Published 23 July 2020
Serendipity stepped in to lead Larry Tsai to Roche. A medical doctor and pulmonary and critical care specialist by training, he spent most of the last 20 years working at small biotech startup companies.
He also had a side gig in recent years in the intensive care unit at Brigham and Women’s Hospital in Boston. When a recruiter called, joining a large company was the furthest thing from Larry’s mind. Truth be told, he probably wouldn’t have returned the call. But then he heard the hiring manager was someone he’d studied under as a third-year Harvard Medical School student.
“That kind of coincidence was enough to get my attention,” Larry says. “I’ve known Mark Eisner for a long time.”
While in school, Larry had a rotation at the University of California, San Francisco, doing pulmonary consultations. He trained under Mark, who is now Senior Vice President, Global Head of Immunology, Infectious Disease, and Ophthalmology Clinical Development at Genentech. Larry was intrigued by Mark’s vision to combine respiratory and rheumatology together under one position at Roche.
So a little over a year ago, Larry became Global Head of Respiratory and Rheumatology Product Development. About six months later, the pandemic arrived.
“Call it coincidence or serendipity,” Larry says. “Here I am, a pulmonary critical care physician, who just happens to be heading the respiratory and rheumatology programme. And by further coincidence, the last company I worked at was an infectious disease company, developing antibiotics. So I had experience in the infectious disease space as well. All of these things just came together and put me in the right place at the right time.”
The sweet spot
He says the thing he loves most about his new role is the opportunity to make an impact on patients' lives and on health care on a global scale.
“Most academic physicians have to make a choice at some point,” Larry says. “Are you going to focus on research? Maybe you’ll discover something that could lead to a therapy and help a lot of patients all at once. But the probability of it happening is really low.
On the other end of the spectrum, you can work primarily as a clinician and a teacher, and you can come home at the end of most days and feel like you helped a patient or a family, or you've taught a student something really important. But the scale is pretty small.”
Larry found his sweet spot in between both of those worlds, taking the most promising ideas and shepherding them through clinical development, regulatory and business processes.
“The goal is to come out at the end of the day with an approved drug, and a set of data that physicians and patients need to understand how to use the drug and how they might benefit from it,” he says. “If you can do that you've added tremendous value to the process that can help a lot of people all at once.”
The catch? It’s easier said than done
“People don't realise that it not only takes a lot of time, but it also takes thoughtfulness, hard work and skill, and often some creativity, to actually get drugs through the development process,” Larry says. “It’s something that I enjoy and have some talent for. That's what I really love about it.”
In an interview, Larry shares his thoughts on running robust clinical trials, what it’s like to be part of an “unblinding” and how he feels about joining Roche just before the pandemic hit.
We’re in a pandemic, and your job is to run a trial. Is it business as usual?
It's tempting to say there's nothing normal about conducting a trial in COVID-19 times, but that's not true. We more than ever need well-designed rigorously conducted trials so we can understand if a drug really works, and if it's safe in a disease, just like we would in any other situation.
Despite the unusual circumstances, that fundamental truth is still there. The circumstances are indeed quite unusual, both because of the urgency of the situation and the dynamic nature of the pandemic. So if you look at the timeline of the current trial, from initial concept to first patient enrolled, it was less than 30 days. This process typically takes months, sometimes years to happen.
We submitted a briefing package to the FDA on a Friday afternoon. We received comments back on a Sunday night. We wrote the protocol in three days and received FDA comments and a Safe to Proceed letter two days after that. It's a great credit to the team and to the FDA for understanding the unprecedented nature of the situation and acting with an appropriate sense of urgency.
Working at small companies, it seems everything's an emergency and the timelines are always extremely truncated, but I've never seen anything like this.”
How has COVID-19 changed the way you approach things?
We really knew very little about the disease we were trying to study, including very basic things like the natural history of the disease and the proper endpoints to study. There was also no established standard of care. So we had to make a lot of educated guesses in designing the study, and we had to design it with enough flexibility to accommodate all of this uncertainty.
Do you feel enormous pressure to find answers that will help patients?
To be clear, a trial has to be a team effort, because it's such a big undertaking. So there really is no individual, not me or anybody else, who is solely responsible for a trial in any sense of the word. Certainly the team and I do feel the pressure of high expectations. But part of my job is to shelter the team from this pressure so that they can focus on doing their work and doing it well, because that's the most important thing.
From a personal perspective, I think my critical care training has been helpful in this pandemic situation. I don't want to sound melodramatic, but in critical care you're often called upon to make literally life or death decisions. You have to be comfortable with that or you can't do your job. I think that helps in terms of managing some of the pressure and expectations.
Can you describe an ‘unblinding,’ when trial results are revealed?
I'm not really a nervous guy, but it can definitely be a really nerve-wracking experience. The whole team works so hard, for so long. From designing the study to getting it approved, getting the study started and then actually conducting the study, it’s just an enormous amount of work. Then it comes down to this single moment where these key tables are revealed to you.
It's a small group of people who initially see the tables. Your heart is racing. Then you look, and either your heart jumps and hits the ceiling, or it drops down through the floor.”
For me, after that immediate response, there’s a kind of disbelief either way – whether it's a positive or negative result. Then I just start digging into the details and looking at all the other tables and checking everything, to make sure everything checks out.
How does it feel if the results are not what you hoped for?
Probably the worst part of a negative readout, and I've been part of positive readouts and negative readouts, is telling the larger team and trying to help manage their disappointment. Of course you're dealing with your own disappointment, but there's absolutely something to be learned from every study. I've always said you can arguably learn more from failures than successes. But that doesn't make it any less disappointing.
