When we think of artificial intelligence, the images that come to mind for many of us are probably somewhere along the lines of Rosie from The Jetsons or Arnold in Terminator. While we’re still most likely several years from sentient household or murder robots, AI is playing an increasingly large role in our everyday lives.
One area that is beginning to see a significant increase in the applications of artificial intelligence systems is healthcare. Massive amounts of clinical health data are becoming increasingly available through clinical documentation, electronic health records, and online medical interactions, as well as new and evolving academic health data which is constantly being generated and updated. This increase in available data coupled with the emergence of new, sophisticated algorithms and software has begun a revolutionizing trend in the healthcare industry that is changing the way we’re diagnosed and treated, how we interact with physicians, and how physicians operate within a clinical setting.
Artificial intelligence is a computing system engineered to allow digital devices to perform tasks without being directly instructed by a human. By utilizing multiple algorithms to sort and analyze data, these systems are able to recognize patterns, make decisions, and even change the way they “learn” when presented with new information. Ultimately, AI is designed to emulate the human cognitive process at an exponentially accelerated rate. In healthcare, essentially this means teaching a computer system to learn and think like a doctor.
Diagnosis and Treatment
One of the most practical yet profound applications of AI in healthcare is in patient diagnosis. Using algorithms that are capable of scouring enormous databases of both structured data from clinical trials and unstructured data from medical journals, AI systems can search for a patient’s symptoms while also considering the patient’s own medical history to come up with the most likely diagnoses. It can then generate a list of probable diagnoses and assist in structuring highly personalized treatment plans based on the patient’s medical history, and considering the latest advancement in medical treatments and their success rates across broad population spectrums.
In a clinical trial of 1000 cancer patients, Watson, a technology being developed by IBM which uses an AI system to diagnose and treat cancer, concluded the same diagnoses and treatment recommendations as oncologists 99% of the time. The far reaching implications of this are particularly important for rural and remote communities where finding a doctor who specializes in a particular kind of disease may be impossible. For example, if there are only two doctors in the country who specialize in a rare, genetic kidney disorder and they are both located in Manhattan while the patient with the kidney disorder lives in Honolulu, using AI diagnoses and treatment plans, a local doctor who is not an expert in the kidney disorder can treat that patient with all of the expertise of the specialist available to them.
Personal Assistants and Remote Access
Personal health assistant apps operate by asking the patient a series of pointed questions about their symptoms and medical history to reach a diagnosis. The apps can then recommend treatments or suggest that the patient see a doctor in person. As this technology becomes more widely available, it could help to see a significant decrease in doctor’s office traffic since patients with minor illnesses or injuries who only require rest, over the counter medication, or home remedies won’t need to make the trip, freeing up doctor’s time to focus on more critical patients.
These “pocket doctors” can also essentially function as a live-in doctor and health coach for chronically sick patients who require round the clock care. By periodically entering their health data like blood pressure, heart rate, blood glucose, weight, activity, temperature etc, and answering questions regarding physical and mental functions like appetite, energy levels, and sleep patterns, medical personal assistants can track a patient’s health and treatment plan, recommend changes to things like diet and exercise to improve outcomes, and monitor for warning signs that something more serious that requires a visit to an IRL physician is occurring.
Clinical Documentation Improvement
Clinical documentation in and of itself is an important aspect in the development and accuracy of medical AI technology because it is the data from clinical documentation that is fed into EMR and EHR systems that produce the robust structured databases AI uses to “learn”. In our next post, we’ll be taking a closer look at the pragmatic applications of artificial intelligence in clinical documentation and how they are maximizing appropriate reimbursements and alleviating physician burnout.
Only for about the last two years has the technology and the data that are driving the AI revolution in healthcare been available, and already the results are awe inspiring and the potential seemingly limitless. While AI still has some obstacles to overcome in the healthcare field the artificial intelligence movement is likely to continue to grow in healthcare, and help improve upon the human endeavor of providing the most comprehensive and effective healthcare possible.