Okay, it’s the year 2500. The civilised world has changed beyond recognition. Almost. There are, you see, still unmistakable reminders of the year 2019. Brexit has still not happened and a defiant VTM (virtual Theresa May) still believes her deal will get through parliament (at the 3,000th attempt, no less). Meanwhile, other areas of life are advancing unabated – mind controlled cars are the norm, Alexa is playing the correct song at the first time of asking and, above all, people are maximising the astonishing potential of their bodies through the ever-accelerating synthetic wisdom of artificial intelligence (AI).
So, how do we arrive in a future where billions of mind- boggling data sets are consumed and interpreted by AI: routinely saving, extending or transforming lives, while developing new pharmaceuticals which bring hope to millions. Getting into the healthcare DeLorean (if you don’t understand that reference, your parents will) and entering a desired year, is all too tempting but, instead, we must lay the foundations of AI in the present. After all, what we put in now will change the snapshot of our collective futures forever.
Barely human
The original latin phrase translated as ‘one for all, and all for one’ perhaps crystallises best what AI should and could ensure in a healthcare context – that ‘nobody is left behind’. It is a noble declaration in the finest tradition of medicine, but beneath it there are bold, futuristic ambitions which – if realised – will send healthcare into light speed, allowing mankind to rage against the dying light for even longer. AI is, in many ways, the ultimate riposteto the dithering of human beings in charge of Brexit – in contrast, its machine precision simply cuts through formalities in pursuit of unrivalled efficiency.
But we can’t get too friendly with AI – it’s not going to go to Nandos with us any time soon (not as a companion, at any rate). Perhaps the greatest attribute of AI is in its innate artifice. We don’t want healthcare AI to replicate the brilliant palette of human emotional intelligence, however tempting that may be.
One of the mistakes we often make when imagining AI – especially through an artistic prism – is that it should somehow adopt the endearing or flawed personality traits of its creators. Look at the androids in the Star Wars films – they are loyal and sentimental. Meanwhile, KITT, the car-integrated AI companion in the television show Knight Rider is sarcastic, and even Metal Mickey derives joy from housework.
None of this would help AI in the real world. Mary Shelley’s 1818 novel Frankenstein provides the starkest warning that creating intelligence by holding up a mirror to ourselves is fraught with peril as its grotesque manifestation wreaks havoc. AI, at its most impactful, is uncluttered, decisive and unburdened by the complexities of a personal life!
Tom O’Leary, CIO at AI pioneers, ICON, explains how exploiting AI lies in its incredible non-human ability to detect subtle patterns within large quantities of data. “AI’s advantage over humans is speed and game-changing accuracy,” he enthuses. “It has the ability to grade cancers – or instances of cardiac disease – according to their risk, distinguishing cases that will kill quickly, and therefore need treatment, from those that can be left untreated.”
Tom believes that the key to maximising AI’s involvement in pharma development also emerges in its ability to draw distinctions and correlations that would otherwise elude human observations. “Critically AI is accelerating diagnosis and helping to ensure patients are treated more immediately, helping to save more lives than ever before,” he says.
For all its clinical and calculated ability to reveal what the human mind could not, it is, however, human beings that remain the conduit through which AI flows, metaphorically and literally. We still hold the sacred palette of instinct, courage, compassion and experience – and, thus, it is the combination of AI and actual people (whether in pharma, science and healthcare, including patients) that is such a compelling prospect.
Future is now
While to many AI is a concept, or something beyond the grasp of comprehension, it is rapidly forming solid foundations in science and healthcare. “Already machine learning is being used to identify patients in A&E who need urgent brain scans,” notes Tom.
Indeed, there have been significant AI advances in several arenas of health, such as ophthalmology, where comparing the performance of AI algorithms is revolutionising diagnosis. DeepMind – the British artificial intelligence group – has built a working prototype of a device that reveals complex eye diseases in real time. “The system can detect glaucoma, diabetic retinopathy and age-related macular degeneration with the same level of accuracy as the world’s leading specialists,” adds Tom.
Meanwhile, sophisticated AI software, provided by the likes of VIZ.AI, is able to detect historically devastating strokes. The technology has been introduced into hospitals in the specific south eastern region of the US known as ‘the stroke belt’ and is improving outcomes for patients by identifying urgent cases.
ICON’s own contribution to the germination of AI has witnessed a string of successful launches in the life sciences, with a particular emphasis on modernising clinical trials. Its ‘Automation Centre of Excellence’ uses AI and Robotic Process Automation (RPA) in order to accelerate clinical trial delivery, while its big data analysis and predictive analytics enable the efficient identification of sites and patients for inclusion in clinical trials.
Through its agreement with Intel, it is also deploying the Intel pharma analytics platform – an edge-to-cloud AI solution that enables remote monitoring and continuous capture of clinical data. In addition, ICON incorporates AI to deliver risk-based monitoring of trial performance as it occurs, while also introducing inpatient- focused technologies, such as mobile sensors, smartphone apps and telemedicine to gather richer data, develop new endpoints and help design novel approaches to drug development.
Direct NHS
While pharma is creating a new generation of clean data through cohorts of disease-specific patient groups and during sophisticated clinical trials, the potential to utilise the NHS’s current and retrospective trove of data (going back over 70 years and covering countless generations) is hindered by its hopelessly fragmented infrastructure, and a long-standing aversion to adopting technology. Remember pagers? The digital message-receiving devices that a few ‘cutting edge’ sixth formers used in the mid-nineties. Unbelievably, these virtually obsolete gizmos are still widely used across the NHS and are a clear indication that our beloved health service needs to arrive in the 21 century.
David Thorne, from Blue River Consulting, has held several roles throughout the NHS over the past four decades and, though optimistic about AI’s potential, he does not underestimate the challenges ahead. “The opportunities are obvious. I hear lots of the what, some of the why, but little that is empathetic on the how,” he says. “AI learns from data, but in the NHS it is poor. If a GP, ward sister or paramedic can’t share data now, how do we deliver a future world that correlates an entire footprint?”
David also believes that getting the entire NHS workforce to evolve is essential to embracing AI. He explains: “Don’t underestimate IT illiteracy or inability to interpret and respond to data in the NHS. The cost of change is key, especially relating to the practical training of a 1.3m workforce.
With David’s observations in mind, it is clear that AI’s wider adoption requires the kind of joint working between pharma and the NHS that has been so successful and mutually beneficial in the last decade or so. The NHS has the labyrinth of data and the pharma industry has the skills to unblock the atriums, creating a centralised database that would revolutionise diagnosis, treatment development, adherence, supply, technology and, ultimately, quality of life.
The data day
A day will arrive in the not too distant future when AI will take its influence to the next level, taking in data from implanted microchips, wearables, smartphones, medical records, live blood samples – even genealogy. It will draw conclusions, discard anomalies, issue warnings, deliver evidence-based predictions and provide indisputable truths about the entire human population or a single individual. The modern patient is central to the flow of this data and, if current trends are anything to go by, they are already in the driving seat.
Tom explains: “The advances in sensory technology can continuously monitor all vital signs – including blood pressure, heart rate, respiratory rate and temperature. With patients having the opportunity to take a more proactive role in the management of their health, there is the potential to prevent large numbers of patients being hospitalised in the future, reducing the costs of care and decreasing the overall burden on healthcare systems.”
Evidently, the possibilities of an AI-driven healthcare future are mesmerising – reach out and we can almost touch them. In the final analysis, however, if we are to fully to embrace AI’s irresistible force, we must also seek to align our existing healthcare planets – only then can we truly get back to the future…
