Key Information

There is something new in the new Common Rule¹: something called “key information.”

Section__.116(a)(4) states: “The prospective subject or the legally authorized representative must be provided with the information that a reasonable person would want to have in order to make an informed decision about whether to participate, and an opportunity to discuss that information.”² Next, section __.116(a)(5)(i) explains: “Informed consent must begin with a concise and focused presentation of the key information that is most likely to assist a prospective subject or legally authorized representative in understanding the reasons why one might or might not want to participate in the research. This part of the informed consent must be organized and presented in a way that facilitates comprehension.”³

This new required component of (most) research consent forms has engendered a great deal of discussion and head-scratching among IRB professionals in the two years since it was enshrined in this final form in the final rule. What is it? How are investigators supposed to write it? How are their IRBs supposed to help them? What sort of “concise and focused key information presentation” will comply with this new requirement? And how does it relate to that new player in the final rule, the reasonable person?

Key information is discussed at some length in the preamble to the final rule. Let's see if that discussion helps.

The preamble first acknowledges that some commentators have argued “that consent forms have evolved to protect institutions rather than to provide potential research subjects with the most important pieces of information that a person would need in order to make an informed decision about whether to enroll in a research study. Instead of presenting the information in a way that is most helpful to prospective subjects — such as explaining why someone might want to choose not to enroll — these individuals argued the forms may function more as sales documents or as a means to protect against institutional liability. … [A] growing body of literature … suggests informed consent forms have grown too lengthy and complex, adversely affecting their ability to effectively convey the information needed for prospective participants to make an informed decision about participating in research.”⁴

This discussion clearly recognizes the well-known potential gap between regulatory compliance and ethics; it is, after all, quite possible for a consent form to meet all of the regulatory requirements, at least nominally, and nonetheless be too long, complex, and difficult to wade through to accomplish the goal of making an informed decision possible. Moreover, when the compliance-ethics gap appears, it is quite understandable for IRBs and investigators to concentrate on compliance, which is generally viewed as far easier to demonstrate to regulators than that variable, context-dependent, squishy ethics stuff.

To address these important concerns, the final rule makes “key information” a new regulatory compliance requirement — and then the preamble proceeds to emphasize its flexible and context-specific nature: “We recognize that how this requirement applies will depend on the nature of the specific research study and the information presented in the informed consent and believe that this requirement strikes an appropriate balance between facilitating the comprehension of subjects of key issues and allowing study-specific flexibilities. In general, our expectation is that this initial presentation of the key pieces of information will be relatively short. This section of the consent could, in appropriate circumstances, include a summary of relevant pieces of information that are explained in greater detail later in the consent form. The requirement that key information be presented in a concise and focused way will require an assessment that is specific to a study and its informed consent.”⁵

The preamble's discussion of key information continues by placing emphasis on the desirability of study-specific tailoring of the length, form, and content of key information — “This flexibility is responsive to public comments recommending against a rigid approach to enable institutions and individuals to tailor informed consents to the circumstances of particular studies.”⁶ It alludes to the possibility of future guidance. It then provides a list of items generally expected to be included in key information, and concludes with a pushme-pullyou-worthy compliance-vs.-flexibility maneuver: “As a general matter, a brief description of these five factors would encompass the key information most likely to assist a reasonable person (or legally authorized representative) in understanding the reasons why one might or might not want to participate in research, as required by §__.116(a)(5)(i) and §__.116(a)(4). However, we recognize that this determination is necessarily fact-specific and that IRBs and institutions may require that somewhat different (or additional) information be presented at the beginning of an informed consent to satisfy §__.116(a)(5)(i).”⁷

And what are the five factors that every investigator and IRB will want to be sure they can list concisely at the beginning of each consent form and thus be in compliance? Surprise: They are:

1. the fact that consent is being sought for research and that participation is voluntary;

2. the purposes of the research, the expected duration of the prospective subject's participation, and the procedures to be followed in the research;

3. the reasonably foreseeable risks or discomforts to the prospective subject;

4. the benefits to the prospective subject or to others that may reasonably be expected from the research; and

5. appropriate alternative procedures or courses of treatment, if any, that might be advantageous to the prospective subject.⁸

These five factors match almost perfectly the first four basic elements of informed consent at _6.116(b). It is not surprising, then, that IRBs might be somewhat confused about how the key information presentation is supposed to differ from the rest of the consent form, or what the preamble's acknowledgment that IRBs may require a different key information presentation might mean in practice. The preamble's discussion of what “concise and focused” means as regards consent form length is, similarly, both enlightening and somewhat opaque. It first describes the key information requirement at __.116(a)(5)(i) as simultaneously specific, detailed, and flexible, which is actually a pretty neat trick, and as striking a balance between facilitating subjects' comprehension of key issues and allowing study-specific flexibilities⁹ — which is an important balance to consider in practice, to which I'll return later.

