My paper was accepted by the Mercatus Center at George Mason University:

Quickly developed therapeutic, prophylactic, and vaccine treatments will be essential to defeat the COVID-19 pandemic. Standard regulatory and development protocols typically yield drug development in 6 to 11 years. However, experimental vaccines and prophylactics against COVID-19 are in development and currently in trials, making deployment after safety trials feasible within months. Though efficacy of these treatments is uncertain, with probability of success, on average, at just 16 percent, efficacy can still be tested in real-world settings by adopting an approach that allows patients to begin drug treatments earlier in the regulatory process.

After phase I trials have been completed and data have been gathered, an expert committee should be convened to advise on the risk, benefit, and chance of success of the novel agent with a view toward distribution to the public through patient choice. There is limited downside risk and a good chance of a positive outcome with this approach. The US Food and Drug Administration (FDA) has the authority to approve pharmaceuticals requiring further trials, but the success of this approach will depend on coordination between sponsoring research organizations, trial designers, and regulators. International pharmaceutical regulators, such as those in the United Kingdom, Europe, or Japan, can also adopt this approach.

At this point, a clear indication of regulators’ openness to this approach would provide certainty for organisations and accelerate development of possible vaccines and treatments. When adjusting for probability, this approach may save hundreds of thousands of lives in the United States and more globally.

More here at this link at Mercatus. Or full PDF here.

Under a traditional approach, a 50 person phase I trial might not be considered enough data to make strong judgments on safety and efficacy. It’s considered exploratory. However imagine…

10 separate Phase I trials (or trials with adaptive design) start in April 2020 across multiple countries. Patient numbers and characteristics (eg age, presence of COVID-19 antibodies, sex) are careful considered. The statistical power of multiple trials is stronger than one or two trials. 10 trials of n = 50 (say, up to 100) could be run in US, China, Italy, UK, Germany, France, South Korea to name a few countries with R&D capabilities with either commercial or non-profit led organisations that could run such trials. Results could be available in July. These data would include safety data, and preliminary measurements of efficacy such as COVID-19 antibodies.

Each interested country could convene its own expert and stakeholder committee. That committee could vote on if tentative approval should be given based on a benefit and risk assessment of the data. This meeting should be held in public and take submissions from stakeholders potentially replicating FDA advisory committee meeting processes or similar. A citizen’s jury process could run given the state of public interest.  

In an optimistic scenario of no safety signals and strong efficacy signals, this would allow a September 2020 launch with high risk populations such as over-65s and healthcare workers prioritised. This is subject to commercial scale up of manufacturing. Funding for this should be given in April in hope of a positive scenario. In a pessimistic scenario, negative safety signals would not warrant tentative approval. The downsides would be limited to the volunteer trial participants and the investments taken.

The cost of running 10x phase I/II trials now is perhaps between $200m to $1bn. Given that this could solve the problem - it seems to be something that governments should be doing or incentivising. Especially vs fiscal bail outs wheich are 1000x this amount.

Now the data might come back negative - in which case there is no quick vaccine. But, if there were very strong efficacy signals and no safety signals then this should be something for governments and regulators to consider.