At SUNY, machine learning in OR scheduling enables big wins
SUNY Upstate Medical University in Syracuse, New York, has 35 operating rooms across multiple locations including academic and community facilities. As with most traditional perioperative departments, it was facing three major issues.
First, low OR utilization despite demand for time. SUNY Upstate was not hitting its desired utilization targets and did not have actionable data for department chairs and administration to take the right actions for improvement. At the same time, it had surgeons and departments wanting more access to OR time.
Second, standardizing systemwide decision-making based on objective data. The provider organization needed to unify decision-making in key areas (block policy to allocate time, block release times and areas for operational improvement) across the academic and community settings to ensure shared best practices and goals while maintaining a high level of visibility into decision-making.
Decisions were questioned often because of the lack of objective data that was clearly defined and understood across the various stakeholders.
Third, visibility into the performance metrics that matter. SUNY Upstate had various reporting sources that were in conflict with each other, creating a distrust in data presented during operational and committee meetings.
To solve its problems, SUNY Upstate Medical University turned to LeanTaaS, which markets software that combines lean principles, predictive analytics and machine learning to transform hospital and infusion center operations.
LeanTaaS’ technology, iQueue for Operating Rooms, promised several attractive outcomes, said Dr. William Marx, medical director of perioperative services at SUNY Upstate Medical University.
“LeanTaaS’ technology predicts when blocks will be underutilized and enables surgeon offices to release them sooner than auto-release deadlines, so a large pool of shared open time can be created,” said Marx. “Like OpenTable makes it easy to book restaurant tables, their tools would make it really simple to see open OR time anytime, anywhere, and ask for it.”
This aspect of the tool would help SUNY work on more cases, thereby improving three key metrics: prime time utilization, staffed room utilization and case volume, he added.
"The data availability and transparency to everyone involved in meeting our goals [have] been critical in our collaboration efforts in policy writing and adoption."
Dr. William Marx, SUNY Upstate Medical University
Transparency was another attractive promised outcome for Marx.
“It would provide cloud-based mobile and web tools that showed the right metrics to the right user on demand,” he explained. “This would alleviate the lack of transparency into data between surgeons and operations and also create a single source of truth to eliminate discrepancies in reports arising from competing sources of information. The transparency has been the best feature as we have introduced the product to our leadership and surgical staff.”
Yet another aspect he liked was what the vendor called “actionability.”
“The tools promised ready access to actionable data to department chairs across SUNY Upstate Medical University so we could hold ourselves accountable, measure what’s important and make decisions, not just debate the numbers,” he said. “One specific such metric is what they call ‘collectable time,’ which is far more actionable than ‘block utilization,’ the traditional metric all ORs have used to right-size blocks, which is far less actionable and defensible.”
Marx was also interested in data hygiene.
“Over the years, we knew we had made decisions and used logic in our EHR that had errors – for example, assigned overlapping block times given to different surgeons. We just weren’t sure how to identify them systematically, and LeanTaaS’ implementation process promised to ferret those out,” he noted.
MEETING THE CHALLENGE
Launched in July 2020, iQueue is being used across SUNY Upstate facilities by OR scheduling, clinic scheduling, surgery chairs and surgeons themselves. It integrates easily with the organization’s Epic EHR and provides a cloud-based add-on the provider can access on any browser, either mobile or web.
“We send iQueue a nightly feed from our data warehouse and a real-time feed in HL7, so the integration is relatively light,” Marx said. “iQueue provides a few powerful tools by using this data in real time.”
One of these tools is OpenTable for Open OR Time. This is actually a set of tools that streamline how the OR can advertise open time, enable and encourage clinics to release time ahead of auto releases, and request it 24/7 when needed. This has led to what Marx called a “mentality of plenty,” instead of one of “scarcity,” where everyone felt they were competing for the same limited OR time.
Another tool is deep on-demand analytics.
“There is a comprehensive set of daily refreshed metrics that help us look at over a dozen powerful metrics – utilization rated, volume, trends, opportunities for improvement in delay, and turnover times,” Marx explained. “The tool allows very easy sharing across campuses, improving data transparency and credibility across all stakeholders.”
