It is the investigator’s responsibility to confirm that potential subjects substantively understand what they are being asked to undertake. If comprehension cannot be confirmed, the person should not be enrolled.
—Michael Paasche-Orlow, MD
On February 24, the Institute of Medicine National Cancer Policy Forum convened a workshop, “Contemporary Issues in Human Subjects Protection in Cancer Research,” in Washington, DC.
In his introduction to the workshop, Steven Piantadosi, MD, PhD, Director, Samuel Oschin Cancer Institute, Cedars-Sinai Medical Center, Los Angeles, said, “In the years since implementation of federal regulations governing clinical research, the number of clinical studies has grown substantially, trials have become more complex, and multicenter trials are common. In addition, the increase in biobanking and genomic analysis has changed the landscape of clinical research, which raises important questions for institutions that conduct research and the patients who participate in it.
“The clinical trials endeavor is in an unquiet state,” he commented.
Advanced Notice of Proposed Rulemaking
Holly Taylor, PhD, Associate Professor of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, and Core Faculty Johns Hopkins Berman Institute of Bioethics, Baltimore, reviewed a number of current issues related to the ethical conduct of human subject research. These included the status of the Advanced Notice of Proposed Rulemaking (ANPRM) issued by the Department of Health and Human Services and Office of Science and Technology Policy recommending changes to the Federal Regulations that guide the ethical conduct of human subject research (45 CFR 46). The notice was posted in July 2011, with the following key topics:
Protecting research subjects from privacy breaches by adopting Health Insurance Portability and Accountability Act (HIPAA) provisions as a universal standard. Specifically, de-identified data collected for non-research purposes does not require consent other than that obtained before data collection. However, data collected for research, with or without identifiers, requires consent.
Calibrating institutional review board assessment according to level of risk. Expedited review was to be revised and simplified including the elimination a requirement for continuing review.
Using a single institutional review board for all sites of domestic multisite studies, with an aim to address the inefficiency of multiple reviews of the same protocol by many local institutional review boards.
Creating a more systematic approach for collection and analysis of data about unanticipated problems and adverse events.
Extending federal protection to all research, regardless of funding source, and harmonizing agency regulations and guidance documents.
The American Association for Cancer Research, Association of American Cancer Institutes, and ASCO responded to these proposed policies. They supported ANPRM in general but were concerned about the need to delineate responsibilities between an external institutional review board and requirements for the local institution to oversee researchers, adoption of HIPAA as the universal rule for privacy protection, and consent requirements in regard to de-identified data.
Their recommendations included: redefine what is and is not “human subjects research,” adopt a new category of “excused” research and institute procedures for implementing it, and forgo consent for use of de-identified data.
The White House’s Office of Management and Budget was expected to issue a revised ANPRM by October 2013 for review by 17 federal agencies. But that hasn’t happened yet and may be stalled for the foreseeable future.
HIPAA Is Everywhere
HIPAA is ubiquitous, and its presence in medical research is no exception. According to Melissa Bianchi, JD, Partner at Hogan Lovells, HIPAA allows protected health information to be used for research with subject consent. A few important notes:
1. Under the new changes to the HIPAA rule, consents no longer need to be specific to the research study. Instead, a HIPAA consent now may permit future research as long as the future research is adequately described such that the research subject would reasonably expect that his or her PHI could be used or disclosed for that purpose.
2. It does not confer compound authorizations (corollary activities such as biospecimen banking) unless the consent clearly distinguishes among various research components, provides opportunity to opt out, and does not involve psychotherapy notes.
Cancer researchers and sponsors, said Ms. Bianchi, now have considerably more flexibility to use protected health information. Nevertheless, HIPAA still takes identification of data seriously, and there are only two ways to de-identify them (ie, to render data anonymous). First, a statistician must determine that the risk of identifying an individual is very small. Second, 18 specific identifiers must be removed, known as the “safe harbor” method.
