Revolution in a Petri Dish: Sustainability in Life Sciences

Innovation in drug discovery is in crisis — new molecular breakthroughs are decreasing while costs are exponentially increasing. At Generation, we try to adopt a systems-level lens to examine key industry shifts, identifying the elements catalysing a more sustainable future. We recognise that human health is determined by far more than just medicine — social determinants, preventative behaviour and care delivery are critical contexts. Still, drugs often act as a last line of defence for patients and we think will always remain a critical part of human healthcare. To us, AI solutions to accelerate drug development are part of the sustainable solution. Aligned with this, we’re excited to announce our $40m investment in BenchSci, which you can read more about here.

In this post, we explain why Generation is turning its attention to the trillion-dollar life sciences industry, and the sustainable solutions we are identifying across the drug development lifecycle.

We are at the beginning of a life sciences revolution — if you are building a company that is building a sustainable solution, please get in touch!

Since 1990, medical breakthroughs have added an astounding 70 million life years. Consequently, the life sciences industry has grown to be the world’s third largest.¹

As life expectancy and our understanding of disease pathology increases, so will the importance of life science and a sustainable approach. Fundamentally, we see unlimited demand for innovation, and are keen to back tech solutions that are fuelling the leading edge of drug discovery without excessive cost.

Life sciences have unquestionably improved our lifespan and life quality. Within oncology alone, we have doubled cancer survival rates and added 5.5 extra years of life expectancy for the average person.² However, making such advances alongside an inefficient healthcare system has come in parallel with societal costs: worsening health equity and access to care, rare diseases left unaddressed and a small pocket of bad actors popularising the villainous ‘Bad Pharma.’

We are convinced that the life sciences industry is on the brink of a sustainability revolution. A fusion of trends is shaping a revitalized field that’s better suited to the future, filled with innovation. To break it down:

1. Disease modalities are more complex and splintered, leading to smaller patient cohorts.
2. Discovery is shifting from small to large molecules and eventually personalized medicine.
3. A tidal wave of clinical and real-world data is giving us richer patient understanding.
4. Advances in cell therapy and biologics manufacturing are bringing new therapies to the clinic.
5. A robust biosimilar market is emerging, accelerating access to novel biologic therapies.

However, these sustainable shifts (while promising) have negative consequences. They are contributing to a sharp reduction in R&D efficiency — an observation known as Eroom’s Law.

Smaller patient cohorts mean longer, costlier clinical trials. Real-world data has increased bioinformatics spend but with unclear demonstrable improvement yet. Plus, as disease modalities complexify, experimentation is getting tougher now that we’ve picked the low hanging fruit. In fact, 80% of scientific work does not advance the study of a given disease.³ Eroom’s Law shows that the cost of bringing a drug to market has doubled every nine years, such that today it costs nearly $3bn per drug — despite cost-down curve declines in technologies such as gene sequencing.

Figure 1: Eroom’s Law, showing the number of new FDA molecule approvals per $bn of R&D spend.⁴

With this rising cost burden in mind, Generation is excited about investing across three major themes that are driving sustainable system shifts across life sciences:

1. Accelerating innovation in drug discovery
2. Broadening access and lowering spiralling costs of clinical studies
3. Supporting new lower-cost commercial models that accompany more complex medicines

Accelerating innovation in drug discovery

The pre-clinical phase of the drug development process is one of the most critical areas to address when it comes to battling Eroom’s Law. Only one in 10 drugs that enter human trials reach the market, after getting through the multi-year pre-clinical process.⁵

For the top-20 biopharma companies, which conduct 75% of global R&D spend, new molecular entity discoveries have been meagre at best. For example, a top-10 biopharma company we spoke to has not successfully developed a new drug in-house for more than two decades, despite spending nearly $100bn in R&D over the same period. Life sciences has a higher R&D cost as a percentage of sales (up to 40%) than any other industry.⁶

Despite these hurdles, pre-clinical and drug discovery is still markedly under-served by technology. This is where we see an opportunity. We believe that there is an enormous opportunity for a suite of AI-powered tools to emerge to help biopharma reverse its efficiency declines and develop innovative medicines faster, at a fraction of the cost. Life sciences needs multiple approaches — from AI biotechs like Atomwise and Recursion to enterprise-grade software platforms supporting the hundreds of thousands of scientists working at large pharma companies. Biopharma are not going to become big data or AI companies on their own — they simply don’t have the in-house capabilities.

This is why we are excited to announce our $40m investment in BenchSci. BenchSci’s ASCEND platform is an AI end-to-end discovery platform for biopharma. ASCEND helps increase efficiency — one biopharma company told us they believed they could avoid $100m of wasted R&D spend per year by rolling-out ASCEND — but what encourages us is that the platform has shown it can accelerate innovation. For example, at one of BenchSci’s top-10 biopharma customers, ASCEND users are identifying 22% more novel targets than non-users.

We believe that the BenchSci team have built a one-of-a-kind technology with the potential to halve pre-clinical timelines, pre-empt efficacy and safety risks, and empower scientists as they hypothesise and experiment on novel compounds. ASCEND allows scientists to ‘join the dots’ — it augments scientists’ experimental design processes, and fills in the gaps often missed in complex and fast-moving fields (especially one where information is spread via the arcane journal system).

