Rock enrollment: Get a boost from health data
Company uses data to find good CT sites
The clinical trial industry's poor outcomes regarding enrollment and trial success are very frustrating to sponsors, as well as to investigators and CT sites.
"I was a sponsor for 17 years in big pharma and small biotechs, and regardless of where I was working or the target indication, I always was frustrated, like many of my colleagues, by the fact that I'd select sites, and they'd do very poorly," says Malcolm Bohm, BSc, MMedSc, president of Trialytics in Plymouth Meeting, PA. Bohm was scheduled to speak about applying health care data to site selection and patient recruitment at MAGI's 2009 Clinical Research Conference West, held Oct. 4-7, 2009, in San Diego, CA.
Trialytics, a subsidiary of SDI Health LLC, contracts with sponsors and others in the CT industry to provide and analyze health care data for the purpose of selecting CT sites and recruiting subjects.
In one study that is typical of sponsors' experience in the CT industry, only 320 out of 440 sites contributed to the study and also met the study's target goals, Bohm recalls.
"So you found that about 30% to 40% of sites basically were a waste of time," he adds. "And this is replicated over and over again."
If anything, the trial site statistics are getting worse, creating a tremendous burden for the CT industry, Bohm says.
It costs sponsors an average of $35,000 to open and close a site that doesn't enroll any subjects, he notes.
"Also, for every site that doesn't contribute, you need to find another one to make your enrollment," Bohm says. "Eighty percent of trials are late by at least one month."
From investigators' perspective, the statistics are just as grim.
"About 45% of investigators in the United States stop doing research every year," Bohm says.
Recruitment driven by data
One key to turning these dismal statistics around is to use available health care data to more efficiently identify and recruit clinical trial sites and to more easily recruit and enroll subjects, Bohm says.
"We use health care data from diagnoses, prescriptions, and procedure data from medical offices and hospitals around the country to determine whether a physician is treating the patients required for a protocol," Bohm says. "We go into our databases online and find, for example, who is treating patients who have intractable epilepsy."
Clinical trial sites can use the same data in a physician's office to identify potential subjects.
The data shows how many patients with this disease are being seen by a particular doctor, which is information a sponsor can use to identify and recruit that physician to be an investigator in a study about intractable epilepsy, Bohm explains.
"We're a niche patient recruitment organization," Bohm says. "We use evidence-based health care data as opposed to more traditional techniques, like media advertising."
Data-driven recruitment practices will be the way of the future if U.S. clinical trial sites are to survive and thrive.
"It's important to find a new recruitment avenue for this kind of enterprise because in the last decade 80% of trials are late, and 30% to 40% are unsuccessful," Bohm says.
While many health care organizations still lack electronic health records, 65% of them have some information technology, particularly in the area of electronic billing. And this is where useful data can be found, Bohm says.
The billing information is made HIPAA-compliant with patient level data that is de-identified, he says.
Sponsors, clinical research organizations (CROs), and clinical trial sites can use the data to make the clinical trial process more efficient and faster.
For example, a clinical trial site can use their electronic systems to conduct a detailed chart review to see if they have patients who meet criteria for a protocol, Bohm suggests.
Investigators can use electronic health care data from their own region to identify physician practices that could be excellent referral sources.
"There are 21,000 research doctors who treat a relatively tiny proportion of the patients available in the United States for clinical trials," Bohm says.
"A lot of trials are globalized and going overseas to recruit patients, but the patients are right here in the United States," Bohm adds. "We're just not doing a good job of asking them if they'd consider a clinical trial."
So if those 21,000 research physicians could use a database to identify doctors in their area who treat the target study populations, then they could send clinical trial staff or recruiters to these physicians' offices to discuss the study and request referrals.
"These are tools that are very simple, highly intuitive, effective, and cheap," Bohm says. "A research coordinator could go to another physician's office and offer to do patient chart reviews so the referring doctor doesn't have to do anything except say that the patient would be a good referral for the study."
Need extra time, money?
Using health care data to identify patients for recruitment could save time and money, Bohm says.
"The opportunity for using this health care data is huge compared with what people traditionally have been doing," Bohm says. "Clinical trials have been delayed and horribly inefficient for years."
But it will take a paradigm shift for study sponsors and sites to change their old habits, he suggests.
Sponsors recruit sites in the same old ways, regardless of how successful they are, and sites recruit patients often in the same old ways, he says.
Bohm estimates that a data-driven process of identifying CT sites, recruiting subjects, and starting trials could make the study three times faster. "And if you get them up and running faster, you have a chance of finishing on time."
For example, a recent chronic obstructive pulmonary disease (COPD) study used health care data for patient recruitment at 110 sites, which were identified through electronic information, Bohm explains.
"The sponsor held an investigator meeting and told sites, 'We're not giving you any patient recruitment money or support, and we expect you to deliver enrollment from your databases,'" Bohm says. "'That's why we selected you.'"
The study finished enrolling subjects four months early, he adds.
"The value to the pharmaceutical company was huge -- to have four to five extra months on the market," Bohm says. "If you're looking at improving enrollment and start-up, why wouldn't you look for alternative ways to conduct these critical steps?"
From sites' perspective, they'll make more money faster if they employ a recruitment process that has fewer screen failures and enrolls subjects faster, he adds.
"Say your target is to enroll 10 patients, and your break-even point is five patients, then you have an incentive to get to 10 patients more rapidly and efficiently," Bohm says.
The clinical trial industry's poor outcomes regarding enrollment and trial success are very frustrating to sponsors, as well as to investigators and CT sites.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content