Biostatisticians draft blueprints for COVID-19 vax trials

Hutch group applies skills honed during decades of HIV prevention trials to bring about safe and effective vaccines for a new pandemic
Dr. Peter Gilbert (left) leads a team of Fred Hutch biostatisticians that pivoted to COVID-19 vaccine trial design. Dr. Holly Janes (right) is a co-principal investigator with Gilbert at the Statistical and Data Management Center for the HIV Vaccine Trials Network.
Dr. Peter Gilbert (top) leads a team of Fred Hutch biostatisticians that pivoted to COVID-19 vaccine trial design. Dr. Holly Janes (below) is a co-principal investigator with Gilbert at the Statistical and Data Management Center for the HIV Vaccine Trials Network.

Photos by Robert Hood / Fred Hutch News Service

A small band of mathematically inclined biostatisticians has quietly carried out an enormously important transformation at Fred Hutchinson Cancer Research Center in Seattle.

For years they have served at the heart of the Hutch-based HIV Vaccine Trials Network, or HVTN, on the unfinished quest to find a vaccine for HIV/AIDS, a pandemic that has taken 33 million lives since 1981.

In coordination with vaccine makers, this team of seasoned Fred Hutch biostatisticians led by Dr. Peter Gilbert develops blueprints for clinical trials, monitors the execution of those plans, and calls the results without fear or favor.

Then came the pivot to COVID-19.

Since last spring, they have been architects of clinical trials poised to validate or dismiss a parade of new vaccines that are the world’s best shot at stopping this new pandemic. Among that team’s admirers is their colleague Dr. Jim Kublin, executive director of HVTN, where Gilbert was named the network’s co-principal investigator in 2004, and has been lead statistician since 2011.

“I’ve been working really closely with biostatisticians for 25 years, and I’ve come, in many ways, to think like they do, and for that I’m fortunate, ” said Kublin, a researcher on vaccines for malaria, tuberculosis and HIV.

He calls the biostatistical team within HVTN “Peter & Co.,” and credits them for building an international reputation for excellence — advancing the field with new statistical methods while applying them in practice in the effort to stop HIV. They are now bringing that experience to bear on the COVID-19 vaccine effort.

Real-time analyses of disease severity

Six months ago, COVID-19 vaccine researchers were concerned about the potential for “vaccine-associated disease enhancement,” an immune response that can result in making a person actually more vulnerable to infection. In the 1960s, its occurrence sank development of a vaccine being tested to block respiratory syncytial virus, a common respiratory illness in children. And it showed up in some animal studies researching a vaccine for SARS, the deadly cousin of COVID-19 that appeared in 2003 — so the concerns were real.

Gilbert and his team built into the COVID-19 vaccine trial designs a series of tests for a “real-time analysis” of disease severity that would spot the problem early if it existed, and they found no sign of it — a huge relief for the global vaccine development effort.

In Kublin’s view, a biostatistician’s most important job for the testing any new vaccine is to produce a lengthy document called a Statistical Analysis Plan, or SAP, which is written for every clinical study.

Dr. Jim Kublin
Vaccine researcher Dr. Jim Kublin is executive director of the HIV Vaccine Trials Network. Photo by Robert Hood / Fred Hutch News Service

“It’s the blueprint of how the study will be analyzed, in all its gory details,” he said. “It’s text and tables, figures and formulas. It says very specifically how this and that will be done, what data sets will be used, what tables will be generated. It lays out exactly what they are going to do, so there is no question about the rigor of the analysis.”

In the esoteric world of biostatistics, where such things really matter, Gilbert’s team is renowned for — yes — their SAPs.

“It’s because of the leadership and expertise that goes into their fantastic SAPs, which they have applied to other clinical trials worldwide, that the Hutch is regarded as such a premier computational and statistical center,” Kublin said. “They also exercise their creative capacity to think outside the guardrails of the SAP to come up with fresh exploratory work. That can lead to radically new ideas.”

HVTN’s co-founder and principal investigator is Hutch professor and acclaimed virologist Dr. Larry Corey. The network is supported by the National Institute of Allergy and Infectious Diseases, headed by Dr. Tony Fauci, who has been Corey’s friend and institutional collaborator in HIV research since the 1980s.

Since its inception in 1999, HVTN has conducted 75 trials, large and small, aimed at developing an effective vaccine against HIV. No vaccine yet has overcome HIV’s frustrating ability to evade it, but the work carries on.

