How Design can accelerate innovation in Pharma

Prompted by Russia’s COVID vaccine, thoughts on Artificial Intelligence, Machine Learning and the future of medicine - and why Design is a necessity to avoid unintended consequences for millions of people.

Russia’s COVID-19 vaccine

There’s been a lot of incredulity in the last couple of days of Russia releasing a vaccine in such a short time frame. The lack of peer reviewed research and published studies has many in the scientific community crying foul. Add to that the recent attacks on the Oxford Vaccine Group’s IT systems by the Russian Cosy Bear hacking group and it all seems very suspicious.

“The Russians couldn’t have done it in such a short time” they say. Right now they’re probably right, at best corners were cut, at worst there’s something more sinister going on.

But… something to we need to bear in mind: whereas previous vaccines have taken years or decades to develop, we should expect newer drugs to be developed at a massively improved pace — so fast that the regulatory hurdles will be the shackles holding them back — until those too are overhauled. Let’s look at why.

The future of medicine

Big Pharma and Big Data are combining forces. Take for example Novartis, a great forward thinking pharma company with Vas Narasimhan at the helm, they have partnered up with Satya Nadella at Microsoft to pioneer an AI Lab to revolutionise medicine through Machine Learning (ML) and Artificial Intelligence (AI), this is a powerful combination, two of the giants joining forces. With recent advancements in ML & AI vaccines and medicines will be developed at an increasingly faster rate.

Machine Learning & Artificial Intelligence

These technologies are already enabling more efficient clinical trials: enhancing the design of clinical trials, choosing a more suitable trial patients, monitoring the adherence to trial procedures and dosages, and analysing the trial results, spotting outliers and erroneous data — all at a rate humans simply cannot do. The same technologies are now being used to supercharge the drug creation process, and using drug repurposing to fast track the creation of medicines to arrive at the prize quicker. As more of the world’s patient data is digitised, the greater the scope for ML & AI to unearth patterns at a vast scale that humans can’t comprehend, and enable an exponential leap in medicine.

Design thinking is a necessity

There’s all this incredible potential, is there anything holding us back? Design — or the lack of it. Let’s take a look back a little — for decades software was built, rather than designed. Incredibly complex interfaces led to billions of hours lost to software inefficiencies, limitations and workarounds. From the turn of this century design fought for its seat at the table, and proved it’s worth. Good software is designed rather than built, and coupling design thinking and good user experience design leads to great business results, just look at the success of Revolut.

Right now ML & AI solutions are built. Datasets are chosen, algorithms are models are developed, and proof of concepts engineered. Inputs go in, outputs come out, once up and running they are essentially black boxes. These systems often reflect the biases of the engineers who built them, sometimes with unintended consequences. To get the most out of these technologies medical decision makers themselves need to become experts in ML & AI — that’s a tall order for doctors and scientists who’ve spent years specialising in their own fields.

Design thinking methodologies allow humans to shape the models and processes in an understandable way. UX design can shine a light into those black boxes, and allow medical experts to view the inner-workings of the algorithms in terms they understand, and can adjust and control as necessary.

Design can enable humans to shape and guide what the machines are literally amazing at doing. Design should play a key role in all ML & AI endeavours, but if there’s one area we should avoid unintended consequences at all costs, it’s the field that touches hundreds of millions of lives daily — medicine — it’s our collective global health.

Let’s shine a light into those black boxes, and make it easy for our medical specialists to guide the machines.

ps - Since this article was written, MIT's Technology Review podcast "In Machines We Trust" had an episode discussing the potential of AI in pharma.