Dr AI: healthcare’s new brain
Massive advances in technology are changing healthcare systems at an unprecedented rate. Last week, I wrote about the data revolution as an enabler of superior, personalised healthcare. As data storage and processing costs plummet, the variety and richness of data, and our ability to analyse it accurately and rapidly, are exploding.
One area of huge new potential in healthcare and other fields is artificial intelligence. AI is a branch of data science that is developing “intelligent” machines capable of crunching massive volumes of data in order to replicate key aspects of human intelligence such as reaching insights, making decisions, interpreting images or making predictions. Machine learning, a subset of AI, enables computers to learn as they work, becoming better and better at carrying out specific tasks.
AI is being applied across many industries and economic sectors, and healthcare is no exception.
Consider how often we turn to “Dr Google” for medical information, only to be overwhelmed by vast amounts of often confusing information. By contrast, AI-based healthcare systems can quickly learn how to respond to consumer questions with relevant, increasingly accurate information, and to provide high-level medical guidance.
Discovery Health’s DrConnect app integrates sophisticated AI technology to give members credible answers to questions about medical symptoms.
As users share details, the AI system learns about what the member is asking, poses additional relevant questions and quickly gives a high-level indication of possible diagnoses ranked in order of likelihood, with advice on whether to consult a doctor and how urgently.
Of course, this type of AI-based symptom-checker can never replace doctors. Rather, in a world in which patients are empowered by the free flow of digital information, AI platforms channel people to credible content and a deeper understanding of the best steps to take next.
Another application of AI is in assisting doctors to make faster, more accurate clinical decisions. For example, AI systems have been shown to be as accurate as highly trained radiologists in identifying breast and brain cancers from scan images. Does this make radiologists redundant?
Definitely not — it’s likely that AI systems will read vast numbers of images very quickly, and identify those that need review by highly skilled and experienced radiologists, allowing their scarce skills to be used where they are most effective. AI systems, with radiologist backup, could also provide desperately needed skills to patients in remote communities who cannot get to a radiologist.
AI is also supporting doctors in highly complex fields, where information accumulates so fast that it is virtually impossible for doctors to keep up. In some of the world’s leading cancer centres, AI systems are being deployed to assist oncologists to identify suitable treatments for complex patients, and to identify clinical trials suitable for their patients. In the pharmaceutical industry, AI systems power the process of discovering new medicines, making human researchers more efficient in their work.
AI systems can also bring dramatic improvements to customer service in healthcare and beyond. Discovery Health’s Ask Discovery chatbot, which is available on our website, uses an AI system that allows members to ask any question, interprets natural language and provides answers with relevant links on the website. Over time, as the system learns, these answers become more accurate.
There are also many new AI applications emerging in the patient/consumer space where AI might, for example, assist patients with chronic or mental health conditions to adhere to medication and better manage the condition. Similarly, advances in AI systems will support doctors as they treat patients, both through decision-support and diagnostic services.
This may lead you to wonder whether we will all be treated by machines in the near future. In my view, healthcare is far too complex and, quite simply, too human for that type of scenario. Instead, we are likely to see a healthcare system in which health professionals and intelligent machines collaborate across numerous aspects of prevention, diagnosis and treatment. This type of system will allow human clinicians to be far more efficient, and to spend their precious time on doing what doctors do best — truly caring for their patients.
Dr Broomberg is CEO of Discovery Health