Federal Government's Proposed Expansion of Regulation of Biospecimen Research Should Be Reconsidered

The U.S. Department of Health and Human Services (DHHS) has announced a draft of reforms to the rules governing human subjects research. Earlier this year, President Obama announced that he would invest hundreds of millions of dollars in precision medicine to “lay the foundation for a new generation of lifesaving discoveries.” ¹ Meanwhile, a bipartisan group in Congress is working with the National Academies to streamline regulations governing scientific research. ² Regrettably, as noted by the DHHS Secretary's Advisory Committee on Human Research Protections in their detailed comments on the proposed rule, some of the DHHS reforms actually expand regulatory burdens on biobanking research, potentially making matters worse at a time when the rest of the federal government is seeking to accelerate research. ³ In particular, the new rule will require informed consent and committee review for secondary research with de-identified samples and may discourage the return of actionable medical information to subjects.

The “Common Rule” applies to most federally funded research involving humans and thus covers much of the human subjects research conducted in the United States. Fundamentally, it requires that patients provide informed consent for any research in which they will participate, and that an institutional review board (IRB) also reviews the project, although exceptions to each requirement exist. Prior research has shown that IRBs can impose substantial delays and heavy burdens on investigators, and that IRB decisions suffer from wide and unexplained variation, often treating the same cases differently.^4–6 In 2011, DHHS signaled that it was considering revisions to the Common Rule, and in 2015, it issued a formal Notice of Proposed Rulemaking (NPRM). ⁷ Given the issues researchers have experienced with the current rules, it is heartening to see DHHS attempting to improve the regulations.

The NPRM includes many reforms, ⁸ including the use of central IRBs, but perhaps the most important changes for precision medicine are oversight of biospecimen research using residual clinical samples. While some biospecimen research collects samples directly from donors, frequently it repurposes samples originally collected in a clinical setting. This surplus material is stored in biobanks for later, “secondary,” use by researchers. Such samples can be identifiable, or de-identified to further protect patients. With the growth of precision medicine, such research will only grow in importance.

This secondary research is distinct from most biomedical research—such as a randomized trial of a cancer drug—since it does not require interaction with human subjects and therefore poses no physical risks to them. Once a de-identified sample arrives in a biobank, the original donor is physically unaffected by whatever scientists may do with the sample. Indeed, courts most often hold that individuals relinquish ownership interest in their biological material once its clinical use is exhausted. ⁹ Nonetheless, courts have been willing to allow donors to enforce fiduciary duties or the conditions under which the donation was made, such as destruction of the sample after a certain use. Researchers have generally been free to include or omit such restrictions in the consent instruments, and the NPRM does not appear to alter that.

In fact, the Common Rule currently requires IRB review of consent documents when samples are being taken for research purposes, but the use of de-identified specimens has not been considered “human subjects research” at all, and thus, scientists need not secure informed consent or submit their research plans for review by IRBs. Brothers and Clayton have usefully called such research “human nonsubjects research.” ¹⁰ The NPRM would change the regulation of secondary research, moving it closer to the collection paradigm, which it leaves largely unchanged.

The NPRM proposes a “broad” consent template form to be used nationwide, rather than requiring individual investigators and their institutions to each create their own convoluted and variable documents, as they now do. This form would describe the biospecimen and private information that will be taken, inform the subject that they may withdraw consent if feasible, and disclose up to 14 other issues. At the end of the day, the net result of the regulatory change will be to add one or more sheets of paper to the stack of materials a patient signs before agreeing to a blood draw or biopsy. For many biobanks, it will simply be different boilerplate language on a form they already use. Still, better to use standardized language, however imperfect, than the varied and complex forms individual institutions currently utilize.

The NPRM also proposes to exempt newly covered secondary research from IRB review, as long as the template consent form is used to obtain the samples. IRB review will be limited to the process used to obtain that consent, typically a one-time examination of an entire healthcare institution's consenting policies, not burdening individual investigators. The exemption from full IRB review is a welcome reform.

Nonetheless, it is worrying to see it paired with new restrictions on the ability of IRB to grant waivers of consent where appropriate. Currently, IRBs may grant waivers of the consent to researchers where the NPRM would require “compelling scientific reasons” to grant a waiver, reasons that would be present only “in very rare circumstances.” ¹¹ We believe that the current waiver provisions are adequate and working well. The new conditions are unnecessarily onerous and should be removed from the final proposal.

It is perplexing that DHHS seeks to expand federal oversight over secondary biobanking, by overturning the longstanding principle that it is not human subjects research. We worry that this reform will not meaningfully resolve the sorts of problems that may have motivated the changes. On the whole, the NPRM expresses two rationales for this regulatory expansion. First, the government worries that it will become increasingly difficult to make data truly unidentifiable. ¹² It is true that donors face a residual privacy risk, although debate continues about its extent. ¹³ Other than requiring biobanks to have reasonable safeguards to protect privacy (which various laws already require), however, NPRM really does nothing to solve that problem. In fact, the language in the proposed broad consent form perpetuates the idea of “nonidentifiable data.” So this rationale is self-defeating.