Once protected health information is de-identified, it is no longer protected health information. It is not subject to HIPAA, and it may be used for any legal purpose.
Limitations of De-identification
According to Bradley Malin, PhD, Associate Professor of Biomedical Informatics, Vanderbilt University, Nashville, however, de-identification is not a panacea. There is always a risk of re-identifying people, and any system can be broken into. The challenge, therefore, is to determine an appropriate level of risk and to ensure accountability for the data.
“We must be reasonable and practical,” he said. “We need to ask how people can be identified, and if they are, what harm can come of it.”
Alice Leiter, JD, Policy Counsel, Health Privacy Project, Center for Democracy and Technology, Washington, DC, distinguished between research and operations. The latter, she said, includes quality assessment and improvement, outcomes evaluation, and population-based studies to improve health, reduce costs, and develop protocols, provided that “obtaining generalizable knowledge is not the primary purpose.”
She added that the Center for Democracy and Technology believes that HITECH focuses disproportionately on the identifiability of data and whether consent is required. “De-identification is an important protection tool, but it is not infallible. Also, consent, though important, tends to be weak in practice.”
Other ways to protect privacy, she suggested, include a more risk-based framework in which publication of study results is not treated as inherently risky, less reliance on consent and more on other models of patient engagement (such as input into the research and sharing results with them), new mechanisms of accountability and oversight, and creation and maintenance of public trust in research.
Informed Consent Problems
Informed consent is a mess and often doesn’t do the job it’s intended for. Institutional review boards require too much unnecessary information, investigators use too much medical jargon, cooperative groups have made consent forms too long and complicated, and industry adds meaningless legalese.
Terrance Albrecht, PhD, Associate Center Director, Barbara Ann Karmanos Cancer Institute, Wayne State University, Detroit, noted that the process of obtaining consent for a trial should take the time necessary to be patient-centered and to have a meaningful, helpful dialogue with the patient and the patient’s family members or companions.
She said that their research has shown that the consent process falls far short of the ideal. For instance, participants often didn’t know that they were consenting to research, not treatment; confidentiality was not explained; nor was the need for a signature. Patients may not have been told that participation was voluntary and that they could withdraw at any time. They may not have understood randomization or what publication means, or even the purpose of the study.
Laura Cleveland, Patient Advocate, Cancer and Leukemia Group B Alliance, agreed that consent forms are too long and overwhelming, form readability levels vary greatly but are generally too high, and there is often no one with the patient while he or she reads the document.
‘Consent Is a Process’
Michael Paasche-Orlow, MD, Associate Professor of Medicine, Boston University, echoed Ms. Cleveland’s remarks and deepened the general criticism. “Literacy levels in this country are usually far below what is demanded of potential research subjects. What’s more, consent is a process, not merely a form presented to potential subjects. Because subjects comprehend far less than is desirable, we ought to shorten and simplify the forms. Even then, readability is only part of the problem; the vocabulary is confusing, if not downright incomprehensible.”
Jeffrey Botkin, MD, Associate Vice President for Research Integrity, University of Utah, Salt Lake City, discussed why some of the problems exist and persist. “Sponsors want to make information about risks and benefits as comprehensive and accurate as possible. They also want to reduce legal liability and develop consent forms they think will be acceptable to institutional review boards. Investigators have the same goals, and it is easier and less time-consuming to write long, complex documents than short, simpler ones.”
Institutional review boards are not subject to regulatory requirements about consent forms, so they too may take the easy way out and overload the forms with institutional boilerplate in legalese.
“Potential research participants are culpable as well,” said Dr. Botkin. “They trust clinicians and investigators to make decisions for them; they’re often embarrassed to admit they don’t understand. They may be afraid of angering or disappointing their doctors—so they say nothing. Moreover, the public is poorly educated about medical research, and patients tend to be passive and dependent, fearful and anxious.”
