (Murphy, 2002).

Another complicating factor is emerging evidence suggesting that, for some conditions, genetic predispositions by themselves may not result in the development of a disorder, unless they interact with appropriate environmental conditions and life events (Thapar et al. 2007). For example, genetically determined variation in the production of the enzyme monoamine oxidase A appears to interact with childhood maltreatment to increase the rates of antisocial behaviours occurring later in life (Beach et al. 2010; Caspi et al. 2002). Similar, albeit controversial, findings have been offered for the impact on depression of serotonin transporter polymorphisms, with increased risk for the disorder reported to be dependent on the occurrence of substantial life stress (Caspi et al. 2003; Munafo et al. 2009), perhaps through epigenetic mechanisms (McGowan et al. 2009; Weaver et al. 2004). Further complicating attempts to identify the effects of gene×environment interactions are findings indicating that people may preferentially seek or create their own experiences, based in part on their genetic endowment, a phenomenon referred to as genotype–environment correlation (Kendler & Baker, 2007). More precise quantification of the genetic contribution to these – and perhaps other – psychiatric conditions awaits the disentanglement and verification of gene–environment effects (Duncan & Keller, 2011).

Given these limitations, in many cases an important part of the consent process will be informing those interested in testing that the results are likely to be indeterminate and result in relatively small adjustments to their estimated likelihood of having, developing or passing on a condition. At present, for most people, risk probabilities will be more strongly affected by past history, family history, clinical factors and demographic variables. In the event that testing is pursued in these situations, the responsible professional must be prepared to face a wide range of patient capabilities related to numeracy and understanding of risk. Given the documented difficulties that patients have in understanding genetic information (Klitzman, 2010b), instruction in probability using simple aids has been suggested (Faraone et al. 1999; Hodgkinson et al. 2001). The development of counselling techniques and decision-making aids will become an increasing priority as the number and range of genetic tests continue to grow (Austin & Honer, 2007). With the anticipated introduction of whole genome sequencing into clinical practice in the next decade (Pasche & Absher, 2011), the urgency of developing means to help cope with the data is even greater.

In making the decision to be tested, individuals must also take into account potential negative consequences. With regard to predictive testing, for example, test results indicating an enhanced propensity for neuropsychiatric disorder can profoundly affect self-image, potentially leading to increased anxiety and depression (Wilhelm et al. 2009). In part, concern over these negative consequences is driven by the stigma associated with neuropsychiatric conditions, which may extend to asymptomatic carriers and family members (Phelan, 2005). Compounding the task of dealing with test results is that an individual with an enhanced propensity for developing a disorder still faces substantial uncertainty that the disorder will develop at all or what its severity will be. The uncertainties are further magnified by the fact that many disorders will appear only later in life and truly effective prevention and treatment are unavailable. If reproductive planning decisions drove the choice to be tested, the results may be equally indeterminate. Thus, genetic information may leave individuals subject to considerable confusion and stress when making decisions about prophylactic interventions, including life-style changes, participation in high-risk research and family planning.

Moreover, genetic information may be put to discriminatory uses. In the US, the Genetic Information Non-Discrimination Act of 2008 (GINA) provides federal protection against discrimination based on genetic information (including family history) in health insurance and employment. However, GINA does not extend to life, disability, long-term care or other insurance or to genetic conditions once they become manifest (Appelbaum, 2010). Physicians can reduce the risk of discrimination against their patients by warning them about the various ways that discrimination could occur, which may lead them to greater caution in seeking genetic information. Altering the source of the test is another option. Test data obtained through a hospital laboratory or other health care entity are likely to reside in an information system that may be accessed by others. An alternative is testing in a private laboratory, from which patients may obtain test results and discuss them with their physicians without including them in the medical record. Of course, genetic information will then not be available to subsequent treaters, with a potential negative impact on patients' health (Klitzman, 2010a).

