The need for Disclosures About the Overall Prospects of Research
The stipulated disclosures found in the informed consent instructions of the referenced EU, Canadian, and U.S. documents do not require disclosures about the overall prospects for the success of trials research candidates are being asked to help make possible, even though it is these prospects that establish proportionality between the benefits to society likely to be produced by trials with the risks they will impose on research participants. This results in informed consent processes that eschew information that actually speaks about those prospects.
Evidence that sheds light on those prospects is readily available and much of it is likely at drastic odds with the motivations that drive many, perhaps even most, people to volunteer for clinical trials. Presumably people volunteer because they want to help shed light on health issues of interest to themselves or others. What does the evidence show the prospects are for generating new knowledge capable of shedding additional light on a drug or device? Ample evidence about the prospects for clinical trials generating new knowledge make two things clear. First, the most likely knowledge to be learned from a trial is that the new drug or device does not work as well as hoped. For example, the failure rate of Phase I trials exceed 90% on average, while about half of Phase II and III clinical trials fail because of lack of efficacy, and another quarter fail because of safety concerns (Harrison 2016). Second, the chances are high that much, and possibly even all, of this knowledge, and thus the benefit gained by conducting the trial, will not be shared with others because the results very likely will never be published in full or possibly even in part (Riveros et al. 2013). Therefore, the “light” of new knowledge that volunteers are motivated to help shed all too frequently never is allowed to shine.
My argument in favor of disclosures about the prospects for a trial’s success is not made in order to suggest that high failure rates in terms of eventual approval of new drugs or devices render a trial somehow unimportant and thus not worth the contributions of research participants. Instead, the point is simply that since disclosures about the prospects for risks in a study are already made and since informed consent processes are supposed to help research candidates weigh possible benefits and risks themselves as they ponder whether to be exposed to risks or inconvenience, it makes little sense not to disclose information about the prospects for benefits that is comparable to the disclosures made about the prospects for risks. Not doing so is tantamount to saying that information about the prospects for risks is key information but information about the prospects for benefits is not. This imbalance in disclosures results in many, perhaps even most, research candidates entering trials not realizing how modest the prospects for success, e.g., increased knowledge, may be, given that the partial, let alone full, results made possible by their participation may never be published. Nor would they know the high likelihood that trials will not result in either advancement to the next phase of research or ultimate approval of the investigational drug or device, which is also information that many research candidates may deem relevant to their deliberations.
The need for Disclosures About Salient Features of Prior Research
Given the importance of proportionality between risks and benefits, research candidates also need disclosures about the prior research relied upon in assessing proportionality. This is especially so for early human trials. After all, the predictive value of informed consent disclosures about risks and benefits is only as good as the prior research that was used to determine them.
Can research candidates have confidence that the clinical trials they are being asked to participate in are based on solid science that establishes not that people think a new drug or device is going to work, because this obviously can never be established in advance, but that there is instead at least a rational prospect that it conceivably might? Absent this rational prospect, there can be no justification for studying either the safety or efficacy of a new drug or device in early trials of it. The evidence suggests such confidence is often ill-advised, but most research candidates will not know this if informed consent processes suppress relevant information about such evidence.
For starters, preclinical studies frequently produce false, as opposed to true, findings. This is due to the fact that most preclinical studies are really hypothesis generating, not hypothesis confirming, studies (Kimmelman et al. 2014). Thus, what can appear to be promising results may in fact be false findings that forecast the likely failure of a trial. This matters because trials can launch in the absence of the follow-up preclinical studies that would be needed to determine if the initial positive result suggesting the potential for benefit was a true positive finding or not. The failure to conduct confirmatory preclinical studies is but one of many well-documented concerns in preclinical research. Others include underpowered studies (Ioannidis 2005), as well as concerns about internal (Crossley et al. 2008), construct (Hackam and Redelmeier 2006), and external validity (van der Worp et al. 2010). Such evidence might shake the confidence of research candidates that the benefits of a study are in line with its risks so it calls into question current informed consent processes that never broach any of these well-known problems in the quality of preclinical research.