Dr. Paasche-Orlow suggested a technique called teach-back. In this way, investigators ensure that potential subjects understand not only the parameters of the research itself but also what their participation means. Saying things like, “I want to make sure we have the same understanding about this research,” or “I would like to hear your understanding of the research project,” can provide an idea of what’s going on in the subject’s mind.
In addition, Dr. Paasche-Orlow recommends asking the subject questions about what has been explained, and letting him or her look at the form. “It’s about comprehension, not memory,” he noted.
“Listen for parroting and technical terminology, and if it’s there, explain further,” he continued. “Ask open-ended questions such as ‘Tell me in your own words what this research is about, and what it means to participate,’” he said.
The investigator should correct misinformation until he or she is certain the subject understands. Moreover, the investigator should take responsibility for the information imparted (“I haven’t explained things clearly”) and should never blame the subject for lack of understanding.
“It is the investigator’s responsibility to confirm that potential subjects substantively understand what they are being asked to undertake,” Dr. Paasche-Orlow told The ASCO Post. “If comprehension cannot be confirmed, the person should not be enrolled.”
He added, “When it comes to the consent process, we need to help our research assistants shift from a model of persuasion to a model of pedagogy.”
Ethics and Genome-Based Research
“The cost of sequencing a single genome has dropped from $10 million to under $10,000, so genomic research such as multiplex panels for cancer susceptibility and genomic tumor profiling has been accelerating,” said Angela Bradbury, MD, Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
This creates new problems, just one of which is what to do with individual research results. “As large prospective cohort studies with banked DNA have become more common, there has been increasing debate over the obligations, if any, to share [individual research results] with research participants,” said Dr. Bradbury.
There are reasons for and against, she continued. Among the latter are the danger of blurring the distinction between research and clinical care, potential for misunderstanding or overinterpretation of clinical significance, compromising privacy, and cost. Among the former are benefiting subjects and respect for their autonomy, as well as acknowledging the increasing interdependence of research and
And sometimes it’s not clear. Some results have been confirmed and are clearly actionable; that is, a clinical intervention is possible. However, deciding what is and is not actionable and has clinical or other utility is not so clear-cut.
Gail Jarvik, MD, PhD, Head and Professor, Division of Medical Genetics, University of Washington School of Medicine, said that an actionable gene is a clearly deleterious mutation for which specific evidence-based interventions can improve health. However, regardless of the results that turn up in the course of research, informed consent should address what is to become of the results, and participants should know that.
In fact, one study revealed that 90% of research participants want individual research results even if no action can be taken. Moreover, 50% to 70% of people who spoke with a genetic counselor wanted to receive individual research results.
Dr. Bradbury said that more thought needs to be given to determining when and if researchers are obliged to return individual research results, how much, if any, depends on context, or if it is permissible at all. Then, which results should be returned, who covers the cost, and is lab confirmation necessary?
Principles and Recommendations
Dr. Jarvik and her team have devised a number of principles and recommendations about return of results to research subjects. These include:
Research differs from clinical care in terms of goals as well as procedures. As a result, the information returned may differ, but clinically valid actionable information resulting from research should be offered to subjects.
Participants have the right to refuse the results—those related to the study and incidental ones—unless the study requires return of results. This should be understood and agreed to at the start.
Resources for research should be directed primarily at scientific discovery. Therefore, there is no duty to look for actionable genomic findings beyond those discovered naturally in the course of an investigation. Researchers should offer genomic results that are valid, medically important, and actionable.
Researchers may be ethically and scientifically justified in returning all genomic information, but care should be taken when benefits and harms are uncertain. There is as yet no best way is to return results, but research ought to attempt to discover the benefits and harms to subjects as a result of receiving genomic information. ■
Disclosure: Drs. Piantadosi, Taylor, Albrecht, Paasche-Orlow, Botkin, Jarvik, and Bradbury and Ms. Bianchi reported no conflicts of interest. Dr. Malin is a paid consultant for Sanofi and Celgene.