AI-enabled value-based care company Navina has released preliminary research showing that adding patient history to ambient scribe technology significantly increases documentation quality.
Navina’s research found that adding historical context improved documentation quality by 18%. It also increased the clinical note’s completeness score from 40.4 to 82.9 on a scale of 0 to 100.
It suggests that ambient scribes alone don’t fully capture necessary context to save doctors time parsing through older patient records nor accurately capture the complexity of chronic conditions.
Navina assessed documentation quality with a subset of the QNOTE clinical note documentation quality instrument, and it assessed completeness on a 100-point scale. The research still needs to undergo the peer review process and is currently available as a preprint on medRxiv.
Yair Lewis, M.D., Ph.D., chief medical officer at Navina and a practicing clinician, said conversations between physicians and patients don’t always include necessary information for evaluating the patient. Information like specific lab values and dates aren’t naturally spoken aloud during a patient encounter, he said.
“So patients [with] diabetes goes to see their primary care doc … it’s not going to be natural for the doctors to say, ‘oh, so, Mr. Johnson, your last a1c, was 8.2 from [this] date, and that improvement compared with 8.7 from [this] date, and your kidney function is [this],’ right?” he explained.
Navina used 354 primary care interactions for patients with diabetes and hypertension. The study uses the two common conditions as a proxy for all chronic conditions.
Lewis said the lack of relevant medical information in the documentation can impact downstream quality of care, have financial implications for value-based care arrangements and fail to save doctors time by using the ambient scribe technology.
“Without incorporating a patient’s full clinical history, including labs, comorbidities, medications and preventive care, important aspects of chronic conditions are left undocumented,” the study says. “These omissions can impact quality of care, increase physician burden and reduce reimbursement accuracy.”
Many major AI scribe companies have been incorporating patient context into their products. Nabla has a context-aware agent that creates a patient summary before the visit. Abridge advertises electronic health record context and a predicted problems section of its clinical note generation technology. Suki has both a pre-visit summary and a question-and-answer function with patients' past medical records to find relevant information. Ambience surfaces patients’ history, labs and notes and offers an AI copilot for the clinician to interact with.
“A lot of time [what] happens is doctors are going to be taking the ambient summary, but then modifying it manually by looking through the record and adding those relevant pieces of information,” Lewis said.
Even though AI scribe companies are moving in this direction, data extraction from EHRs is difficult, Lewis explained. He said provider organizations should consider what parts of the EHR data a vendor is able to tap into, whether they’re connected to health information exchanges and how patient history is combined with the eventual clinical documentation.
Navina’s AI technology surfaces patients’ historical medical context for providers at the point of care. In July, it teamed up with ambient scribe company Nabla to add AI scribing to its platform.
Lewis said in the future, the company wants to expand the study to include thousands of encounters examining a variety of chronic diseases.
“Now that it's actually semiquantifiable, I think it's just going to be another sort of headwind to make more companies work in that direction,” Lewis said of the impact of the research.