How Taking Smart Risks Makes Drug Development Less Speculative

Originally published by Forbes Technology Council

Drug discovery is a risky business.

It’s common knowledge that over the past decade, the cost of producing a new drug has increased dramatically, ranging between $314 million to $2.8 billion per new drug. Such high costs and the time it takes to bring a drug to market are forcing the pharmaceutical industry to take smarter risks.

As a whole, drug development has been declining, made all the more so during the Covid-19 pandemic, where in 2020, the FDA approved just 53 new drugs. This has led to "Eroom’s Law," which is Moore’s law in reverse. Moore’s law states that the speed and capability of computers can be expected to double every two years. Eroom’s law was coined to observe "that the number of drugs being approved decreases year on year while the costs increase."

The good news is AI is helping alleviate this decline by creating innovative techniques for mapping drug structures and decreasing the cost of scientific development.

Through breakthrough genome editing tools like CRISPR and AI-powered models and applications, technology innovations in drug discovery are proving their ability to actually derisk many aspects of the drug development process, predicting drug toxicity more accurately, repurposing existing molecules for novel therapies and identifying better drug candidates faster.

Calculating Drugmaker Risk

As drugmakers pursue an approved drug, they are under enormous pressure to make the right decisions along the way—which projects to fund and which to terminate. Everything from the quality of available data to the complex pathophysiology between the disease and target and the high ethical and material costs of indeterminate animal studies must be considered and weighed as factors in the drugmaker's overall risk analysis.

Taking Smarter Risks To Reduce Risk

Even in cutting-edge, research-driven industries like healthcare, there’s a lingering hesitancy to engage new technologies that disrupt established internal structures and proven methodologies. In many ways, this caution makes sense. When a drug fails to pan out, years of effort (and millions of dollars) are sunk.

Given this risk calculation, not surprisingly, some drugmakers are wary of introducing new technologies in place of more well-established scientific methods (Theranos is perhaps the most high-profile and enduring novel technology cautionary tale). And the industry has a history of managing its risk in other ways: relying on time-tested R&D tools, squeezing service providers to reduce costs and changing internal structures to reduce data siloing.

But as more and more successful use cases for technology-driven methods in drug development emerge, and as regulatory agencies continue to weigh in on their use favorably (with the recent passing of the FDA Modernization Act, drugmakers can increasingly feel assured that their investments in technology like computer modeling will be accepted by the FDA as a viable means of compliance), the risk calculus for technology and drug R&D is changing.

Now, drugmakers find that taking a risk on new technology can actually help them reduce their risk across other aspects of the R&D pipeline, as novel tools and applications help them make better decisions earlier in the process and improve overall translatability.

For example, in drug R&D, data is often siloed, decentralized, disorganized and incomplete—and may not appear ready for modern, technology-driven research processes. However, machine learning and deep learning (DL) can work with a range of different data sets (smaller volume and better-organized data sets for ML and larger, raw, and even unstructured data for DL), helping break down data silos, detect unseen patterns and make connections with incredible speed and accuracy.

And today, you don’t need the backing of an Internet giant to access these groundbreaking resources. Machine-learning models and other AI-enabled technologies are on the rise, helping extract more accurate predictions across a broader range of variables, even anticipating a molecule’s safety and efficacy before it reaches clinical trial. Investing in these solutions can ultimately lower risk exposure in drug discovery, streamlining operations, improving accuracy, quickening rollout and reducing costs every step of the way.

Three Tips For Evolving While Mitigating Risk In Drug Discovery

Drugmakers and other research organizations hoping to evolve their approach to drug discovery while also mitigating their risk should keep these three best practices in mind.

1. Educate internally: Work to break down organizational knowledge silos and create more opportunities for cross-functional education and training around novel technologies and AI skill sets.

2. Establish better benchmarks: Look beyond time and resource spending for other ways to benchmark the ROI of new technologies (e.g., did you increase your understanding of biological mechanisms? Did you develop better product profiles? Did you automate operational processes?).

3. Partner with experts: Leverage third-party partnerships to help mitigate risk exposure while supporting organizational transformation. Look for specialists and experts who can complement and augment internal resources, aid scalability and generally advance the drugmaker’s goals and objectives.

It’s impossible to totally eliminate risk in drug discovery. However, AI-driven approaches and other breakthrough technologies are continually proving their worth, helping create tangible efficiencies and better outcomes for drugmakers working to move more therapies successfully from bench to bedside. In drug discovery, that’s a risk worth taking.