Getty Images/iStockphoto
Humans Top Digital Scribes For Better EHR Documentation
Clinicians scored digital scribes well behind hybrid models and human scribes when it came to EHR documentation.
Human scribes still trump digital scribes in EHR documentation, according to a study published in the Association for Computing Machinery Digital Library.
A digital scribe is not currently capable of widespread adoption because of a lack of accuracy, completeness, and medical synthesis, clinician respondents said.
Healthcare organizations are beginning to adopt and implement digital scribe solutions to ease the burdensome EHR documentation process. A digital scribe aims to automate the clinical summary during a patient-physician interaction.
But that’s not exactly the case, recent data showed. In a simulation with 24 clinicians, researchers let participants interact with anonymous digital scribes from behind a one-sided mirror. Researchers analyzed clinician interaction with the scribe, perception about note quality, and workflow impact.
The digital scribe method utilized a transcribing service and then leveraged the digital scribe to extract and summarize the transcript to generate clinical notes. The hybrid method leveraged the transcribing service, used the digital solution to summarize and generate clinical notes, and then a medical resident to edit the notes. The human method had a medical resident write clinical notes during the consultation.
Research revealed the digital scribe scored the lowest average score of the three methods. The hybrid model finished second, while the human method was the clear clinician favorite.
“When participants reviewed Machine notes, they found accuracy, completeness, relevance, and comprehensibility to all be lacking,” wrote the study authors.
“They reported the system erred in places where information was clearly and specifically articulated during the encounter. These may include inaccurate labeling of pertinent information, incomplete description of overtly discussed symptoms, repetitive, non-relevant or incoherent writing, and disorganized content that made it difficult to follow and understand what happened during the encounter.”
Clinicians added they could not trust or tolerate an inaccurate or incomplete scribe system because it could lead to patient safety issues.
"This is just not helpful, there's a lot I would add,” the anonymous clinicians agreed. “I feel like it's not going to capture the correct information. I would prefer to write my own notes; it would have been faster than editing."
Accuracy was a significant issue for the clinicians.
"Some of the negatives were put as pertinent positives and positives as negatives,” wrote an anonymous clinician. “It's unsafe. If I miss something and then it says something I didn't intend, that could change my management of the patient.”
Clinicians expressed mistrust with the hybrid model after it did not complete the SOAP note correctly, which clinicians use to describe the subjective, objective, assessment, and plan of a clinical encounter. One respondent said the subjective section was acceptable, but it missed the note’s objective, assessment, and plan portions of the note.
Overall, clinicians said the hybrid model was inconsistent, and it was full of different factors to “help me push towards or away from the diagnosis.”
On the other side of the spectrum, the clinicians overwhelmingly preferred the human method.
"That's pretty good,” the anonymous clinicians wrote. “That covers a lot of what I was looking for and looks like an actual note I'd take. I would use this if the system consistently produced this quality."
Another physician applauded the human method’s accuracy and completeness.
"I think the fact that it was able to capture a lot of the questions I asked, that was really helpful and saved me time,” the physician noted.
To be clear, separate data has shown human scribes to be high-cost. Human scribes also experience high burnout, resulting in a high turnover rate for the role.
Although the digital scribe scored the lowest, research revealed roughly half of the participants expressed optimism with its potential to boost documentation efficiency and patient engagement.
“It’s unfortunate the way encounters are now where you have to type while you talk to the patient, or else you’d be spending hours doing your notes,” an anonymous clinician said. “Patients don’t like it. They would much rather you sit and just have a conversation with them. I don’t want to either, but it’s to make sure my documentation is complete.”
Clinicians recommended digital scribe developers improve accuracy and relevance, completeness, and medical synthesis.
“We argued that the limitations of current language processing technology make digital scribe systems impractical for actual clinical use,” concluded the study authors. “However, physicians are optimistic about using the digital scribe system in the future to improve patient engagement and documentation efficiency.”