Next, however, the preamble sets forth the expectation that “this initial presentation of the key pieces of information will be relatively short” — “a summary of relevant pieces of information that are explained in greater detail later in the consent form.”¹⁰ To illustrate that what “concise and focused” means is highly study-specific, the preamble offers the example of a complex oncology trial with a 20- to 25-page consent form, and states that a key information presentation of “no more than a few pages” would be appropriate, whereas 10 pages would not satisfy the “concise and focused” requirement: “[I]nstead of needing to mention every reasonably foreseeable risk, … this beginning section of the consent form should identify the most important risks, similar to the information that a doctor might deliver in the clinical context in telling a patient how sick the chemotherapy drugs will make them, but with a particular emphasis on how those risks are changed by participating in the study.”¹¹

At this point, the regulated community might be forgiven for being a little confused. Nonetheless, some academic medical centers immediately started developing key information models, genuinely searching for that balance between creativity and compliance. This is very much to their credit, but it also reflects the reality that the long path taken by the Common Rule revisions, from ANPRM to final rule, informed and was informed by two complex and ongoing dialogues: one between OHRP and the Common Rule agencies, and the other between OHRP and the regulated community.

An exhaustive examination of the massive body of Federal Register commentary along the pathway to the revised final rule is far beyond the scope of this essay, but it would probably yield insights worthy of someone's dissertation. As regards key information, though, research ethics scholars going back at least to Jay Katz,^{Reference Katz12} along with many in the regulated community, have long wanted to offer a short consent summary focused on the information most needed by potential subjects, but compliance concerns intervened. Thus, as a result of the discussion process, everybody knew that something like this was coming, and many stakeholders took — and continue to take — responsibility to help shape the final product.

The Reasonable Person

Now, at least, we know that research consent forms must be rethought, so that they can begin with “a concise and focused presentation” of “the key information most likely to assist a reasonable person … in understanding the reasons why one might or might not want to participate in research.”¹³ This passage from the preamble to the final rule specifically links key information to the reasonable person. But how did the reasonable person get in here?

The reasonable person is not really a new actor. Since the final Common Rule revisions were published, IRB professionals and others have been asking questions like “Who is reasonable? Who gets to decide that?” Those are actually the wrong questions. Instead, the reasonable person standard seems intended to parallel the application of the reasonable person standard to informed consent disclosure and decision-making in the clinical context, albeit fitted into the prospective context of research oversight. The reasonable person in the revised Common Rule is thus not directly related to the reasonable person as actor in ordinary negligence case law; it has a more complex pedigree, following from the reasonable person as decider in informed consent case law.^{Reference Capron14}

The reasonable person began as a hypothetical actor in negligence law and was transformed into a hypothetical decision-maker as part of the process of moving the failure to obtain informed consent in clinical medicine from battery to professional negligence. As exemplified in Canterbury v. Spence, subsequent case law, and state statutes choosing the reasonable patient standard over the professional standard of disclosure, this shift required the reasonable person to become a split personality, governing both (1) what information should be disclosed as material to the patient's decision whether to agree to what the physician recommended and (2) whether the patient's agreement would have been withheld if undisclosed information had instead been provided.^{Reference King15} When the reasonable person standard appears in the clinical context, then, it governs both the forward-looking determination of what information is material and thus should be disclosed, and the backward-looking determination of whether specific material information should be considered dispositive under the circumstances. Because the research context is entirely forward-looking, the reasonable person standard can apply only to what information counts as material, not also to determining what should be dispositive. But as __.116(a)(5)(i) provides, it also applies to the way material information is presented, in order to enhance the decision-making process of potential subjects in a specific way: toward understanding the reasons each might decide to join a study or not to join it.