In addition to the quality concerns ascribed to preclinical research generally, additional problems with particular types of preclinical studies raise concerns that pertain in turn to specific types of clinical trials. Research candidates in cancer trials for example, would likely value disclosures pertaining to whether mislabeled cell lines were possibly used in preclinical studies since they can lead to erroneously launched clinical trials. Preclinical studies meant to investigate esophageal cancer, for example, may have actually been conducted on cell lines from a different type of cancer (Tadich 2016; Boonstra et al. 2010).Footnote 3 Evidence about external validity for certain classes of research can prove equally problematic (Pound and Ritskes-Hoitinga 2018). For instance, no drugs shown effective in preclinical studies in rodents to combat either Alzheimer’s disease ( Sarawitz 2016) or stroke (Dirnagl 2016) have ever shown any success in people. Surely some research candidates for such trials would find such information pertinent to their deliberations (Gotzsche 2014).
Similar concerns can apply to preclinical research for specific Phase I trials. Consider people who volunteer for them. Even though they are informed that the trials are meant either primarily or exclusively to study safety, one knows from the previously referenced research about unwarranted optimism and overestimation of personal benefit—a phenomenon that is commonly referred to as therapeutic misestimation (Horng and Grady 2003)—that many volunteers nevertheless believe that new drugs may help them.Footnote 4 What, then, might they make of the fact, if it was disclosed to them, that regulatory approval agencies such as the U. S. Food and Drug Administration (FDA) may have only reviewed safety data from preclinical studies (FDA)? In other words, what would research candidates make of the fact that drugs can enter human trials without any preclinical data supporting efficacy? As regards preclinical evidence about new investigational drugs that are approved with both safety and efficacy data, those data may diverge significantly from patient expectations.Footnote 5 For example, ALS patients may enter a Phase I trial with hopes for greater life-expectancy should the new drug be effective, yet the efficacy endpoints used in preclinical studies may not have looked at survival benefit at all, looking instead at matters such as gait (Hefferan et al. 2012).
Informed consent forms fully compliant with the disclosures historically required by regulations typically disclose no information related to any of this evidence, despite its clear relevance to decisions about joining a clinical trial. Regulations have not required disclosures about what is either known or unknown, either generally or specifically, about the prior research used to judge potential risks and anticipated benefits, nor why the information about that research is thought reliable enough for a REC to use (Kimmelman and Federico 2017). It is hard to square such non-disclosure practices with the weight of all the assembled evidence.
The need for Disclosures About the Potential Social Value of Research
Turning now to later phase trials as opposed to early ones, a new set of disclosure issues pertaining to a trial’s importance emerge. A requisite amount of social value has long been recognized as an ethical requirement for the conduct of clinical trials (Emanuel et al. 2000). Informed consent processes typically devote little more than cursory information about this value for any given trial. Regulations can stipulate that there be a description of benefit to participants and/or “others,” but they do not require any disclosures about the strength of prospects for that benefit.
This is problematic when one recalls that the proffer of an informed consent form to a research candidate is an implicit endorsement that it is important that the trial get conducted. Such endorsements are strengthened by the background frame effect (Charuvastra and Marder 2008; Bubela et al. 2009; Caulfield 2018) that results from dominant hyped discourses that situate clinical trials in the pursuit of breakthrough discoveries that can cure diseases and improve the quality of lives (Caulfield 2018). Yet, basic facts about many later phase clinical trials reveal the poor fit this background frame has with them.Footnote 6 Consider evidence pertaining to the various commercial purposes that drive countless clinical trials, and thus what research candidates are being asked to help make possible. These purposes rarely if ever get broached in informed consent processes. Evidence about two kinds of trials and the commercial aims behind large numbers of each illustrates why this is problematic.
The first type is known as head-to-head trials. They compare the effectiveness of competing drugs. Frequently, such trials are “non-inferiority” trials that study whether a new drug is “not significantly worse” than a currently available efficacious treatment (Palmas 2018). As researcher Walter Palmas who also writes about the ethics on non-inferiority trials describes, such trials, by design, tolerate some loss of efficacy compared to the current standard of care for the research candidates randomized to the test drug in hopes that this loss of efficacy will be offset by some improvement, such as greater tolerability, safety, or convenience. At times, “the clinical outcome of interest is severe, such as death,” highlighting both the ethical complexity such trials can have and the importance of their proper design (Palmas 2018).
