Albany, Georgia-based Phoebe Physician Group, a subsidiary of Phoebe Putney Health System, serves a largely rural, 41-county area where caregivers say patients are finding it increasingly acceptable to miss doctor visits.
PROBLEM
That left the physician group with an overall absentee rate of 12%—more than twice the Medical Group Management Association’s national average of about 5%. In urban markets, for example, provider organizations can pay for taxi coupons; but in south Georgia, that’s not an option. Automated text messages and reminder calls didn’t assist.
Phoebe Physician’s size, the enormous markets of its clinics, and the difficulty of staffing in rural areas only exacerbated the problem. Constant turnover and minimal staff experience led to inconsistent scheduling, double-booking, and variable appointment confirmation practices.
“You think your staff is sending these reminders, but often they aren’t or they’re not doing it effectively,” said Matthew Robertson, chief administrative officer at Phoebe Physician Group. “So we decided to explore how AI technology could make it easier to see more patients while minimizing disruption to providers and improving the patient experience.”
APPLICATION
It all started with a conversation with Berkeley Research Group. Phoebe Physician staff informed BRG of the issue and that text and call reminders were not working. And that the organization needed to remove the human element to free up staff time and ensure their work was actually being done.
“BRG came up with an AI tool, Tool Development and Piloting MelodyMD, which they developed with Trajum ML,” Robertson explained. “The tool uses machine learning to analyze years of patient visit data and predict the likelihood that a given patient will not show up for their visit.
“When new patients are scheduled, MelodyMD communicates with Phoebe Physician’s appointment system to analyze patient absences and automatically creates an adjacent appointment if the likelihood of absence exceeds established thresholds,” he added.
MET THE CHALLENGE
The tool’s creators examined data points to identify those with the strongest correlation to a patient’s likelihood of no-shows. These included patient demographics, provider specialty, appointment turnaround time, previous visit history, and insurance. As data on recent patient visits was added, the creators continued to refine the model.
“One of the key things we worked on over time was making sure that double bookings were limited on a daily basis,” Robertson noted. “That is, making sure that only patients who were likely to not show up were considered for double bookings. The exclusions were then applied to specific clinics and types of appointments.
“As we implemented and tested the model, we made changes to the reminder process to improve communication with patients and ensure our team had enough time to fill newly released appointment slots,” he added.
He added that the AI tool also enabled the organization to measure performance and implement improvements at the following levels:
-
Patient access. Provides regular monitoring of service utilization, number of absences, completed appointments, cancellations, cancellations within 24 hours and rescheduled appointments.
-
Command management. Enables regular monitoring of referral volumes, patient attrition and retention rates, as well as distributor and competitor volumes.
-
Supplier Scorecard. Provides regular monitoring of relative work value units, visit types, assessment and management coding, average number of visits per session, median scheduling days for recent and established patients, absence rates, and payer mix by provider.
-
Physician/provider productivity. Provides regular monitoring of work relative value units, visit type, assessment and management coding, and details of current procedural terminology by provider.
-
Personnel outside the service sector. Ensures regular monitoring of full-time pay, productivity and overtime to ensure staffing is aligned with demand.
RESULTS
From January 2023 to February 2024, Phoebe Physician saw an average escalate of 168 appointments per week. This represents approximately 7,800 additional appointments and $1.4 million in recent net patient revenue.
Although people still miss appointments and rural stereotypes about making medical appointments are still a barrier, double-booking has helped significantly reduce the impact of such incidents, Robertson said.
ADVICE FOR OTHERS
“It’s definitely important to talk to providers early in the process and be transparent about what’s going on,” Robertson advised. “You want to be clear about the problem, the goals, the potential impacts that you expect from the AI technology. When we talked to our 20 primary care providers and told them they were having 84 absences a day, they were shocked. It helped them get the hang of trying out a recent solution.
“You also have to consider health equity,” he continued. “For example, how can you ensure that the AI tool doesn’t introduce implicit bias into appointment scheduling? How can we improve the algorithm to ensure that underinsured and uninsured patients, or those with certain types of insurance, aren’t disproportionately affected by ongoing double-booking, potentially increasing office wait times?”
Of course, the quality of artificial intelligence depends on the quality of the data on which it is based, he added.
“We worked with BRG to organize and organise three years of patient data, which we could then use to develop an effective model,” he concluded.