The trusted source for
healthcare information and
Helping break the logjam: What IRBs can do
Think like a business; gather data on problems
In trying to streamline the IRB process, institutions should look to other industries as an example, says David Dilts, PhD, director of the Center for Management Research in Healthcare at Vanderbilt University in Nashville, TN.
Dilts, who has been studying the process of activating clinical trials, says he's struck by how little many IRBs know about their own activities — how long they generally take, why things are sent back to investigators, etc.
"IRBs will know what happened to a particular study, but not to the collection of studies," Dilts says. "For example, if I was to go into an institution and ask 'Can you tell me the number one reason why things get sent back from the IRB? Using hard data.' The answer is usually no, I can't tell you. That's phenomenal to me.
"I'm an old manufacturing person and it's like saying, I run a repair shop and 60% of all the cars that come in here have a broken fender but nobody ever asks if fenders are bad. I know what happened to a particular car, but I have no overall global picture."
IRBs often can't say for sure whether the changes requested are more scientific in nature or just to address administrative or clerical concerns, he says. IRB officials often think they know what the problems are, but lack hard data to back it up.
"I believe in evidence-based medicine," he says. "If you're the head of an IRB, collect evidence on exactly why things got sent back. Were the top five things dealing with truly scientific issues or were they dealing with administrative issues? And if you can't answer that question, how are you going to get any better?
Beware of 'scope creep'
Other ideas IRBs should consider to help smooth the path of research:
• Looking for the value added. Dilts says IRBs should scour their processes for steps that don't add value to the final study and simply waste valuable staff time. What would happen if those steps were eliminated or automated?
"There are some activities that you need to have no matter what, just because they're part of the logistics of the system," he says. "But you need to ask yourself, why exactly are we doing all this paperwork? Why do we keep doing these steps? And often the answer is either A) because we always have or B) this is a really important step because this is what I do."
• Consider the customer, in this case, the investigator. IRBs should be able to provide information that investigators can use to make their plans. Dilts notes that an IRB often can't tell an investigator how long to expect the process to take.
"If I'm doing this really straightforward placebo-controlled design, should I expect it to go through the IRB one time, two times, three times?" he says.
"It's okay to say, 'A study that includes X is going to take a long time.' What I object to is the fact that an investigator has no way to know how long it's going to take to go through the system. If you ordered a product from a department store and they say it might be in tomorrow or it might be in three months, it all depends. How in the world do I plan for 'It depends?'?"
Another suggestion would be to have someone review the submission package ahead of time, in order to help an investigator with easily-fixed problems before the IRB sees it, says Norman M. Goldfarb, managing director of First Clinical Research of Palo Alto, CA, and editor of the Journal of Clinical Research Best Practices. "Don't just wait until the meeting and then say, "Your package is broken."
• Be wary of 'scope creep,' or the tendency of an IRB to delve into the science of a protocol under review rather than focusing on human subjects protection issues.
Dilts notes that IRBs do have to consider the science, because it would be unethical to sign up subjects for a study that's rooted in bad science.
"There's a fine distinction between whether it's a good study or whether it's a study you would or wouldn't do," he says. "If it's a good study, then let it go."
• Have a global perspective. Dilts and Goldfarb both say that IRBs need to consider the larger clinical trials process and try to work better within it.
For example, Dilts notes that different organizations often have different systems for tracking a study.
"How many different tracking numbers are there on this trial?" he says. "You've got a scientific review number, you have an IRB number, you have a budget number — I can understand that. But show me the central cross-listing of all those numbers."
Goldfarb says that because slowdowns in one part of the clinical trial process can affect other parts of it, all the players — the IRB, contracting, budgets — should make information available as to where a study stands, so others can catch up if they're falling behind.
Dilts says that as one part of the process may improve, bottlenecks may occur in other parts.
"If I fix the IRB problem, I might then be constrained by the grant," he says. "If I fix the grant problem, I might then be constrained by the case report forms. So what happens is as you systematically fix each of the problems, it gets better, but not by the magnitude of fixing the one problem itself."
Ultimately, the IRB needs to be part of the team, from a business standpoint, Goldfarb says.
"Obviously you have to keep yourself separate from an approvals point of view, but from a business process point of view, you need to be working together to make the whole process as efficient as possible," he says.