Taking a brave peak into the future of healthcare doesn’t mean looking that far into the future. Soon, we’ll be living in a world where healthcare services are delivered directly or indirectly by virtual doctors, accessible anytime and anywhere.
Looking back fifty years, when computers for personal use first started to come into public view, a Business Week article predicted that the office of the future will be paperless. Record-keeping, transactions, and managing documents would all go digital, a radical idea at that time. This is now our daily life, the wild conceptions of scientists drive the development of technology, and this trend is set to repeat going into the future.
Taking a brave peak into the future of healthcare doesn’t mean looking that far into the future. Soon, we’ll be living in a world where healthcare services are delivered directly or indirectly by virtual doctors, accessible anytime and anywhere. We no longer need to physically travel to a hospital, queueing for your turn, or worry about contracting infectious diseases during an epidemic.
Healthcare services become so pervasive that they can be carried out by the facilities you’re living in, the devices you’re wearing, a local station or even a healthcare booth delivered to you by self-driving cars or unmanned aerial vehicles when more specialised medical equipment is needed. But what about those working in the medical profession? Where are the clinicists, the doctors, the surgeons? They will form a highly-optimised pool of resources, they can be on duty anywhere, whether at home or on the move. They may use their specialised terminals to perform a routine health check and diagnosis, or, more radically, carry out surgical operations on the other side of the planet, as if chatting with a friend over the phone.
We don’t need to wait another fifty-years to see the above vision turning into reality. In fact, looking at the emerging technologies reported in the press and literature, the key enabling technologies needed to achieve this are all in our hands already. For example, 5G/6G communications, robotics, mixed reality and virtual reality, and artificial intelligence (AI) are at least in the reaching distance of this generation.
The Covid pandemic has profoundly changed how people live and work, with virtual meetings becoming the new normality. The world shut down physically during lockdown but a lot of jobs can continue virtually; one essential infrastructure to enable this is fast broadband infrastructure, such as 4G or enhanced home Asymmetric Digital Subscriber Line (ADSL) service. They provide reliable and good quality of service for real time video conferencing. If we were ten years back facing the same pandemic, these would all be impossible. 5G and beyond provides even faster, more reliable communications, with extremely low latency. The ultra low latency with end to end delay less than 10ms may not make any difference for our video calls as our eye takes at least 20 milliseconds to respond to changes, but such improvement is so critical for high precision tele-operated equipment, to allow large volume signals to be exchanged within milliseconds. Imagine a surgeon is in their office in London carrying out an operation on a patient in a ward in Hull, the surgeon must be able to see exactly what’s happening in the ward in real time without any delay to be able to operate the robotic arms and instruments with precise movement and haptic feedbacks. This demands ultra low latency and high throughput communications.
Robotics will play essential roles in the future of healthcare to realise the eventual delivery of service to assist people or treat patients. Not to mention the highly automated industrial robotics, we are seeing robotics in our life already: the robot vacuum cleaner at home, the robot chefs that automatically prepared and delivered food at the Beijing Winter Olympics, the reception robots in hotel and banks, the humanoid robot Pepper. Many minimally invasive operations have been carried out by surgical robots, such as the Da Vinci robots. New research has also allowed robots to express facial emotions, bring the human-robot distance closer. AI has empowered robots to perform many challenging task autonomously, however for complicated or unplanned critical scenarios, we still rely on the model of human in the control loop. Robotics will play assistive roles, rather than rolling out to all services completely.
Mixed reality is a combination of (completely) virtual reality with interaction with the physical world. With mixed reality technologies, highly immersive virtual user experience can be created.
According to a poll with over 600 GP responses, over 70% of general practice consultations are conducted remotely and over 80% of GPs felt they can deliver appointments without unnecessary follow-up as an in-person appointment.
Many GPs prefer consulting in person, particularly when delivering care to patients with complex health needs. Remote consultation can make it harder to pick up on ‘softer’ cues, which can be helpful for making diagnoses. Now, not only visual contents can be virtualised, but also touch, temperature, texture, and even smell can be virtually presented and experienced remotely to users thanks to the development of a variety of sensors. Indeed, the Internet-of-Medical-Things (IoMT) forms a new promising healthcare paradigm through connected sensors, wearables and clinical systems that can improve the quality of medical care. Integrated with mixed reality, a digital twin of a patient can be created in a virtual world, and interact with a digital twin of a doctor, hence even the ‘softer’ cues can be picked up. Furthermore, not only the clues which are observable by doctor, but also the ones that can only be measured by medical devices will be presented lively to doctors. The doctors see readings, graphs along with the patient in the virtual world, and even the suggested treatment and highlighted abnormalities worthy attention if any.
“Last but not the least, AI, especially deep learning, will play a role in the entire system.”
Broadly speaking, AI covers machine learning and deep learning. Machine learning is a branch of AI and deep learning is a subset of machine learning. Some AIs are built by sets of manually devised rules, reacting upon inputs mechanically to generate outputs, such as expert systems. Scientist know the outputs produced by a given input in advance, with the more popular ones being put through a “training” process for the machine to learn the rules. The learning process is a kind of algorithm to tweak the values of parameters and the weights to link them within given ranges by try and fail searching.
One simplified example, to find the right size of lid for a pot, the machine learning process would be to try on a lid, and if it is too small, then increase the size slightly, try again, until the lid can fit on the pot. As one key element of deep learning, neural networks try to mimic how we learn by using millions or even billions of artificial neural nodes organised in layers, with interlinks to connect the nodes across layers or even within layers. Although it’s mysterious how exactly humans learn, it’s generally believed there are billions of neurons in our brain, and those neurons store information and are connected with other neurons via Axon. Then those connectivity and neurons react to their environment.
So for machines to “learn”, they need to tune the values of millions of parameters and weights to fit the data given to them to “train” the algorithms. The rise of deep learning in recent decades is mainly attributed to the improved computing hardware such as CPUs, GPU and TPUs, some smart approaches to tune the parameters automatically fuelled by tons of data that can be used to train the AIs. For example the GPT-3 model has a capacity of 175 billion parameters, the order of magnitude of neurons a human brain has.
Dr. Cheng currently forms a vital part of the Latus Health advisory board, and is playing a key role in our ambitious plans to revolutionise the industry through ai technology. Latus Health’s development of a fully remote occupational health service has additional benefits including a dramatic reduction in CO2e emissions and an increase in occupational capacity.