The rapid spread of COVID-19 has sparked an unprecedented global search for new treatments. In March, the U.S. Food and Drug Administration launched an emergency program to find new treatments for COVID-19. The FDA now oversees some 70 active clinical trials, with more than 200 in the pipeline.
At Stanford, the FDA has approved trials of several promising therapies. Biostatisticians across campus have played a crucial role in designing many of these trials, working closely with the Stanford Medicine to create robust research plans that will satisfy the stringent demands of the FDA.
“I’ve consulted on many clinical trials over the years,” said Tze Lai, the Ray Lyman Wilbur Professor of Statistics. “The difference between COVID-19 trials and previous consultations is the urgency, because everybody wants an answer quickly.”
To get FDA approval, investigators must present a reliable research plan that exposes as few participants as possible to experimental treatments that may cause harm. Statistics can help achieve that goal by providing researchers mathematical tools to help validate their hypotheses, eliminate biases, and determine the sample size and endpoint of the trial.
Since February, Lai and his colleagues have consulted on a half-dozen COVID-19 trials at Stanford, including a proposal to treat patients with angiotensin-(1-7), an anti-inflammatory peptide found in healthy human beings.
“Angiotensin-(1-7) helps reduce inflammation of the lungs, heart, and other tissues” said Kevin Grimes, a professor (teaching) of chemical and systems biology. “But the coronavirus blocks the enzyme that makes angiotensin-(1-7), taking it out of commission.”
The inability of infected patients to produce angiotensin-(1-7) could increase their susceptibility to serious lung diseases, such as acute respiratory distress syndrome, the leading cause of death in COVID-19.
“My hypothesis is that restoring angiotensin-(1-7) peptide to people with serious infections will help them recover,” Grimes said. “That’s why we’ve proposed a clinical trial to administer the peptide to hospitalized patients with low oxygen levels who are not yet on a ventilator.”
To help design the trial, Grimes turned to Lai and his colleagues at the Stanford Center for Innovative Study Design (CISD).
“They’re phenomenal,” Grimes said. “I asked them if they could help us. Within a day I had five of them on a Zoom conference. That says a lot about what drives them.”
Before submitting a proposal to the FDA, investigators must first win approval from the university.
“Stanford has its own rules,” said Lai, co-director of CISD. “We help the investigators write the protocol, which is a very important part of the process, because it has to go through institutional review.”
Using statistical analysis, the CISD team provided Grimes and his colleagues recommendations on key design issues, including which patients to target and how many would be sufficient for a randomized clinical trial. They settled on 100 active patients and 50 on placebo or control.
“Human beings are a very challenging research subject,” said CISD co-director Philip Lavori, emeritus professor of biomedical science data and, by courtesy, of statistics. “With COVID-19, there is pressure on the FDA to get fast approval for drugs. But we need to get the right answer, not just a fast answer.”
With the help of the CISD biostatisticians, the proposed design for the angiotensin-(1-7) clinical trial received university approval and has been submitted to the FDA for review.
“Not having input from Phil, Tze and their team would have made this process much more challenging,” Grimes said. “They’re brilliant statisticians. Every time I work with them, I thank God we have them as a resource at Stanford.”
To read all stories about Stanford science, subscribe to the biweekly Stanford Science Digest