Cognitive reasoning – a dose of stability for drug development

Friday 22 September 2017


AI might be accelerating drug discovery, but cognitive reasoning can offer the pharmaceutical industry much-needed reliability and traceability.

Drug research and development is an extensive, costly, and highly regulated process: most new drugs don’t reach the market for around 12 years after their initial discovery, and the entire process can cost in excess of £1.2 billion.

Much of the expense in drug R&D is incurred through the laborious process of testing initial leads – most of which are deemed unsuitable, and have to be discarded. Yet, stringent but necessary approval processes also make it difficult to progress a drug to the next stages of testing, and to achieve market readiness, at pace. Authorities like the FDA rely on data submitted by sponsors to decide whether a drug should be approved – the need for thorough access to all results within the testing process makes drug approval an extremely data heavy operation.

This slow process is not just costly in the lab: the longer a drug’s journey to market, the more expensive it becomes. In developing countries, this is particularly problematic because consumers are often only able to afford a partial course of a drug – severely limiting its therapeutic success.

Who patents wins – but AI isn’t just about acceleration

To keep pace with the growing demand for access to new – and affordable – drugs, the pharmaceutical industry needs a long-term, scalable AI solution which can streamline its clinical development and regulatory processes, without compromising standards.

Major players in the industry are already working with external AI companies, as well as developing their own internal capabilities – in only July of this year, GlaxoSmithKline unveiled a new $43 million deal with a company in the AI field – and tools like IBM Watson for Drug Discovery are aiming to accelerate breakthroughs by analysing, identifying, and prioritising new targets for drug development.

These processes of data management, lead generation with High Throughput Screening (HTS), and lead optimisation are the main focus of many AI solutions currently disrupting the pharmaceutical industry. Now, researchers can use AI to: interpret research data; discover innovative hypotheses; narrow down a more accurate selection of potential compounds much earlier in the process, and rapidly optimise these leads.

AI is accelerating the pace of drug development, saving vital time and money. Ultimately, companies can patent their products – and consumers can access life-changing new drugs – much more quickly.

But as AI rapidly takes on these increasingly complex roles, it is equally important for researchers to accelerate their own understanding of how developmental decisions are (and should be) made using AI-powered automation.

So, how can companies ensure that they are making the most of AI solutions in drug development?

Optimising AI – What can cognitive reasoning do for pharmaceuticals?

With increasingly stringent regulations from governing bodies, it is essential for researchers to provide a thorough explanation for the development of their decisions. Without this, approval processes and the delivery of a drug to the market are vastly impeded. If a drug is removed from testing, regulatory authorities, funding sources, consumers, and researchers need to know why. And if a drug is deemed successful and advanced to the next stage of testing, the same rationale applies.

Factors that delay an approved drug being brought to market could be automated with AI, like the task of determining the price of a drug in international markets. By rapidly analysing data about the prices of previous, similar products, a cognitive reasoning system could easily take over this manual process – accelerating a drug’s delivery to market, without compromising standards. Companies could also improve transparency and consistency in their pricing, as automated reasoning could clearly highlight the justification for a drug’s value to regulatory bodies and consumers.

The most serviceable AI solutions can work in tandem with researchers to help streamline these human processes, allowing data to speak for itself, and removing excess human subjectivity. Yet, when a fully traceable rationale is provided for each decision, researchers can remain in control of the process, allowing them to note interesting anomalies and easily support their findings with thoroughly supplied explanations.

AI has the potential to boost the progression of innovative new discoveries, by freeing up more time and money to undertake these ‘riskier’ therapeutic ventures.

Easy to use, scalable decision making

It’s becoming apparent that accelerated drug development is of minimal use to us if we cannot understand the rationale behind rapidly automated decisions.

With Rainbird’s AI-powered cognitive reasoning, experts are replicating and scaling their human process of decision making in specialised fields. Researchers can model their knowledge and the rules that they use to make decisions – such as when to continue testing a lead, and when to stop – allowing Rainbird to keep track of the reasons why certain decisions are made, and follow the same human reasoning process.

Regulatory bodies, like the FDA, acknowledge that ‘sometimes the benefits and risks [of a drug] are uncertain and may be difficult to interpret or predict.’ By using cognitive reasoning to provide fully traceable, automated decisions, the scope for standardising and improving the efficiency of approval processes is vast – encompassing everything from automating the calculation of pricing for international markets, to providing reasoned decisions that justify why a particular drug should continue to be trialled.

Improving efficiency by accelerating R&D is clearly a priority for the pharmaceutical industry – ultimately enabling vital drugs to reach the market faster – but the need for reliable, human-like auditing and traceable decisions will always remain at the forefront of research and approval processes.

With intelligent, automated reasoning, approval processes can be streamlined and accelerated, enabling vital new drugs to be brought to market sooner, and at a lower cost to the consumer.

To find out more about how Rainbird’s automated decision making is already benefiting regulated industries, take a look at our case studies, here.