Regulatory demands, quality reporting requirements and value-based bundled payments have prompted provider documentation initiatives to improve accuracy, yet these programs often exacerbate the already enormous administrative burden on physicians.

To address both documentation accuracy and physician satisfaction, GBMC Healthcare implemented a clinical decision support documentation-improvement program that uses artificial intelligence-enabled clinical intelligence embedded into speech-to-text dictation tools and EMR workflows.

Physicians are offered automated clinical insights in real time as they document so the complete patient story is captured.

Using the robust capabilities of GBMC's EMR, the physician's documentation workflow has been united with coding and other revenue integrity workflows to reduce retrospective documentation queries and rework.

"The result is more detailed and accurate documentation, yielding better patient outcomes, greater efficiency for the physician," explained Dr. Neri Cohen, lead clinical informatics liaison and vice chair of innovation for information and technology in the department of surgery at the Greater Baltimore Medical Center.

Cohen, who will speak on the topic at HIMSS21 in Las Vegas, said that integrating the AI-enabled clinical insights into the clinician's standard workflow on the same platform at the same time as the documentation occurs is the least disruptive way to ensure complete and accurate clinical and administrative documentation. 

"Utilizing speech-to-text tools, for example, clinicians are able to produce highly specialized content – notes or documents – more accurately and in half the time compared to standard data input tools like keyboard and mouse," he said.

Some of the things GBMC is measuring are reduction in retrospective coding queries, reduction in turnaround time to respond to retrospective coding queries, reduction in the number of "unable to determine coding queries," reduction in "discharged but not finally billed" open charts, increased physician engagement, increased satisfaction of coding specialists and identification of additional opportunities for documentation improvements, Dr. Cohen said.

"Physicians document in real time, describing the clinical details of the patient's circumstances in clinical terminology," he explained. "To meet documentation requirements for specific regulations, quality reporting and billing – either fee-for-service or value based payments – accurate description of the same clinical circumstance requires specific administrative terminology, such as language from ICD-10 or HEDIS code sets."

He said when coding specialists are not able to assign proper administrative terms from the physician's clinical documentation, the patient's chart cannot be completed and the record cannot be closed.

In response to a retrospective delayed coding query, the physician must go back into the record and amend the clinical documentation to include the necessary administrative terminology.

"This exercise, responding to delayed coding queries, is critically necessary to complete the administrative requirements for complete and accurate documentation in administrative terms, but is of no clinical benefit to the patient and is additional nonproductive work on the physician," Dr. Cohen said.

Dr. Neri Cohen will address the potential for AI-driven clinical insights at HIMSS21 in a session titled "Documenting the Patient's Story: AI-Driven Clinical Insights." It's scheduled for Thursday, Aug. 12, 2:30-3:30 p.m. in Caesars Alliance 315.

Nathan Eddy is a healthcare and technology freelancer based in Berlin.
Email the writer: nathaneddy@gmail.com
Twitter: @dropdeaded209