Artificial Intelligence (AI) was consigned to big screen science fiction movies. But now the it can be used to make healthcare personalised.
Artificial intelligence (AI) is an umbrella term for technologies no longer restricted to following precise codified instructions given by human programmers; instead AI can learn to solve problems by itself!
AI systems can build their own assumptions, spot trends, and create insights from massive datasets. It’s a fast-moving sector.
In 2010, Google developed driverless cars. In 2011, IBM’s Watson supercomputer beat the human champion at TV quiz show Jeopardy. By 2017, a 19-year-old boy called Joshua Browder developed a language AI chatbot from the comfort of his own bedroom and named it ‘DoNotPay’.
To date it has helped over 375,000 people claim back $9M in unfair parking fines. As AI races ahead, the question of human superiority in pattern recognition, language, prediction and even judgement is being shaken to the core.
In terms of healthcare AI, here are some regularly reported quick-fire predictions:
The AI health market is expected to reach $6.6 billion by 2021, growing at 40% CAGR,
AI in medical imaging and diagnosis is projected to exceed $2.5 billion market size by 2024,
Health AI is predicted to make $150 billion annual savings for the US healthcare economy by 2026.
As AI disrupts the healthcare market, there will be winners and losers, just as there have been in the finance sector. Healthcare is becoming dominant for AI investments, attracting 15% ($748M) of all funding in 2016.
Of all companies valued over $1bn, 40% are in healthcare and the AI ‘healthcare unicorns’ include:
Flatiron (raised $175M, at a $1.16Bn valuation in 2016) a platform that extracts, connects, and structures cancer data using machine learning to fuel the fight against cancer.
iCarbonX (raised $154M at around a $1Bn valuation in 2016) reached unicorn status within six months. Their pledge is to create a complete health ecosystem for an individual, based upon genomics, proteomics, transcriptomics, immune response and analysis of gut bacteria and lifestyle factors. They leverage machine learning to create actionable, individualised insights.
BenevolentAI (raised $100M at a valuation of $1.8Bn in 2015) use AI to combat the mismatch between healthcare information (a new scientific paper comes out every 30 seconds) and human’s ability to process it. The firm has had early success through discovering a new compound, which has since been shown to delay the progression of motor neurone disease. It’s also had some success with finding a new drug for Amyotrophic Lateral Sclerosis (ALS).
Investment in ‘Healthcare AI’ is growing faster than ‘AI as a whole’, albeit from a low base.
How is AI shaping healthcare?
AI is evolving fast and clinical experience highlights it will change our current care pathways. For the naysayers out there, the future needs the combination of man and machine to cope with the pincer movement burden of an increasing ageing population with more long term conditions, alongside an acute on chronic staffing shortage.
This blog considers how AI is shaping the ‘pre-primary care space’ as a way to relieve the strain at the front door of primary and secondary care.
This can be divided this into three areas:
1) Internet searches:
The most traditional method has been to type in a medical query into ‘Google’ and return a momentary cancer scare!
2) Symptom checkers:
These are used as an on line second opinion. Some symptom checkers also provide triage information, such as has been running since 2000, started by the founder whose daughter Isabel suffered a near fatal misdiagnosis.
Since 2012, Isabel has been programmed with AI advanced natural language processing to cover 6,000 disease symptoms and is focused on helping people double check their diagnosis.
3) Personal health assistants:
These platforms leverage AI to discover information that a person may not have spotted as a symptom, so are a bit more like taking a clinical history. Other AI technologies track personal symptoms over time without ever being admitted to a hospital.
Examples of companies in this space include Medopad, Your.MD, Ada, Buoy, Doctor Care Anywhere and Babylon.
Unlock and unleash Healthcare AI potential
Healthcare has lagged behind in terms of AI. Companies such as Netflix and Amazon already harness masses of available consumer data to feedback personalised tempting offers, direct to the end user.
If healthcare AI follows the same trajectory, it will create a new wave of opportunity for businesses, investors, and customers.
In the short-term, efficiency savings and data optimisation will predominate, increasing profits, and creating new revenue streams. In the longer-term, truly personalised healthcare could become a reality.
One thing for sure: AI is here to stay and as we age we shall become increasingly grateful for it.
By Dr Michelle Tempest (@DrMTempest)
Partner at Candesic, healthcare management consultancy