The Healthcare sector all over the world is now bracing hard for the revolution that is nearly here. The uncharted waters of Artificial Intelligent Control Systems are being explored by MNCs and tech-geeks all over the world. It might be prudent of me to say this, but it will be the healthcare sector that will be benefitting most from this plague. But I have my own reasons. As an undergrad Biomedical Engineering Student I’m slowly realizing the hurdles in processing biosignals. For starters, they are so small, they’re virtually non-existent as we plunge deep into our tissues. Second, the little signal that we manage to record will be filled with noise which would take very expensive systems to clean. and third, all the difficulties that we’ve yet faced in this industry will be a shrug for a truly intelligent system.
You would have been living under a rock if you haven’t heard the story of an American teen who wrote out an algorithm that can detect subtle cues of deadly diseases. No doubt an impressive feat, this is in fact made possible due to the smart crafting of Machine Learning and such Artificial Intelligence algorithms. They are so powerful that a Whiz Kid from New York can work magic with her hands. Such systems that make use of Big Data in healthcare, the mass data that are available in public domain – ECG, EEG, EMG, etc – are becoming increasingly common, particularly among startups and college projects. The feature extraction from a large pool of raw data, nearly – well, definitely impossible for an ordinary human cognition, is what AI is taught to do. And they do that very very well.
But using such exotic algorithms are not a new thing in the healthcare industry. The giants of Biomedical Engineering have been employing Artificial Neural Networks, Fuzzy Logic, and Genetic Algorithms in the instruments, especially in Image Processing. The CT (Computed Tomography), MRI (Magnetic Resonance Imaging) or PET-CT (Positron Emission Tomography, boy, do I love saying that) are basically supercomputers that house the most powerful Image Processing Artificial Intelligent algorithms. The CT, MRI or PET-CT images obtained from those exotic and expensive scans will be bathed in noise when it reaches the processor component attached. These computers that employ perhaps the worlds most powerful Machine Learning and Pattern Recognition algorithms help in filtering out the noise helping millions of people from the early diagnosis of terminal diseases.
Making these machines smart would also mean bringing the cost down of diagnosis and treatment. The GE Healthcare, who is a pioneer in using such technologies, employ such tools even in low-cost instruments like ultrasound scanner, fine-tuning them to detect signal artifacts that are generally thought far too sophisticated for these low-end pieces of equipment. Thus by bringing the cost down, we are able to make instruments affordable to India and its rural market. This also helps the hospitals to turn away from refurbished goods, which will be nearly banned in India.
The development of Pattern Recognition algorithms has contributed greatly to the growth of such feature extraction from signals. The same program that powers facial recognition on your Facebook Wall, helps people diagnose and prevent the onset of terminal diseases. These PR algorithms are also employed in ECG monitors, Holter Monitors by various manufacturers to identify arrhythmias, irregular rhythm of heartbeats, peculiar to several heart diseases. The studies aiming at an earlier detection of diseases would mean, effectively bringing the headcount of 1.7mn deaths per year in India alone.
Apart from analytics and diagnostics, these tools would also mean the end of conventional therapies. The population of the world still thinks drugs are the ultimate cure to any problem. I am not rooting for hippy eastern stuff, but I’m speaking of the therapies that are evolving around the world to treat psychosomatic disorders. Although drugs are undoubtedly effective in most cases, we always ignore the long-lasting effects of injecting these foreign chemicals into our body. Alternate therapies like physical therapy, music therapy, gene therapy, etc would be more effective for chronic diseases like spondylitis or a migraine, where the disease blur the line between psychosomatic and real. In fact, while comparing the MRI artifacts of a portion of patients suffering from back pain to those that are neutral, its found the artifacts doesn’t show a significant deviation. But consumption of drugs for curing such a condition would mean disrupting the already diluted rhythm of the body. But by comparing the physiological characteristics and therapy description of patients from a large database, an intelligent algorithm can actually recommend the most effective therapy useful for the patient. Using such systems would mean, the end of an era of misdiagnosis, mistreatment, and malpractice in the medical field.
So far I have discussed how the Artificial Intelligence and tools like Machine Learning or Neural Networks have boosted the healthcare industry especially in Imaging and Diagnostic arena. But let’s look what the true capability of Artificial Intelligence in Healthcare Industry would be. For starters, the same old losing jobs, but here it would be physicians or paramedics, one of the top paid jobs and highly skilled in nature. It would mean medical degrees worth a lot of money going worthless as the paper its printed in. It might seem scary but every day 1 in 200 patients who spend their nights in a typical US hospital will be dead due to medical error. This statistics is terrifying, but it also ushers the need to implement smart solutions in this field.
With the advancement of Nanotechnology, technology can create nanobots that can be injected into our body, continuously monitoring our health and updating our logs in our physicians hard drive. Such a vision of future is far away, but it is coming to us nevertheless. These bots can actually find out the early onset of diseases, fight them before they reveal themselves and even the capacity to edit our genes to get rid of hereditary disorders. With most of the surgeries and therapies performed within the body, the hospitals will rethink about their role in healthcare.
Employing Big Data in healthcare also mean predicting the onset and extent of a pandemic before it even begins. It would also mean the optimization of sales of drugs where the illness would hit next.
Machine Learning and other tools of this age are proving to have massive application in the healthcare industry. In a few decades, the conventional way of diagnosis and treatment would have changed for sure. Since we have found our brains too humbling before the immensity of data of biological world, we literally invented and intelligence that can rival the challenges of evolution.