As healthcare professionals face increasing demands, many doctors find themselves balancing patient care with administrative tasks, especially in the latter hours of the day. In addition to addressing patients’ health concerns, doctors often spend valuable time recording detailed notes for billing and documentation purposes.
This dual responsibility, which includes managing electronic health records and completing other administrative duties, can lead to feelings of exhaustion and uncertainty about the quality of care provided.
Experts warn that the growing administrative burden is contributing to physician burnout, raising concerns about its impact on patient care.
This scenario is increasingly common among healthcare professionals and has been linked to significant burnout.
A survey published in the National Library of Medicineon the impact of administrative burden found that two-thirds of the 1,774 physicians who participated reported that administrative tasks hindered their ability to provide the best patient care.
On average, physicians spent 24% of their office time on administrative duties, and those who devoted more time to these tasks reported lower career satisfaction and higher burnout levels.
The burden extends beyond office hours as well, with 48.2% of physicians in 2023 reporting burnout, often attributed to administrative work outside regular hours.
Notably, 20.9% of physicians said they spent more than eight hours per week on electronic health records outside of office time. While these findings highlight concerns, there are efforts to improve efficiency as technological advancements continue to shape the healthcare industry.
In the past year, Dr. Dong-Gil Ko, Dr. Umberto Tachinardi, and Dr. Eric J Warm from the University of Cincinnati (UC) have been working on an innovation to use artificial intelligence to improve the administrative work process in healthcare, specifically telehealth billing.
As the primary researcher behind this project, Ko’s primary purpose was shifting telehealth’s current model to prioritize physician expertise over physician’s time spent on the task.
In a podcast by the Ko Lab, the speakers discuss how the current model of telehealth-billing doesn’t consider all the different aspects that go into it.
“The current system doesn’t really capture the complexity of medical decision-making when you take it to a digital space… it’s always like ‘I pay for this service I get the service’ but with healthcare, something so complex, weren’t not just paying for the doctors time but the years of training and the ability to take all this information and make a decision about your health,” said Ko in the podcast. “And that doesn’t change because it’s on a screen and that’s where this research comes in.”
Recognizing the complexities of medical decision-making, Dr. Ko aims to shift the focus of telehealth billing from response time to the physician’s expertise. Under the current system, billing is based on how quickly physicians respond to secure messages, which can lead to inequities, as more experienced physicians may respond faster but are compensated less.
Ko’s new system, which incorporates artificial intelligence, evaluates a physician’s expertise upfront to ensure fairer compensation.
Looking ahead, Ko hopes this AI-driven approach will also assist with other telehealth tasks, such as predicting payment needs before a patient’s question is submitted.
This model, he explains, could help physicians dedicate more time to complex cases while improving the overall physician-patient relationship.
“We don’t want to replace doctors with AI, it’s how do we use these tools to make the doctor patient relationship better, so if AI can take care of the quick ‘do I need a refill’ questions, doctors can spend more time on the really difficult cases, so it’s finding that sweet spot right where technology and human touch can work together and this feels like a good step,” Ko said.
To support their project, the researchers conducted an experiment using machine learning models and mixed-effects modeling, as outlined in their article Secure Messaging Telehealth Billing in the Digital Age, published in the Scholarly Journal of Informatics in Health and Biomedicine.
They trained eight classification machine learning models using electronic health record audit log data, focusing on non-urgent patient messages. The mixed-effects modeling helped assess the significance of the findings.
The results indicated that the machine learning models had a positive impact on billing for secure messaging telehealth services, aligning with the goal of prioritizing physician expertise over response time.
With what’s already been done and plans for this project to be launched in 2025, thoughts arise on how healthcare will continue to evolve in the future and whether artificial intelligence will or will not have a significant role in it.
For individuals currently pursuing an education to go into medicine, artificial intelligence is a topic that is made aware of early on so students can understand to what extent it can be used as of now in their professional careers.
Madison Powell, a first-year medical student at the College of Medicine, discusses how they had discussions on artificial intelligence from the beginning of their education.
“It’s definitely something that we’ve talked about since the first week of medical school in terms of using AI in notetaking for physicians after they see patients because the amount of administration time that physicians have versus the time, they have to actually see patients is a big contributor to physician burnout,” Powell said. “So, I’m interested to see how it [AI] grows as we’re still very early.”
For those currently in a professional career position such as professor Heather Nester who teaches classes for undergraduate students figuring out their professional path, she believes that AI does have the potential to make a significant impact in healthcare moving forward.
“If done ethically and correctly, I do believe AI has the potential to revolutionize many industries. In healthcare, I could see AI being used to analyze medical images, creating personalized patient plans with cultural competency resources included, use predictive analysis to take proactive measures for patients and the greater society, scheduling appointments for patients, etc,” Nester said.
“AI can continue to make a significant impact as long as it is supporting the work of professionals so simple tasks can be streamlined and more focus can be placed on patients.”
As discussions about the role of artificial intelligence in healthcare continue among students and professionals, some organizations have already begun experimenting with AI in clinical settings. Sutter Health, for instance, is participating in the American Medical Association Health System Program, aiming to improve the future of medicine by providing enterprise solutions for leadership, physicians, and care teams.
The organization is exploring ways to reduce the time physicians spend on electronic health records outside of office hours, with AI being one potential solution.
AI is being utilized to generate visit notes following physician-patient interactions and to draft responses to patient messages, which physicians can later review and revise. While the system is still in the testing and refinement phase, it offers a glimpse into the potential integration of artificial intelligence in healthcare.
Regardless of how technology evolves or whether AI becomes a larger part of administrative tasks in healthcare, the central objective of healthcare remains clear of providing quality patient care.
“Keep quality patient care at the forefront of the goal always, I think there’s value in having a person whose experienced lots of different things and a person who knows you as an individual, because every patient is different and not everything can be summed up in an algorithm, so I think experience and human connection is not replaceable,” Powell said.
link