Don’t be confused by the rules governing adverse event reporting

Rules confusing you? Here’s some advice

Reporting rules for adverse events in clinical trials only seem straightforward in the case of death or injury requiring hospitalization. Otherwise, those involved in clinical trials have to figure out how soon they need to report, what information needs to be included in the report, and who should get it.

There is help available. Take the on-line tool developed by Jeremy Wood, PhD, a communications consultant based in Merion Station, PA. Using a series of yes and no questions, the tool — located at — guides users to a determination of who needs to know about an event and how soon, based on the severity of the occurrence. For instance, a several-step inquiry about a trial involving an investigational new drug (IND) that results in a serious and unexpected adverse reaction results in the following final message:

"According to the SAEtool, the study sponsor should inform FDA [the Food and Drug Administration] and all participating investigators about this serious and unexpected adverse experience in a written IND safety report. The notification should be made as soon as possible and within 15 days of the sponsor’s initial receipt of the information."

At every step, there is a link to any relevant federal regulations — in this case, the Code of Federal Regulations, Title 21, Volume 5, or the FDA’s April 2002 regulations on investigational new drug safety reports.

Wood says he developed the tool after doing work for an institutional review board at the University of Pennsylvania. "I was asking people about what forms people involved in research might need, and the answers were always along the lines of, That’s a complicated question.’ I did a tool to help sort it all out," he says.

Some of the users of the tool have had Wood link it to the specific forms researchers would need to fill out given a specific set of events. It is one more step that larger institutions are taking to simplify the many steps researchers have to go through and do it in a paperless way.

Another area of trouble involves those adverse events that may not be unexpected and may not be serious, but are important for researchers to note, particularly if they happen more frequently than those researchers expected. Oftentimes, such reports do not come to the attention of researchers for months.

A group of cancer researchers at the Mayo Clinic in Rochester, MN, developed a way for quick — indeed, nearly real-time — reporting of these events. Initially reported in the Journal of the National Cancer Institute, the tool was reprinted most recently in the Journal of Clinical Oncology.1

While there are some common-sense suggestions that could probably be extrapolated to any study, one of the creators of the tool, Daniel Sargent, PhD, director of cancer center statistics at Mayo, is quick to note that this was developed for cancer trials.

That said, Sargent notes that there are a wide number of reporting requirements for adverse events. With INDs in Phase I trials, every single event has to be reported, "even if it’s just a sneezing fit by one person," he says. "But if you get to a Phase II trial of an existing drug, only life-threatening events resulting in hospitalization or death are required."

But research by Sargent and his colleagues turned up events that were expected, not individually isolated, but occurred at greater frequency than was expected. "If things are happening a lot more often than you thought it would, even if it’s not fatal, we need to know about it and take action."

They developed a single-page form that could be faxed in — or filled out electronically — that includes the kinds of information the researchers need. "The existing expedited reports were cumbersome," says Sargent. "We didn’t need that level of information."

Among the information included on the tool:

  • the protocol number;
  • the patient ID number and initials;
  • the date the clinical research associate was made aware of the event;
  • the dates of the event occurrence;
  • the type of event;
  • whether it was a grade four or grade five event;
  • whether the event was, in the opinion of the reporting party, related to the study medication;
  • whether hospitalization was required;
  • if yes, the date of admission;
  • the reason for admission — toxicity (of study medication), prophylactic, or other.

The information goes into a central database daily, and a program generates an automatic e-mail to the study chairperson, the data monitor, and the statistician about any new reports. "It’s real-time toxicity monitoring for us," says Sargent. "Every day when I open my e-mail program, I get a report. If there are no events, I don’t get any e-mail. If there are events, I can get the information and determine what, if anything, we can do. It takes days to react, not months."

As an example of the kind of event he is talking about, Sargent suggests a subject who gets diarrhea that wasn’t bad enough to cause hospitalization, but because it was six or eight loose stools a day, it had the potential to be life-threatening. Or a patient who had a low blood count that wasn’t accompanied by fever or chills. "If too many of our patients have low counts, then it’s only a matter of time before one develops a fever and we are in real trouble. This helps us be aware in a much more timely manner."

The tool has certainly helped in the three years it has been in use. "We know that 75% of the relevant events are reported within three days of occurrence, and we know about them within five days," he says. "Compliance isn’t perfect, but it is good and getting better as we continue to do training."

Investigators have been very positive about the tool, and the National Cancer Institute has commended Sargent and his peers on the tool several times, he says.

"It also has a real impact on the way we do things," Sargent concludes. "There have been six studies in the last three years where these reports have resulted in protocol changes. In five of the six cases, we were able to reduce the toxicity of the drug. In the sixth, we were not, so we shut that trial down."


1. Goldberg RM, Sargent DJ, Morton RF, et al. Early detection of toxicity and adjustment of ongoing clinical trials: The history and performance of the North Central Cancer Treatment Group’s real-time toxicity monitoring program. J Clin Oncol 2002; 20(23):4,591-4,596.