IC experts advocate improvements to process
But proposals in ANPRM may not entirely solve readability problems
In the proposed revision of the Common Rule, the Department of Health and Human Services (HHS) responds to years of complaints about informed consent documents – that they're too long, too complicated, too full of boilerplate and risk management language.
The improvements proposed in the revision of the Common Rule address all of those points:
— mandating that certain content be included in consent forms, with more detail than is currently required and that other content be removed;
— limiting the length of certain parts of the forms;
—providing more detail about how information should be presented (what information belongs at the beginning of the form, for example, or what belongs in an appendix);
— reducing locally required content that doesn't protect subjects and primarily is intended to shield the institution;
— proposing creation of consent form templates, giving institutions a format that they can be confident will pass regulatory muster.
Those who have studied the challenges of providing appropriate consent say these proposals aren't new.
Michael Paasche-Orlow, MD, MPH, associate professor of medicine at Boston University School of Medicine and a nationally recognized expert in health literacy, says he assisted in an initiative by the National Cancer Institute that is taking on many of the same goals, including creating templates for consent documents.
Furthermore, he says it's not even the first attempt by NCI.
"I'm friends with people who were involved with the NCI informed consent effort that happened about 15 years ago," Paasche-Orlow says. "And they had some of the same recommendations back then that they did this last time.
"The issue is how come no one followed their recommendations?"
One obstacle is the sheer complexity of modern clinical trials, says Mark Hochhauser, PhD, a readability consultant who serves on an IRB at North Memorial Medical Center in Robbinsdale, MN.
"It used to be that you would see a study that involved, say, an experimental drug vs. a placebo," Hochhauser says. "And you have a clinical trial now that has maybe three or four research parts to it. That's complicated and hard to understand even for IRB members."
He says it's hard to limit the length of documents when there are so many required elements that have to be communicated.
"You've got somewhere around 20 topics that have to be covered in the consent form. You can't ignore 15 of them and just do five. You have to do all of them."
He says that removing some of those elements to focus on what patients truly need to know would be an improvement.
Adding to the murk of an informed consent document is that many groups involved in its development often have goals in addition to informing subjects, Paasche-Orlow says.
"The organizational goals have to do with risk management," he says. "Until there's a much stronger hand from the regulators, the risk managers for the organizations are going to fear changing what they're currently doing. Because what they're currently doing is considered acceptable, appropriate and defendable."
Paasche-Orlow says previous efforts at simplifying informed consent had input from regulators, but the agencies were reluctant to give templates their official "approval." He says he's happy that the ANPRM provides for the creation of templates.
But he and Hochhauser say templates won't by themselves solve readability problems.
Hochhauser notes that a template can only provide a general outline of the information in a specific study.
"The problem is that the template will start off with some sample language and it's to be filled in by the researcher," he says. "That's where it all falls apart. It's the to-be-filled-in part that winds up being written in incomprehensible legal and regulatory language."
Paasche-Orlow notes that even when forms actually are written plainly, it may be hard to get certain concepts across to subjects.
As an example, he points to language he was testing to be included in an informed consent toolkit developed by the Agency for Healthcare Research and Quality. He was trying to find the simplest way to tell subjects they would not be compensated for injuries they might suffer in a study.
"It turned out that actually it was very common for people to assume that we were saying the opposite," Paasche-Orlow says. "Because, in fact, the assumption is, 'Of course, if I'm injured in your study, you would pay for that. How could you not pay for that?'"
Hochhauser and Paasche-Orlow say there are steps HHS could take that could improve the informed consent process.
"The form itself is just a piece of the puzzle. There are a lot of things that even with best language in your documents will need to be supported by a process," Paasche-Orlow says.
• Use real forms: Instead of templates, Hochhauser would like to see HHS take existing drug and device trial consent forms and simplify them, to give researchers a better jumping-off point for their own forms.
"I don't think a template is going to do it," he says. "People need to see real consent forms from real projects. Change any of the proprietary language, and show it in 'before' and 'after' versions — the before version, which is long and complicated and hard to understand, and the simplified version.
"And say, 'Feel free, if you're a researcher, to take this simplified form and adapt it so that it fits your needs.'"
• Allow for flexibility: Giving developers of consent forms flexibility to get their information across in different ways may not necessarily shorten documents, but can make them clearer. For example, Hochhauser argues for a table of contents in consent documents. Since participants rarely read the whole form, he says, it makes sense to direct them easily to the most important parts.
Paasche-Orlow promotes the use of tables and figures and images to help explain difficult concepts. But he's found opposition to adding such tools, because incorporating them would require further IRB approvals.
He notes that on the NCI project, he suggested creating simple software that would illustrate rates and percentages by generating a field of, say, five out of 100 dots to signify five percent.
"They said: 'Oh, that's way too complicated -- we'd have to have all the IRBs approve the software.'"
"When you're dealing with complicated concepts like percents and rates, which are broadly misunderstood, the adult education literature shows that there are ways to improve how we talk to people about these things," Paasche-Orlow says. "Really, the whole system has to be much more flexible and amenable to the use of various kinds of teaching aids."
• Evaluate understanding: The ANPRM asks whether, in certain types of studies, investigators should assess how well potential subjects comprehend the information before allowing them to enroll.
Paasche-Orlow says this would be an appropriate requirement, particularly in concert with a shorter form that focuses only on the most important information.
"This is not like some sort of memory test where you say, 'How long is the third study visit?' That's not the crucial component here," he says. "The crucial components are things like understanding that this is voluntary. That this will not impact their relationship with their doctor. These are the critical things.
"Having a shorter template can substantively help the process, by focusing on the ethically relevant ideas. And that would put us in a good position to say, 'And yeah, confirm comprehension of those things.'"