Artificial intelligence (AI) has no universally agreed definition. It roughly includes computing technologies that are similar to processes related to human intelligence; for e.g. reasoning, learning and adaptation, sensory understanding, and interaction. (1,2) At present, most AI applications are limited, since they can only perform specific tasks or solve pre-defined problems. (3) Recently, AI has been gaining significance in the field of healthcare. The AI industry is estimated to be worth $6 billion dollars by the year 2021. (4) A recent McKinsey review has projected that healthcare would be one of the top 5 industries to involve AI. (5)
Artificial intelligence aims to emulate human cognitive functions. It is transforming healthcare with the help of abundant healthcare data and ever growing progress in data analytics. Major AI techniques include machine learning methods (ML, for structured data) and natural language processing (NLP, for unstructured data). Important therapeutic areas that deploy AI tools include cancer, neurology and cardiology. (6)
Rapid technological advances in the world of life sciences are leading to ever more diverse and unstructured data, thus expanding the number of sources and the volume of accessible data. However, better decision making and long term business value are achievable by transforming these widely available data into actionable insights. For instance, marketing and medical affairs teams often deal with an increasingly complex network of external stakeholders and market influencers. In this case, AI can help faster obtain more comprehensive insights about key stakeholders at lower costs. (7)
Various data sets represent different stakeholders of interest. The scientific profile of these data sets provides a sound idea of the expertise; whereas, previous collaborative regulatory arrangements might offer profound understanding to support key decisions. Identification and extraction of this data from the hugely different, siloed sources is a herculean task, so is manually structuring the diverse information and mapping each activity to the right stakeholder, even at a small scale. Applying the same process for the global landscape and then re-doing the same thing each day to keep the data fresh and updated is unimaginably difficult. In these instances, if applied correctly, AI can structure, refresh and update actionable insights to improve decision making. The power of automation and AI can facilitate efficient use of huge volumes of information in decision making, instead of wasting time to just gather, structure and analyse the data.
The global medical affairs functions have seen a significant shift in the performance in last two years. The days of limited stakeholder engagement with the existing network are gone already. In many progressive companies, a data-driven culture has taken over, which goes all-out to produce better, most unbiased results. Such an approach provides a huge competitive advantage today and lays the foundation for a structural competitive advantage for future. Furthermore, implementation of AI can provide speed, operational excellence and long-term competitive advantage; thereby saving millions and decreasing costs. The advances of machine learning and AI open up a range of new possible approaches that, when applied, can create positive feedback loops to different parts of your organization. (7)
However, in spite of the attractive benefits of AI technologies, their real-life large scale applications are still facing obstacles. The first hurdle comes from the regulations, since they lack standards to assess the safety and efficacy of AI systems. This can be overcome by the guidance provided by the US-FDA for assessing AI systems. Another hurdle is data exchange which originates from the need of AI systems to be constantly trained by data from clinical studies in order to function properly. Also, continuous data supply can pose critical issues in further development of the system; post the installation of an AI system after initial training with historical data. (8)
Artificial Intelligence has the A) ability to analyse large and complex data sets in healthcare, and B) plenty of potential ranging across functions like discovering new medicines through to analysing journals, reports, trial results, conference proceedings and determining the best treatment for specific patients. (9) One of the predictions for AI is that it will greatly improve the efficiency of the drug discovery process, reducing both the 15-year timeline and $1 billion plus costs in getting a drug to market. (10) However, to achieve success in this rapidly growing AI environment will require implementation of tools of computer science (i.e. statistics, probability, and logic) to understand the complexities of healthcare datasets as well as the concepts of Big Data, deep learning and cognitive computing.
Become an Certified HEOR Professional – Enrol yourself here!
- Artificial intelligence technologies. Engineering and Physical Sciences Research Council.
- Artificial intelligence (AI) in healthcare and research. Nuffield Council on Bioethics. May, 2018.
- Preparing for the future of artificial intelligence. US National Science and Technology Council. 2016.
- Frost & Sullivan. Artificial Intelligence & Cognitive Computing Systems in Healthcare, 2016.
- McKinsey & Company, Artificial intelligence: The time to act is now. January, 2018.
- [Jiang F, Jiang Y, Zhi H, et al. Artificial intelligence in healthcare: past, present and future. Stroke and Vascular Neurology 2017; 2: e000101.
- Jansson F. 3 ways to boost marketing and medical affairs with artificial intelligence. March, 2018.
- Graham J. Artificial Intelligence, Machine Learning, and the FDA. 2016.
- Trends in medical affairs.
- Cox D. Artificial intelligence in healthcare. August, 2018.