Disaster Planning Primer
Katrina survivors offer tips on preparedness
Think of all possible scenarios
At least for the next few years the clinical trial industry will view three names as turning points: Katrina, Rita, and Wilma.
The three major hurricanes to hit the Southern United States in the late summer and fall of 2005 will be cited as the reasons sponsors have begun requesting clinical trial sites to provide standard operating procedures for disaster planning, some experts predict.
"Because of the hurricane disasters we’re suggesting clinical trial sites could do more to prepare," says Alicia A. Pouncey, MEd, managing director of Aureus Research Consultants of Metairie, LA. Aureus Research performs contract monitoring of clinical trials.
"My fear is the sponsor will require more disaster preparedness and that will add more hours to a person’s week," Pouncey says. "And the person having to carry it out is the research coordinator, who doesn’t have enough time anyway."
Ideally, the sponsor and IRB will share in the responsibility of preparing for disasters, Pouncey says. "The challenge we saw in the Hurricane Katrina disaster was the fine line between a company that wanted data and the one concerned about protection," Pouncey says.
An unexpected consequence of a disaster, such as Hurricane Katrina, has been the reality of decreased revenues coming in to an institution, says Kenneth Kratz, PhD, director of the Office of Research Services and manager and co-chair of two IRBs at Louisiana State University Health Sciences Center in New Orleans.
"Obviously, much of our clinical trials and other sponsored research have been put on hold because the investigators either don’t have some place to perform the study or there isn’t a subject population out there, and there are no new patients to enroll in the study," Kratz explains.
When principal investigators have been proactive and found a new location to conduct their studies, this means the research office is called to help them draft amendments to their original IRB application, Kratz says.
How will you exchange information?
Pouncey, Kratz, and Tom Davis, PharmD, chief executive officer and president of Odyssey Research Services, a site management group in Bismarck, ND, offer these suggestions for clinical trial sites to consider when drafting a disaster plan:
• Anticipate major communication problems.
As the aftermath of Hurricane Katrina showed, communication can be a profound problem following a disaster. The only way Kratz and his staff could communicate, since most telephones, cell phones, and Internet access were down, was through text messages, Kratz says.
"All communications networks related to the city weren’t functioning. So if you had a New Orleans area code the communication was extremely unreliable and calling those numbers was very difficult for a number of weeks," Kratz recalls. "But we could send text messages to people on cell phones."
Meantime, the university’s buildings in New Orleans were flooded on the first floors and the servers located there were inaccessible. Nonetheless, the university set up an emergency web site with rudimentary capability, Kratz says.
"We have e-mail service finally re-established in rudimentary fashion after three weeks, and that was a big hurdle," Kratz says. "After four weeks we could dial 504 area code numbers."
It took several weeks for the university to find adequate space for the research office, Kratz says.
"The most frustrating thing is that in one of these situations everybody is spread out all over the country," Kratz says. "So when I was able to establish some communication over the Internet and by telephone, we put up notices on the emergency web site, asking investigators to contact me with regard to the status of their projects."
It’s a good idea to establish a voice mail system with out-of-state service prior to, or immediately after, the disaster, Pouncey suggests.
"I think that’s part of the planning process, deciding how to communicate with employees post-disaster," Pouncey says.
Staff can use the voice mail number to let sites know where they are located and how they can be reached, and the same toll-free number can be given to subjects and patients, she says.
Another communication strategy is to designate prior to the disaster, when it is possible, where employees will meet in the event that all telephone and Internet communication is impossible, Pouncey says.
When a hurricane or storm knocks out power and telephone service, but does not impact road travel, this strategy gives everyone an opportunity to meet and discuss in person the next steps they will take.
Likewise, trial sites should find out in advance where each employee plans to go during the evacuation and what their contact information will be, Pouncey adds.
It’s important to communicate frequently with staff and allow employees to talk about what happened to them, Pouncey says.
"We held a teleconference staff meeting three to four weeks after the storm," Pouncey says. "I was communicating with people daily, whenever they could get to me and to them."
At the teleconference, staff shared their new addresses, described their living arrangements, and provided contact information, she says.
"It’s some level of comfort to re-establish communication with peers and colleagues," Pouncey says. "A week after the storm I was [evacuated] to a daycare center, and someone sent me their time sheet from last week."