Even when definitive results are available, some people will prefer not to shoulder the burden of knowing their future or the risk of conceiving children with a propensity for a disorder and will forgo the prognostic capabilities of modern medical technology (Kass, 2004). For other patients, negative consequences will be of less importance than the availability of clinical interventions or greater diagnostic certainty. In a review of the literature, Marteau & Croyle (1998) found that only about 10% of those at risk for Huntington's disease, for which there is no effective intervention, elect to undergo predictive testing. However, among women at risk for breast cancer, a disease for which preventive measures and treatment are available, about 50% choose to have susceptibility testing; similarly, about 80% agree to be tested for familial adenomatous polyposis, for which effective prophylaxis exists.

Some authorities in the field, including the influential Nuffield Council on Bioethics (1998), have taken the stance that the decision to offer or to recommend genetic testing should turn on two factors: the availability of effective interventions for the tested disorder; and the predictive accuracy of the test (see also Burke et al. 2001). This approach has led major professional organizations to take the position that apolipoprotein E (APOE) gene testing for susceptibility to late-onset AD is not clinically indicated (Goldman et al. 2011). The test is only modestly predictive – studies indicate a roughly three-fold risk for AD with one copy of the E4 allele and a 12-fold risk for those with two copies. Approximately half of those with late-onset AD have one or more copies of the E4 allele, which is present in about 25% of normal controls (Verghese et al. 2011). Thus, many people who carry the E4 allele will not develop AD and may suffer undue worry from test results and many of those who test negative for it will develop the disorder and may be falsely reassured. Also, of course, prophylactic interventions are not available. Of note, APOE4 is among the most predictive polymorphisms detected in association studies of common variants in neuropsychiatric disorders, underscoring the limitations of testing for these polymorphisms.

Recent results from a controlled study suggest that some concerns about the effects of disclosure of APOE status – and by extension perhaps other polymorphisms predisposing to disorders with similar characteristics – may be exaggerated, at least over the short term. In the REVEAL study, Green et al. (2009) investigated the adult children of patients with AD. The control group received a risk assessment for developing AD based on age, gender and family history. The study group received the same information plus their APOE genotype-specific risk. No differences between groups were found in anxiety, depression and test-related distress at 6 wk, 6 months and 1 yr. However, those who tested positive for the E4 allele had a significantly higher level of test-related distress when compared to those who tested negative, although the former group was no less likely to say that they would undergo testing again (Zick et al. 2005). Whether the results from this highly selected sample will apply to the general population remains unknown.

Approaches to the ethical challenges

Given the current status of genetic testing for neuropsychiatric disorders, what principles should guide clinical practice? We have a preference for involving patients in the decision-making process regarding potential testing and, when testing is performed, providing information regarding results to the fullest possible extent. As part of this decision-making process, patients will need to be made aware of the risks as well as the benefits of testing (Finn & Smoller, 2006). This is consistent with respect for the decisional autonomy of patients, even when their choices are contrary to professional recommendations. How best to do that without engendering unnecessary anxiety that can compromise decisional abilities is an important area for ongoing research; models include the REVEAL study described above (Green et al. 2009) and Wilhelm et al.'s (2009) study of disclosure of data related to depression risk. Research is also needed to help clarify patients' abilities to understand the complexities of genetic testing and to identify methods to improve understanding.

To guide their patients, clinicians will need to be knowledgeable about emerging genetic findings. Indeed, in all but the largest referral centres, this burden is likely to fall on physicians directly involved in caring for patients, since the very small number of genetic counsellors (approximately 2400 certified genetic counsellors in the US; American Board of Genetic Counseling, 2011) means that they generally will not be available. At present, most clinicians are not prepared to perform this task. There is evidence that physicians, in general, are not well informed about genetic testing (Menasha et al. 2000; Baars et al. 2005). A survey of psychiatrists revealed a low level of knowledge regarding psychiatric genetics, with less than one-quarter feeling competent to discuss genetic information with patients and families (Finn et al. 2005). In the mental health field, many patients are treated by non-physician clinicians, who are likely to have even less knowledge of genetics. Major educational efforts will need to be addressed to the relevant clinical disciplines, especially given current evidence of the inadequacy of training in genomics in psychiatric residencies (Hoop et al. 2010; Winner et al. 2010).

