One of the striking features of the current COVID crisis is the lack of disease specific treatments – ventilators support the function of the lungs but only the immune system of the body can actually clear the virus from the body.

Medics the world over have gone back to traditional methods to find potential medicines that could be used to improve the course of the disease and lessen its impact. In many cases, doctors work through complex pattern recognition that requires extremely subtle interactions with patients to elucidate the symptoms and physical signs of disease and weigh them appropriately against experience where no two patients are quite the same. This subtlety is why it takes so long for doctors to train and also is why non-clinically trained millennials in silicon roundabout pushing AI algorithms for healthcare might not be as successful as their investors might appreciate.

Typically doctors begin by noticing anecdotal associations that might be as simple as two or three patients doing better than expected and coincidentally receiving a particular drug. These observations are then broadened by discussions with colleagues and through the ‘invisible hand’ the myriad of associations are organically distilled down to identify hypotheses worthy of formal testing.

The recent announcement of a number of formal UK COVID trials - including the testing of malaria medicines in COVID – is notable because of the extremely rapid progression to significant enrolment after only a month. It is doubly exciting in the context of a post-Brexit Britain – the country has effectively demonstrated an extremely valuable capability for rapid recruitment of patients from a universal healthcare system that could be easily applied to other disease states.

The significance for industry is the uniformity of patient care diagnosed by world class physicians using modern diagnostics. Patient care is not affected by insurance status or wealth – which leads to varied and sub-optimal care that makes it harder to detect treatments that work. Patient data can also be accessed from all of the patients’ records – something which is difficult if not impossible in the likes of America, where patients often change health insurers as they change jobs, and where there is no obligation and often no ability for private insurers to share information. Britain already has the significant repositories of patient data along with the genetic and other data required.

Any new medicines regulator can further extend this advantage by enabling innovations such as ‘adaptive regulation’ which provides a framework for selected patients to be able to access experimental medicines that have shown good results, but where further clinical trials are required before the medicine can be used by less-specialist doctors for most general patients. This would provide support for patients – particularly those with rare and life threatening conditions – as well as important support for Britain’s biopharma industry. Could we see a world in which every patient care episode is generating data on real world effectiveness of medicines?

The next steps are exciting but will require Britain to develop it’s own approach to data usage – care must be taken not to simply sell the access to internet giants in Silicon Valley but instead to insist on measures that would develop the local industry. Robust safeguards are required for patient data with transparent understanding about data ownership. We should rethink publication to ensure the natural desire for academic publication does not lead to results being published in journals that cannot be accessed by the public, or which inadvertently give away intellectual property.

Updated (13/04/2020, 6.02pm): An earlier version of this post incorrectly listed the author. Dr Smith authored this article. See here for more information on this blog.