Can pharma competitors collaborate, share and win in the end?

A week ago, I finally reunited with my younger brother who took a train to visit me in NYC from Rochester University where he has just finished his first year. The night he arrived, Boris and I went up to celebrate his belated birthday on the roof deck of the building I live in, and – as it usually goes – dove into a 4-hour long conversation about all kinds of topics.

He was curious about my work in pharma as well as my involvement in data science community and asked lots of questions. Unlike a year ago, when he and I disagreed on questions around innovation and achieving progress through failure, this time – to my surprise – he was taking my side a lot more than I had expected. Until we got to a question about competition.

When I told him that I have recently joined the board of PMSA, a non-profit that brings life science analysts together to improve upon the practice of data science and machine learning, he couldn’t believe such an organization existed.

“Are you telling me you work alongside your competitors and share secrets with them?” he asked.

“Not quite,” I responded. “We don’t share secrets, rather, we share best practices in data science, so that we all bring positive change to our organizations.”

“But what good does it do to share? Aren’t you afraid your competitors would steal your ideas and run with them leaving you behind?”

For someone lacking direct experience in the worlds of pharma and data science, Boris’s questions were fair. I crafted an analogy I thought could help.

Imagine you are the first to build a bicycle. Of course, you may want to patent it to protect your invention, but still decide to keep all the knowledge of making it to yourself. It may take others months or years to build their own bicycles but everyone would still ride bicycles. Now, imagine you came out and shared best practices of vehicle manufacturing, process automation, scalable production, etc. Even if you didn’t give out a specific recipe for building a bicycle, you shared knowledge that can help your industry. Let’s say, someone figured it out and built a bicycle and a scooter, technically leaving you behind in the “personal propulsion” industry.

“Exactly!” Boris exclaimed. “They are now ahead of you.”

“Yes,” I said. “But now everyone is riding bicycles and scooters. And don’t forget, you weren’t sitting on your hands all that time. You were listening to other competitors sharing their best practices, and perhaps now you’re already designing a new bicycle. Maybe one that is good on rough terrain. Or maybe you figured out how to increase the range of power available with the use of gears. This may sound like sharing secrets, but it’s all temporary. At the end, sharing is how everyone wins.”

Boris understood the logic, but didn’t really agree with me. As I pondered his point of view, I knew he wasn’t alone.

I once had a conversation with a long-time friend of mine at a PMSA conference about sharing and open source. We talked about how great it would be if all commercial data providers joined forces (at the request of pharma and biotech companies) and contributed their data to a consortium that would tokenize and link it for collective use. The idea originated sometime in 2014, and even got some traction with the NIH. But it was shocking to the data providers because data was the largest source of their revenue. Once shared, their business could be at risk. Suffice to say, the initiative faded away pretty soon after it was conceived.

I couldn’t stop thinking that it can work – as long as everyone’s incentives are aligned. Like in my over-simplistic example with a bicycle. But what should those incentives be?

Fast forward to a few days ago. I was invited to moderate a Virtual Think Tank organized by SmartLab Exchange R&D about the state of AI in drug discovery. In my preparation I’ve started doing research on global open source AI initiatives and came across MELLODDY, short for Machine Learning Ledger Orchestration for Drug Discovery. Launched in 2019, MELLODDY became a platform that combines data from chemical libraries of ten pharmaceutical companies and aims to use machine learning in order to predict which compounds could be promising in the later stages of drug discovery and development. Ten competitors came together and contributed arguably their most sensitive data assets to reach a common goal – discover effective medicine to save and improve people’s lives.

So how can this possibly work? 

Estimated total global spending on pharmaceutical research and development 2010 - 2024

Source: Statista

By 2024, it is projected that the global pharma R&D spending will reach $213 billion, up from $129 billion in 2010.

In less than 15 years our industry will increase its spend by the equivalent of market cap of an international company like Starbucks.

And this is without any guarantee of a sizeable payout since so few compounds make it to clinical trials and overall success rates are very low.

What’s also shocking is that despite enormous progress made in the Big Data space, despite ever increasing rate of processing power and cheapening technology to crunch petabytes of data pharma industry possesses, R&D costs keep climbing.

The concept behind MELLODDY is simple yet brilliant. In exchange for data (held by the platform in strict confidentiality), pharma companies get access to infrastructure, processing power and AI technology they themselves may not yet have. And while they are working to digitize their R&D efforts, they have three years to tap into their competitors’ data in order to find compounds that could give them a head start. In this case incentives are aligned – give me what I want, in exchange for something I have but don’t utilize to its full potential.

MELLODDY is still an experiment and is set to conclude by 2022, but we’re already seeing companies like Amgen, AstraZeneca, Bayer, GSK, Janssen, Merck and Novartis participating in the platform alongside prominent research institution and AI computing experts.

Even though Boris and I didn’t reach a consensus on the importance of collaboration between competitors, I’m hopeful that more and more people start asking the same questions including “How might we join forces with our competition in order to make progress as an industry?”

All in search to make much needed medicine for patients faster and more cost-effective. Could competitors collaborate, share and win in the end?

Let me know what you think in the comments.