The anticipated introduction of whole genome (or exome) sequencing to clinical medicine will complicate the process of obtaining consent and returning results to patients. It will not be possible to fully inform patients about the possible findings and their implications, since literally any component of the genome could be involved. Moreover, many of the findings are likely to have unclear import for the patient's health and reproductive choices. One approach may involve helping patients to decide what kind of information they would like to receive by stratifying consent, e.g. asking them whether they want information that is clinically actionable, has reproductive implications, may affect life planning, etc. The National Genome Research Institute (2011) has funded a set of studies to examine return of genomic data from research settings, which should provide guidance for the clinical context as well.

Prenatal genetic testing

At present, prenatal genetic testing is available through pre-implantation genetic diagnosis for couples utilizing in vitro fertilization (Klitzman et al. 2008), but is more commonly performed on samples obtained by amniocentesis or chorionic villus sampling for chromosomal abnormalities (e.g. Down's syndrome) and single-gene conditions (e.g. Huntington's disease, Fragile X syndrome). However, the development of technologies for the detection of foetal chromosomal abnormalities in maternal serum (Chiu et al. 2011), with indications that reliable detection of genetic polymorphisms may be possible (Lo et al. 2010), is likely to encourage greater use of prenatal testing (Greely, 2011).

Preliminary studies suggest there is substantial interest in prenatal testing for other neuropsychiatric disorders such as bipolar disorder (Jones et al. 2002; Smith et al. 1996; Trippitelli et al. 1998), schizophrenia, alcoholism, attention deficit disorder and depression (Milner et al. 1998). The high level of expressed interest may reflect a lack of understanding of the likely value of predictive testing for these conditions. For example, a genetic test that suggests a doubling of lifetime risk for developing schizophrenia – higher than is currently feasible (with the exception of 22q11 deletion syndrome and rare, highly penetrant variants within families) – would raise the probability from 1% to 2% – a high relative risk, but still a low absolute risk. Even if prospective parents vary considerably in how heavily to weigh the risk of neuropsychiatric illness in their decisions to abort pregnancy, this level of predictive power is not likely to provide much helpful information to them.

Genetic testing of children

Genetic testing of children may be diagnostic or predictive and each scenario has different ethical implications. Diagnostic genetic testing as part of an evaluation of an existing condition is the least problematic, but can raise issues regarding impact on family members. For example, testing for Fragile X syndrome facilitates diagnosis, treatment planning and future reproductive decisions. However, identification of a child with Fragile X syndrome may have unwelcome implications for unaffected parents who carry the premutation (a smaller number of trinucleotide repeats than are found in clinical cases); males may be susceptible to tremor/ataxia syndrome later in life (Leehey, 2009) and females to premature ovarian failure (American College of Obstetricians & Gynecologists Committee on Genetics, 2010). Hence, parents should be aware of the collateral information that may become available before making the decision about testing.

Even more serious ethical concerns emerge in the context of predictive testing. Parents may seek to have their asymptomatic children tested for susceptibility to disorders that might appear later in life, invoking their right to know their children's vulnerabilities and to make family decisions (Tarini et al. 2009). They may have an interest in ensuring that they have some children free of the risk of developing a specific neuropsychiatric disorder and may seek testing as part of family planning. However, these interests are pursued at the expense of the child being tested, who will receive no immediate benefit and whose choice whether to be tested later in life will be pre-empted. When a child is identified as being susceptible to a future disorder, parents may allocate family resources for education to the detriment of the at-risk child or seek to protect the child in ways that can affect development or, perhaps more subtly, self-image. Finally, test results may lead to discrimination in its many manifestations.

The Nuffield Council on Bioethics (1998) and other thoughtful commentators (Clark, 1994) have suggested that parents should not have their children tested for unpreventable adult-onset disorders; these decisions should be made by the children, once they reach adulthood. Parents should only be allowed to test when effective therapeutic interventions are available during childhood. The simplicity of this analysis, however, may be complicated by research on gene×environment interactions. Well-meaning parents may seek genetic testing with the intent to shield at-risk children from exposure to environmental circumstances that could interact with genetic vulnerabilities to trigger neuropsychiatric disorders. Further understanding of gene×environment interactions and prophylactic interventions may unsettle the current ethical resolution of this issue.