Invoking the reasonable person in the research context thus resembles the shift in clinical informed consent law to focus on the decisions made by reasonable patients, rather than to consider only the actions and choices of their physicians. Applying the reasonable person standard to informed consent in research ought therefore to remind IRBs and study teams of an essential aspect of informed decision-making that is fundamentally similar in the clinical and research contexts: The consent form and process in both treatment and research are intended to support and enhance the patient's and the potential subject's decision-making capacity. Thus, reasonableness is not meant to be a fixed or testable characteristic of potential subjects. “Who is reasonable?” is not the most salient inquiry; “What is reasonable?” is far more important.

“Reasonableness” is a deliberate chosen term, distinguishable from “rationality” with its illusion of mathematical objectivity. “Reasonable” acknowledges not only reasoning and context but also the inescapably emotional components of human decision-making. Reasonableness is the central characteristic of the decision-making model that informed consent is meant to bring to fruition; indeed, the key question to ask is “Do this consent form and process help potential subjects understand the reasons on which to base their decisions about participating in this study?” All people faced with decisions about research participation need help to make reasonable sense of the choices they are offered.

The reasonable person standard is therefore a guiding principle about how to frame and conduct a genuine engagement with potential subjects, especially if they are also patients: “Here are the things we think you should think about; here is what's different about being in this research study from being treated for your condition; now let's talk about what's important to you, and what else might help you decide to join this study or not.” The reasonable person standard is the new Common Rule's way of turning the form and content of consent toward decision support, in order to enable potential subjects to say either yes or no to participation based on their understanding of the reasons they might want to participate or not. This necessarily shifts the content and form away from both institutional self-protection (with its extensive, exhaustive liability-limiting detail) and research promotion (featuring the subtle manipulations of word choice and emphasis designed to lead potential subjects toward saying yes and away from saying no).

How Key Information and the Reasonable Person Speak to Each Other

I promised to return to the preamble's argument that the concise and focused presentation of key information strikes a balance between facilitating subjects' comprehension of key issues and allowing study-specific flexibility. One of the most important and, arguably, groundbreaking aspects of key information is its focus on the reasons that potential subjects might have for deciding whether to join a study. This study-specific tailoring to what might be important to a given subject population permits and encourages consultation with the relevant population, and with the experienced subjects who, as Rebecca Dresser has eloquently argued, belong on IRBs and study teams and should be involved in human subjects research in a wide range of other ways.^{Reference Dresser16}

Yet investigators and IRBs must be aware of a potential problem that could flow from this kind of population-specific inquiry: What should happen if members of the relevant study population tell the study team, “You have it all wrong; we don't need to know A, B, or C. There's nothing good out there for our condition, so all we need to know is where to sign!” This response is far from implausible. The study team and the IRB could respond in several ways. They could say “Sorry, you need to hear items 1-5 in the regulations, whether you want to or not.” Or they could say “Okay, the regs say that we can be flexible, so let's leave all that out of key information for this study.” The first response might be regarded as paternalistic maintenance of the status quo: “We know better than you what you need to know.” And the second response both ignores the preferences of some minority percentage of reasonable potential subjects and leaves others vulnerable to the insight of hindsight when they learn about what they thought they didn't need to know.

So I recommend a third way of thinking about this: Key information ought to include information that reasonable people who are potential subjects should want, whatever other information they want.^{Reference Powers17} This formulation errs on the side of information inclusion rather than exclusion, but that default is acceptable because it nonetheless requires the IRB and the study team to consider carefully what information reasonable potential subjects should want and do want under the relevant circumstances. It does not endorse a comprehensive listing of everything in the preamble's five factors; instead, it mandates a genuine examination of the fit of the five factors to a given study, and then invites an examination of what else potential subjects may want to know under the circumstances. Finally, and perhaps most important, it endorses an ongoing process of mutual education among IRBs, investigators, potential subjects, and experienced subjects about (1) the reasons for including certain information and (2) good ways to explain the importance of included information.

I acknowledge the need to explain why this formulation is at all distinctive. Certainly, it is true that potential subjects want and need to know some things that are not always well addressed in consent forms. The list of such things could be quite long and is necessarily specific to the study and the subject population. But what makes “the information that reasonable potential subjects should want” any different from “We already know that you need to know the Five Factors”? The answer lies in the conundrum of the therapeutic misconception.

