You may recall the IBM supercomputer Watson obliterating Ken Jennings in Jeopardy! some years ago and thinking we’ve reached the pinnacle of artificial intelligence. But new research on Watson’s inability to independently learn new information reveals its limitations in the medical field as decision support software. Medical technology publication Healthcare Informatics recently published a piece describing how engineers must manually upload information to the computer’s database. With the rise of machine learning, Watson’s dependence on human overseers leaves it now comparably among the most basic forms of artificial intelligence in the field.
The Root Of The Problem
The problem lies in the sheer enormity of information in the medical field. Not to mention that terminology and clinical documentation protocols continue to evolve without standard instructions for the protocols. Now more than ever, with the implementation of ICD-10 compliance and other clinical terminologies important for Quality like SNOMED, coding systems that contains tens of thousands of codes overall, the prospect of loading and organizing all relevant logic that incorporates the knowledge bases in the medical field individually is time consuming and thus unsustainably expensive due to labor costs.
For these reasons, using Watson as a resource for clinical decision support and knowledge discovery ultimately runs contrary to the fundamental advantages of artificial intelligence. Because Watson must depend on humans to adapt to this vastly expansive assortment of information that is often in flux, it is ultimately unfit for the job when compared to newer approaches which do not rely on domain experts but rather advanced knowledge engineering.
Why Hiteks’ Machine Learning Capable Solutions Suite By Insight Is Distinguishable From Watson
Meanwhile, medical informatics company Hiteks Solutions has picked up where IBM left off. Hiteks’ Insight: Revenue Rescue Suite approaches knowledge base development in a practical way that initially does not try to eliminate the need for a physician, nurse or CDI Specialist. Hiteks’ CDI software utilizes the latest in machine learning artificial intelligence to offer gentle suggestions at the moment the documentation is created in the EHR to improve CDI and quality which improves billing and provides a strong ROI for introducing the Hiteks software. Our cutting edge medical software evolves automatically without the need for human intervention. As a result, we have managed to enhance virtually all aspects of health care that once caused relentless tension and frustration for patients, providers, hospital billing and insurers for years.
Enhanced Patient Care
With Hiteks’ AccessRX, providers now have the resources to anticipate how contraindications for medications may affect each patient before and during the point of care. This leaves providers better equipped to prescribe the correct medication and dosage the first time around. This would not be possible with Watson, as it lacks the capacity to learn independently a patient’s chart and anticipate its effects based on the patient’s medical history.
Enhanced Clinical Research
Imagine if a physician diagnoses a patient with a rare disease for which important and useful research is still ongoing. It would be in the patient’s best interest to have access to those who are interested in developing a cure and would presumably be willing to consent to participating in a study in pursuit of the cure. Until now, researchers were unable to reach most patients who were eligible for their clinical research studies because they simply did not know where they were. With Hiteks’ AccessCT, which learns automatically who is eligible by drawing from a central EHR database, suggests participants for studies on a national scale. Again, this would not have been possible with the capacity for machine learning.
Streamlined Health Administration
ICD-10 was implemented as a means of improving the available financial claims data from a provider, which requires clinical documentation improvement (CDI) by providers to ensure their documentation is compliant. But due to its expansiveness, such improvement can only take place with the assistance of a software that can store and analyze it in real-time. Hiteks’ CDI software, ConcurDI does just that, and it does so within less than a second. Without it, abiding by ICD-10 protocol manually would take months, leading to complications in provider compensation and patient reimbursement by insurers. What’s more, ICD-10 is subject to change as new research comes out. To depend on humans to adapt with each evolution in the coding system would once again be overly time consuming and ultimately defeat its original purpose.
Learn more about Hiteks’ Insight: Real-Time Revenue Rescue suite, powering hospitals across the nation with lightning speed machine learning, by dialing 212-920-0929.