The way businesses operate and use technology is changing drastically – more companies are now relying on AI and Machine Learning to keep up with industry expectations and stay ahead of the curve, continue to innovate and create an agile business model.
Many health and pharmaceutical firms have been closely analysing the capabilities of AI and machine learning to help with drug discovery and analyse the possibilities of new and effective treatments. The impact of COVID-19 has significantly affected businesses around the world to shift their priorities to help users and clients adopt to new norms. The crisis may be a catalyst for change, galvanising public and private sectors to collaborate to address the health challenge. This has been a turning point for health startups, who have had to focus on innovation and technology to fulfill challenges in an uncertain climate.
How health startups use artificial intelligence
AI has the potential to transform clinical decision making through utilising the vast amounts of behavioural and biographical data that is generated across the health system.
An example of a healthtech firm using AI is Exscientia; by fusing the power of AI with the drug discovery process, the company is able to automate drug design, surpassing conventional human approaches.
The company explains that their systems learn from both existing data resources and experimental data from each design cycle. The principle is similar to how a human would learn, but the AI process is far more effective at identifying and assimilating multiple subtle and complex trends to balance potency, selectivity and pharmacokinetic criteria.
As a result, the AI-driven process is more likely to achieve the end goal and to do this more rapidly and efficiently than traditional human endeavour.
Benevolent AI are another firm using artificial intelligence to identify and analyse the possibilities of already approved drugs being effective in treating coronavirus. Pando, a healthcare messaging service, has quickly developed a forum for the current crisis that allows medics to share vital information quickly and securely.
London-based artificial intelligence company Deepmind, which was founded in 2010 and acquired by Google in 2014, have utilised their recently-published deep-learning system AlphaFold in the hope that they can feed into wider research around the virus.
Alderley Park-based John Overington, CIO at Medicines Discovery Catapult commented, “I think there will be so many areas of business and science that AI will affect. Clearly in the patient care side of things there are already big advances made in image based diagnostics, and patient stratification. I think there will be a lot of opportunities around IoT devices and integration of monitoring devices into phones and watches, and these data sources will hopefully lead to early diagnoses and interventions.
He continues: “Earlier in clinical trials there will be additional impact, in the design and running of adaptive trials, and analysis of patient groups. However, for me personally, the most interesting advances are likely in preclinical research, where the connection of smart algorithms to automation could transform research. We’re already seeing great advances from many groups in compound design, and retrosynthesis.”
Waqar Ali Shah from Chief.AI who are based at Bruntwood SciTech’s Manchester Science Park believes that the vast amount of data being generated by patients and equipment in healthcare systems around the world lends itself well to leveraging by machine learning and AI.
He says, “Artificial intelligence can increase healthcare efficiency by great orders of magnitude, shaving hours off radiology and physician labour. Algorithms trained on historical drug interactions can predict new uses for existing medicines, taking drug repurposing and discovery to a whole new level. However, the disadvantage of AI is that prediction algorithms are only as good as the data they are trained upon.
“Therefore, it is critical for the industry to ensure that the machine learning datasets made available to algorithm developers are relevant and well curated to the disease area. It is also essential to create global data sharing agreements and protocols to enable the free flow of quality training data to machine learning developers around the world.”
He continues, “The same can be extrapolated for almost all business areas outside of healthcare. Once relevant, quality data is brought to bear, there is no limit to the business efficiency that can be achieved, freeing up human efforts to look at those rare, truly complex problems that are hard for a computer to solve.”
The value of healthcare and AI
According to ‘Rude Health’, a new research report on the AI in Healthcare sector, “AI has potential applications across the entire life sciences value chain, including drug discovery, clinical trials and patient-care, in addition to potential improvements in speed and efficiency of company operations.
“The global market was worth $2.1bn in 2018, with exponential growth to $36.1bn predicted by 2025, at a CAGR of 50.2%.
“However, AI presents various new challenges, and the pharmaceutical industry has highlighted many technologies in the past that promised to drive productivity, but nothing has yet worked on a large scale. Nevertheless, the authors believe that AI is likely to become a greater differentiator in the next 5-10 years and the report presents case studies and real-world examples of the benefits it could provide.”
Farid Khan from Chief.AI adds, “Currently it takes pharmaceutical companies about 10-15 years and $1 billion to develop and market a drug. The costs are largely due R&D, marketing and the testing of the drugs in people called ‘clinical trials’.
“The use of AI will never solely rely on technology to predict drugs that work- we will always need to test- or at least demonstrate this in laboratories (‘in vitro’). This is because we do not have the means to model a human from a molecular perspective. Perhaps we will one day but that will be a long way off!
“Today it is possible to predict the desirable properties of drugs and look at ‘repurposing’ which has a great advantage in fast tracking treatment for diseases, but we still have to demonstrate the ‘theory’ to practice. That’s why partnerships with academic institutes are important.”
The healthcare sector is particularly primed to benefit from AI because it deals with large amounts of data on a daily basis. Companies should begin building their AI infrastructure now or risk falling behind. The report states: “AI-experts are not an unlimited resource, especially given that most of the talent in this sector naturally gravitates towards traditional IT and technology companies. Even in companies dedicated to AI in healthcare, AI-experts typically only comprise around 15% of staff.
“Another important motivator for healthcare companies building AI infrastructure now is the entry of Big Tech into the healthcare field. Google, Microsoft, Amazon, Apple and Facebook have already announced different initiatives to enter the healthcare space – companies which will have fundamental advantages in terms of AI capabilities.”
There has also been a significant rise in healthcare companies employing artificial intelligence – and these look set to be increasingly attractive investment options as technology advances, rising data generation, the growing number of deals, and the competitive advantage AI can provide all combine to raise the potential performance of these businesses and the returns investors can expect.
Overington adds, “The availability of relatively low cost compute resources, and public data from resources like PubChem and ChEMBL have allowed the development of a large number of novel algorithms and their application in both large companies, and in small companies. I think the UK really has been an excellent location for innovation here, and some of the real advances are being done here.”
Euan Cameron, PwC UK Artificial Intelligence Leader, said, “We’re currently piloting a new scheme. Employees are using wearable technology for the duration of the pandemic to enable us to track biometric, cognitive and psychometric data.
“Machine learning will help us identify patterns on overall stress levels and cognitive performance. Not only will this anonymised data help us assess how people are coping with work, it will help us to see how we can then manage things better – not just now but with remote working in the future.”