Close this search box.

AI and Machine Learning: how tech driven healthcare is driving the industry

Richard Low | AI and Machine Learning
Richard Low

In 1989 Praxis embarked upon the development of a unique type of AI-based Electronic Medical Record technology that has been progressively fine-tuned over the over all these years. Today, AI-based Praxis is rated number one both in usability and physician user satisfaction. This is not because Praxis is a better built software, but rather because the Praxis “concept processor” resolves time consuming process of medical charting using artificial intelligence that learns from its user.

To most people, AI in medicine envisions an expert-type program knowledge that arrives from outside of the provider delivering information at the point of care.  Technologies such as IBM’s Watson come to mind. The Praxis AI technology, however, is internal to the mind of the provider; it works with the mind of the provider and empowering the provider to practice efficiently. Praxis is currently the fastest and highest quality way to chart medicine today, hands down.

The concept behind the AI in Praxis EMR is that medicine is an art and not just a science.  The idea that medicine is an art-form rather than an exact science may feel initially strange to non-physicians. People may assume that medicine is a science that makes use of precise logical rules similar to those found in mathematics, engineering and accounting.  The practice of medicine is not like that at all.

Our medical art deals with uncertainty, probabilities and unknowns, it works with with gray areas of knowledge. Even after two hundred years of remarkable technologic and scientific achievements, medicine today continues to be based primarily on personal experience and human skill. No two-doctors practice the same way; no doctors think the same way.  Indeed, with the advent of computers, if medicine were not an art form, then doctors would probably not be needed.

Algorithms could be set up so computers could practice all by themselves. A patient would access a program. explain his or her clinical issues, answer a few questions, be subject to a physical exam performed by expert robots which exist today in other areas, input ordered laboratory and other studies into the system, and…presto…the computer would generate a perfect diagnosis and advise on appropriate treatment every time. In fact, a computer could handle surgical procedures with perfection and speed if programed correctly if medicine were an exact science. It is not.

This vision seems obvious, but it is not likely to happen in the foreseeable future. The human creative perspective, which is impossible for a computer to mimic successfully, is very much a part of the medical practice every day.  Yes, a computer can follow a logical process perfectly, but it cannot follow reason, to say nothing of intuition, and therefore it fails where logic defies the clinical condition.  It is in this circumstance where humans prevail over the computer.

Medicine deals with a different type of knowledge. Often, we are called upon to make clinical decisions based on inadequate information, sometimes without even knowing the cause and mechanism of the disease process involved. Thus, personal experience makes a huge difference to the practice of this unique art. No matter how many books one reads, without having been exposed to patients for a significant time period—years perhaps—it is not possible to practice based on book learning or intelligence alone.

Yesterday, this personal training was performed by a mentor, an older physician who would share experience and wisdom with his young protegee.  More recently, formal medical residency programs have been set up to perform this training over years of training. Developing intuition and experience takes much time and effort.

So even the simple migration from the paper record to its electronic equivalent by using macros and picklists to speed things up actually bogged down physicians even further. Today, the average provider in primary care spends twice as much time with their computer, and they do with their patients.

On average, they waste about two hours a day playing data-entry clerk into a computer. This has become a major source of clinical error, physician stress and burn-out. Ironically, the computer is turning our physicians into robots.

Understanding this subtle difference in how medicine works in real life allowed us to create a very different solution.  If medicine is learned from personal experience, why not set up the charting in the same manner, by learning from the provider’s own previous writings to help him/her not only chart dramatically faster but also think clearer? This is the basis of the Praxis Concept Processor, a neural network type of software that instantly finds and retrieves the text of the closest possible encounter that the individual provider has handled in the past in relation to the case being encountered.

This approach makes use of the fact that that some clinical cases are very rare indeed—they may have never been encountered by this provider—whereas others present daily and even hourly, which does not mean they are not complex and do not demand precision and care. Here is exactly where the computer becomes extremely helpful, as it never forgets anything it is taught. In essence, the doctor is charting backwards. Instead of writing what was done, one does what is written. The use of previous personal writing as a checklist to assist in the approach to a clinical event flies against what the hard sciences teach us. However, this approach is highly effective. Let’s face it, one may disagree with someone else, but it is very hard to argue with oneself.

Therefore, if a new case presents that is identical to one seen in past, the use of the previous chart on an identical case not only speeds up data entry dramatically, but more importantly, ensures that nothing is forgotten, that all the t’s are crossed, and the i’s are dotted. Not only is the charting done instantly, but all the orders are carried out at once, the prescriptions written precisely and electronically sent to the pharmacy, the laboratory orders are generated, the instructions and handouts to the patient are handled, the referrals, the reschedule and the bill are all generated at the same time as the note, all done in a single instant step. And if the provider disagrees with anything in the past for a similar patient, the reason can only be that either the doctor made a mistake handling the previous case—medical knowledge may have improved in between, an actual error from before may be found— the current case should be handled slightly differently. In the first case, the correction ensures that random errors will never recure, whereas in the second case, the system will learn for the future.  In other words, the more Praxis is used, the faster and smarter it becomes for its user for all cases, and the less time is wasted charting, and the less clinical errors are made.  Here the basis of the AI is the doctor’s own knowledge and approach. Based on this unique approach, a huge set of clinical possibilities arises.  Clearly, this is a different approach than the old “dead-paper-record paradigm” and one that progressively improves the quality of medicine and decreases clinical errors, physician stress and burn-out, saving thousands of hours of wasted paperwork charting a year.

Indeed, AI in medicine can be thought of in ways different ways: Empowering an art-form is certainly one of them, and that is what Praxis EMR is all about.



Related Posts

Copyright 2023 © Insightscare Magazine ( a Digital Ink brand ) All rights reserved.