Artificial intelligence is the technology that enables a digital device to recognize, understand, and respond. In the case of machine learning, you can add to that the ability to make decisions and learn to change its “thinking” and behavior as it analyzes petabytes of data received from the cloud. Here are seven ways AI and machine learning are transforming healthcare.1. Killing Occam’s Razor
Occam’s Razor or as it is sometimes referred, the Law of Parsimony, is the idea that given multiple solutions to the same problem, the most simple solution is correct. In medicine, the term diagnostic parsimony is often used. For example, when a patient presents with several symptoms, according to parsimony, the proper diagnosis would consist of the fewest possible causes that could explain every symptom.
As early as the late 80s, scientists were beginning to apply computer algorithms to test this diagnostic theory. Most recently , AI has proven effective in not only proving the fallacy of diagnostic parsimony but also proving the counterargument known as Hickam's Dictum, which states:
"Patients can have as many diseases as they darn well please,” John Hickam, MD
Because AI and machine learning have no element of human assumption and the capacity to interpret inhuman amounts of data, more accurate diagnosis of multiple and unrelated diseases can be reached. This is especially helpful in diagnosing disease in the elderly and immunocompromised who are more likely to experience multiple unrelated diseases at the same time.
2. Precision Medicine
There is a common saying in medicine, “treat the patient, not just the disease.” It referred to the tendency for practitioners to become so focused on treating the disease process that they failed to see the nuances of the disease that were specific to the individual patient. Now, with personalized diagnosis and precision medicine, it is possible to actually treat the patient’s disease, not just the disease? How are they doing this? By applying deep learning algorithms to the vast stores of data collected by medical facilities over the span of decades, analyzing them, then creating meaningful reports based on real-life patterns. Now, using proteomics a more specific type of DNA mapping along with AI and machine learning, advances are being made in the treatment and prediction of diseases at the most intricate level imaginable.
3. Virtual Health Assistant
Currently in beta testing is the AI technology that acts as a personal health coach. Developed by NextIT, this virtual clinician functions similarly to a customer service rep at a call center. Think —Suri meets Dial-a-Nurse. The digital assistant can ask questions, take notes, and provide some advice all the while streaming the data to the healthcare provider for quick follow-up.
4. Virtual Fitness Coaches
The Cafewell app uses IBM technology to understand its users’ health and fitness goals. Then it compares the patient’s data to previously analyzed data to find and recommend a complete plan of diet, exercise, encouragement that has the best statistical likelihood of success.
5. Automated Cancer Treatment
Everyone's favorite AI Jeopardy Contestant —Watson, is at it again. Scientists at IBM now have the CareEdit tool which is being used by oncologists to create best practice guidelines. It works by analyzing massive amounts of data including past treatment regimens and their outcomes to formulate a “clinical decision support system” which helps physicians prescribe the best plan of care for the individual patient. The benefits are twofold. Not only does it improve patient outcome and survivability but also reduces overall costs.
6. Use of Selfies to Diagnose Depression
The science of facial recognition, analysis, and recognition through the use of AI is fascinating stuff. Last year, a team of researchers from Harvard and the University of Vermont took the technology to the next innovative step by applying the technology to pictures found on the social media platform, Instagram. They developed a machine learning program that with 70% accuracy, did identify users suffering from clinical depression. To put this in perspective, previous studies have found that physicians, using traditional in-person assessments, were able to correctly diagnose depression 42% of the time.
7. Google Spots Eye Disease
The Google-owned AI company, DeepMind, is working on technology set to help doctors diagnose certain eye diseases by analyzing eye scans and medical records of more than 1.6 million patients, then creating algorithms to create a screening tool that will help doctors with early diagnosis. The two diseases the team is working on are diabetic retinopathy and macular degeneration, which are both leading causes of blindness worldwide.
“80% of health data is invisible to current systems because it’s unstructured.” IBM Watson Health
All of these innovations in healthcare tech are due to the ability of AI and machine learning technologies to collect massive amounts of data, analyze that data and then structure it in a meaningful way. While AI and machine learning will never completely replace healthcare providers, it is already disrupting the industry; a process that always leads to more innovation.