As the Covid-19 pandemic unfolds across the globe, physicians are scrambling for treatments that can effectively treat patients sickened by this novel coronavirus. The antiviral remdesivir has gained some traction, as has the anti-malarial medicine hydroxychloroquine, a drug touted by President Trump as a “game changer,” despite scant evidence of its effectiveness, while others, such as Dr. Anthony Fauci, have sounded an alarm on it.
Reckless use of a drug can cause devastating side effects, including death, while the unchecked virus, or an unbridled immune response to it, can also be deadly by causing the body’s organs to shut down.
There’s another scenario, a middle ground, in which the focus is on helping patients while capturing outcome data to learn from all patients on all treatments all the time. With such a system, instead of frontline doctors deciding to try experimental drugs on the fly, they could tap into the collective experiences of thousands of doctors around the world to learn what has worked and what hasn’t for similar patients.
In the cancer research and treatment community, the treatment of one patient can inform the treatment of others. Real-world data are routinely collected from tens of thousands of cancer patients through services like CancerLinQ and Flatiron, as well as data deposited in publicly maintained databases, including one run by the nonprofit organization I founded, Cancer Commons. Doctors can mine these AI-driven databases to advance cancer research and learn how treatments — some of which are already approved by the Food and Drug Administration for other uses and can be used off label — have worked for similar patients in the past.
Patients with advanced cancers often exhaust standard therapeutic options. For them, treatment data are scant and no one knows the optimal treatment. Such patients are routinely treated with off-label drugs and novel drug cocktails. However, these individualized N-of-1 experiments are uncoordinated and their results seldom reported, so little is learned.
Cancer Commons, together with its for-profit spinoff, xCures, has built a global rapid-learning network that captures this real-world data and uses it in real time to inform the treatment of the next patient. The platform coordinates treatments across patients to maximize collective learning and prioritize the development of therapies that appear to be working. By tightly integrating cancer research and clinical care, this network enables pharmaceutical companies to slash the time and cost of developing drugs, helps physicians and payers make better decisions, and offers patients the possibility for superior outcomes.
This same rapid-learning approach can be used to improve treatment recommendations for patients with Covid-19, thousands of whom have already received dozens of off-label and compassionate-use therapies: hydroxychloroquine, remdesivir, sarilumab, and ivermectin, to name just a few. In the absence of a proven therapy, many patients with Covid-19 will receive experimental drugs outside the confines of randomized controlled trials. Some will recover, others will not. We can learn from all of them.
Beat19, the first Covid-19 patient registry, was created for this purpose. We have pivoted the Cancer Commons/xCures platform to track treatment decisions, treatment rationales, and outcomes among Covid-19 patients, along with demographic, behavioral, and environmental data. Powerful analytics sift through this information, as well as through physicians’ notes and preprints, to inform critical treatment decisions in real time — and before trial results are available.
These data can also be used to answer important questions posed by public health researchers, such as whether Covid-19 patients with high blood pressure who take an ACE inhibitor fare better than those who take a beta blocker. Answers to simple questions like this can save lives.
Meanwhile, hundreds of clinical trials — nearly 700 of them, according to ClinicalTrials.gov — seeking to answer whether hydroxychloroquine or numerous other drugs being thrown at Covid-19 are beneficial have begun around the world. These trials are important, of course, but will take months or even years to recruit patients, analyze data, and publish results in peer-reviewed articles in academic journals — time that people with Covid-19, which kills or causes irreversible lung damage in days or weeks, do not have. If the system I propose here was fully operational, the answers these trials are seeking would, in many cases, already be known.
The Covid-19 pandemic represents both a challenge to traditional clinical research and an opportunity for transformational change. Supplementing — and possibly even replacing — traditional clinical trials with a global learning network adapted from cancer research is fundamentally better along every dimension — faster, cheaper, more agile. The advantages of doing this have already been demonstrated in cancer care. Now it’s time to show the world that a flexible, decentralized approach can drive clinical research at warp speed against an aggressive virus that has encircled the globe.
Marty Tenenbaum is the founder and chair of CancerCommons.org and the founder and executive chair of xCures Inc.