We can’t help but recall humanoids of Spielberg-movie lore or the post-apocalyptic imaginings of The Matrix Trilogy when we hear terms like artificial intelligence and machine learning.
What many of us fail to realize is that like human intelligence, there exist a consistently expanding set of variables and applications attributed to AI, applied in industries like healthcare without worrying about the deep ethical questions of creating artificial beings.
It’s not uncommon for healthcare IT applications, like revenue cycle management systems, to feature rule-based intelligence nowadays. After all, rule-based systems are finding practical applications outside of their traditional finance-based bounds, such as domotics and in some military capacities.
Dynamic rule-based systems capture the user’s knowledge within a set of rules, each of which encoding bits of that knowledge. So, the system contains information about certain facts and objects that must be true in order for the rule to potentially fire. This signifies an unprecedented level of sophistication in medical billing automation.
Still, rule-based intelligence is only a checkpoint in the history of health IT. For instance, although improvements have been made to make rule-engines easier to interface – as evidenced by the Blaze Advisor Rules Engine and the Java-based ILOG JRules – these systems still represent one of the simplest forms of artificial intelligence used in any industry.
So, for revenue cycle management systems to become smarter and more sophisticated, they’ll likely move up in the artificial intelligence hierarchy. We’ve outlined some of the directions RCM systems can take and other AI features they’ll likely incorporate below.
Machine Learning/Natural Language Processing (NLP)
Natural Language Processing for RCM systems doesn’t represent unexplored territory, seeing as a number of strides have been made towards implementing said technologies within RCM systems. It just hasn’t gone mainstream yet.
What kind of benefits does NLP represent for RCM systems? Standalone NLP applications are impressive as they are – we’ve seen what a limited degree of NLP has done for EHRs for note taking. See: Nuance.
As is, NLP will save practices time by being able to intelligently input data like ICD codes and patient demographics verbally into a rule-based system.
Natural language processing can save time and energy, and may entail a freer flowing RCM system where doctors, billers and practice staff can make notes on the revenue cycle – be it regarding progress made or questions about, say, a delayed claim.
In other words, NLP can strip away some of the rigidity of data entry and make it seem natural. Improvements in reimbursements undoubtedly follow. But what about NLP technologies used in tandem with automated reasoning?
As its name implies, automated reasoning technology helps produce applications that allows computers to reason automatically. Combine it with NLP and you’re a step closer to Watson, IBM’s famous question-answer supercomputer.
Automated reasoning’s roots lie in the development of formal logic, and entails systems can process complex proof systems and recognize some logical fallacies. By employing a combination of NLP and automated reasoning, RCM systems will develop the ability to make judgment calls on individual patients.
For instance, instead of automating collections based on how much time a claim has been out, automated reasoning systems can decide on a logical approach to the amount of time that has elapsed, which may depend on service, nature of consultation, etc.
This adds a bit of a human touch to RCM systems, which leads to our third prediction.
Netflix has received much recognition for its renowned ‘recommendations’ engine, which mines consumer data to figure out what movie titles you may like, thereby enhancing its competitive advantage.
But what role would data mining play in RCM systems? Well, data mining is already used in customer relationship management (CRM) applications to predict whether to contact prospects based on their likelihood of responding to an offer.
The same can be for revenue cycle management. Claims can be followed up on depending the likelihood of response, thereby enhancing automated reasoning and overall practice diligence.
Furthermore, it’ll be easier to determine claims and rate of reimbursement based on your history with the patient and the payer, mining both their histories separately and combining said data to enhance your practice’s bottom line.
It’s a brave new world for revenue cycle management. Make sure you keep up.
What kind of technologies are you embracing at your medical practice?Tweet