Broadening access and lowering spiralling costs of clinical studies

Human clinical trials are the most resource-intensive step in drug development. Timelines have lengthened as protocol complexity increases, while clinical site capacity reaches its upper limits and patient identification becomes even harder.⁷ Unintended by-products include:

• Worsening participant diversity: people of colour represent just 2–16% of US trial participants⁸
• Elevated patient drop-out rates: on average 30% of patients drop out of studies⁹
• Non-enrolling trials: 80% of trials fail to meet enrolment timelines¹⁰

Biopharma is starting to take notice of trial diversity, and technology companies are here to help. Inato is building a network of under-served clinical trial sites, while H1 is connecting physician profiles to clinical ops teams looking for principal investigators who can supply more diverse patient populations.

Clinical sites and CROs (Clinical Research Organisations) are also embracing technology to broaden patient eligibility and enrolment — platforms like OneStudyTeam (owned by Reify Health) are digitising the relationship between sites and biopharma, and managing the shift to hybrid ‘decentralised’ trials. Meanwhile, companies like Medable are managing entirely digital-driven decentralised trials, a reality that for some therapeutic areas is already possible and could be up to 13x cheaper.¹¹ Unlearn is applying AI to patient data to build ‘digital twins’ of participants, reducing the enrolment burden, time sink and cost of control arms.

Our portfolio company, SOPHiA GENETICS, has built a platform powered by machine learning and artificial intelligence that extracts actionable insights from genomic, radiomic, clinical and biological data from over 750 clinical institutions worldwide, and is helping biopharma companies expedite patient selection for clinical trials by providing actionable, real-time, real-world data.

Supporting new lower-cost commercial models that accompany new innovative science

The life science industry spends hundreds of billions of dollars on sales and marketing each year¹², however much is wasted when medications are not effective for patients with specific genetic, racial or health realities. In the past the world has viewed therapeutic success through a black-and-white lens of ‘it works!’ or ‘it doesn’t!’ Today we know that it takes years of work gathering and analysing real-world evidence on drug usage and effects to really understand the impact a given medication has in the context of other human and environmental factors.

Real-world evidence is not new, but companies like Aetion and ConcertAI are helping biopharma make sense of how medicines perform in the wild (turning real-world data into real-world evidence), shortening time to regulatory approval and accelerating expansions. On the data side, Datavant and Komodo Health are working with payors and life sciences to map disjointed patient experiences into complete journeys, accelerating research and commercialisation that maps to unmet need. This data will be critical in the shift to value-based care, where finding the right treatments for the right patients is directly tied to reimbursement.

We believe that because the types of diseases and corresponding medicines are evolving, so too should the ways medicines are commercialised. In a world where 95% of rare diseases lack a FDA-approved treatment, we need a new way to get drugs to patients. We are moving now towards a blurring of lines between clinical trials, personalised ‘n=1’ medications and ‘in the wild’ usage, where old primary care sales rep models of the past are less relevant and less effective.

We also need to make medicines more cost effective for patients, i.e. ensuring that lower cost to produce filters through to the end-user. Drug prices in the US have increased by 11% per-year for the last 13 years.¹³ Part of this increase is attributable to the consolidation the pharmacy benefit managers who control supply between life sciences and pharmacy distributors. We believe technology also has a role to play in driving down end-user costs, through new models of payment and distribution. Cost Plus Drugs is building a direct-to-consumer generics platform to cut out middlemen, while Phil is building a therapy deployment platform for specialty pharmaceuticals.

What’s next?

Generation is at the beginning of investing in solutions addressing this life sciences revolution. We are envisaging a sustainable future where life sciences can:

1. Align incentives to effectively match demand for medicine (the need) with supply (drug development);
2. Spend R&D dollars effectively through better clinical trials, efficient early stage experimentation, cross-collaboration of ideas and innovative use of structured data;
3. Bring down the cost and hopefully price of medicine in the long term by being better actors through more transparent supply chains that cut out the middlemen;
4. Improve access to innovative medicines for those in emerging markets — this should be through clinical trials and broader distribution schemes; and

If you are building a company that is driving to our vision of a sustainable future, please get in touch!

[1] http://valueofinnovation.org/

[2] https://80000hours.org/2012/08/how-many-lives-does-a-doctor-save/

[3] https://www.nature.com/articles/d41586-021-03736-4 — read more on the irreproducibility crisis

[4] https://www.researchgate.net/figure/Erooms-law-the-number-of-new-molecules-approved-by-the-US-Food-and-Drug_fig4_326479089

[5] https://www.bizjournals.com/boston/blog/bioflash/2016/05/report-suggests-drug-approval-rate-now-just-1-in.html

[6] https://www.efpia.eu/publications/data-center/value-to-the-economy/rd-spending-as-a-percentage-of-net-sales/

[7] https://www.centerwatch.com/articles/25033-trend-of-longer-trial-timelines-is-likely-to-continue#:~:text=Clinical%20trial%20timelines%20have%20been,to%20continue%2C%20if%20not%20accelerate.

[8] https://hbr.org/2021/06/addressing-demographic-disparities-in-clinical-trials — despite comprising 39% of the US population

[9] Primary research conducted by Generation

[10] https://www.clinicaltrialsarena.com/marketdata/featureclinical-trial-patient-recruitment/

[11] https://www.appliedclinicaltrialsonline.com/view/new-data-reveals-dcts-yield-better-clinical-trial-roi

[12] Based on public filings

[13] https://jamanetwork.com/journals/jama/fullarticle/2792986 — adjusting for the characteristics of these medicines