Moreover, DHHS makes clear that the consent requirement will be imposed only prospectively. That is, samples collected before the promulgation of the broad consent form may still be used in future research without needing to secure a new consent from the donor, or undergoing IRB review. This is a regulatory compromise: while it preserves the usefulness of previously donated samples, the policy implicitly reveals that there is no overwhelming ethical problem with continuing use of samples that lacked original informed consent. DHHS would be better to wait and regulate problems as they emerge, rather than rush in now with an unsatisfactory compromise.

Second, the government says that, “a growing body of survey data show that many prospective participants want to be asked for their consent before their biospecimens are used in research.” ¹² We worry that polls and focus groups are not ideal domains for deliberative and informed decisions. As the Secretary's Advisory Committee on Human Research Protections stated, “public attitudes cannot be the sole determinant of public policy in this complex domain.” ³ Policymaking requires a broader perspective.

DHHS may be motivated by particular controversies, such as the Havusupai case, which involved genetic research on samples taken from members of a Native American tribe in Arizona. ¹⁴ In this case, which led to litigation, there was allegedly a regrettable mismatch between subject expectations and researcher intentions. The forms in the Havusupai case provided a broad consent to “study the causes of behavioral/medical disorders,” while the subjects said that they believed the samples would be used only for diabetes research. Importantly, this prominent case involved an explicit consent form and benefitted from IRB review. It is not clear at all that NPRM efforts to expand the use of consent forms and IRB review will avert a future such controversy, or protect the larger public legitimacy of biobanking research.

Another plausible motivation for the reform is a noted recent book about the HeLa cell line, based on cells taken from an African American woman named Henrietta Lacks. ¹⁵ That book has also captivated bioethicists and the broader reading public, with its compelling narrative of profound racial and social inequity, contrasting the poverty of Ms. Lacks and her family with the billion dollar biomedical science complex. While it is objectionable that Ms. Lacks did not sign a consent form decades ago, it is implausible that such a piece of boilerplate paperwork would have sanitized the ethical concerns in her case. Such a document would not possibly have informed Ms. Lacks about the potential value of her cells, which we now know only in retrospect.

Worse than being ineffectual, the NPRM expansion of the definition of human subjects research could backfire. By establishing the principle that even de-identified secondary research requires informed consent, and then requiring a standard all-encompassing consent form, the NPRM invites people to wonder why they were not truly informed about the ways that their samples would actually be used. The same polling data cited by the NPRM shows sizeable numbers of respondents saying that they want to be recontacted and asked for consent for each study that will use the biospecimen, which may have been donated a decade previously. ¹⁶ And some ethicists suggest that individual consent to each study is the only way for subjects to be truly informed. The NPRM gives credence to these preferences, but the proposed boilerplate form does not satisfy them. Nor could it: it simply is not foreseeable today what scientists may want to do with our biological samples a decade hence. We need to rely on smart governance and regulation, not the ritual of informed consent.

We are pleased that the NPRM avoided the urging of many ethics commentators to take a more prescriptive approach to informed consent. ¹⁷ These ethicists argued that the consent should contain built-in limitations on the use of the sample, based on expert opinions and poll results about what sorts of future research some people may or may not like. Such a static approach could have set back scientific research in these areas that seemed controversial at first, many years before the science would actually be done. This method would also impose some individuals' ethical concerns upon donors who may or may not share them. The NPRM's broad form thankfully avoids these problems.

Another concern: the NPRM reduces any incentive that investigators have to de-identify samples—since the work will be regulated as human subjects research either way. From the public health perspective, research on identified samples is advantageous, because it may produce incidental findings, which are medically actionable if returned to the donor or her physicians. In this study, however, the NPRM creates another problem. Under the proposed system, to remain exempt from the vagaries and burdens of IRB review, investigators must also plan not to return research results to individual donors. We worry that this creates a perverse incentive for investigators to not return results, even if they would be medically actionable. This is especially concerning given the ongoing vibrant debate about what to return, when, and how to return it. ¹⁸

Just as the NPRM proposes a standardized consent form, we suggest that the final rule also create, or defer to a specialized body to create, national protocols for return of results. ¹⁹ Although such protocols could not cover the full variety of research activity, investigators who work in areas covered by these standard protocols would not need to submit to the busy and unpredictable review of local IRBs.

Ultimately, we agree with the NPRM that, when clinicians or researchers take a biomedical specimen, they should notify patients that it may be used for research purposes generally, and give the patient the option to decline. The reason is distinct from that offered by the NPRM, however: it is because the patient has a right to decline consent for the giving of that sample, while it is still hers and situated in her body. Later, however, once the specimen has been alienated from the person, and especially if de-identified, the concepts of “informed consent” and “human subjects research” are inapposite.

The NPRM represents an encouraging attempt to update human subjects regulation. However, it should be revised before being finalized. Secondary biobanking research need not be swept into the ambit of human subjects regulation and IRB review. Instead, DHHS should simply require informed consent for the taking of samples, as is already done in research settings and is typically already done in clinical settings. DHHS should also provide a national mechanism for return of actionable medical results, which will not itself unduly burden investigators or create rampant local variation.