When designed and conducted appropriately, such trials can have immense social value because they can tell us whether one of those drugs is actually substantially better than the other. Sadly, a recent study of head-to-head trials showed that most of the trials actually had negligible social value (Fiacco et al. 2015). The study consisted of a random review of 319 randomized clinical trials published in 2011 that were head-to-head comparisons of drugs and biologics. Most of them (82%) were funded by industry. Only 3 of the trials were deemed to be “truly antagonistic” comparisons that were capable of detecting actual greater social value. Nevertheless, the study found that all but 2 of the industry-sponsored trials reported favorable results that, as is discussed shortly, can be used to boost marketing efforts and future prescribing practices for undeserving drugs.
Seeding trials, which are another type of marketing study done to familiarize physicians with a drug, raise concerns similar to those of head-to-head trials. A recent study that looked at 194 drug trials, also published in 2011, in 5 leading medical journals documents this (Barbour et al. 2016). The study was carried out to examine whether trials might have been conducted in order to get physicians to prescribe a new drug. To see if this was so, the authors carried out a study to test their concerns about the frequency with which clinical trials get done in order to boost a drug’s marketability to physicians. They completed the study by looking at publications of trials that were suspicious for marketing purposes whose results were published in the New England Journal of Medicine, The Lancet, Annals of Internal Medicine, PLOS Medicine, and BMJ. They concluded that 41 of the 194 trials were found to be “suspected marketing.”
They did so because the suspected marketing studies tested drugs on patients from an average of 171 different geographical areas, whereas trials deemed not to be marketing trials recruited from just 13 different geographical areas. When one considers how much higher the number of prescribing physicians is who get exposed to the study drug in the trials suspected of being conducted for marketing purposes compared to the number of prescribing physicians in those studies not suspected of being done for marketing purposes, the true purpose behind the suspect clinical trials is apparent. Nor should one fail to note the fact that all the suspected marketing studies were industry funded.
Such evidence shows that people are being recruited into trials to advance the financial interests of the sponsors, since physicians who recruited their patients into the trials would be more likely to prescribe the study medication to their future patients after completion of the trial. This impact is greatly compounded when one considers that other physicians will read about the studies in prestigious medical journals and may change their prescribing practices as a result. One reason they may be more likely to do so is because the articles are often tainted by what is known as “attributional bias,” wherein the involvement of academic researchers is highlighted in the publication while the role of the industry sponsor gets downplayed (Matheson 2017). Another is that the studies are written about in such a way that the marketing influences on them are “difficult to identify by the average [physician] reader” (Barbour et al. 2016).
Current informed consent processes facilitate all of this by disclosing partial information, such as a “study is being conducted in order to see how well a new drug works,” while simultaneously concealing the overriding purpose of the study: produce data for high-profile reports that will be used to influence physician prescribing practices to choose one competing and potentially inferior drug over another. This concern is borne out by a recent study that looked at the informed consent forms of clinical trials designed to assess non-inferiority hypotheses in antibiotic trials comparing two competing drugs against each other. The study, which reviewed informed consent forms for 78 randomized trials that enrolled almost 40,000 volunteers, found that almost all of the consent forms failed to accurately describe the purposes of the study (Doshi et al. 2017).Footnote 7
All of this evidence is deeply troubling. At a minimum, hundreds of thousands of people (Fiacco et al. 2015) volunteered for all of these trials. These large numbers of volunteers are due in large part, no doubt, to two obvious features of the trials’ informed consent processes. First, the regulation-compliant processes failed to flag any information suggesting suspect design of the trials that would be needed to counter other information in the informed consent forms that implicitly endorses the need to conduct them. Nor should one assume that volunteers were informed that, by volunteering for the trial, they would be contributing to a sponsor’s aim to get study results published in high-impact journals where they would have an outsized influence on future clinical guidelines that will drive prescribing practices to the possibly inferior drug (Fiacco et al. 2015). Furthermore, it should be noted that non-disclosure about these kinds of commercial considerations is at odds with at least one published study showing that research participants want to be told this kind of information (Cook and Hoas 2011).
