Artificial intelligence in healthcare is an overarching term used to describe the use of machine-learning algorithms and software, or artificial intelligence, to emulate human cognition in the analysis, interpretation, and comprehension of complicated medical and healthcare data. Specifically, AI is the ability of computer algorithms to approximate conclusions based solely on input data. What distinguishes AI technology from traditional technologies in health care is the ability to gather data, process it and give a well-defined output to the end-user. AI does this through machine learning algorithms and deep learning.
So it stands to reason that a Canterbury Business School lecturer with a background in computer science and artificial intelligence is making significant strides to help speed up the development of medicines. Dr Pan Zheng has been working on this research for a year with colleagues in China, Malaysia and Spain. “I did study science, but business schools are so multidisciplinary. I’m working in information systems. It is a discipline that is always trying to help the management and operations of organisations work smarter and more efficiently,” he explained.
Because “it costs millions of dollars to produce a new medicine and takes a long time before it can be released to the public”, Dr Pan is researching a novel way of speeding up the process using machine learning and AI to assist drug repurposing. This approach to drug repurposing saves time and money, making it quicker and cheaper to make medicine publicly available as well as reducing the cost for the end-user.
At a molecular level, the chemical structure in medicine reacts with the protein structure of a disease to suppress or stop it, providing a cure. He is studying the mapping from the chemical structure to the protein structure to construct a machine learning model that predicts the possible relations. In biochemistry term, it is called binding affinity. There are existing databases that keep such information. Machine learning models can be created and trained using these data sets. When there is a new protein structure, i.e., a disease, input to the model, it will recommend some existing medicines that could be possible cures is confident Dr Pan.
None the less, as with all machine learning models, it won’t be 100%, but with continued improvement, he is optimistic to get to 80%, 90%, 95% accuracy. Essentially, he explains, the more information that goes into the model, the better the solutions it will present – quicker and more accurate.
Another medical application for this research is to use machine learning and AI to test for Alzheimer’s disease, which is often done on paper with a doctor. Additionally, it could free up the doctor to see more patients. From an AI perspective, Dr Zheng says they are trying to liberate humans from tedious work by giving it to machines.
AI could contribute more than $700 million of value and savings per year to the New Zealand health system by 2026, according to the AI Forum of New Zealand’s 2019 report, which says New Zealand’s district health boards are looking at a $500 million annual deficit. Executive Director Ben Reid acknowledged that the health sector in New Zealand is facing challenges. These include increasing demand, rising consumer expectations, and the pressures of an ageing population. These factors are straining the health workforce, increasing costs and limiting access to care.
AI’s contribution, Reid says, to the NZ health system could rise to between $1.6 to $3.6 billion by 2035. AI promises to bring significant clinical, workforce and cost benefits to the health sector, as well as personalise medical care, opines Reid. It can help with predicting disease and injury and mine vast quantities of literature for research insights.