Harnessing Technology to Address Physician Burnout

Mar 1, 2024

Rebecca G. Mishuris, MD, MPH, MS, Chief Medical Information Officer and VP, Digital at Mass General Brigham, believes that understanding the contributors to burnout is critical to designing impactful interventions. She recently published a study in JAMIA Open, proposing a novel metric that can both identify risk of burnout and measure the impact of burnout interventions.

In this Q&A, she discusses how Mass General Brigham is using technology to help measure physician workload, prevent burnout, and make an impact on both physicians and patients.

Q: Could you give us an overview of the factors leading to physician burnout?

RM: Burnout is multifactorial. Things like control over one's schedule, teamwork in a clinical setting, inefficiencies in workflows, and the electronic health record (EHR) all contribute to it.

There are increasing demands on people's time in health care — from the ever-expanding amount of clinical data to increased inbox messages to other aspects of a physician’s job outside of clinical care. Physicians may have their clinical job, but they may also have another role in the hospital, as a researcher or administrator. They have family responsibilities or home responsibilities. Burnout happens when there’s a mismatch between demands and resources, whether those are institutional or personal.

Q: How do EHRs contribute to this problem?

RM: I think the EHR is the face of many other aspects of clinical care. That's everything from billing requirements to external reporting needs to external quality measures to communication with other care team members. Some of that is critical to safe effective care, but some of it is administrative burden or inefficiencies in workflow. So, it's not actually the EHR itself that is the sole driver of burnout, it's all of those things combined that are coming at you through the EHR. Bill Gates, I’m paraphrasing, said if you automate an efficient process, you will highlight that efficiency, but if you automate an inefficient process, you will exacerbate those inefficiencies. We see this in the EHR, particularly related to in basket management.

That being said, the EHR itself can contribute to burnout if it's not well designed. If instead of one click, I must do 10 to accomplish the same task, that’s going to add to my negative experience when using the EHR, and by work burden overall.

Q: How does burnout negatively impact providers and their patients?

RM: A burned-out physician is more likely to suffer from substance use problems and mental health issues like anxiety, depression, suicide, suicidality. They tend to have poorer emotional and physical quality of life and have broken relationships with friends or family.

A burned-out physician is also more likely to be depersonalized and have poor interactions with the people around them, which negatively affects patient experiences. They can suffer from impaired memory and attention spans, which leads to an increase in medical errors. They’re also more likely to leave either their practice or the field of medicine altogether, meaning that we have fewer physicians to take care of patients. This creates even greater access problems, diminished productivity and reduced contributions to the field as a result of burnout. It is an incredibly important, impactful and prevalent condition, unfortunately.

There’s no silver bullet to solve burnout, but we’re hopeful that [ambient documentation] will be a game changer.

Rebecca G. Mishuris, MD, MPH, MS
Chief Medical Information Officer
VP, Digital
Mass General Brigham

Q: Could you tell us how you’re using technology to address burnout at Mass General Brigham?

RM: One of the things we’re excited about is the use of generative AI for ambient documentation. The idea is that a secure app on my phone would listen to my clinical visit and draft my notes for me. This means I don’t have to spend the time to do it myself between clinic visits, or at 8 p.m. at night, or four days later. It also means I can pay full attention to my patient and not the EHR while I’m with them.

We’re about to launch a pilot of 500 physicians and advanced practice providers using ambient documentation at Mass General Brigham. There’s no silver bullet to solve burnout, but we’re hopeful that this will be a game changer for both provider and patient care experiences. The early, anecdotal feedback is that people are loving it. Physicians are able to look at their patients during visits instead of having to type away as they talk. And that’s what it really is about for me, to be able to figure out how we can use technology to deliver care better and to deliver better care.

Q: Are there any measures Mass General Brigham is implementing to ensure the responsible use of AI?

RM: It’s incredibly important as this technology becomes more widely spread that we understand its limitations and ensure guardrails around its use. Over the summer, we launched an AI Governance Committee at Mass General Brigham to develop a framework for the responsible use of AI.

We want to ensure that concerns around issues such as equity, privacy, transparency, and security are addressed, and that vendors are held responsible for the performance of the technology itself. For instance, it’s important that the ambient documentation works just as well for providers and patients who don’t speak English as their primary language. We also require that our vendors delete the recording used for ambient documentation once we've created the note because we don't want a recording of somebody's voice out there, for any number of reasons.

We want to make sure that these technologies are having a benefit for our patients, for our providers, for our health systems, for society. And so, we're starting in the low-risk space of reducing administrative burdens before we get into more directly impacting patient care.

Q: How do you see the role of AI in health care evolving over the next 5–10 years?

RM: I have no idea! I hesitate to predict what might happen that far out, but I think in the next 3–5 years, we will drastically change how we deliver care. Compared to 10 years ago, there is exponentially more medical knowledge and information available to healthcare providers today. With large language models and AI in general, we’re able to harness all of that information and make it actionable for an individual provider and patient.

I don’t think it’s coming to practice tomorrow, although plenty of people are hard at work already to develop and test this, but there will be a day when AI will be able to take a patient presenting a constellation of symptoms and help me refine a differential diagnosis and treatment plan. We must be good stewards of the technology as we roll it out, be cognizant of its limitations, and use it in a way that promotes better care — higher-quality, safer, more equitable care.

Read Dr. Mishuris's study