The study of novel therapeutics in humans presents many ethical dilemmas, for example when testing the safety of therapeutics that may later prove ineffective in clinical settings. Nita Farahany noted that as therapeutic development for nervous system disorders moves forward, ethical parameters are needed to ensure not only good science, but also the protection of research subjects. These ethical concerns are particularly germane when experimental therapies move to human trials without efficacy having been established in appropriate animal models, she said. She argued for ethics to be viewed by researchers not as a roadblock for preventing innovation, but in partnership with science, with conversations about ethics preceding discussions about regulatory clarity. Farahany suggested that alternative regulatory and ethical approaches may be needed to ensure that research can progress without putting people at unnecessary risk. Moreover, she said it is imperative to move forward in this manner because research that causes significant risk can set entire fields back by decades or more. Regulatory models, Farahany added, should be built on ethical constraints, ethical justifications, and ethical safeguards.
Throughout the workshop, many participants highlighted two key themes related to ethics. The first relates to the identification of safeguards that may be needed before moving into human trials in the absence of predictive models of efficacy. The second is the need to ensure adequate capacity and consent from vulnerable populations, including children, individuals who may be compromised intellectually, as well as patients who are willing to subject themselves to considerable risks when therapeutic alternatives are not available.
Although many companies have adopted ethics boards to advise and guide them with regard to the process of scientific discovery, Jonathan Kimmelman, associate professor in biomedical ethics at McGill University, argued that the key to getting companies to do the right thing is to create incentives. For example, regulations, ethical oversight, and poli-
cies at academic medical centers can help to establish a set of expectations and incentives for companies to make data more available, he said.
ENSURING SAFETY IN FIRST-IN-HUMAN TRIALS
A basic tenet of all ethics codes, whether applied to clinical or preclinical studies, is that benefits must outweigh harms, said Kimmelman. He outlined six core points relevant to early-phase neurological drug trials conducted when evidence is scant.
First, said Kimmelman, establishing an ethical basis for conducting trials requires engagement of preclinical evidence. This evidence may take different forms, often involving in vivo models designed to recapitulate some aspects of the disease. However, as discussed throughout the workshop, such models are often not available for nervous system disorders. Nonetheless, preclinical evidence in support of a study may exist in the form of mechanistic research that examines fundamental aspects of pathophysiological processes. The important point to remember, said Kimmelman, is that a clinical trial should be initiated only when preclinical studies support a well-grounded hypothesis and when the trial has a high or reasonable probability of clinical application. Then, even if the study yields a negative result, something meaningful will be learned. The alternative “fast and frugal” method, where a broad range of targets with a low probability of success are scanned, results in high failure rates, such as those that have been seen historically for late-phase neurological trials, he said. However, according to Lucie Bruijn, chief scientist with The ALS Association, studies of investigational treatments should be allowed to go forward in a disease for which there are no meaningful treatments, even when the probability of efficacy is low for patients with terminal and untreatable diseases such as amyotrophic lateral sclerosis (ALS), as long as there is no evidence of toxicity.
For his second point, Kimmelman asserted that investigators are not making full use of available preclinical evidence. One reason for this is a lack of rigor in preclinical research, especially a lack of prespecified hypotheses and power calculations to prove those hypotheses, he said. The consequence of these shortcomings is substantial publication bias, a vast overestimate of treatment effects in preclinical models (Sena et al., 2010), and a serious replication problem (Steward et al., 2012).
Third, Kimmelman argued that human protection policies for persons who lack competence do not make sense with regard to early-phase trials because it is virtually impossible to know at this stage of development
whether a study poses a minimal risk, a minor increase greater than minimal risk, or an even greater risk that is offset by the potential of therapeutic benefit.1 Indeed, Rebecca Dresser, Daniel Noyes Kirby professor of law and professor of ethics in medicine at Washington University, said there is no consensus among ethicists about how to classify first-inhuman studies. Could one ever be certain there is minimal risk or even a minor increase over minimal risk? She said there have been efforts to redesign first-in-human trials to minimize risk (including the reliance on expert opinion coupled with any preclinical evidence available), yet this question remains unanswered. Farahany asked whether there might be a different set of safeguards that could be implemented, such as adaptive regulations. Moreover, she cited the need to balance risk assessments against ethical obligations to find solutions that can benefit individuals. Not going forward with research and thus not identifying beneficial therapies presents ethical concerns, added Farahany.
