According to data from Statista, global pharmaceutical sales now stand at some one trillion US dollars. The United States is the largest single pharmaceutical market by far, contributing more than US$450 billion in revenue, while Europe generated around US$214 billion.
Cutting-edge technologies such as AI, blockchain and robotics are steadily transforming drug R&D, pharmaceutical logistics and drug distribution by leveraging extensive data and improving inefficient workflows.
Deep Learning in Drug R&D: Neural network models trained by large-scale chemical compound datasets can enhance target selection and lead compound screening (LCS) research, reducing screening time and improving drug discovery efficiency.
Intelligent Pharmaceutical Logistics: Taking advantage of machine learning, IoT and other advanced technologies, pharmaceutical logistics innovations include quick loading, scheduling systems, and path optimization. Blockchain technology meanwhile can make the entire drug circulation process traceable.
Automatic Drug Distribution: Medical robots can substitute for humans in automatic drug delivery, sorting and dispensing tasks. By using advanced voice interaction and other technologies, robots can also directly interact with humans and provide drugs to patients.
Medical Data Mining: With data mining, NLP and other AI technologies, researchers can analyze large amounts of medical information such as drug data, literature data, medical patients and Internet data, adopting correlation analysis and generating valuable insights.
Intelligent Technology Usage Scenarios in the Pharmaceutical Industry
In the pharmaceutical industry, intelligent technology usage scenarios can be divided into two main categories: Drug R&D and Drug Market. In the field of Drug R&D, AI is used primarily in drug discovery tasks such as lead compounds discovery, target screening, etc. AI’s role is more diverse in the Drug Market, which includes medical robots, pharmaceutical logistics, pharmacy automation, drug compliance and so forth.
Panasonic — HOSPI-R: Using pre-programmed maps and recognizing its surrounding environment via sensors, the mobile HOSPI-R bot can autonomously route-plan and transfer materials such as drugs and specimens in hospitals. If destinations or layouts change, information can be updated in real time. The bot is designed to accelerate and navigate carefully to ensure safe and stable deliveries.
Xiaohe Intelligent Technology — VV-BOX: This is a dispensing device with different pill compartments which automates the drug dispensing process and ensures medications are taken in line with prescriptions. Moreover, if a patient forgets to take their drugs at the prescribed time, VV-BOX can discern whether a known person or clinical staff is nearby and alert them.
Pfizer — AI-Powered Drug Discovery Platform: Along with its existing relationship with AI biotech company XtalPi for crystal structure prediction (CSP), Pfizer is also developing an AI-powered software platform for accurate molecular modeling of drug-like small molecules. The goal is to optimize the process of computation-based rational drug design and solid-form selection.
IBM — Watson for Drug Discovery: Not all AI in healthcare efforts proceed as planned. Reports published in April 2019 suggest IBM will stop sales of its Watson AI software for drug development. It’s believed IBM will turn its AI efforts to clinical development products, exploiting a huge network of healthcare partners that includes pharma giants Pfizer and Novartis.
Trends and Limitations for Intelligent Technology in Pharmaceutical Industry
Although there are many merits to applying intelligent technology in the pharmaceutical industry, pharmaceutical-related data is currently relatively difficult to collect and utilize efficiently as it is typically scattered across different companies, institutions, and hospitals. This is an issue that could be addressed as part of the current global dialogue on data privacy.
AI technology in drug discovery research now mainly focuses on small molecular compounds discovery and has difficulty dealing with biomacromolecules for complex structures. Meanwhile, the scale of the existing compound databases also needs to be expanded to adapt to the increasing demand for large-scale drug compounds sample training.
Overall, the outlook for intelligent technology in pharmaceuticals remains positive: The drug target discovery process is being accelerated by literature data mining, while transfer learning is being applied to algorithm model training in drug discovery to solve generalization problems associated with small sample compound datasets. Quantitative calculation combined with AI meanwhile is also being used to speed up the process of solid phase screening and drug design.
Author: Haitian Fu | Editor: Michael Sarazen