The Therapeutic Misconception

Potential subjects of biomedical or behavioral research are in many circumstances at least somewhat likely to confuse research and treatment, either because they are insufficiently aware of the difference or misunderstand the information provided about research participation, or because they are patients who are given — or at least not disabused of — reason to hope that the research intervention is likely to provide them with significant direct benefit. When Paul Appelbaum and colleagues first described the therapeutic misconception — that is, misconstruing research as treatment — they attributed this misunderstanding to all stakeholders, but focused primarily on the therapeutic misconception of research subjects.^{Reference Appelbaum, Roth and Lidz18} Historically, in order to minimize the likelihood of the misconception, many IRBs have preferred that consent forms significantly downplay their discussions of potential benefit, limiting them to vague and uninformative statements, such as “You may or may not benefit” or “Personal benefit cannot be guaranteed.”^{Reference King19} This vagueness, however, fails to explain what sort of benefit is at issue in a given study; thus, providing more detail about the nature, magnitude, and likelihood of direct benefit may be far more informative and therefore better able to correct misunderstandings. To give a simple example, “benefit” might mean reduction of symptoms, or it might mean “cure;” only a more detailed description of potential benefit can enable potential subjects to interpret it appropriately. Moreover, clarifying the distinction between research and treatment helps to inform even potential subjects who are convinced that the research is their treatment, by emphasizing uncertainty and explaining the data gathering that is intended to reduce it. In this way, more detailed discussion of potential direct benefit can help both to correct the therapeutic misconception and to minimize the temptation to overestimate direct benefit.^{Reference King, Henderson and Churchill20}

Countering the therapeutic misconception and the overestimation of direct benefit is increasingly difficult, however, for several reasons. First, emerging biotechnologies don't always demonstrate predictable dose-response relationships; thus it is often less than clear how the potential benefits and risks of harm should be described and discussed in first-in-humans and other early-phase research studies. A paucity of good preclinical data — attributable to many factors, including the lack of good animal models — contributes to this uncertainty, and can contribute to inflated expectations about study participation.

Second, changes in translation and study design have increased the likelihood of potential direct benefit in many trials. For example, research with a surgical component (which includes much regenerative medicine research) can be challenging to conduct on subjects who are not patients. Thus, enrolling only patients as subjects in early-phase research increases both the expectation of, and perhaps also the potential for, direct benefit. This development mirrors the challenges already posed by phase I oncology trials to accuracy, clarity, and honesty in the consent form and process.^{Reference Koyfman, McCabe, Emanuel and Grady21}

In addition, gene-based intervention research is increasingly conducted with children and other treatment-naive patients as the first subjects. This is in part because genetic interventions, if effective at all, are more likely to be effective early in the disease course, and effective levels are often easier to achieve in younger, smaller patient-subjects. Both of these factors also increase expectations of direct benefit.^{Reference King and Cohen-Haguenauer22}

Third, learning health care systems, community-engaged research, quality assessment and improvement activities, and system-based designs like cluster randomization, which bypasses individual decision-making, have all helped to blur the distinction between research and treatment, deemphasize information-sharing and the consent process, and suggest that most (if not all) health-related research interventions are equivalent to “new treatments.” Comparative effectiveness research offers one useful example of the difficulty. When two standard or commonly used treatments are compared in a clinical trial, it is tempting to characterize participation in that trial as without added risks of harm beyond what is known about both treatments, and thus as meriting little or no disclosure. However, a meaningfully specific consent form and process should at least explain that the treatment to which a patient-subject is randomized in the trial might be different from what the physician or hospital would recommend outside the trial context.^{Reference Feudtner, Schreiner and Lantos23}

Finally, both science journalism and popular media, increasingly in search of clickbait, are more likely to contribute to the hype about experimental interventions that are being studied than they are to support and promote understanding of the uncertainties that necessarily play a role in medical advances. Social media hype plays a significant role as well; for example, discussion of the Charlie Gard case in media and social media,²⁴ Change.org petitions for access to expensive and/or unproven interventions,²⁵ and GoFundMe efforts to raise the large sums of money sought by shady stem cell clinics^{Reference Snyder, Turner and Crooks26} all support heightened expectations of direct benefit that are often unwarranted.

