Digital Health
Technology
Debating the Utility of EHRs AI and Other Digital Investments
I had just finished giving a presentation to a room of hospital CEOs on the current AI landscape when one of the executives raised her hand.
“Is AI really going to be a thing in health care? We’ve spent over a decade installing EHRs in our system and all they have done is increased cost and burnout. And with cybersecurity threats, I feel AI is only going to make things worse,” she said.
I responded by noting that yes, technology investments can be daunting as they not only put pressure on organizations financially but can also disrupt workflows and other processes as we figure out ways to integrate and optimize new technology. It’s also clear that cyber risk is a growing problem that we don’t have an easy answer for. I then asked the CEO if, hypothetically, she could wave a magic wand and remove the EHR and other digital systems from health care and revert to paper-based processes, would she?
Without hesitation, she responded with an emphatic “YES.” It was at this point that another CEO also chimed in to say that he would do the same.
I’ll admit I was caught off guard by this response. Not only because I had just given a stellar presentation on AI that should have addressed all their concerns (kidding), but because despite the cost, integration headaches and other risks associated with digital investments, we have seen that most health care stakeholders we speak to are optimistic about technology’s role in supporting the future of health care. At the time, I decided to play it safe by acknowledging that they had a right to feel skeptical about AI’s impact and frustrated by their historical challenges in using EHRs, but we ultimately ended on, “Let’s agree to disagree.”
But after thinking about this exchange some more, I have thoughts.
What does the data show on EHR use?
There’s no denying that physicians are spending too much time in the EHR during and after work hours. A recent study of over 200,000 ambulatory care physicians at 396 organizations found that physicians spend an average of 5.8 hours per 8 hours of scheduled patient care actively using the EHR (eg, documentation, chart review, orders, inbox). Most of that time is spent doing documentation or chart review, with a significant share occurring outside of standard clinic hours. Specialties like infectious disease, endocrinology, nephrology and primary care show the highest EHR use times.
Despite the increase in EHR burden over time, there are ways to help mitigate these challenges. A separate analysis of over 2,000 family medicine physicians found that greater software usability was associated with higher EHR satisfaction, such as making it easier for users to find relevant information or reducing unnecessary alerts. EHR usability was also affected by other efficiency strategies such as using scribes, templated text and voice recognition or transcription tools. The study found physicians with highly usable EHRs had a 38.5% chance of being very satisfied with their EHR, whereas those using two or more efficiency strategies had a 51.4% chance of being very satisfied. The study’s authors also note that there are other factors beyond usability that impact EHR burden, including policy, reimbursement and organizational support.
Will AI reduce or amplify EHR challenges?
The introduction of more powerful large language models and other generative AI capabilities has brought renewed interest in health systems seeking to reduce EHR-related burdens for their staff. While still early days, initial pilots have produced promising results across different tasks:
- Northwestern Medicine has been testing generative AI within its Epic workflows to assist with clinical documentation, finding that users saw a 24% decrease in time spent drafting notes and a 17% decrease in after-hours work.
- Mile Bluff Medical Center in Wisconsin is using Meditech’s Google-powered AI search and summarization tool to quickly pull, synthesize and organize structured and unstructured data in medical notes, helping clinicians save an average of 5–10 minutes of previsit prep time per patient.
- Stanford University School of Medicine evaluated using generative AI for drafting replies to patient inbox messages, finding that while the tools didn’t alter reply or read times, there were statistically significant reductions in both physician task load score derivative and work exhaustion scores.
Examples like these help to shed light on where AI can make an impact—but to be clear, we still need more rigorous studies to fully understand its potential.
What would it mean to revert to analog processes?
Let’s return to the thought experiment of magically removing the EHR, AI and other technologies in health care. Would clinician burdens ease overnight? Would privacy and security risks subside? Would care quality and coordination improve? Would manual workflows enhance the patient experience? The predigital era wasn’t that long ago—we can recall what those days entailed:
- Manually sifting through stacks of paper records to review or update patient notes
- Deciphering illegible handwriting, increasing the risk of medical error
- Needing physical transport of information across departments or sites of care
- Requiring a phone call to schedule every visit and issue appointment reminders
- Lacking infrastructure for virtual visits, remote monitoring or home-based care
- Lacking advanced analytics for predictive insight or personalized medicine
Digital health has certainly introduced new costs and complexity to health care, but technology will be a necessary component to ensuring that the health care industry can keep up with other aspects of the 21st century. Paper forms, landline phones and fax machines can only take us so far.
Where do we go from here?
In our Sg2 research we have continued to promote the idea that digital transformation has no endgame. Digital initiatives are never “done,” they just evolve, and so the best way to navigate ever-shifting business priorities, hype cycles and competitive threats is to have a disciplined approach that prioritizes the fundamentals. That means not only continuing to optimize EHRs and other features of your organization’s IT operational backbone, but also being thoughtful about matters of governance, embedding a transformational culture and working to build strategic partnerships with external stakeholders. Setting that foundation will make it easier to adapt to this fast-paced (and increasingly AI-enabled) digital environment and support your various initiatives over the long term.
Additional resources on optimizing the EHR:
- The American Medical Association’s STEPS Forward open-access library of tool kits and playbooks
- Vizient blog: Optimizing Electronic Health Records (EHRs) to Improve Clinician Efficiency, Business Decision-making