Physician burnout is a major concern in the United States. In 2018, researchers with Medscape surveyed 15,543 physicians across 29 specialties. Over one third (42 percent) reported being burned out and 56 percent indicated that a high volume of bureaucratic tasks, like charting and paperwork, contributed to their burnout.
Becker's Hospital Review recently spoke with Tim Ruff, vice president of solutions management at 3M Health Information Systems, now together with M*Modal, about ways that technology can address the physician burnout problem. He discussed how new solutions are reducing the amount of time clinicians must spend in front of EHRs and billing systems and increasing the amount of time available for patient care.
How speech recognition and artificial intelligence can liberate clinicians from bureaucratic work
Voice interfaces like Siri or Alexa have become the norm in the consumer world, but this type of speech recognition isn't a good fit for the healthcare sector. Although Siri and Alexa are easy to use, they are designed for quick questions and brief verbal responses.
"Clinicians and physicians don't have time for extended conversations with a computer," Mr. Ruff said. "The volume of patient data out there makes it very difficult for a Siri- or Alexa-type voice interface to provide a useful, short response. Technologies like artificial intelligence are needed to quickly access the right information."
3M’s speech recognition solutions incorporate machine learning and deep neural networks. These technologies "learn" over time, which increases the accuracy of speech recognitions. "Good speech recognition on the dictation side of the EHR allows physicians to spend less time on documentation. This translates into reduction in the administrative burden and burnout," Mr. Ruff said.
In addition to excellent speech recognition, artificial intelligence provides clinicians with proactive and context-specific "nudges" that improve documentation quality. As a physician creates a note, either by typing or through dictation, 3M M*Modal Fluency Direct’s computer-assisted physician documentation (CAPD) functionality offers suggestions which help create a fully compliant document.
"A simple example is specificity around heart failure. It's not sufficient to simply say that a patient had heart failure," Mr. Ruff said. "The doctor must specify the acuity and type of heart failure. If that information is omitted from the documentation, our system detects it and prompts the physician for more information."
Both speech recognition and AI reduce the amount of time clinical documentation specialists must spend on follow-up to ensure that documentation is accurate and complete. When documentation is correct from the beginning, physicians spend less time overall on documentation and billing-related tasks.
Healthcare is different – It's critical to handle administrative tasks in safe and user-friendly ways
The 3M M*Modal Fluency platform has been designed to address challenges unique to healthcare, such as patient safety. According to Mr. Ruff, "Our technology can understand very complex directives, but we want to be sure that users didn't misspeak, omit an important detail or were somehow misinterpreted. To be efficient and safe, the system requires a confirmation step."
In addition to speech recognition and advising capabilities, 3M’s virtual assistant technology offers data lookup functionality and automation of the EHR, so clinicians can visually see important information. These features ensure efficiency and safety.
As healthcare technology evolves, it will create a surge in physician productivity
Today, speech recognition systems require explicit directions — users give a command and then wait for the system to do the right thing. In the years ahead, the next big productivity leap will come from ambient intelligence that listens in on conversations and automatically determines what should be done.
"A good human assistant intuitively knows what to do based on a conversation between a physician and a patient," Mr. Ruff said. "We are developing the ability to listen in on physician-patient conversations and use speaker separation to extract relevant facts from both sides of the conversation."
This new wave of speech recognition and conversational AI technologies will enable 3M to infer certain tasks from conversations between patients and clinicians. For example, if a physician talks about a certain medication, an order may be entered automatically into the system. Alternatively, if the patient mentions a certain health condition that the physician fails to document, the system can categorize the information automatically.
Emerging applications of technology have the potential to dramatically reduce the administrative tasks that contribute to physician burnout. "There will be a time when the system does most of the administrative work, without the clinician telling it what to do. It will just listen in and perform the right actions. It takes time and data to train these types of systems, however. It's an evolutionary process," Mr. Ruff said.