Eight months into the global coronavirus pandemic, the life sciences industry is ramping up drug research in previously unprecedented ways, investigating existing drugs as well as potential new therapies and vaccines to treat and prevent Covid-19.
The rush to research, however, has resulted in some haphazard, poorly designed, and costly Covid-19 clinical trials, as demonstrated by a STAT analysis.
What we need is a coordinated effort, something like the one used in the U.S. space program. That effort began on July 29, 1958, when President Eisenhower signed legislation forming the National Aeronautics and Space Administration. NASA not only brought together the best and brightest minds from academia, industry, and government, but it also focused the required infrastructure and resources on the audacious goal of landing the first human on the moon.
As the epicenter of U.S. civil aerospace research and development, NASA was charged with ensuring that the country’s scientific and engineering resources were used as effectively as possible. It worked closely with other government agencies to avoid duplicating efforts. NASA began with three research labs and two test facilities, but quickly expanded to include military and academic research organizations, leveraging a variety of brainpower and assets.
We have an opportunity today to provide similar NASA-type leadership to conquering the pandemic through an initiative that will accelerate the discovery, development, and availability of critically needed treatments to combat the novel coronavirus: an open-sourced clinical neural network to optimize and accelerate treatments and vaccines.
An open-sourced global clinical neural network, born of a partnership between government agencies, academic institutes, clinical trial sponsors, data analytics leaders, and others, would enable access to and sharing of clinical research and best practices on a previously unprecedented scale. This substantial set of information, complete with universal errors and collective insights related to Covid-19 research being conducted around the world, would be available in real time and free of charge.
Researchers could leverage this virtual neural network at any time and from any place to learn from ongoing and past clinical development efforts. As a result, drug development collaboration could be dramatically enhanced. We would be able to course-correct clinical trials in real time as new data become available. Research efforts would be more organized and efficient, as well as less costly, avoiding the duplication of common mistakes.
The overall effect would be to accelerate the development of therapies to cure and prevent Covid-19, enabling the medical community to defeat this devastating global pandemic with previously unmatched agility and speed.
A prototype of sorts for a clinical neural network is currently operational. Earlier this year, the White House, the Allen Institute for AI (AI2), and leading research groups created the Covid-19 Open Research Dataset (CORD-19), a publicly available database containing more than 130,000 scholarly articles about Covid-19, SARS-CoV-2, and related coronaviruses.
Recognizing the imperative to expedite Covid-19 research, my AI-focused company quickly developed a semantic search capability to help the global research community use this database even more effectively. We used the most advanced natural language processing technology to apply a deep-learning reading comprehension model, trained explicitly for pharma, to CORD-19. This highly accurate search engine, which is free to use, shows every article header and text snippet that contains the search term entered.
On any given day, it receives between 20,000 and 30,000 queries. Researchers, doctors, and even the general public are using this tool to get the most enlightening answers and insights from the scholarly papers in the database. Other data analytics companies, including Amazon Web Services and Google, are leveraging AI-powered analytics to improve access to CORD-19.
Imagine the potential and power of a next-level clinical neural network, one that encompasses all of the stakeholders and their previous and ongoing Covid-19 research efforts. The result of such a partnership would accelerate drug discovery and development not only for Covid-19, but ultimately for drugs to treat every other disease and condition.
NASA offers a vital lesson about the extraordinary impact of collaboration between scientific and engineering minds: bringing together engaged experts gets things done.
To borrow a phrase associated with the Apollo 13 lunar mission, “Failure is not an option.”
Suresh Katta is the founder and CEO of California-based Saama Technologies.
Yes, that’s what we need! A bloated Wasington bureaucracy that can never die with a budget several times larger than the NSF! What we don’t need is an Elon Musk, exposing how much money you can save when you have a lean, out-of-the box approach.
And of course, how is it going to be ready in time? Just writing agendas for all the meetings is going to take you into next year. If you include the PowerPoint presentations, we’ll be two or three pandemics down the road before this thing is ready.
What we should have done a long time ago (so it would have been available now) is require all federally funded researchers to disclose ALL their raw data, and make it available with their published papers. We already do the latter but not the former. If all that data were available, other researchers (including amateurs and hobbyists) would be able to throw their own analytical tools and methods against it. Maybe one of them would discover a relationship the original researchers missed. We might actually learn something that brings us closer to the goal.