Cutting Time Without Cutting Corners: COVID-19 and Vaccine Development
Cutting Time Without Cutting Corners: COVID-19 and Vaccine Development
I am very likely to get a COVID-19 vaccine as early as it is (legitimately) available to me and recommended for someone of my circumstances. I am not terribly worried that the incredible rush on the development of the vaccines is causing important corners to be cut and creating unprecedented risk. My comfort does not come from inside knowledge of any one vaccine development program (although I have reviewed some programs), nor from connections to insiders, but rather from considerable experience with and knowledge of the vaccine development process. I know that there are many ways to shorten development without compromising safety. I know that the trials underway are very large – large enough to surface key safety information. And, I am observing important behaviors on the part of pharmaceutical manufacturers and regulators that suggest that they are behaving as they usually do when developing vaccines. These are all signs that suggest that – once approved – these vaccines have a high likelihood of being as safe as most vaccines – and vaccines are generally pretty safe. What follows are a few details about why I think so:
Ways to shorten vaccine development timeline without compromising quality (i.e., safety and efficacy)
The world record for development of a vaccine is four years and that was for the childhood vaccines against mumps. The story is great and I can highly recommend a book by Paul Offit that tells that story and others. More normally, the average vaccine takes about ten years, and (at the outset) has less than a 10% chance of success . However, in recent times, a lot has happened to make the world more ready to develop a vaccine and we have some important experience from previous coronaviruses that matters as well. Now we can develop a vaccine quickly, we just need unsurpassed alignment from stakeholders around the world and a willingness to invest billions of dollars – much of which will never deliver a return on investment. There is a very nice New York Times article that covers many of the points that I cover. I talk about them differently, but for a second (and much more interactive) opinion, here is the link.
Use Existing Data
Usually, the first approximately four years – Phase One – of vaccine development is spent getting to the point where a vaccine candidate is safe enough and likely enough to be useful that it is ethical to try in healthy humans. You may recall that this happened very rapidly with SARS-CoV-2 and that is because of three things. First, a lot of research already existed on closely related coronaviruses, especially the SARS (SARS-CoV-1) and MERS (MERS-CoV) viruses, so the risks and benefits of those vaccine candidates helped to inform the development of the SARS-CoV-2 candidates. Secondly, the SARS-CoV-2 vaccine candidates use very similar platforms to vaccines that have already been studied in humans. They were nearly the same vaccine, so fewer studies needed to be done to show safety, and strategies for demonstrating likely efficacy (including animal models that are compelling) were also already developed. Finally, a set of organizations including the Coalition for Epidemic Preparedness Innovations (CEPI) has spent years investing in early stage vaccines – getting them ready for trial in humans, establishing platforms that can work for a “disease X”. The work of these institutions was critical to funding the first and second points on this list.
This is where you spend billions – knowing that many of those investments will not be needed. CEPI and several major government entities (e.g., the Biomedical Advanced Research Development Authority, or BARDA, in the US) have all invested hundreds of millions or even billions of dollars on vaccines before the technology is proven. In vaccines, this means investment in manufacturing plants to produce the vaccine as well as investment in massive clinical trials to show efficacy and safety in large populations quickly. Phase Three clinical trials for a vaccine can easily cost over three hundred million US dollars. At the same time, so can the cost of manufacturing facilities to make the vaccine. These investments are specialized: very little can be repurposed to produce other vaccines or products without considerable additional investments.
When I first worked in pharma, a drug company in the US sent in a new drug application via multiple trucks sending tens of thousands of pages of reports and tables to the FDA. Now it is done electronically, but the amount of data is still enormous – in fact much greater. For our current emergency, the regulators in Europe and the US are taking an unprecedented amount of information and reviewing it long before the entirety of the application is complete. Like the monetary investment, the human resource investment is unprecedented. Regulators are reviewing parts of applications for vaccines that may never turn out to work. That type of risk of finite resources is not worthwhile in normal times, but these are not normal times.
The trial sizes of the first vaccines that look promising are enormous. While vaccines trials often include thousands of patients, these trials include tens of thousands. The reason is that they are endpoint driven. You need a certain number of cases of COVID-19 to occur before the study will have enough data to accurately show whether the vaccine has an effect. This means training hundreds of sites and thousands of clinical researchers on the handling of the product, selection of patients, and reporting of data. Having very large trials increases the probability that you will get results that are statistically significant faster.
Vaccine development is highly multidisciplinary. You need biologists, bioengineers, statisticians, clinicians, and many other subdisciplines to review the work on a vaccine – both to do the work and also (as a regulator) to review the work. Traditionally, a lot of this work is done serially – especially the regulatory review. Sometimes, serial review is necessary. Statisticians, for instance, need to prepare analyses and reports for clinicians to review. Normally, each handoff from one discipline to another means another place where the work must wait to get to the top of the queue. All these specialists are working on many different programs at the same time. Right now, however, every discipline is prioritizing every morsel of work that is COVID-19 vaccine related. In Phase Three, this type of prioritization can save weeks or even months.
There are a few other observations that I have had about the lead programs that give me a sense of confidence. The most important is that manufacturers are behaving in ways that are consistent with usual behavior for clinical trials. Most importantly, senior leaders of the pharmaceutical companies are making vague and/or slightly inaccurate predictions about when they will get trial results. This is encouraging because we CEOs like to be precise and accurate, but vaccines trials like these are endpoint driven, which means completing them and doing the analysis depends on when trial subjects get sick, not when CEOs want to be done. It is good that the CEOs are not forcing the science. Another good sign is that trials are stopping for brief periods of time when adverse experiences happen so that outside safety monitoring boards can analyze the situation without compromising the integrity of the trials. In this time of intense public and government pressure, waiting for a complete safety review is not easy to do.
I am quite comfortable that the big pharmaceutical manufacturers and the regulators are behaving normally where they should maintain standards and integrity and are behaving in unprecedented ways in terms of how they use their resources and those of others.
These are unprecedented times and with it comes the opportunity do something truly unprecedented – which is to safely vaccinate large populations only a year after the disease is even identified.
 Pronker, Esther S et al. “Risk in vaccine research and development quantified.” PloS one vol. 8,3 (2013): e57755. doi:10.1371/journal.pone.0057755.