All this evidence begs for justification for continuing to use informed consent processes that sanction silence about it. Such justification would have to show that it is permissible to use research volunteers to pursue undisclosed commercial purposes even when this might lead, for example, to higher prescription prices for a non-inferior or, depending on the extent of flaws in trial design, a possibly inferior drug. It would also have to explain why these commercial purposes and potential personal cost and health outcomes are not among the “elements” of research that research candidates have an interest in knowing. Finally, it would need to show that partial, and arguably quite misleading, disclosures about a trial’s purpose that mask the major reasons why studies are conducted are somehow considered capable of helping research candidates make “understanding and enlightened” decisions as they ponder why they “might or might not want to participate” in the trial.
Counterarguments
No doubt some will raise questions about the expanded disclosures advocated here on the basis of the aforementioned longstanding concerns about the informed consent process itself. For instance, why call for more disclosures when no one disputes that informed consent forms are already too long and opaque? Others may question why disclosures about the matters highlighted here are thought capable of producing the desired effects, given what was mentioned before about what is known about the limited effectiveness of disclosure in general (Schneider 2005).
Such concerns mainly involve worries about the effectiveness of informed consent itself rather than the specific disclosures presented here for reforming it. So, due to space considerations, they will not be addressed in depth. Instead, readers can note that the informed consent changes being advocated need not result in voluminous disclosures. For example, a disclosure for a Phase I trial might read “Since this study is what is known as a Phase I study, there is only about a 5% chance that the drug will eventually be found to be effective.” A sample disclosure for an industry sponsored head-to-head trial suspicious for being done primarily for marketing purposes,Footnote 8 should a REC decide to approve such a trial despite those suspicions, might read “This study is what is known as a head-to-head trial that compares two different drugs meant to treat the same condition. This study is also being paid for by the manufacturer of one of those two drugs. Often with these kinds of trials, the company paying for the study designs it in such a way that it will produce results suggesting that its drug may be preferable, even though the trial’s design may not be capable of establishing this. They do this to boost the appeal of their drug to physicians in hopes that they will start prescribing it rather than one of the other competing drugs. We do not know whether this particular study is like that because the company paying for the study would not provide our committee with the information it would need to determine whether or not the study was designed in a way to produce misleading results.”Footnote 9
Readers should also entertain the possibility that, given the extent to which the information found in the advocated new disclosures is at odds with the background frame about research and the presuppositions that research candidates can bring with them to the consent process, these disclosures are likely to stand out in the informed consent process compared to other content in it. If so, the information might help to make headway against both therapeutic misconception (Appelbaum et al. 1982; Kimmelman 2007), which refers to the phenomenon of research participants failing to appreciate how research differs from clinical care, and previously referenced therapeutic misestimation, and thereby increase the effectiveness of informed consent.
There is a more germane concern about the proposed new disclosures that warrants a more fulsome response, which is that the informed consent process was never meant to endorse unlimited disclosures. That is why there is a threshold that needs to be met in order to trigger disclosures. For example, there is no need to disclose every conceivable risk that might occur during a clinical trial. Instead, thresholds of severity and/or prevalence often guide disclosure practices. Acknowledging the legitimacy of such thresholds, an explanation about why the recommended disclosures trigger them is in order.
Consider a hypothetical situation. What if hospitals did not check physician credentials, meaning that at times unqualified surgeons performed surgeries at hospitals? If a REC was reviewing a clinical trial that involved surgery, it would want to both guard against this risk and disclose it to research participants. Of course, in reality such possibilities are never actually disclosed by RECs because licensure and credentialing requirements render risks that impostors may find their way into the operating suite so remote that it would be silly to disclose such a rare occurrence. But the same assurances often cannot be made about the provenance/quality of prior research. Thus, the risk is far from remote that flawed prior research will be relied upon during the REC review process. Depending on the nature and severity of the flaws in the prior research and publications about it, harms similar in severity to those that unqualified clinicians can cause are possible.