For all of these reasons, an IRB's determination that saying less about potential benefit is preferable is not a solution for the problem of the therapeutic misconception. It could of course be argued that the therapeutic misconception is not, in at least some of these new circumstances, a misconception at all. Yet even if the distinction between research and treatment is genuinely blurrier now than it has ever been, it is still true that receiving treatment as part of a research study suggests some potentially important differences in data-gathering and use, and perhaps other relevant design features. That makes for something different from ordinary treatment, and warrants disclosure about those differences, at the very least — even if some of these differences are becoming ubiquitous as learning health care systems grow.

Information and Consent

What makes the research informed consent process, and its memorialization in the consent form, so important that it warrants the addition of key information? Well, for one thing, there is at least some evidence that the consent process follows the consent form.^{Reference Henderson, Davis and King27} It is unusual for a study team member to tell a potential subject: “We have to say that in the consent form, but you can ignore it, because here's the truth.” Instead, clear, cogent, and accurate information in the consent form shapes a clear, cogent, and accurate discussion in the consent process. It is worth remembering that informed consent encourages self-scrutiny by the physician-investigator.^{Reference Capron28} The mere exercise of putting clear information on paper and saying it out loud — not just mechanically cutting and pasting from the protocol into the consent form template — is an exercise of moral imagination and relationship-creation with research subjects.

Study teams and IRBs could learn even more from research subjects if they routinely asked them to contribute their views of the consent form and process. For instance, a simple exit interview or survey could help investigators learn a great deal about how consent forms are understood and about barriers to and facilitators of a good consent form and process.^{Reference King29} In fact, some research into assessing patient-subjects' research experiences has begun, using standardized surveys administered through patient portals.^{Reference Kost, Lee and Yessis30} More would be even better. It is possible to envision a standard starting question set, designed for short open-ended responses:

• Did the consent form and process prepare you for being in this clinical trial? Why or why not?

• Did you have other questions? What were they?

• Did you need/ask for more information during the trial? Why or why not?

• What other information could have helped you if you had known it sooner? Why?

• What information didn't you need? Why not?

• What should have been explained more clearly? In more or less detail?

• What were the most important things for you about the trial/your experience in the trial? Why?

• What were the least important things for you about the trial/your experience in the trial? Why?

• How would you explain this trial to somebody thinking about joining?

An exit survey like this doesn't directly inform future trials, of course. Still, it seems to me that learning how now-experienced research subjects view the consent form and process after participating could teach study teams a lot about how to approach the consent form and process for the next trial, whatever that trial may be.

SACHRP Key Information Commentary

Even if I have convinced some readers that the inclusion of key information makes sense, it's understandable that IRBs and investigators want guidance about it. So, in December 2017, the Office for Human Research Protections (OHRP) and the Food and Drug Administraiton (FDA) asked the Secretary's Advisory Committee on Human Research Protections (SACHRP) a set of questions, the answers to which could potentially form the basis of a joint guidance on the new key information requirement. The final version of SACHRP's commentary, which includes recommendations, was submitted to the Secretary of the Department of Health and Human Services in November 2018.³¹

The commentary tries hard to address the compliance-vs.-flexibility problem, emphasizing that SACHRP sees “these new consent requirements as providing an opportunity to fundamentally change and improve the consent process and the consent form” while admitting that “the best solutions are not immediately apparent” but that “the new requirements provide the regulatory mandate and the flexibility to test and implement substantial improvements.”³² However, in responding to OHRP's questions about what should be included in key information and how it should be discussed, SACHRP's answers are probably best characterized as “it depends,” as they focus more on encouraging the regulated community to be creative, in particular by organizing information differently, than on enumerating specifics.

The commentary emphasizes that the key information summary should be regarded “as an opportunity to orient, guide, and assist potential subjects in the decision-making process,”³³ and argues that it “should lead to new ways of organizing and presenting the required elements of consent, and … to the inclusion of new information … in order to best facilitate informed decision making.”³⁴ It concludes by noting the flexibility inherent in e-consent and similar new models, calling for ongoing empirical research and collaboration, and observing that “the Common Rule agencies and the regulated community have a significant opportunity to make the informed consent process better for research subjects.”³⁵ Thus, the commentary reinforces the message that creativity is welcomed, but this may not satisfy those in search of more specificity and certainty.