When COVID-19 burst onto the scene in last winter, and Fauci was looking for a way to run massive, credible and rigorous trials of potential vaccines, he turned to Corey and the experts at HVTN for help.

'Harmonizing' COVID-19 vaccine trials

During that frenetic spring, Fauci, Corey, and National Institutes of Health Director Dr. Francis Collins hammered out a plan for the COVID-19 Prevention Network, or CoVPN. The fastest and most efficient path was to bring together the veterans of HIV vaccine studies at HVTN to coordinate the big company-run trials needed to test a variety of unproven coronavirus vaccines.

The strategy was to “harmonize” trials of as many as seven different vaccine makers, asking a common set of questions, setting a common set of goals, and measuring that success with a common set of tests carried out by independent laboratories. Time was of the essence.

By June, CoVPN was up and running, with Corey directing the operations center overseeing vaccine trials. HVTN staffers flipped their primary assignments to COVID-19. Peter & Co. were tasked to work with drugmakers to assure different vaccines made with different technologies could be tested and fairly compared. Bringing their hard-won HIV vaccine expertise to bear on a new pandemic, the HTVN crew transitioned to trials that are bringing hope to a jittery world.

“What the HVTN statistical center has done for so long is really very similar to what we are doing for COVID-19,” Gilbert said in an interview from his home, where he has been working since March. “It’s just a pivot, and that experience helps me kind of see what’s needed, so I keep rolling with it.”

In November, mirroring the performance of the first COVID-19 vaccine — developed independently by the Pfizer-BioNTech partnership — the first CoVPN trial of the similar Moderna vaccine showed an astonishing 94% efficacy.

pull quote photo

'I just had some kind of fire to work on this disease, and then it was just kind of natural to meld math with that, because that is what I knew.'

— Dr. Peter Gilbert

“As someone who has specialized in genetic diversity my whole career, I’m used to HIV vaccines failing, because the vaccine is too far away from the viruses. These vaccines are on top of the virus,” Gilbert said. “We definitely knew we were in a totally different realm than HIV.”

In late February, a new single-dose vaccine developed by Johnson & Johnson will likely win an endorsement for an Emergency Use Authorization by the US Food and Drug Administration, which in December gave the go-ahead to the Moderna and Pfizer dual-dose vaccines.

Ultimately, drug companies are responsible for the designs of their studies. While Pfizer charted its own path and privately funded its vaccine trial, the Moderna and J&J studies have been carried out in collaboration with CoVPN, which is also coordinating design and analysis of U.S.-sponsored trials of the AstraZeneca and Novavax vaccines, expected to be completed this spring.

Detail-oriented people

No stranger to biostatistics, Peter Gilbert is the son of Drs. Ethel and Richard Gilbert, an earlier generation of University of Washington statisticians.

“My mother is the famous one,” Gilbert said. She was a UW assistant professor in the 1960s and has won numerous awards for her work on the risk of radiation-induced cancers at Pacific Northwest National Laboratory and the National Cancer Institute.

A math major at UW, Gilbert said his parents did not push him toward any direction. But during his college days, he became interested in HIV/AIDS.

“I wanted to make a difference,” he said. “I just had some kind of fire to work on this disease, and then it was just kind of natural to meld math with that, because that’s what I knew.”

While a knack for math is common in his field, Gilbert said that the common denominator among his colleagues is meticulous attention to every part of a problem.

“We’re detailed oriented people,” he said.

The legacy of UW Department of Biostatistics looms large at Fred Hutch and HVTN. The department’s reputation was built by giants in the biostatistics field including Drs. Norman Breslow and Thomas Fleming. Dr. Ross Prentice, who — like Breslow and Fleming — has an important biostatistical test named after him, is former vice president and director of the Hutch’s Public Health Sciences Division and retains his long ties to the UW program as a biostatistics professor.

Before he launched the virology program at Fred Hutch, Corey built his career in HIV science at UW, and he continues his close relationships with the university’s faculty in his current vaccine work. When the coronavirus struck America first in Seattle, the expertise was there, too, waiting for it.

“I think of HVTN as a kind of machine that was poised and ready to go,” said Hutch biostatistician Dr. Holly Janes. She is a co-principal investigator of the Statistics and Data Management Center at the network along with Gilbert and Dr. Yunda Huang.

