Improving Processes Within Health Care Systems Through AI, Natural Language Processing

By Julie Gould 

We recently spoke with Joe Ferro, president of abeo Technology Solutions and founder of TRUSTEDi10, regarding the use of artificial intelligence and natural language processing within health care systems.  

Mr Ferro is a long-time entrepreneur and senior health care executive with more than 25 years of health care informatics experience. His career has been focused on deploying and optimizing technology to assist health care institutions with proper, appropriate and compliant reimbursement. In conjunction with his teams, they have designed and deployed pioneering web-based health care technology to hundreds of hospitals, thousands of providers and many payers.

Using your technology as an example, can you highlight the importance of artificial intelligence in health care? 

In today’s health care environment, we are overwhelmed with data. Today’s EMRs store a tremendous amount of information on each patient as well as copious payment rules, contract parameters, documentation tables, and so on. The health care community has known how to store, retrieve and report on data for a long time. The root problem is how to turn health care data into actionable information.

A purposely engineered artificial intelligence (AI) platform like TRUSTEDi10 provides a computer-assisted coding expert on the user’s desktop. The health care system now has access within its EMR to an intelligent tool that makes connections across large data sets, applies rules and conditions, and produces actionable information for the user instantaneously. TRUSTEDi10’s learning AI allows bulk health care data to become actionable information without requiring outside expert assistance.

How do programs with natural language processing help improve processes within health care systems?

EMRs create detailed medical records that are easily accessible to a provider. The problem is that medical records are now regularly exceeding 10,000 pages. When properly deployed, natural language processing (NLP) in a system like TRUSTEDi10 “reads” the data and can allow the user to instantly identify specific items, thoughts, and conditions. When a provider needs to see actionable, relevant information from within their EMR, NLP can be a vital tool used to quickly identify it within the chart so the AI can then provide the actionable information.  

What are some of the major benefits of using these types of technologies in health care systems?  

I see AI and NLP as a means to provide expertise directly to a user, which makes their job easier, their process more efficient, and their output more accurate. The ability to deploy AI and NLP to a user’s desktop, integrated within their EMR, is incredibly powerful. Picture a medical record that contains all the information from a patient’s current medical procedure. This record will have automated output from devices, results from vitals, all notations, provider notes, nursing notes, lab results, discharge instructions, and so forth. All of this information is typically presented in a semi-structured way: the first service is the first entry into the chart.

The identification and review of the action items on this chart are cumbersome and time-consuming; however, AI and NLP can automatically re-organize the patient chart into actionable views customized for each user. When reviewing the chart, the provider might want to see vitals and outcomes first, with discharge instructions highlighted. The patient account staff, however, might desire the procedure codes and diagnosis codes with descriptions first.  

The use of this technology can also be applied to identify clinical documentation inconsistencies. Many EMRs have labor-saving components built-in, which include multiple provider templates. A provider may select a labor-saving template that defines the patient’s vitals as “All Normal,” while the chart clearly shows high blood pressure. This minor documentation issue would cause billing and payment issues and follow-up work for many parties. TRUSTEDi10’s technology will identify this documentation issue while the provider is charting and notify them of the discrepancy immediately.  

What are some of the challenges? Specifically, what challenges are faced when implementing these technologies into practice? 

The multitude of technical and clinical systems that a provider needs to learn and stay current on is among the most significant challenges surrounding AI and NLP technologies. The setup and maintenance of many disparate systems present staffing and maintenance challenges. Additionally, keeping all systems in sync and integrated with each other presents yet another obstacle to overcome. Proprietary infrastructures further increase that integration obstacle.

I strongly believe that new technology must work from within the hospital or practice infrastructure. Billions of dollars are being spent on advanced EMRs, so hospitals and practices want to ensure that they are maximizing the return on their EMR investment. It’s essential that providers and staff work from their EMR platform. Outside vendors benefit the hospital or practice by supporting their existing EMR. This support saves training time, increases productivity through system consolidation, improves security with fewer passwords needed, and does not add computer clicks to users’ work process.  

Is there anything else you would like to add?  

It is an exciting time to be in health care IT. AI opens a new world of efficiencies that will make health care more productive, efficient, and, ideally, safer. My development team at abeo Technology Solutions is working on new projects to take our TRUSTEDi10 application to the next level, so stay tuned for more developments in the future.

Joe Ferro is the President of abeo Technology Solutions and Founder of TRUSTEDi10. abeo provides revenue advisory, services and technology solutions that help to drive optimization, compliance and efficiency surrounding hospital-based specialties like anesthesia. TRUSTEDi10 is a computer-assisted coding application using AI and NLP technology for emergency department E/M level coding.