From emergency department pilots to campus-wide rollout, rural facilities share hard-won lessons on AI implementation, including leadership education, steering committees, safe experimentation, and using ambient tools to ease provider workload
When everyone else is zigging, that’s a good time to zag.
Such was the case for the Arizona-based Mount Graham Regional Medical Center when it implemented an ambient listening tool, powered by AI, in its facility. While many hospitals are focused on bringing this to clinics, this hospital started with the emergency department (ED), according to Justin Millar, the organization’s CIO.
His comment was part of a webinar hosted by the American Hospital Association (AHA) titled “Building the AI-Ready Rural Workforce: What Leading Hospitals Are Doing Now.” The session provided first-hand accounts of AI implementations from four executives across three different rural facilities.
AI to improve patient charts with less effort
Ambient listening tools promise to help doctors with charting by listening to patient interactions and summarizing the conversation for later notation. Mount Graham turned to DAX Copilot by Microsoft for this capability, and Millar says it made a real difference by reducing the “cognitive workload” for the facility’s ED physicians.
On any given day, ED doctors have approximately 70 interactions with patients. That’s a lot of detail to remember for charting later. With ambient listening, it was no longer necessary. Instead of “rushing to a computer” to enter notes after every interaction, the tool enabled providers to focus on the patient.
This shift brought tangible results: charts got better, and doctors were less fatigued at the end of the day. It also cut the “pajama time” – the time providers spent at home, off duty, on nights and weekends, finishing charts.
Getting to this point wasn’t quick. Millar said Mount Graham has been experimenting with DAX Copilot for about two years. Now that they’ve figured it out, they are currently rolling out licenses for all providers in the facility.
Stay in touch by subscribing to our email newsletter.
We will never share or sell your email address.
A rural health blueprint for approaching AI
The webinar effectively provided a blueprint for any AI implementation and included lessons learned from Mount Graham and other rural facilities. These include Artesia General Hospital in New Mexico and Indiana-based Reid Health. Below are some of the lessons we took from the webinar.
1. Start with leader education
Education is crucial for level-setting what AI can and cannot accomplish. Leadership buy-in has long been essential for implementing new technology – proactive efforts to educate leaders are a sound way to get a head start.
Artesia CIO Eric Jimenez started by simply sharing a series of short educational videos created by Microsoft as part of its Rural Health Resiliency Program. The videos were created in partnership with the Georgia Rural Hospital Association and are freely available. Each one covers a different aspect of AI in a rural setting and runs about 25 minutes long.
Importantly, Jimenez was careful to follow up. “Hey, did you have a chance to review that?” he’d ask. When they hadn’t, he’d stress the quality of information and how reviewing it would facilitate his IT department in helping its departments identify practical opportunities.
Millar, the CIO from Mount Graham, found that simply making rounds and engaging with different departments on a regular basis went a long way. He’d routinely spend 15 minutes with different leaders across the organization. He’d ask them about their IT issues and evangelize possible AI solutions, which he believes will be more transformative than even EHR.
2. Consider forming an AI steering committee
One speaker shared an anecdote about an emergency doctor showing up to work with Meta Glasses. The intent was good; that physician was merely being proactive in finding a way to use AI to assist him in his job. However, it brought a plethora of risks, including HIPAA concerns.
Healthcare facilities need to get ahead of this by developing a governance program. The interactions Millar had in his “rounding” helped him foster stronger relationships throughout the hospital. Eventually, they formed an AI steering committee to address this from a policy perspective.
Some of the key governance areas mentioned on the webinar that an AI steering committee should address include:
- Access
- Ethical use
- Guardrails; and
- Cybersecurity.
Importantly, implementing a governance policy isn’t about banning shadow AI, like a doctor bringing personal tools to work. Rather, it’s designed to establish the structure to help rural healthcare put AI in place for the benefit of doctors and staff.
3. Identify well-defined problems AI might solve
Another task for the steering committee is to identify well-defined problems that AI might solve. Starting with a problem focuses these projects on real needs to produce tangible results. Aside from charting and notes, some of the other projects participants mentioned on the webinar were:
- Pre-registration of patients for appointments;
- Insurance verification and co-pay processing;
- Revenue cycle management – especially timely invoicing; and
- Assistance with imaging and MRI analysis.
This is just a sample of opportunities from one group of facilities – there are many effective use cases for AI in healthcare.
4. Develop training and a safe spot for experimentation
Training is essential to remove the fear of AI and help staff learn how to use it effectively. The speakers described a blend of options from formal hands-on training to facilitating natural interactions.
Millar described role-playing with clinicians in the early stages of their ambient AI implementation. “I would actually go and do a hands-on training with them, and I would interact with them as if I were the patient,” he said. This provided the physician with the opportunity to get familiar with the tool and build confidence in a safe environment.
Ben Wells, the president and CEO of Reid Health, emphasized safety in experimentation and implementation. He said his IT team “put together a controlled data environment” that provided a “safe spot” to upload information they wanted to use in AI. This allowed the team to “play with it” while mitigating the risk of information “getting out into cyberspace.”
Several speakers noted that customer service was a big part of employee training with AI. As healthcare increasingly relies on technology, it’s important to remember that medicine is about helping people – we’ve got to keep the “care in healthcare.”
5. Implement pilot programs first and build on success
All of the speakers described a process that started by choosing high-impact opportunities, focused on a well-defined problem and tested with a pilot program. They learned as they went along and are now building on that initial success for broader implementation:
- Mount Graham Regional Medical Center started with ambient listening in the emergency department and saw good results that turned doctors into advocates. They are now implementing the project campus-wide.
- Reid Health began with ambient listening for clinicians, proved that it was a valuable addition, and rolled it out for the nursing team. The hospital is now considering the next department to expand to, which will “probably” be “physical therapy next.”
- Artesia General Hospital initiated a project focused on administrative tasks, such as pre-registration and insurance verification, observed the progress, and focused on continuous improvement. The hospital is now looking at implementing it for physicians.
An edge for rural healthcare recruiting and retention
Since Intelliworx develops applicant tracking, onboarding, and credentialing software built for small and rural hospitals, we’d be remiss if we didn’t report the impact these speakers said AI was having on recruiting and retention.
These tools are proving to provide relief for provider burnout, and word is getting out. Millar noted that his hospital recently searched for a cardiologist. The “first question” that “every single one of them” asked was “Do you have an ambient listening service?”
That anecdote makes it clear to him that doctors want these tools. Candidate enthusiasm grows “when they find that a small rural health facility has it,” while many of the larger urban facilities do not. “I think for us being rural health, we are a little bit more nimble and able to do things that better serve our providers and our patients and our employees.”
There may well be an opportunity for more rural healthcare facilities to do more zagging.
* * *
See the software for yourself! WorkReady is designed specifically for rural healthcare recruiting: Contact us for a no-obligation demo.
If you enjoyed this post, you might also like:
Lost billings: The financial impact of provider credentialing process in healthcare