Janes said that HVTN operates differently from other clinical trial coordinators in that it taps its biostatistical expertise at the earliest stages of study design.

“What biostatisticians bring to the table is that we pose testable questions. Our training helps us take scientific questions, translate them into very precise terms and measurements, and develop analytic strategies to address them,” Janes said.

Correlates of protection

Because variation is a constant in biology, it is a complicated task to discern how results of tests involving a sample of a population can be applied to everyone. Meaningful signals in study results need to be separated from random, meaningless data — what statisticians call “noise.” So, it is important from the start to design a study of sufficient size that the final results will be definitive and can’t be attributed to chance.

More complicated is to structure studies so they can tease out useful biological measurements — such as specific levels of antibodies or infection fighting blood cells — that coincide with success in preventing disease. These are called “correlates of protection.”

Baked into the design of these first COVID-19 trials are requirements for companies to store specimens, such a blood and nasal swabs from participants, so independent labs under CoVPN can find these correlates. By identifying and using these correlates as substitutes for the time-consuming and tragic process of tallying rates of disease and death, next generation COVID-19 vaccine trials can be smaller and quicker.

“It will mean those future trials will need just several hundred people and take three months, instead of 30,000 people and six months,” Janes said.

Dr. Jessica Andriesen
Dr. Jessica Andriesen is associate director of operations for HVTN's Statistical and Data Management Center. Fred Hutch file photo

Fast, definitive trials will become vital as the virus, through small genetic mutations, builds up resistance against the current crop of vaccines. So far, the vaccines appear to closely match their targets, which are tiny structures on the surface of the coronavirus’ distinctive spikes. Our immune system — trained by vaccines — can recognize and attack those features. But the virus is changing in ways that are worrisome.

Already, genetic variants with slight changes to the spike surface have emerged that may erode the protection afforded by new vaccines. They may continue to prevent severe disease and death, but no longer stop symptomatic infections. Yet compared to constantly shapeshifting HIV, the COVID-19 virus is much more stable, and vaccines likely can be tweaked to adapt to these slower changes.

Dr. Jessica Andriesen has a unique perspective on the inner workings of vaccine clinicals trials. A Fred Hutch senior staff scientist who is associate director of operations at HVTN’s Statistical and Data Management Center, she serves as a kind of ambassador shuttling between the biostatisticians and the researchers at clinical laboratories and at trial sites.

“I make sure they have what they need to get their jobs done,” she said.

Andriesen helps the CoVPN team juggle the streams of information, tasks and demands that flow from the NIH and the FDA, from the companies carrying out the trials, and the teams at Fred Hutch analyzing the data.

“What makes biostatisticians unique are the complexities of what they are solving,” she said. “They have a respect for problems and love solving them. … What would make us exhausted seems to energize them. When you watch them think, you can see the gears turning.”

A self-described “data nerd,” Andriesen said that she shares with her colleagues a drive to make sure the science can get done as rigorously as possible. In these COVID times, she values the sense of urgency that hovers over everything they do.

“We need the statisticians to do all the analyses, and we need to have our conversations up front. Saving time literally saves lives right now,” she said.

Like her teammates, she recognizes the historical nature of the work they are doing, but they also have a personal stake in it.

“I’m excited to be involved so we can meet our deadlines,” Andriesen said, “and so my kid can go back to school.” 

Sabin Russell is a former staff writer at Fred Hutchinson Cancer Center. For two decades he covered medical science, global health and health care economics for the San Francisco Chronicle, and he wrote extensively about infectious diseases, including HIV/AIDS. He was a Knight Science Journalism Fellow at MIT and a freelance writer for the New York Times and Health Affairs. 

Are you interested in reprinting or republishing this story? Be our guest! We want to help connect people with the information they need. We just ask that you link back to the original article, preserve the author’s byline and refrain from making edits that alter the original context. Questions? Email us at

Related News

All news
Q&A: Dr. Larry Corey at the center of COVID-19 vaccine research Fred Hutch scientists will play key roles in 'harmonizing' clinical trials June 4, 2020
Numbers don’t lie, but sometimes our brains do Avoid confusion about statistics in medical research with these hints and resources February 13, 2020
Are we doing diet and nutrition research wrong? A Q&A with public health researcher Dr. Ross Prentice on how to assess and improve studies of diet and chronic disease January 28, 2020

Help Us Eliminate Cancer

Every dollar counts. Please support lifesaving research today.