Don’t be surprised if your next doctor is a bot
ARTIFICIAL intelligence (AI) has become a transformational force in healthcare by providing countless opportunities to use technology to implement more exact, efficient, and impactful interventions at the optimal moment in a patient’s care. Some of the biggest technology companies (including Google, Microsoft, and IBM), as well as several AI start-ups, are currently driving the proliferation of AI in healthcare.
Artificial intelligence is increasingly taking on the diagnosing and decision-making role of humans in healthcare. Bots, or automated programmes, are likely to play a crucial role in the future in discovering cures for some of the most difficult-to-treat diseases and disorders. Thus, do not be surprised if your next doctor is a bot.
And if you think that a bot as doctor is a bit far-fetched, consider that American researchers detected cardiac arrhythmia with 97 percent accuracy on wearers of a smart watch with AI, allowing early treatment to avert strokes.
British scientists created AI software that predicted heart attacks better than doctors do, while researchers from Harvard and Vermont (in the US) developed a machine-learning tool (a type of AI that allows computers to learn without human programming) to identify depression by studying Instagram posts, thus making early detection of mental illness possible.
Due to the magnificent capability of AI to analyse large sets of clinical data, research publications, and professional guidelines, it has the potential to aid the diagnosis of disease and to assist in decisions regarding treatment.
By using big data on cancer patients and the treatments used in each case over many years, AI is able to suggest treatment options for distinctive cancer cases based on the most successful treatments of the past.
The advantage of AI in diagnostics is early detection and increased accuracy, thus assisting healthcare to move away from reactive care and getting ahead of chronic diseases, costly acute incidents, and the sudden worsening of the condition of patients.
AI, and in particular machine learning and deep learning, is used in medical imaging and has shown promising results in detecting conditions such as pneumonia, tuberculosis, breast and skin cancers, as well as eye diseases.
AI proved to be better than humans when analysing MRI and CT scans or X-rays by comparing medical images with those of millions of other patients, noticing details that a human may miss. In fact, AI made the next generation of radiology tools possible, that are accurate and detailed enough to substitute the need for tissue samples in certain cases, thus enabling “virtual biopsies”. In cardiography, AI is used to analyse echocardiography scans to detect heartbeat patterns and diagnose coronary heart disease.
Even more astonishing is that robotic tools controlled by AI have been used to carry out specific tasks in keyhole surgery, such as tying knots to close wounds, identifying distances or specific body parts or to steady the motion of robotic limbs when taking directions from human controllers. In orthopaedic surgery, AI-assisted robotics can analyse pre-op medical Bloomberg records to actually guide the surgeon’s instrument in real-time. It can also use data from actual surgical experiences to suggest new surgical methods.
Research involving 379 orthopaedic patients found that an AI-assisted robotic technique resulted in a fivefold reduction in surgical complications compared to when surgeons operated alone.
AI-assisted robotic surgery could also generate a 21 percent reduction in orthopaedic patients’ length of stay in the hospital, following surgery, due to less complications and errors. In the future, machine and deep learning could be used to allow robots to autonomously master and execute surgeries.
Another major development in healthcare is predictive analytics and clinical decision support tools that are powered by AI. New York University researchers used AI software to analyse medical and lab records to accurately forecast the commencement of numerous diseases and conditions including stroke, type 2 diabetes, and heart or kidney failure. Researchers in the UK have even developed an AI tool that identifies developmental diseases by analysing images of a child’s face.
The AI algorithm can discover distinct features, such as a child’s jaw line, eye and nose positioning, and other traits that might denote a craniofacial abnormality. Presently, the AI tool can match the facial images to more than 90 disorders. AI is also increasingly seen as a means for identifying depression and other psychological illnesses, by identifying unapparent patterns. An American researcher, Jessica Ribeiro, found that by scanning medical records AI could predict with 80 to 90 percent accuracy whether someone will attempt suicide as far as two years in advance.
Many neurological AI tools have been developed that examine speech patterns to predict psychotic incidents, as well as discover and monitor symptoms of neurological disorders such as Parkinson’s disease.
Neurological diseases and trauma to the nervous system can cause some patients to loose their ability to speak, move, and interact meaningfully. Brain-computer interfaces, assisted by AI, could reinstate those essential experiences thus significantly increasing quality of life for patients with sclerosis, strokes, locked-in syndrome, or the half million people worldwide who experience spinal cord injuries every year.
There is no doubt that the possibilities of AI in healthcare are vast and that they would in future easily surpass our most creative imagination.