Providers Prefer Manual Nurse Scheduling Over Predictive Analytics
Provider organizations using automated staff and nurse scheduling tools are not leveraging the predictive analytics or machine learning capabilities, KLAS reports.
Source: Thinkstock
By - Despite vendors offering predictive analytics and machine learning to support staff and nurse scheduling, provider organizations prefer manually developing their work schedules, a new KLAS report shows.
The survey of executives, scheduling directors, and other leaders from over 300 provider organizations revealed that only about one-half of respondents using three of the nine vendors with predictive scheduling are leveraging, or trying to leverage, their solution’s predictive analytics function for staff and nurse scheduling.
“Those that have been scheduling for some time feel that human expertise is still more valuable than artificial modeling,” the health IT research firm stated in the report.
Nurse and staff scheduling are daily challenge for hospitals and other healthcare facilities. Common scheduling obstacles include having too few or too many nurses on hand, staff not working enough hours, staff working too much overtime, nurses with skills not appropriate to the needs of a specific group of patients, staff burnout, and staff calling out of work.
Optimizing nurse and staff schedules is vital to running a provider organization and ensuring patients receive the highest quality care. Providers turn to automated scheduling solutions to help directors and managers create schedules that fit their workforce and workload.
READ MORE: Gaining Visibility into Healthcare Workforce Management Cuts Costs
To help provider organizations optimize scheduling, staff and nurse scheduling vendors are increasingly offering more advanced analytics tools as part of their solutions. Predictive analytics and machine learning tools promise to optimize scheduling to minimize labor costs, maximize staff availability, and meet patient demands.
However, the advanced analytics tools are proving to be more work than they are worth, KLAS reported. Few respondents said they were using their scheduling solution’s predictive analytics tool and the few that are reported accuracy issues.
Only one of the nine vendors studied had more than 15 clients using their predictive scheduling function. The vendor was OnShift and the company received just 7.3 out of 9 points for accuracy, which was slightly below the market average of 7.4 points.
Other vendors analyzed received the following marks for predictive scheduling accuracy:
- Schedule360, 8.1 points
- ShiftWizard, 8 points
- GE Healthcare (ShiftSelect), 7.8 points
- Change Healthcare (McKesson), 7.4 points
- Cerner, 7.2 points
- GE Healthcare (Staffing and Scheduling), 7.2 points
- Kronos, 6.4 points
Accuracy and other issues with predictive scheduling may be stemming from a lack of comprehensive data within a provider organization, survey participants noted.
READ MORE: Addressing Productivity, Labor to Bend the Healthcare Cost Curve
“Respondents often point out that predictive scheduling is only as good as the data it consumes and the users who manage it,” the report stated.
Gathering and cleaning the right data for staff and nurse scheduling may be the key to not only adopting predictive analytics and machine learning for scheduling, but saving on labor costs through the advanced analytics tools.
Mercy health system in St. Louis, Missouri saved $4.3 million in 2017 by using a prescriptive and predictive analytics platform as part of their scheduling solution with SAP Health. The platform allowed the health system of 30 acute care hospitals, 11 specialty hospitals, and over 800 practices and outpatient facilities to fill scheduling gaps in a timely manner and standardize scheduling workflows across its facilities.
But provider organizations had to put in some work to realize the benefits of predictive scheduling. Mercy’s System Vice President of Operations Mike Gillen explained to RevCycleIntelligence.com that gathering the right data was key to optimizing nurse scheduling.
“It took us between 60 and 90 days to work through all those processes of getting the right data, and keep in mind that the data came from multiple applications and systems within Mercy,” he said. “Some data was coming from our staffing and scheduling software or staff applications. Some was coming from our HR applications. Other information came from some of the application support systems we have for tracking when people go on vacations, take PTO [paid time off], time off, or FMLA [Family and Medical Leave Act].”
READ MORE: Hospital Utilization Management Can Reduce Denials, Improve Care
Gillen also had to look at how the health system presented data to nurses and scheduling managers to make sure the information was actionable for scheduling.
“It really required us to not only look at the data and the analytics, but also the workflows of how nurse leaders were able to see the information and understand how they could affect changes in schedules to improve our use of core staff and drive down the need for agency and incentives,” he said.
Once Mercy had the right data and scheduling presentation, the predictive analytics tool significantly saved the health system by reducing its dependence on outside staffing agencies to fill their scheduling gaps. The tool also saved the health system time and increased scheduling efficiency, Gillen reported.
At this point, leveraging predictive analytics and machine learning may not be worth the time for many provider organizations. However, as data aggregation and health IT interoperability improves, provider organizations may have the information they need to truly deploy their scheduling solution’s advanced functions.