Surgeons get personalized weekly texts and messages pushed to them to keep them abreast of their performance. SUNY Upstate has been able to shift cases from one campus to another and improve utilization of others. It also has seen how some surgeons have actually had duplicate blocks on the same day. The on-demand analytics has helped to redesign block-scheduling patterns so they coincide with surgeon and OR availability.
Then there is data-driven accountability, Marx said.
“We have been able to adopt a new way of looking at how to measure usage of block time and hold ourselves accountable (‘collectable time’). This has made the conversations in our decision-making bodies like the OR committee a lot more data-driven and actionable,” he said. “There has been strong adoption and engagement from leadership, which continued to expand to all levels: business office, surgeons and MSG schedulers.”
By cultivating a strong partnership between SUNY Leadership and the LeanTasS team, SUNY Upstate has seen continued active user growth and engagement, he added.
“We now have a Tiger Team at SUNY Upstate Medical University, assembled from key stakeholder groups, relying on iQueue to provide real-time data to the team to assess progress and drive improvements in efficiency,” Marx noted. “The data availability and transparency to everyone involved in meeting our goals [have] been critical in our collaboration efforts in policy writing and adoption.”
In just a few months (since the tools went live in early July), SUNY Upstate has been able to see significant success. The key results achieved with iQueue include:
- Absorbed significantly higher case volume during business hours. “We have seen a 3.4% increase in weekly volume of cases done within existing capacity and in business hours,” Marx explained. “This is hugely important for any system, since OR volume is a key driver of hospital performance.”
- Increases usage of OR minutes during business hours. Correspondingly, SUNY Upstate has experienced a 5.5% increase in the number of OR minutes used during business hours. Since business hours are when ORs are staffed, that also means the organization is using its fixed and variable costs better, instead of doing cases into the night. Since costs are fixed, improving OR utilization has helped reduce nonproductive time.
- Increased utilization across the board. SUNY Upstate has seen a positive impact on key utilization metrics, including prime time and staffed room utilization with a 2% and 1% improvement, respectively, in under 3 months.
- Increased “release proactivity.” Before iQueue, offices were releasing time on an average of two days ahead of EHR auto release. With iQueue, the organization increased its block release time to five days, and block release has actually increased to 12 days ahead of auto release across the health system. This is very important, Marx said, because now offices that need time have a lot more runway to plan for cases and get patients access into the OR sooner to get cases done.
- Data/EHR hygiene. Increased visibility of the data stored in the EHR has led to improved EHR hygiene, which was an unexpected benefit, Marx noted.
“As a result of all this, we have been able to collaborate across the system to adopt a systemwide block policy with stretch goals. Our goal is to improve block utilization from around 50% to 70% by January 1, 2021, and we are well on our way to getting there,” he said. “This will be a big forward step for us.”
ADVICE FOR OTHERS
Marx has three pieces of advice for his peers who may be exploring this kind of technology.
“Math works,” he stated. “Predictive analytics- and machine learning-based systems are going to be instrumental in taking healthcare from making decisions based on ‘tribal rules set by committees’ or gut feel supported by ad-hoc analytics to a data-driven approach that makes useful predictions and prescribes the right actions.”
Surgery has to move to adopting more objective data and 21st century tools to make critical operational and clinical decisions, he added.
“Fear of change is natural, but the status quo is worse,” he continued. “We have all seen what has happened to industries like retail, banking, airlines, transportation. Those players that use data win, and others that don’t fall behind.”
Healthcare has to embrace innovation and change, and there are excellent options out there to work with organizations that truly get healthcare problems and will work with executives to solve them, he said. Data transparency has assuaged some of the anxiety that surgeons have about access, he added.
“And COVID-19 can accelerate your thinking and decision-making,” he concluded. “Telemedicine is a great example of how such shocks to the system can accelerate the pace of adopting new tools. Use that to your advantage and take action.
“Because no matter what happens in Washington, reimbursement levels aren’t about to go up, and surgical volumes aren’t going to come down,” he added. “So we all need tools that help us do more with less – increase patient access, improve the patient experience and lower unit delivery costs.”
Without such tools, he said, healthcare executives and clinicians are shooting in the dark and making sub-optimal decisions every day.