OpenEvidence has released artificial intelligence-powered medical coding capabilities built into its clinical AI assistant.
The new feature, called Coding Intelligence, provides automatic suggestions for Current Procedural Terminology (CPT) codes, evaluation and management (E/M)-level recommendations, and supports medical decision rationale and ICD-10 diagnoses written directly into notes, the company said.
The tool, available on OpenEvidence Visits, provides coding derived from clinical documentation and is automatically applied at the end of every physician and patient visit, executives said.
“Without any additional work, OpenEvidence can generate a concise rationale for CPT + E/M recommendations, which truly captures the complexity of your practice and saves you time while in the ER,” said Ania Bilski, MD, vice president of clinical AI at OpenEvidence.
OpenEvidence, an AI-powered medical search engine, is growing rapidly, supporting approximately 18 million clinical consultations from board-certified physicians in the United States in December. The company says OpenEvidence is currently actively used daily by more than 40% of U.S. physicians, on average, in more than 10,000 hospitals and medical centers across the country. The company has expanded from clinical search to other clinical workflows, exposing it to more direct competition from other healthcare AI companies. In August, the company unveiled its Visits feature, a clinical AI assistant that transcribes patient visits. Last month, OpenEvidence made its AI-integrated doctor dialer capabilities more widely available, taking over Doximity’s core business directly.
Modern medical billing is complex and time-consuming. OpenEvidence executives say that with tens of thousands of billing codes and multiple ways to code the same visit, the complexity of billing makes it difficult for physicians to receive appropriate reimbursement without taking their focus away from patient care.
Most coding solutions today are very “outdated”. In fact, even today, most coding is still done by doctors themselves, nurses, or outsourced to third-party agencies that employ humans to do the coding,” Zachary Ziegler, co-founder and CTO of OpenEvidence, told Fierce Healthcare.
Many of today’s AI-powered solutions simplify the process and allow for simple mapping to these common codes.
“In reality, there are currently tens of thousands of codes between CPT and ICD-10, many of which represent complex, nuanced, and context-dependent situations. Coding intelligence is integrated directly into the visit, inferring the entire visit record and the final clinical record generated to provide a comprehensive understanding of what was done, what diagnosis was reached, and what treatments and tests were ordered. “We find the exact evidence using the same technology that allows us to find the exact evidence when we use the complete set of codes that are appropriate for a particular visit,” Ziegler said.
OpenEvidence’s coding tools can report an accurate and complete set of codes that represent the entirety of what happens during a visit for all physician billing methods, including E/M codes, including complexity-based or time-based procedural CPT codes using medical decision making and supporting ICD-10 diagnosis codes, Ziegler said.
According to OpenEvidence, Coding Intelligence provides E/M-level recommendations with a complete medical decision rationale already written.. OpenEvidence automatically generates breakdowns from clinical notes, so whether a claim is due to complexity or time, the basis is already documented and included in the record.
Coding suggestions include uncommon procedural codes that can be easily overlooked in complex cases. OpenEvidence has built coding intelligence to ensure the correct codes are submitted, minimizing the time doctors spend on paperwork while minimizing revenue loss, executives said.
The company has also designed a coding tool that provides automatic CPT code sequences to maximize reimbursement. Under Medicare’s multiprocedure payment reduction policy, all procedures after the first procedure are reimbursed approximately 50%. The order in which the codes are listed determines how much the procedure will be reimbursed. OpenEvidence displays the expected RVU value along with each CPT proposal, allowing the code to be properly ordered before a claim is released, executives said.
OpenEvidence also applies CMS Correctcoding Initiative (CCI) rules to verify all code combinations before making recommendations, Ziegler said. Submitting incompatible code pairs is one of the most common causes of claim denials and compliance flags in medical billing.
“By running all proposed code sets through our CCI rules engine, Coding Intelligence ensures that only valid, billable combinations are presented to physicians. This eliminates all kinds of errors that human coders and simpler AI tools routinely make,” he said.
