Mankind is within reach of collecting unprecedented amounts of data about patient behaviour and physiology. The question is do we know enough about human biology to use this data to help patients.
Last week, Pfizer and IBM announced that they will team up to collect data from wearables and improve diagnosis and treatment of neurological diseases such as Alzheimer’s. In the same week, there were two other announcements: One from a data science competition hosted by Kaggle, and one from Brown University, of machine-learning algorithms performing at par with human beings in detecting disease based on MRI and ultrasound scans. This is especially significant since there is an acute shortage of high-quality radiologists in most countries, and the analysis of medical scans (any medical data, for that matter), is an extremely laborious and time-consuming job.
The question that these announcements pose, for the world in general and the medical profession in particular, is whether we are at the brink of a technological revolution in the field of medical research and practice, which might literally sweep doctors and other medical professionals off their feet.
The clairvoyant writings of Lewis Thomas, a renowned biologist, linguist and polymath, in the 1960s and ’70s, offer a rich history of the medical profession and some reference points to judge where we are today. Thomas described how the medical profession stepped out of its infancy only in the 1930s, before which medical interventions were largely palliative, since there were few proven ways of curing disease. It was only with clinical practices introduced by the steely Dr William Osler in 1937, of simply observing disease progression instead of intervening, that vast quantities of invaluable data started accumulating, and biological research on causes and, consequently, cures of diseases began, which then led to the discovery of vaccines, antibiotics and surgical interventions.
Today, we stand at the brink of another watershed moment. The advent of smartphones — the most rapidly adopted technological innovation in the history of mankind — has ushered in a new era of data collection and human monitoring, which makes Osler’s methods look medieval. With novel wearable or implanted sensors that connect to a smartphone or other gateway device, researchers and healthcare companies have access to vast amounts of data related to the activities and physiological parameters of billions of healthy and sick subjects. Some biosensors and accompanying display technologies make it possible for users not only to monitor their physiological parameters such as ECG and EEG in real-time, but also to alter their biological signals or rhythms by way of biofeedback and neurofeedback. In fact, Thomas happened to witness the birth of these revolutionary technologies towards the end of his life and wrote in bewilderment about the implications of these technologies. He, however, added this cautionary note: “One ought to feel, I know, elated with the prospect of taking personal charge… My trouble, to be quite candid, is a lack of confidence in myself. If I were informed tomorrow that I was in direct communication with my liver, and could now take over, I would become deeply depressed. I’d sooner be told, 40,000 feet above Denver, that the 747 jet in which I had a coach seat was now mine to operate as I pleased.”
Thomas’s question is almost as relevant today, as it was in the 1980s. Yes, we are now within technological reach of collecting unprecedented amounts of data about patient behaviour and physiology. But do we really know what to do with it? Do we understand enough about human biology to use this data to provide real value to patients who most desperately need it? Or will this path be lit by the patterns that emerge from the data itself?
Thomas’s fears are shared by Siddhartha Mukherjee, a man he inspired four decades later to study and write about the mysteries of medicine. Mukherjee’s insightful book, The Laws of Medicine, published in 2015, ends with the ominous line, “Malcolm Gladwell wrote that the [political] revolution will not be tweeted. Well, the medical revolution will not be algorithmised.”However, the first of the three laws described by Mukherjee highlights the value of prior information about a patient’s habits and history and the importance of Bayesian analysis of data in arriving at the correct diagnosis. The second law discusses the importance of detailed studies of the data corresponding to clinical outliers in discovering new facts about the human body. Finally, the third law talks about the need for eliminating human biases in scientific studies. In many ways, these laws do not negate the value of algorithms at all. Instead, they underscore the importance of designing smart algorithms (as opposed to dumb aggregational ones) that ask the right questions, and take prior information and uncertainty into account.
A 2013 study at Oxford University calculated the risk of automation of different professions within the next 20 years. It predicts that there is only one to two per cent chance of a “health professional” or a “medical practitioner” being replaced by machines (as opposed to a 99 per cent risk of automation for a “telephone salesperson” or a 97.6 per cent risk for a “financial accounts manager”), putting the two in the top 20 safest professions among a total of 366, which is in line with the thinking of Thomas and Mukherjee. Yet, it also predicts that there is an 85 per cent chance of a “healthcare practice manager” and a “medical secretary” being automated in the same time period, which indicates the first step in what Eric Topol describes as the radical transformation of medicine in his recent book, The Patient Will See You Now. Topol, a renowned cardiologist and the current director of the Scripps Translational Science Institute, writes about the ongoing transfer of power from doctors to patients and says he knew the world had changed when an ECG was emailed to him by a patient with the subject line, “Iam in atrial fib, now what do I do?” What remains to be seen is how profound the impact will be on the determination of causes and cures for diseases, and the delivery of treatment.
The likes of Pfizer and IBM appear to have their money on Topol, while the medical community remains largely as sceptical as Thomas and Mukherjee. Ultimately, it might be innovative, hybrid combinations of technology and human expertise, sweetened by empathetic and personalised care, that provide maximum value to the patient.
Written By: Manav Bhushan