Pharmacogenomic profiling

Pharmacogenomic profiling refers to the study of genetic influences on drug responsiveness, including propensity for adverse reactions. Polymorphisms may lead to variable drug response by altering enzymatic metabolism or the sites of drug action. The promise of pharmacogenomics is that pharmaceutical drug development may be improved and treatment personalized based on individuals' genetic information. For example, identifying the genetic basis for treatment resistance could lead to the development of new therapeutic agents or clinicians may be able use pharmacogenomic profiling to select medications that minimize the risk of adverse reactions and maximize the likelihood of treatment response (Wilke & Dolan, 2011).

In many instances, pharmacogenomic profiling raises few ethical concerns. But in some cases, genetic markers used to profile drug response will convey more problematic information. For example, APOE profiles that provide information about risk of AD may help to predict treatment response to lipid-lowering statin medications (Nieminen et al. 2008). Thus, information that has been generated for pharmacogenomic purposes for the treatment of one condition conveys important risk information – that patients may not want to know – for another. With genetic research progressing rapidly, it is almost certain that some genetic information obtained in a low-risk context will later be found to have more profound implications (Netzer & Biller-Andorno, 2004). During the informed consent process, patients who are to undergo genetic testing for any purpose should understand that the results could later be found to have collateral implications.

Duties of physicians to patients' families

Because genetic information may have significant consequences for patients' relatives, questions have been raised about whether physicians have a duty to warn family members when the results of genetic tests indicate they may be at risk. US courts have addressed this issue in a handful of cases that suggest what the appropriate ethical contours of such a duty might be.

In Safer vs. Pack (1996), Ms Safer sued Dr Pack, her father's physician, who had treated him for ultimately fatal colon cancer due to familial adenomatous polyposis, which is usually an autosomal dominant disorder. Those with the disorder develop multiple benign polyps at a young age; progression to malignancy is inevitable unless the colon is removed. Dr Pack had not told the family (Ms Safer was aged 10 yr at the time of her father's death) about the underlying condition or its heritability, information known to the medical community at the time. Unaware of her risk, Ms Safer took no precautions. At the age of 36 yr she was diagnosed with familial polyposis and metastatic carcinoma of the colon. The New Jersey Superior Court held that a physician had a ‘duty to warn those known to be at risk of avoidable harm from a genetically transmissible condition’. The duty to warn family members, as articulated in this case, would appear to trump patients' interest in confidentiality. The Florida Supreme Court, in Pate vs. Threlkell (1995), articulated a narrower duty, one that did not sacrifice patients' right to privacy. Under Pate, which involved a claim brought by the family of a patient with medullary carcinoma of the thyroid, the physician must only warn the patient of the risk to family members, leaving it to the patient's discretion whether they are informed.

To date, only one court has addressed information derived from genetic testing. In Molloy vs. Meier (2004), the mother of a daughter with Fragile X syndrome claimed that the child's physicians had failed to inform her about the risk for subsequent conceived children. The Minnesota Supreme Court held that ‘a physician's duty regarding genetic testing and diagnosis extends beyond the patient to biological parents who foreseeably may be harmed by a breach of that duty’.

These cases suggest that courts are likely to find a duty to warn family members based on three factors: (1) the likelihood that family members are at risk; (2) the severity of the potential consequences; (3) whether effective interventions exist. As previously discussed, genetic testing for neuropsychiatric disorders is likely to involve alleles that result in modest increase in relative risk, with the exceptions noted above. Psychiatric conditions generally are treatable, although neurodegenerative conditions are not, but in either case preventive measures are not currently available. Thus, at present, it will usually be difficult to justify warning family members directly, particularly if this would involve a breach of patient confidentiality. However, it is good practice to advise patients of any increased risk to their family members, as this allows patients to decide whether and how to share information with them. The practice of providing information and placing responsibility in the hands of patients is supported by the public (Lehmann et al. 2000) and genetic counsellors (Dugan et al. 2003) and may evolve into the standard of care.

