By Warren Wang, Senior Vice President and President, APAC
The concept of artificial intelligence is exciting for any industry, but it shows considerable promise in healthcare. With healthcare providers now collecting vast amounts of data from physicians and patients, there is a huge opportunity to use AI and machine learning to improve healthcare delivery outcomes and save lives.
While AI as a concept has been around for hundreds of years, it's finally begun moving beyond the basics of simple computation to actually begin mimicking human decision-making processes. Across every industry, we're seeing AI and machine learning applications being able to do some of what humans do, at a much faster rate. Because of the huge potential for AI in healthcare, it's no surprise to learn that investors see these technologies becoming a vital part of the healthcare ecosystem.
As an indication of the private market, a 2018 report from CB Insights found that healthcare AI startups have raised $4.3B USD across 576 deals over the past five years, topping all other industries in AI deal activity. The EU has also recently announced their intentions to invest $24B USD into AI by 2020, and with BIS Research predicting the global market for healthcare big data is set to hit $68B USD by 2025, AI is likely to represent a sizeable proportion of that market.
In terms of where AI is showing real results in healthcare, AI-assisted diagnosis has plenty of ground-breaking examples:
- A Stanford University project fed 130,000 images of skin lesions into an AI software program. By training and testing the software against the opinions of 20 qualified dermatologists, the software developed the ability to accurately identify the most common types of skin cancers.
- DeepMind from Google have developed an AI that screens retinal scans for conditions such as glaucoma, diabetic retinopathy and age-related macular degeneration.
- Oxford University have been developing AI systems for interpreting echocardiograms, which are ultrasonic scans of the heart. The system can detect changes invisible to the eye, and improve the accuracy of diagnosis by up to 20%
AI and machine learning systems will always have a degree of error involved in their predictions, so their real value will only be realised when paired with medical expertise. While AI can automate areas where we have clean, consistent data to analyse, and help us to identify outliers, human judgment will always be essential. But AI will be able to drastically reduce the amount of burden and time it takes for human beings to make diagnoses and determine the best treatments. There are also game-changing ramifications for the development of many new drugs, vaccines and medicines.
At Boston Scientific, we are continuously exploring new applications for AI systems to benefit clinicians and patients. We also collaborate with research organizations and early-stage ventures to identify new potential uses for AI and machine learning. Our annual Connected Patient Challenge awards USD$50,000 in in-kind support to winners who can find new ways of leveraging digital health solutions to advance patient care.
One thing is for sure - AI and machine learning are only just beginning to come into their own in terms of revolutionary capability, but the true revolution won't be in the immediate future. All of humanity is set to reap the rewards of these new capabilities, and this is particularly the case for healthcare providers and physicians, who will have faster and better access to the right insights for informing excellent care provision to patients.
This article first appeared on LinkedIn.