Within AI, sub-categories include:
- Machine learning (ML) – enables a computer program or algorithm to develop the ability to learn a new task and perform that task without explicit assistance via programming instructions.
- Supervised machine learning (SML) – similar to standard machine learning, SML is based solely on input data to generate a predictive model.
- Deep learning (DL) – operates as a subset of machine learning that learns through an artificial network of unstructured data.
- Natural language processing (NLP) – a computer program that can understand human speech drawing from computer science and linguistics.
- Virtual screening (VS) – involves the predictive capacity to examine large library volumes in silico.
The intersection between AI and pharma executives
In 2020, OpenText surveyed 125 pharmaceutical executives to determine how familiar each respondent is with AI technologies within their industry. The survey results revealed that an interest in AI increased to 85% in 2020, up from 47% in 2018, when a previous similar survey was conducted. Approximately 75% of respondents indicated they intend to or plan on using data scientists’ analytics centers of excellence. Unlike with previous OpenText surveys, respondents indicated that issues with regulation and promotional content shy rocked to the top of the least organized areas, as opposed to document management and processing of regulatory submissions, which were well-defined areas.
The benefits of AI in pharma
The incorporation of AI into the pharma industry provides a number of tangible advantages. In particular, AI can:
- Speed up processes by managing large volumes of data.
- Used to reshape the key steps of trial design, especially when it comes to clinical trials for neurological therapeutics.
- Optimize manufacturing by replacing older processes that typically rely on human intervention or input, eliminating the possibility of human error.
- Identify marketing techniques that attract customers.
- Optimize the manufacturing of drugs and related products by reducing the need for human intervention.
- Reduce waste and design time.
- Improve quality control, predictive maintenance, and reuse.
Barriers to implementing AI in pharma
Further findings from the recent OpenText survey revealed, the percentage of companies looking at next-generation technologies dropped from 23% in 2018 to 19% in 2020. Other revelations from the survey indicated that many pharma executives think AI is “too complex” or they “don’t know where to start” to implement the technology. This may be attributed to a lack of understanding about the opportunities and limitations of AI in pharmaceuticals.
Another barrier to adoption is company infrastructure. OpenText issued the same survey two year prior and found 27% of respondents reported that their companies were still mostly paper-based or operated within siloes. In 2020, 35% of respondents reported the same conditions.
Recruiting and retaining data scientists with an AI skillset also presents another hurdle for widespread adoption. Companies will need to further invest and learn to retain top talent in an effort to boost their own capacities.
Finally, an absence of a properly connected regulatory framework on data collection is missing for many organizations. While formats differ widely, they are largely not compatible with each other, or aren’t digital at all.
How pharma is currently using AI
In terms of pharma’s practical applications, they are abundant. For instance, at MIT a team used ML to help identify a new drug that can target antibiotic-resistant bacteria. The Coalition Against Major Diseases presents another good example. The collation utilized a clinical trial simulation tool to target Alzheimer’s disease. The tool which incorporates computational components for drug modeling can be used for clinical trial design.
Pharmaceutical companies like Johnson & Johnson, Pfizer, and Novartis have recently announced initiatives to use the latest AI technology, which combines human experts with augmented intelligence to translate large amounts of patient data.
Reduction in cliental clinical trials is another prime example. Bayer, one of the largest pharmaceuticals and life sciences companies worldwide, is harnessing the power of AI to conduct virtual clinical trials through home delivery of medical supplies, e-consent forms, and telemedicine. Other devices, such as wearables provide access to live data on a constant basis, eliminating the need for frequent in-person visits.
In particular, when we think about practical applications in the realm of health, medical devices like wearables are very much well-suited than in the past to provide data in a digital and accessible format, making them excellent candidates for AI to utilize their information, given the large volumes of patient data that is generated every moment of the day.
A look forwards the future
COVID-19 has served to both propel and expose life sciences organizations to the areas that need improvement and the tools that can help them adapt while maintaining business continuity. Current trends that were at the infancy stages of consideration have only been accelerated in application during COVID-19. Specifically, unlocking the power of data, platform interoperability and patient centricity have become paramount.
AI within the global healthcare ecosystem is likely going to increase. COVID-19 is not the first global health calamity the world has seen and nor will it be our last. However, what this pandemic has done is stimulate a discussion and interest given all of the benefits AI can offer. In particular, it may help shift practitioners and executives’ way of thinking into adoption of new transformative technologies that will likely spur further research and development.
The use of AI in pharma doesn’t stop within drug manufacturers, but also harnesses others within the industry like biologists and biopharma scientists. Together, all practitioners can unitize AI and ML to shape the future of life sciences. Companies that have this mind and begin the journey of transforming their practices and process will be better equipped to manage the fallouts of the next global health crisis when it strikes next.
Ferdi Steinmann is responsible for the LS global industry strategy development at OpenText.
Notes
1. Sinhasane, S. (2019, October 4). Artificial Intelligence in the Pharmaceutical Industry – An Overview of Innovations. Retrieved November 20, 2020, from https://mobisoftinfotech.com/resources/blog/artificial-intelligence-in-the-pharmaceuticalindustry/
2. Harrer, S., Shah, P., Antony, B., & Hu, J. (2019). Artificial Intelligence for Clinical Trial Design. Trends in Pharmacological Sciences, 40(8), 577-591. doi: 10.1016/j.tips.2019.05.005. Retrieved November 20, 2020, from https://www.sciencedirect.com/science/article/pii/S0165614719301300
3. Anyoha, R. (2017, August 28). The History of Artificial Intelligence. Retrieved November 20, 2020, from http://sitn.hms.harvard.edu/flash/2017/history artificial-intelligence/
4. Harrer, S., Shah, P., Antony, B., & Hu, J. (2019). Artificial Intelligence for Clinical Trial Design. Trends in Pharmacological Sciences, 40(8), 577-591. doi: 10.1016/j.tips.2019.05.005. Retrieved November 20, 2020, from https://www.sciencedirect.com/science/article/pii/S0165614719301300
5. Devoy, M. (2019, December 19). Artificial Intelligence – Technology-Driven Disease Prediction to Advance Patient Care. Retrieved November 20, 2020, from https://pharma.bayer.com/artificial-intelligence-technology-driven-disease-predictionadvance-patient-care
Credit: Google News
