Administrative work takes up bulk of week for clinicians, medical office staff: Poll

Most clinicians and healthcare office personnel are spending much of their week, often most of their week, dealing with administrative burdens.

Aashima Gupta, Google Cloud’s global director of healthcare strategy and solutions, says generative AI could help reduce clinician burnout and let physicians focus on their patients.

A new report from Google Cloud and the Harris Poll offers more findings on the heavy toll of administrative tasks on the healthcare workforce. The report also finds optimism that generative AI can help reduce burdens on clinicians and support staff, easing not only bureaucratic headaches but also aiding in retention. Google Cloud released the report early this morning.

The Harris Poll, which conducted the survey, found that clinicians spend nearly 28 hours per week on administrative duties. The burden is even higher for medical office staff, who say they spend 34 hours per week on administrative tasks. Claims personnel spend 36 hours per week on such tasks, according to the report.

Most healthcare providers and payers who participated in the survey said the time spent in administrative work is tied to burnout and difficulties in retaining staff. The report found 82% of clinicians and 81% of medical staff said they felt symptoms of burnout.

Most respondents said they feel positive about the potential of generative AI to help reduce administrative burdens, including 91% of healthcare providers and 97% of payers.

• Read more: Doctors report high burnout and low impact of wellness efforts

In a media call organized by Google Cloud, healthcare leaders also expressed their enthusiasm that generative AI could help their staff.

Michael J. Schlosser, MD, senior vice president of care transformation and innovation at HCA Healthcare, said the health system has surveyed its clinicians to get a sense of the problems that were adding to their stress at the workplace.

“Administrative burden was absolutely number one on the list,” Schlosser said. Clinicians said they need to have more time back and more time thinking more deeply about their patients. “We really took that to heart,” he said.

HCA Healthcare began focusing on a key area of concern: the handoff of patients from one nursing shift to another. Across the HCA system, there are about 400,000 handoffs each week, or 21 million annually. Those handoffs involved sharing vital information about the patient in a brief summary, so the nurse taking over understands the patient’s needs and areas to watch.

“This is an incredibly important time in care delivery,” Schlosser said, adding, “Nurses spend a lot of time manually collecting data to have that concise summary.”

HCA has tried other remedies to ease the process, and Schlosser said, “Gen AI is the tool I think that’s really going to crack it for us.”

In recent months, HCA has been testing generative AI models that could provide concise but useful summaries for nurses in handing off patients at shift changes. Nurses gave good feedback on the summaries, and said 90% of the AI-enabled summaries were helpful and could replace some manual work, Schlosser said. He noted that the model was designed with clinical workflows in mind, so they would be adopted and effective.

HCA plans to begin rolling the AI model out to test in five hospitals by the end of the year, and is hoping to begin rolling it out more broadly in 2025. He said it will require rigorous testing to ensure it can be used in millions of patient handoffs each year.

• Read more: Even with AI, healthcare must be in the humanity business

Tony Farah, MD, the chief medical officer and chief clinical transformation officer of Highmark Health, discussed how the organization has used Google Cloud’s AI solution to try and improve a process he admitted is not very popular: prior authorization. With prior authorization, clinicians and health systems need to get approval from insurers before moving ahead with a treatment plan.

Payers approve the vast majority of authorization requests, yet Farah acknowledges they are spending too much time in approving those requests. So Highmark has begun automating some authorizations.

“That has started to reduce significantly the burden on the payer side,” Farah said.

In early use of automation, providers are getting approvals within seconds, as opposed to waiting days and weeks, which has an impact on patient care, he said.

Google Cloud has also introduced its Vortex AI Search for Healthcare, which utilizes Google's medical large language model to help healthcare workers find the information they need in patient records.

The company also announced its Healthcare Data Engine is now available internationally, including all regions where Google Cloud’s Healthcare API is available. Providers and payers can use the Healthcare Data Engine to build an interoperable data platform and gain new insights from patient data.

Aashima Gupta, Google Cloud’s global director of healthcare strategy and solutions, said she’s excited about generative AI reducing burdens on clinicians and staff. She also noted that healthcare leaders are concerned about both burnout and the chances of human error in manually dealing with data.

“It’s not Gen AI replacing doctors,” Gupta said. “It’s Gen AI letting doctors be doctors.”

• Read more: Building trust with the use of AI in healthcare