Compliance Corner

QI program manager cites common mistakes of CTs

Watch documentation, protocol amendments

A good way to find weak areas in clinical trials (CT) research is by gathering information through a quality improvement (QI) program about what CT sites have done wrong.

After reviewing 60 different studies as part of an educational and quality improvement program, the QI program at Children's Hospital Boston in Boston, MA, highlighted some of the most common types of CT research problems.

"When I do a review I say, 'I want to see warts and all, all the mistakes,'" says Eunice Yim Newbert, MPH, manager of education and quality improvement program at Children's Hospital Boston.

QI reviewers gain a pretty clear picture of where clinical research educational needs come up short once they've analyzed the findings, she says.

Newbert wrote a report about the first 60 reviews for the hospital's administration, and it's this analysis that provided information about trends, including the things CT sites do wrong.

Here are some examples of how CT sites can avoid the more common mistakes:

• Document protocol deviations: "If a site missed the subject's visit that's something they need to document and explain," Newbert says.

For instance, if a CR coordinator missed a data point on a laboratory sample, then the documentation should note the reason why it was missing. An answer might be: "It's missing because the subject left early," Newbert says.

"If the principal investigator decided that it's not worth doing the blood draw, and that decision is in the best interest of the patient, then this should be explained," she adds. "If you don't document it somewhere it doesn't make sense."

Study documentation should be written clearly so that anyone who reads it doesn't find gaps in the narrative or errors.

"If you fill in the gaps then there won't be questions at the end of the study," Newbert says. "You don't want to have questionable data, so fill in the gaps, cross-out errors, write memos to the files, and remember what the hospital's deviation policy is."

• Amend protocols when mistakes keep occurring: Sometimes the problem isn't with the person who deviated from the protocol, but with the protocol itself.

For example, Newbert recalls reviewing a study in which a survey given to participants often was not completed.

"We did an investigation and found they had five long surveys to do at every visit, and a lot of [participants] couldn't finish all of them," Newbert says.

This could be handled as a protocol deviation and documentation could explain how each participant who failed to finish the survey had run out of time. Or the overarching problem could be addressed instead.

"We said, 'Now that you have so many of these deviations, rather than document 20 deviations, why don't you amend your protocol,'" Newbert says. "We try to offer a simpler solution to a problem and explain the benefits of doing it this way."

So CT investigators changed the protocol to allow families to bring some surveys home, and investigators obtained permission from the IRB for this change, Newbert says.

Often, it's the principal investigators will think the deviations are minor and won't impact their study outcomes.

But if it's okay to have incomplete surveys 20 out of 100 times, then the question has to be asked, "Do you need this entire survey at all?" Newbert says.

"And if you do need it, then let's find a simpler way to get this done," she adds. "Sometimes you can't make a protocol change, but it's worth investigating."

• Improve informed consent process: A common mistake among CT sites involves the informed consent documentation and process, Newbert says.

For example, CT sites often keep expired informed consent forms around, and these could be used mistakenly.

The solution is to keep the most recently approved informed consent (IC) form in an electronic library. This library or documentation form site should be the only source for copying new forms.

"We don't make 100 copies of the informed consent form and stockpile them," Newbert says. "We have people go to the library by typing in the protocol number and principal investigator's name and then make one or two copies as they need them."

It's when people say they think they have a copy in their binder and they use the IC form they find in there that they often make mistakes because the binder might include a number of versions of the IC form, she explains.

CT staff could keep a shortcut to the documentation library on their desktop computer, and the library Web site could be accessible from any computer, Newbert adds.

"This process has dramatically reduced the number of expired informed consent forms we see," Newbert says. "We still ask for a hard copy for the regulatory binder, but for new forms we ask them to go to that Web site and copy them per subject."

Another part of informed consent documentation that can cause problems involves the signature on the IC form.

In pediatric research, the parent or guardian signs the IC form. But sometimes, the person who is bringing the child in for medical services is not the legal guardian. And if the IC form doesn't specify who the adult signing the form is there could be a problem, Newbert says.

"A lot of people forget to specify the relationship of the parent or guardian because of the assumption that the person is a mother or father," Newbert says. "But a lot of times it is the [grandmother] or aunt and, in one case, the stepmother."

Although the stepmother or aunt or grandmother might have raised the child, this doesn't mean they are the legal guardians, she explains.

"We tell the research staff that you need to document who the person is who signed the informed consent form and then make sure that person is the legal guardian," Newbert says.

• Match signatures and dates: Another IC documentation issue involves making certain subjects and guardians both sign and date the IC form.

The subject's signature and the date signed should be accurate. Sometimes the subject will forget to sign the form, and someone else will add a date that is earlier than when the person could have signed the form, Newbert explains.

"So people write in the wrong date and no one pays attention," she says. "That's the kind of small things we come across."

Or perhaps the subject's parent accidentally signs the witness line and not the guardian line, and the principal investigator says he knows that it is the parent who signed the form, Newbert says.

"But if a parent signs the witness line, and someone is looking at this form three years later, no one will know that this person is the parent," Newbert explains. "So you need to write a note to the file to show that you recognize that the person who signed the witness line is the parent."

This extra documentation clarifies the situation and minimizes the number of questions someone might ask down the road, she adds.