To claim that prevalence of this risk falls below threshold triggers ignores the extensive evidence that exists about research quality. It would also mean that research candidates have no alternative other than to have blind faith that RECs can and will check the provenance of the research that was relied upon to assure that there is a proportionate risk/benefit ratio, even though such careful checking is currently outside the norm of most REC practices. This reveals genuine prospects that assessments about a trial’s value, risks, and benefits may be less reliable than research candidates would presume, which means that the research itself may be less valuable or more risky than appears, and that the risk/benefit ratios may be less favorable than a REC has determined. Such information is surely germane and thus should be counted within the “key information” needed to enable “understanding and enlightened decision[s].” The fact that RECs have not historically been specifically instructed by regulations to either inquire about or disclose such matters to research candidates does not trump the legitimate interests those candidates have in knowing about them.Footnote 10
Readers may grant the above points but disagree about the need for additional disclosures since consent forms routinely state that there may be “unforeseeable risks” in the research. Hence, little is gained by the proposed disclosures. But the risks that the proposed disclosures are intended to mitigate are foreseeable rather than unforeseeable and disclosures about them to research candidates will help them in their risk/benefit assessments. Simply telling people that there may be unspecified, unforeseeable risks masks the nature and extent of them, rendering any weighing of benefits against risks incomplete.
A final counterargument claims that the proposals set forth here are far too modest. This counterargument is suggested by prior commentaries concerned with the proverbial elephant in the room referenced at the outset. It contends that, rather than disclosure, what is needed is an end to REC approval in the first place of trials tainted by some of the ethical problems reviewed here. For example, in order to avoid the risks and waste posed by unnecessarily duplicative clinical trials, bioethics scholar Julian Savulescu and colleagues proposed in 1996 that RECs require investigators to conduct systematic reviews in order to determine whether or not there were prior studies that had already answered the questions to be investigated by a proposed trial (Savulescu et al.1996). Since then, bioethics and preclinical research expert Jonathan Kimmelman and colleagues have recommended that RECs “condition approval of any trial delivering putatively active drug doses on positive preclinical confirmatory studies” (Kimmelman et al. 2014). Most recently, philosopher Kirstin Borgerson has made an even more sweeping proposal by calling for RECs to only approve studies that exhibit “scientific integrity” (Borgerson 2014).
There is no doubt that these other remedies would prevent many of the problems highlighted here but, as I mentioned at the outset, they will remain in abeyance until such time as deep-seated practices, such as unreliable preclinical research practices or unimportant industry-sponsored trials, are reduced. Resolution of these matters is a massive undertaking the research community and governments have largely resisted to date. So long as this persists, and so long as current informed consent disclosures remain unchanged, then thousands upon thousands of people will continue to volunteer for clinical trials oblivious to the existence of the kinds of problems reviewed here that may be tainting any given trial. Reformed informed consent practices, that start with ethically adequate disclosures, is really the only viable option at this moment for guarding their interests.
The Common Rule “key information” and “reasonable person” informed consent revisions are the ideal catalyst for pursuing that option. The new disclosure standards could produce much needed transparency about the potential for the range of ethical deficiencies discussed here. And RECs are well-positioned to create this transparency. They are capable of, comfortable with, and good at gathering information, information that could subsequently be disseminated to research candidates through the re-imagined informed consent process envisioned in the Common Rule changes. So, persuading RECs to make to research candidates the disclosures advocated here should be an achievable goal.
The potential power of the transparency that would result from the new disclosures will improve the informed consent process by better aligning it with the original disclosure aims of the Nuremburg Code. It will also provide critical information to many REC members and compliance officials that the informed consent process is entrusted to. They likely are not sufficiently familiar with the evidence reviewed here and thus do not appreciate its effect on the trials they have historically been approving. This also means that they likely have not sufficiently considered to date the extent to which most research participants would value knowing this information about the trials they are volunteering for. No other explanation for why people would approve or enroll in such problematic trials seems plausible. Though empirical study would ultimately need to determine the scope and impact of the increased transparency that would result from the proposed new disclosures, it could very well improve current troubling portions of the clinical trials landscape. Again, the Common Rule revisions have set the stage for instigating it.Footnote 11
It must be noted, however, that transparency is no magic wand. Concomitant, or even preparatory, steps will be needed for it to achieve its potential. Probably the most critical is REC member education so that they can become familiar with the range and extent of issues discussed and alluded to in this essay and, more importantly, the ethical implications of them. This will be a large undertaking but at least part of the infrastructure needed to complete it is already in place. Many RECs have education and training requirements for their staff and members and these new topics could be integrated into them.