Microsoft Unveils AI Copilot to Help with Clinician Burnout

Amid increased competition in the clinical assistance world Microsoft has unveiled Dragon Copilot, its new AI-backed assistant for doctors, which it hopes will not only change interactions between clinicians and patients but protect overworked doctors from burnout.

Microsoft already has a big presence in physician assistance tools through Dragon Medical One, which uses speech recognition software to document patient information, and DAX Copilot, an AI-backed clinical assistant and decision support tool.

Dragon Copilot builds on this by not only combining the two technologies but adding generative AI. Microsoft says this will not only save time and make the patient experience better but also allow clinicians to draw on information from external sources like the FDA to further guide clinical decisions.

Lifting the Burden

In practice, the new technology could allow a physician to automatically write up clinical notes and referrals and ask Dragon Copilot the correct course of treatment.

Yesterday, Microsoft gave the example of a doctor asking if a patient should be screened for lung cancer and receiving answers with links to other sources including the Centers for Disease Control and Prevention.

Touting the benefits of the technology Microsoft said that clinicians reported saving five minutes per consultation, while 70% reported reduced feelings of burnout and fatigue. 62% of clinicians also said they were less likely to leave their organization and the tech company reported that 93% of patients reported a better overall experience.

Microsoft intends to launch Dragon Copilot in May in the US and Canada before rolling it out to the UK, Germany, France and the Netherlands later in the year.

Increased Competition

Microsoft’s entry into the world of AI-back clinician assistance has come via its $16bn acquisition of Nuance Communication in 2021, the creator of Dragon Medical One.

But Microsoft is not alone in this highly competitive space with companies like Abridge and Suki, also looking to get a foothold in an industry that is rapidly growing in an attempt to capitalize on reducing burnout among healthcare workers.

However, having the full weight of a multi-billion-dollar company behind it is bound to help Dragon Copilot in any competitive process. Not only does the group have more marketing muscle, but its technology also seamlessly integrates into its hardware and software, which is used by many businesses, providing a useful competitive moat.

Stopping the Burn

Microsoft’s latest invention comes amid growing healthcare staff shortages and increasing number of professionals leaving healthcare settings due to burnout due to both clinical and administrative work.

A study by rival tech company Google showed doctors reported spending nearly 28 hours a week on administrative tasks, with almost nine hours alone spent dealing with documentation. The study also revealed that doing administrative tasks contributed to feelings of burnout for 82% of clinicians.

So having tools like Dragon Copilot could help with retention – and if its claims over efficiency play out, doctors having more time to spend with patients. This in turn argues Microsoft should lead to improved quality of care, job satisfaction and retention.

Proceed with Caution?

There have, however, been concerns about hallucinations in generative AI devices in healthcare. A study led by Cornell University last year showed that transcriptions created by OpenAI’s Whisper, used by a company called Nabla, showed hallucinations in medical conversations.

Additionally, a group of researchers at the University of Pennsylvania last month published a study in JAMA Network Open questioning whether AI ambient scribes did decrease clinical note burden. While the paper found the surveyed 46 clinicians using Dax Copilot did save time, clinicians’ feelings about the technology were mixed, with some finding ambient scribe-generated notes accurate and detailed and others finding them overly error prone.

There are also questions around data security and potential biases. In light of this Microsoft has been keen to emphasize that Dragon Copilot is built on a secure HIPPA-compliant framework. The group has also highlighted Dragon Copilot’s ability to learn from user interactions to improve accuracy.

This mitigation looks wise, as ultimately the success of Dragon Copilot will depend on patient trust, and the technology living up to its promises around workloads, for it to be fully embraced.