Paging Doctor R2

Co-Authored by Paul Hartley
It is undeniable that technology-enabled devices have had a great impact upon our ability to diagnose, treat and care for the sick and infirm. However, we must be cautious to not get caught up in the flashing lights, bells, whistles and web-integrated gadgetry of it all. Many people have a dangerous penchant for placing unwavering faith in technology and, in the case of medical devices, this kind of blind trust can lead to great harm. At times, it is crucial to ground our confidence in technology to remind us – particularly in a field so vital to us as health – what these devices are and are not capable of.
What can technology do?
Fundamentally, any piece of technology is created to accomplish a particular task that, given unlimited time and resources, a human could accomplish. The most important thing we gain from assigning work to a device is speed—a machine can simply do many things faster that we can. However, beyond speed, we also gain a way to reliably do a task over and over again. Repeatability, or the successful execution of the same task, is the other major benefit of technology integration.
However, there is an important differentiation between what technology can do and how we use it. What something does is defined by the target goal inherent to the task it has been created to accomplish – in the case of most health technologies: increase speed and repeatability. What we use the technology for is a broader matter pertaining to how the technology integrates and plays a role in a much larger diagnostic process. In the case of health and medicine, technology use can be seen as falling into a number of high-level categories:
  • Expand The Senses: An augmentation or enhancement of basic human senses to allow for more detailed, thorough or less harmful inspection of the human body.
  • Communicate Information: Methods for transferring information quickly and seamlessly between points to allow for remote patient diagnosis, collaborative analysis and multi-device clustering.
  • Redistribute Responsibility: Shifting medical actions between doctor, patient and device to increase control of the process by the patient and free up valuable time of medical professionals.
  • Accurate, Consistent Data Collection: The tracking of reliable data at regular intervals to improve the quantity and quality of information used in the input of diagnosis, ensuring a better picture of an individual’s health.
  • Data Analysis & Prediction: Processing large amounts of aggregate health data across multiple users to spot trends and patterns within large populations and improve the accuracy of personalized diagnoses.
  • Active Intervention & Treatment: Integrated with data collection and analysis, these technologies play an active hand in treating and influencing the human body through continuous feedback loops.
What can’t technology do…yet?
Yet as powerful as the technologies that live in the above list are, it is essential to understand the limitations of our current technology for two key reasons.
First, we must understand the limits of the trust we should currently place in medical devices. Just because we are capable of creating technology that can perform a task does not always mean that we should. Creating technology for technology’s sake may be well and good in the realm of gadgets and games, however, with a subject as critical as medicine, we must ensure that our technology solution is truly superior to old ways and does not create more problems than it solves.
Second, it is essential to understand the limitations of our technology so that we know where to go next. Without knowing what stands in our way, we will never know how to get around it. As it stands, here are what we see as three of the greatest challenges we must overcome to push the envelope of medical technology:
Adaptation to Unique Scenarios: Current technology is effective because it is usually designed to do one or two things: search for patterns and give basic responses. This is a very binary approach to technology that works for our current state of technological infancy. To progress to a state where our devices can adapt to unique scenarios and deal with outliers in addition to the mainstream, we need two key things: smarter AI and more data. As software intelligence grows in complexity, it will learn to decipher more difficult problems, but in order to learn, this AI will need vast amounts of data to build an understanding of the different scenarios it could encounter in order to interpolate and extrapolate upon existing knowledge and deal with each unique case.
Holistic, Integrated Bodily Monitoring: Part of treating each unique case will come down to recognizing that the human body can not be treated in isolation. It is a vast, interconnected system and for devices to improve diagnosis and treatment, they must understand a holistic view of who we are. In addition to preventing catastrophes due to unwanted side-effects of certain treatments, collecting more data points on our bodies on a consistent basis will provide insights into the body that have never been seen in the medical community. To reach this state, we need a simultaneous improvement in our sensing, communication and multi-variable analysis capabilities.
Emotions and the Mind: Perhaps the holy grail of holistic monitoring would be to access the knowledge and mysteries stored in the human mind. Generations of medical professionals, philosophers and leaders have alluded to the interconnectedness of body and mind. As such, if we aspire to create a robotic doctor, we first need to have the technological capability to analyze and understand the human body as well as the mind. Interestingly, this is less a technology problem and more concern of psychology and biology. Our current understanding of the human brain is juvenile in the grand scheme of its complexity and many of our current remedies for ailments of the mind are based theories, intuitions and an intangible human connection. Before we can hand such great responsibility to a machine, we must be certain in our comprehension of the mind and leave nothing to guesswork. While we have no doubt that artificial intelligence will one day achieve sentience, we question whether a robot will ever be able to develop a gut; hunches and intuitions simply don’t mesh well with circuits.