Ethics in genetic research

Genetic research into neuropsychiatric disorders poses many of the ethical issues found in the clinic, but some issues peculiar to the research setting also arise, including whether consent must be obtained for subsequent use of de-identified samples in other research (Rothstein, 2010), whether community consultation or even consent should be obtained prior to research that may have implications for discrete population groups (Greely, 2001), whether participants' should be given access to their research data (Wolf et al. 2008) and whether investigators have ongoing duties to subjects if subsequent analyses indicate risk of a disorder for which preventive interventions exist (Wolf et al. 2008). We lack the space here to consider even this truncated list of issues in detail, so we focus on the question of return of research data to subjects, which has attracted enormous interest.

In the early years of genetic research, when most findings were of uncertain clinical significance, investigators generally did not inform participants of their personal genetic results, which generally would have been of little use. Today, however, with a growing number of loci associated with increased risk for the development of disorders – and in some cases clinical management – there is an evolving consensus among bioethicists that subjects who desire to receive genetic information should be given access to at least some of the findings (Bredenoord et al. 2011). Moreover, as whole genome and exome sequencing for research purposes become more common, the amount of data will skyrocket and the likelihood that some of it will be of interest to subjects will increase (McGuire et al. 2008).

In the most widely adopted formulation, return of research data is viewed as part of a reciprocal obligation of the researcher, in exchange for the subject's participation and trust (Richardson, 2008). Studies of subjects' preferences regarding access to information generated in genetic and genomic research have found consistent interest in knowing the results of these analyses, especially when actions can be taken to protect subjects' or relatives' well-being (Murphy et al. 2008; Wendler & Emanuel, 2002). In keeping with these preferences, a growing number of federal agencies, expert panels and authors have recommended that at least some genomic data be made available to subjects; although there are substantial differences in the specifics of these recommendations, there is general consensus that data should be offered when they have clear, actionable implications for subjects' health or behaviour (e.g. Fabsitz et al. 2010; National Bioethics Advisory Commission, 1999; National Human Genome Research Institute, 2010).

Even with this consensus, however, it is inevitable that line-drawing problems will arise. How substantial a risk and how effective an intervention must exist before disclosure is warranted (Biesecker & Peay, 2003; Wolf et al. 2008)? The fact that most associations of genetic loci with psychiatric disorders have not been subsequently confirmed provides sobering evidence of the difficulty of these judgements. The time frame for verifying efficacy of some interventions may be very long and research designs may be prohibitively difficult.

Moreover, there may well be reasons for disclosure even in the absence of prophylactic interventions. Couples may find information about carrier status to be valuable for reproductive planning. Persons at risk for developing AD, for example, may use genetic information to make financial decisions, purchase insurance or make other personal arrangements. As the earlier discussion of clinical testing suggests, a focus on medical interventions simply does not do full justice to individuals' concerns.

The final resolution of the problem of return of research data will be impacted by another set of issues, namely, the burden on investigators to scan, interpret, confirm and communicate results. To the extent that substantial resources are diverted from the research effort to this arguably collateral obligation, an argument can be made for significantly restricting its scope (Klitzman, 2006). Another important question relates to how meaningful informed consent for return of genomic data can be obtained from subjects when the extent and significance of potential findings are unknown. Final resolution may await results of studies under way to assess the costs and consequences of returning genomic data to human subjects.


Research on the genetics of neuropsychiatric disorders is unfolding a panorama of rich complexity. It is not yet possible to foresee the full ethical, legal and social implications of the emerging data.

But already it appears that genetics is headed out of the realm of research and into the clinic. As genetic testing assumes its place in clinical practice, it is inevitable that a more diverse range of clinicians will find a broader set of applications for testing and more innovative clinicians will be sought out by patients who are motivated to be ‘early adopters’ of genetic testing. New questions and ethical concerns are likely to emerge in this fluid and uncertain era in neuropsychiatric genetics. The full involvement of the public in these issues will signal the maturation of the ethics of genetic science.

Statement of Interest



This work was supported in part by funding to Dr Appelbaum from the National Human Genome Research Institute (1P20HG005535-01).


View Abstract