Complex Data Need Simplified Trials, Expert Argues
Clinical trials professionals will have to be laser-focused as emerging technology tempts them into a fool’s paradise of big data, a top industry analyst tells CenterWatch.
Machine learning, artificial intelligence and even consumer electronic apps are offering troves of data that would have been unimaginable even a couple of decades ago, says Ken Getz, an associate professor at the Tufts Center for the Study of Drug Development.
But if sponsors, sites and investigators aren’t careful, they can lose control of their studies and see knowledge dissolve in information.
“It is a paradox of our age — right at the time when our complexity has risen so that it’s having a negative impact on our performance, we’re at the same point where we can collect so much information on every aspect of our trials,” Getz said after presenting at the inaugural Metrics Champion Consortium conference last week. “It’s very tempting, isn’t it, to just collect more data? We’ll just let the machine do the thinking for us. And that’s also dangerous.”
Data and information technology have already made trials incredibly complex, he says. According to a Tufts analysis, the average trial has seen the total number of endpoints rise 86 percent since the beginning of the century. The number of eligibility criteria has risen by 61 percent, the total number of procedures performed in a trial by 70 percent and the number of countries involved in a trial has risen 100 percent.
It’s no coincidence, Getz argues, that drug development costs have risen as dramatically as trial complexity. The average price tag for developing a drug was around $1 billion in 2003, according to Tufts data; a decade later, the average tab had nearly tripled.
Perhaps most troubling is that barely a quarter of those costs have been directly tied to development, Getz says, noting that almost a fifth of them stem from time wasted sifting through irrelevant data.
The amount of “non-core” data has risen by 78 percent since the early 2000s, Tufts reports.
“It’s all around optimizing and utilizing the information really wisely and effectively,” Getz says. “In this age where there’s just this explosion of data, how do we use it really intelligently?”