Matthieu Legrand, MD, PhD, who has been a principal investigator on numerous traditional clinical trials, believes in their value. “But sometimes,” he says, “those trials take far too long to affect daily clinical practice, which has such a big impact on outcomes.”
That concern has spurred Legrand’s interest in what are known as pragmatic trials. Conducted in daily clinical practices with typical patients and relaxed eligibility requirements, pragmatic trials reach a broader group of patients and compare multiple patient management strategies in real time so clinicians can adapt to how a population is responding.
In collaboration with colleagues in the UCSF Department of Anesthesia and Perioperative Care, Legrand is currently developing pragmatic trials aimed at more precisely defining how best to use vasopressors in the operating room and, in the future, in intensive care units.
Accelerating Treatment Advances
A practicing intensivist and anesthesiologist, Legrand’s more traditional research has focused on the development and validation of diagnostic biomarkers for sepsis and acute kidney injury. The two conditions, which are so prevalent in the ICU, present complex decisions about the use of vasopressors, vasodilators and fluids.
“It’s important to think about how we apply these treatments, in which order and how to adapt our strategies to how patients respond,” says Legrand. He, Romain Pirracchio, MD, PhD, and their teams at the Peri-Operative and Critical Care Outcome Research Group – whose work combines electronic health record and other point-of-care data with bio-statistical analyses and machine learning in the hope of transforming and personalizing patient care at the bedside – will develop the pragmatic trials across all sites where the department practices.
“We are clinicians first, deeply anchored in clinical practice, so our goal is to quickly and reliably translate research into clinical practice,” says Legrand. “We hope the addition of machine learning will enable us to become super clinicians, who integrate a lot more information than we could before so we can keep learning and safely testing different strategies while practicing standard of care medicine.”
Some Unique Challenges
Pragmatic trials bring with them challenges that standard trials do not. For example, Legrand says that in pragmatic trials it’s important to weigh the noise introduced by broadening inclusion criteria against the ability to much more quickly gain insights into which population is more likely to respond to what treatment.
Machine learning introduces another challenge. “Studies using machine learning in these settings studies raise complex technical, ethical and psychological questions,” says Legrand. “That’s why we need to give clinicians freedom about whether to follow – or not follow – the recommendations generated.”
And because these trials take place in everyday practice and require that providers with diverse expertise participate as investigators, it is essential to work collaboratively with colleagues throughout the process.
“That’s why we explore questions that providers believe will have an impact on their patients,” says Legrand. It is also why he and Pirracchio have consistently sought input from colleagues as they’ve made choices about what to study and how to do it. In return, they’ve been thrilled with the level of enthusiasm for proceeding.
Building on Tradition to Change The Game
None of this is to say Legrand has given up on standard clinical trials.
“Standard clinical trials will always be important, but they typically take a decade or so to complete. I’m convinced that the pragmatic trials we’re proposing can be done within a year or two and will add an enormous amount to our knowledge and our ability to refine treatment strategies quickly,” he says. “In that sense, I believe these trials can be game-changers.”