The question about human protection regulations for early-phase trials leads to Kimmelman’s fourth point, that the ethics of prelicensed drug trials involve factors beyond human protection. Clinical development, he said, involves a fragile coalition of patients, clinicians, industry scientists, regulators, public funders, and academic medical centers (London et al., 2010). All of their needs and concerns must be considered for a successful drug development enterprise, said Kimmelman.
Returning to the general issue of the need for increased rigor in preclinical studies, Kimmelman emphasized that early-phase studies should be designed so that negative clinical trials are informative. He called for more attention to secondary analyses with prespecified hypotheses, correlative studies (e.g., studies in cerebrospinal fluid), standardized protocols to maximize the interoperability of data, and biobanking.
Ethical issues go beyond the welfare of human subjects, and benefits to society must also be considered given the investment and resources that go into clinical trials, said Kimmelman. He said one of the most important ethical problems in neurological drug development is the inaccessibility of data. Farahany agreed, saying that if a trial is allowed to go forward, there may be an ethical obligation to share data, including failed research results. Information sharing does not necessarily have to be accomplished through peer-reviewed literature, she said, but can be embedded in regulatory mechanisms that enable preclinical data sharing. Bruijn
___________________
1 Additional discussion of vulnerable populations is provided in the next section.
added that just publishing a dataset may not be enough. Data also need to be interpreted correctly and in a way that patients can understand.
Kimmelman reiterated that to be ethical, the potential benefit of a study (e.g., increased generalized knowledge) must outweigh the harm. However, although reporting guidelines exist for clinical trials, such guidelines for the most part are nonexistent for early-phase studies, except perhaps in the neuro-oncology field (Chang et al., 2005). In a systemic review of novel neurological drugs approved by the Food and Drug Administration (FDA) between 2005 and 2009, Kimmelman and colleagues found that among licensed drugs for neurological indications (which presumably had positive clinical studies), only 55 percent of trials in the translation trajectory had been published within 5 years; for unlicensed drugs (negative studies), the corresponding number was 22 percent. Kimmelman suggested that regulators, the National Institutes of Health, and academic medical centers all could establish policies requiring publication of results within a reasonable time frame of 6 months to as long as 2 years. He added that the FDA does not require prospective registration of Phase I clinical trials; thus, there is no public record of these activities. This sort of registration could be the first step to improving reporting, said Kimmelman.
Todd Sherer raised more ethical questions related to the risks and benefits of clinical trials and disclosure. When a study establishes a subset of a population based on genetic or clinical characteristics so that a genetic or other risk factor may be uncovered, how much information—such as individual research results, which some participants may want—should be disclosed and how much should be added to the consent process to inform participants without adding substantial additional burden? Farahany noted that the Presidential Commission for the Study of Bio-ethical Issues made recommendations in this regard in their report Privacy and Progress in Whole Genome Sequencing.2 Bruijn added that the ALS community has struggled with this question in part because patients hold widely different opinions. For example, in studies where presence of an SOD-1 mutation is an inclusion criteria, some patients and families want to know the results for planning purposes, while others do not want to know. Dresser said the research community has developed a general guideline of clinical utility for researchers to ask whether disclosure of the information would make a difference in the person’s health? However, she added that for some patients, the concept of personal utility goes
___________________
2 For more information, go to http://bioethics.gov/node/764 (accessed November 18, 2016).
beyond clinical benefits, and she suggested that trial participants should be treated as partners and educated about the limitations of the information that is being disclosed.
Kimmelman suggested several policy changes to help ensure that first-in-human trials are conducted within an ethical framework. These include
- Establishing policies at academic medical centers that require investigators to publish studies within a period of 6 months to 2 years from completing primary outcome data collection as a condition of allowing investigators to recruit patients;
- Establishing policies requiring that animal studies used to provide the evidence base for initiating a clinical trial include a prespecified hypothesis, power calculations, and identification of outcome measures; these should be shared in a public database; and
- Enforcing compliance with reporting standards regarding the design and interpretation of studies.
CONDUCTING TRIALS IN VULNERABLE POPULATIONS
Ethical guidelines regarding clinical trials for children allow studies under three conditions: (1) they pose minimal risk, (2) they pose greater than minimal risk along with the prospect of direct benefit to subjects, or (3) they pose a minor increase over minimal risk, Dresser said.3 This latter category allows more leeway for children with preexisting health problems who face higher risks in their everyday life, and also includes research that does not offer direct benefit, but presents only a minor increase over minimal risk. Dresser said there is another important category in which studies that do not meet these requirements can be approved by the Secretary of the Department of Health and Human Services or the FDA Commissioner and Institutional Review Board after consultation with appropriate experts and a public comment period, if they advance important knowledge related to the condition and are conducted “in accordance with sound ethical principles.”
