Digital Adoption by Life Sciences Companies

Digital Adoption by Life Sciences Companies

Studies on digital adoption in healthcare companies have shown that the uptake of companies has been enthusiastic with blockchain and IoT but slow in manufacturing and supply chain, resulting in limited outcomes. Zymergen was an American biotech company founded in 2013 that applied genomics and machine learning to design organisms that produce chemicals. Although the company positioned itself as a tech-driven company and focused on its discovery platform, it neglected its biomanufacturing and commercialisation efforts. As a result, the company finally wound down for a much smaller value than expected and was acquired by Ginkgo Bioworks. An industry survey in India, a bespoke study conducted by Cognizant, found that while 71 per cent of life sciences organisations had planned manufacturing as the primary focus for digital innovation, only 10 per cent had moved beyond planning to implementation. A study of the Indian pharma industry found that seven out of ten digital projects fail to deliver results despite heavy investments. The reasons for projects not yielding the planned outcomes include poor adoption, legacy spreadsheets still in use and redundancy of digital tools. At times, over-investment in technology platforms without sufficient focus on process change, workforce skills and data readiness has also been the reason for failure to realise the benefits from tech investments as envisaged. The key takeaways from these examples include following through tech investments with robust execution plans. With tech access being available to all companies, only those firms who move with speed and adapt their processes to changes in technology, market needs and regulatory requirements would be able to stay ahead of their competition. Recognising this, many organisations have started inducting AI in their business models to create compelling value propositions for their customers.

The Immersion of AI in the Life Sciences Business

From 2010 onwards, AI and machine learning have started to get integrated into drug discovery and healthcare systems.

A number of devices, telehealth and wearables have been introduced and adopted by customers.63 Companies such as Pfizer and Moderna could develop Covid-19 vaccines with the help of mRNA platforms.64 AI is supporting scientists in sifting through large datasets to identify new drug targets and predict molecule-target interactions. Traditional drug discovery takes ten to twelve years or more and involves huge investments. AI helps in accelerating the process and reducing costs. Further, with the explosion of data through imaging, real-world data, biomarker data and genomics, the use of AI has become extremely helpful in studying and analysing the complex data. GenAI is being used to design molecules and predict properties. In clinical trials, AI helps to select appropriate patient cohorts, predict drop-outs and optimise the overall trial timelines and operations. It is also helping in mining real-world data to identify patterns and side effects for different population segments.

PandaOmics, the AI platform of Insilico Medicine, helps pharmaceutical companies to accelerate target identification, biomarker modelling and compound generation by leveraging large datasets derived from patients and sample sets.65 NantWorks, BostonGene, ExpanseWorks and CarisMPI are among the leading companies that are using AI engines for multiomics data integration and target identification.66 There are also Indian start-ups that are building businesses on the strength of AI applied to the omics/systems-biology domain. Proteomica, a spin-off from the Indian Institute of Technology Bombay, is an example of a venture that is blending AI with omics. Proteomica specialises in developing digital health solutions for personalised, evidence-based wellness utilising proteomics and multiomics.67 Sravathi AI is another Indian start-up that is working on the continuum from target discovery through to drug design in the systems biology/omics value chain.68 Vgenomics is using AI along with genomic data to accelerate diagnosis and target discovery aimed at rare diseases.69 Life sciences companies are using AI in manufacturing in biologics production with sensory data, prediction of equipment failures, process cycle time reduction and optimisation of processes, and in supply chain optimisation with demand forecasting and inventory optimisation. Pfizer has reported that the AI-powered manufacturing processes aimed at detecting anomalies in production, reduction of process cycle time and yield improvement have resulted in an increase in throughput by 20 per cent and reduction in cycle time by 25 per cent. Pfizer could also deliver the oral medication for Covid-19 faster because of the deployment of AI-enabled optimisation in manufacturing and supply chain.70 Novartis has deployed digital twins of its manufacturing lines by using AI/machine learning with the partnerships built with Amazon Web Services, thus making production more robust, reducing downtimes.71 Apart from pharma companies, specialist software companies such as Blue Yonder support drug manufacturers in their supply chain optimisation and play a significant role in risk reduction, lowering of cost and improvement of responsiveness with their solutions.72 AI is being used in the design and development of devices and analysis of images. Siemens Healthineers has integrated AI into its imaging systems and diagnostic workflows, which enables the generation of consistent imaging across sites and standardisation in diagnostics.73 With the democratisation of mobile access, medical device manufacturers such as AliveCor have been able to democratise access to diagnostics, reduce hospital costs and detect medical conditions earlier with their AI-enabled portable devices.74 Pathology is an important part of diagnostics. Use of AI has been improving the speed, throughput and accuracy in pathology that help with treatment pathways. Ibex AI provides diagnostic solutions that help pathologists analyse tissue slides for cancer efficiently.75 The extensive use of AI in diagnostics would help developing countries like India with cost-effective wearables and devices and treatment for hitherto unaddressed diseases. The potential for further AI integration is significant in the entire value chain of drug discovery, manufacturing processes and supply chain and in the device design and diagnostics support systems. However, the stakeholders engaged in designing and implementing AI-driven solutions recognise that the quality and access to data is complex and that data may not be standardised. The methods used by AI models for arriving at decisions have to be made transparent as they concern the health and safety of consumers. Regulatory approvals would be required for new AI models, and they also need to be validated through appropriate channels. Ethical concerns, bias creep and patient data privacy are concerns that are yet to be addressed by all the stakeholders concerned while building AI models. While there is a huge potential for deployment of AI in life sciences in India, which could lead to customisation for different population sets at low costs with genomic diversity, data volumes and patient populations being large, there are still challenges of data infrastructure, interoperability and regulations that need to be addressed.