Odyssey doesn’t have a back-up answering service in the event of blizzards because the company has never had a telephone service disruption, Davis says.
"We have locations in Asia, South America, and China—all connected," Davis says. "We use an IP phone system over the Internet, and we can connect with Argentina by dialing four digits, so if Odyssey’s phone lines go down, we still have the ability to have some communication with the rest of our sites."
• Make data and supply protection a priority.
Clinical research organizations have large numbers of samples and massive amounts of data on site, Pouncey notes.
"At a Phase I site, you might be the only site and study, so if you lose that data, you lose the entire trial," Pouncey says. "But if you lose data from five or 10 sites out of 500, it won’t have the severe impact it would have had in a phase I study."
One phase I facility in New Orleans had samples on site and wasn’t able to move them before the hurricane, she recalls. "But they maintained those samples via generator and even risked life and limb to get them out after the disaster, when they retrieved them on dry ice and shipped them to a second location," Pouncey says.
A clinical trial site could have data backed up at a secure server off-site, but the site would have to make provisions for secure access and privacy considerations, Pouncey says.
"All the sites we currently work with appear to have paper data intact, some because the records were stored on the second floor," Pouncey says. "We have not been able to determine the status of their electronic records."
However, the supplies that were kept on site during the hurricane posed a problem, she says. "None of the sponsors will allow a drug sitting there during the storm to be used," Pouncey says.
"A number of sites are doing studies with controlled substances, and when you have a disaster with the mayhem of New Orleans, do you want controlled substances sitting there in a facility?" Pouncey says. "They may be in lock and key, but what kind of challenges do these present to a site?"
Since the supplies on site were unusable, clinical trial staff have to manually reorder supplies and explain the situation to suppliers, Pouncey adds.
The IRB files at Louisiana State University Health Sciences Center were converted to a new software system last January, and so they were using a version of the program that was housed on a server in New York, Kratz says.
"We’ve been accessing all IRB information over the Internet, and that’s been extremely helpful for us," Kratz says. "We were able to get the PIs’ names and contacts."
If the institution starts a clinical trials office, part of it would include having a management software package that could act as a repository of subject information, and this could be housed in another state, Kratz notes.
Odyssey’s computer systems are located in various locations and are constantly backing up each other, Davis says.
"If one is flooded out, the other should be OK," he says. "And we have a calling tree for notification of individuals who need to be called if there’s a system problem."
Also, the company has access to institutions that have generators so if there is a prolonged power outage, company officials can make arrangements to have their supplies and samples at those locations, protecting them from being ruined, Davis says.
• Consider privacy issues when storing data and contacting subjects.
In the event of a disaster as devastating as Hurricane Katrina, some privacy practices and rules seem inconsequential.
For example, a cancer center at Louisiana State University had a large patient database on a server located in a downtown building. After the hurricane, the staff couldn’t obtain access to the server, Kratz says.
"What this may argue strongly for is that at least in our situation we might want to think about having a large central repository of information and have it housed on a server someplace else," Kratz says.
"I’ve always been reluctant to suggest this because of privacy issues and the potential for compromising the privacy of individuals, but this experience argues strongly that we might want to go in that direction."
A lot of people might say that if all data were electronic then the problems associated with the recent hurricanes wouldn’t happen, Pouncey says. But sponsors shy away from discussing storage of patient identifiers, largely because of HIPAA, Pouncey says.
Before Hurricane Katrina, no one was forced to think about how sites would communicate with subjects after a disaster, Pouncey explains.
Now it’s important for sites to include this in their disaster plan.
For example, one strategy might be to post a message on the weather channel, prior to a hurricane or blizzard, that asks subjects to contact a clinical trial site as soon as they can after the storm, Pouncey suggests.
Then if the power goes out, sites could place radio ads, because many people will have battery-operated radios, she says.
The operating principle should be to do the right thing, putting subjects’ safety ahead of regulatory privacy issues, Pouncey says.
"These are individual people who were giving to us by providing their bodies for us to observe, and HIPAA is there for their protection," Pouncey explains. "But let’s protect that individual body in this situation, which might be more important than protecting their identity."