Why does it take so long to test a medicine?
It goes back to the complexity of the process. Three main things contribute to this: First, there are just long lead times in terms of how long it takes to design a study, and how long it takes to get them then approved by regulatory authorities. It's not just national authorities and local authorities, it's individual sites and ethics committees that need to approve it. That process typically takes months, sometimes years, and that's before you can even get started.
Second, when designing and conducting a study, you need to have a well-defined patient population in order to maximise the signal-to-noise ratio, and maximise the probability of success.
Because of course, you always want to have a positive study if you can. There's always this balance where the more carefully and narrowly you define your patient population, the harder it is to find those patients and the longer it takes to enroll the study.
On the other hand, if you broaden things too much, then you end up with a lot of noise that actually makes it hard to detect a signal, which means that the study needs to be larger, which of course also leads to longer enrollment times. It can be very difficult in clinical trials, trying to find that balance.
Third, enrollment varies from country to country and region to region. It may be a cultural thing as much as anything, but I think especially in the United States there's a lot of reluctance for patients to participate in clinical trials under normal circumstances. During COVID it's a little different because there are so few approved therapies. Because of the desperation of the situation, it's perhaps easier to enroll patients in studies.
Why is it so difficult to find patients for studies?
Let’s use antibiotics as an example. Most people agree that we need new antibiotics to help fight multidrug-resistant infections. But, in the United States, the rate of multidrug resistance is relatively low. That means most patients can be treated with a regular antibiotic. Why would they take a chance of enrolling in a clinical trial with a new antibiotic if there's already an antibiotic with a known safety efficacy profile, that is likely to be effective for them?
When there are available treatments and there isn't desperation, people are reluctant to enroll in clinical trials. They see it as an unnecessary risk. I understand that, but at the same time that’s the problem when we talk about what we need to do to prepare for the next global health emergency.
You've got to stay ahead of this. By the time you actually have multidrug resistant organisms spreading around the world, or by the time the next pandemic virus is here, you're playing catch up. It's too late. So you need to have those drugs being developed and tested ahead of time.
What does clinical evidence mean to you?
The term “clinical evidence” can be a lot of things. To me the quality of the evidence is key. For COVID in particular, where the natural history is so poorly understood, you can only get a robust answer to questions about safety and efficacy by doing well-designed and rigorously conducted clinical trials, so you have high-quality, unbiased clinical evidence that comes out of it.
In critical care, when you make an intervention like trying an experimental drug and the outcome is not good, you tend to blame the disease and say, “Ah, that patient was so sick. It's not that the drug didn't work, it's that there was nothing that we could do for that patient.”
Whereas if you make an intervention and then the outcome is good, you tend to say, “Ah, the intervention I made really worked!” So there's sort of this inherent bias that all physicians probably have to some extent, and I think it particularly applies to critically ill patients.
If you don't have that rigour, the well-designed rigorously conducted, ideally double-blind, randomised controlled trials to minimise or even eliminate that bias, then you can't get a robust answer to these critical questions.”
How are you and your team doing?
For the first few months in particular, the team was working late every night and every weekend, and that clearly wasn't sustainable. Of course there are moments of conflict and moments of frustration, but I think everybody understands the importance of what we are doing. That's what has kept everyone going more than anything. The reality is that you can have a long career in this industry and not have a chance to contribute to a program that's as important as this one, with its potential impact on so many patients.
Everybody recognised very quickly that this was such an unusual, unique circumstance. One of the things that makes this company special is that everyone really believes in this mission of trying to bring transformative medications to patients with unmet medical needs. It's hard to argue that there's more pressing medical need right now in the world than COVID-19.
What has surprised you?
I come from a small company background where I'm used to doing things a bit differently, with a little more flexibility, occasionally cutting a few corners to get things done.
When I came to Roche, one of my concerns was that this would be a very rigid organisation where everything has to go through 10 different committees before you can actually do anything. The company has gone through some transformation over recent years and has tried to become more agile, and has tried to make the governance process a little bit less cumbersome and more efficient.
So maybe it has worked well in some ways, and not so well in other ways, but I do think that this experience is a success story for that transformation. When the pandemic hit, the team was prepared to move quickly and make decisions that in the past might have required complex governance and higher-level approval, and then was ready to be really agile and deal with the consequences and adjust if things weren't going the way they should.
Will this change the way you do things in the future?
I don't think we should always do things this way. It was an extreme circumstance that required this kind of disruption. But the fact that we can do this is an important lesson. If we're ever faced with a similar situation, we know we can do it again.
The importance of robust clinical trials
Since just four months ago or so, our understanding of COVID and its terrible complications has grown tremendously. Every day it seems like there’s another study being published, including a lot of data about existing drugs coming from case studies and small controlled and uncontrolled studies, but despite all the data that we have now, we still don't have answers to fundamental questions about the safety and efficacy of some of the drugs being used to treat COVID. One study comes out and shows something that looks quite positive, another comes out that shows something not so positive.
Each of these studies have their own flaws, their own limitations. The reason we don't have answers to these fundamental questions is because only well-designed rigorously conducted studies can provide these kinds of answers. But the studies are difficult to do even under normal circumstances.
First, they've got to be carefully designed and approved by regulatory authorities. Then you have to find study sites, assemble the study staff, and train them very carefully to make sure that they're following the protocol and that they're not introducing bias. Bias skews the results and makes the results either misleading or at the very least difficult to interpret. On top of that, everything has to be meticulously documented and verified. You can imagine in the context of a global pandemic this becomes exponentially more challenging.