___________________
3 For more information on ethical guidelines concerning clinical trials for children, including definitions of risk categories, go to https://www.hhs.gov/ohrp/regulations-andpolicy/guidance/special-protections-for-children/index.html (accessed February 1, 2017).
Children are considered a vulnerable population and need special protection for different reasons, including lack of understanding and cognitive and social immaturity, said Dresser (de Melo-Martin et al., 2012). They may also be pressured by parents or others (either implicitly or explicitly), are more vulnerable to long-term effects of adverse events, and are more likely to benefit from future improvements, which a risky trial might negate, she said. Although there are no special rules for adults who are unable to make their own decisions, many ethics advisory groups have recommended guidelines similar to those governing pediatric research. Some groups recommend a possible exception that would apply to persons with dementia who previously signed a research advance directive indicating their willingness to expose themselves to higher risk as a means of finding better treatments in the future, although such advance directives are not addressed in current research regulations.
For equitable subject selection, Dresser said the prevailing view is that sponsors should first try to enroll adults capable of informed consent. If there is scientific justification for enrolling a different population, such as children or adults with dementia, efforts should be made to select older children or mildly impaired adults over those less able to participate in choice. When a study has no reasonable prospect of direct benefit to subjects, the decision should ideally be made by the person who will bear the risk, said Dresser. Direct benefits should be supported by objective evidence, she added. Collateral benefits, such as additional health assessments, that can accompany a trial are not generally thought to be sufficient to support enrollment of subjects.
INCLUDING PATIENT PERSPECTIVES AND NATURAL HISTORY OF DISEASE IN ETHICS DISCUSSIONS
Bringing the patient perspective into ethics discussions early in the drug development process can be very valuable, both to educate investigators and to help avoid downstream consequences, said Bruijn. Dresser concurred, adding that, in her opinion, the experienced research subject is an underused resource. Ramona Hicks, chief scientific officer for One Mind, added that because participation in clinical studies is usually done voluntarily with no compensation, factoring patients into a team serves as a way of rewarding them for their contribution to the study.
John Krystal suggested that in addition to assessing patients’ understanding of a study and their tolerance of risk, it might be useful to assess
their motivation for participating. The ethics of including patients in a study might be different if they think they will personally benefit versus having an altruistic motivation of advancing knowledge so that someone else might benefit. Farahany agreed, suggesting that it would be useful to carefully evaluate ethical guidelines for patients with these different motivations, especially when the risk data and preclinical data are different from more conventional studies. The emergence of personalized medicine, for example, requires a rethinking of standards and guidelines, she said.
Farahany also mentioned a study at Duke University’s Science, Law, & Policy Lab that generated empirical data to inform law and policy on questions such as the risk tolerance of individual patients, as well as the empirical results of setbacks. Empirical data, she said, can ensure that regulations and ethical guidelines are not based on false conjecture.
According to several workshop participants, longitudinal and natural history studies can also provide information useful in ethics discussions for certain conditions. For example, most studies for traumatic brain injury are done in adults even though the consequences and the response to treatment may vary across the age spectrum, noted Hicks. Even for conditions with less lethal consequences, natural history studies can be informative, added Steve Hyman. For example, he recalled a pharmacoepidemiologic study from decades ago that showed that one in a thousand children in America between the ages of 3 and 5 were getting Ritalin despite an absence of clinical trials data. Given that natural history data suggested that young children with severe symptoms had a poor trajectory, the decision was made to conduct a study in this population, rather than simply advising parents that there were no data to support this indication. Indeed, the study showed that Ritalin was effective in these children, although less so than in older children, and that these young children had a lower threshold for side effects. The important lesson from this study, said Hyman, is that in a knowledge-poor environment, an understanding of disease trajectory may be useful in deciding whether to move forward with clinical development.
Ethical considerations also come into play in relation to clinical trials of people with terminal, untreatable diseases, where vulnerability arises out of desperation. For example, Farahany mentioned the controversy around Right-to-Try legislation,4 which could allow patients to gain access to untested or minimally tested drugs. While recognizing the des-
___________________
4 For more information, go to https://www.congress.gov/bill/114th-congress/housebill/3012 (accessed November 18, 2016).
peration that some patients feel, the field needs to determine under what conditions this might be appropriate, she said.
This page intentionally left blank.
