Can Predictive Analytics Have a Moneyball Effect on Healthcare Staffing and Scheduling?
Remember the 2003 book (and later film adaptation in 2011) by Michael Lewis that brought the hype of big data analysis to the sentimental game of baseball? For many, Moneyball opened their eyes to the possibilities and valuable insights that were waiting to be unlocked by predictive analytics, sparking an obsession with data that has expanded to many industries.
Healthcare organizations have been the latest to join in the analytics craze. When used smartly, predictive analytics can improve the accuracy of numerous types of forecasts. And when you’re considering patient care outcomes, accurate forecasts are invaluable.
In Lewis’ wildly popular book, the romanticized story of how Oakland A’s General Manager Billy Beane and Assistant General Manager and economist Paul DePodesta built a playoff-bound team by valuing statistics over instinct resonated with the masses. It was an inspiring underdog story that allowed the A’s to compete against big-budget teams by using statistics in a way that had not been done before in the game of baseball.
At odds with the traditional, subjective method of scouting players, the solution embraced by Beane and DePodesta was influenced by a school of baseball statistical analysis known as sabermetrics – a type of advanced statistical analysis that crunches data from player performance – which allowed them to see the hidden potential in undervalued players.
Healthcare organizations are beginning to take notice of the hidden value in their data. But while advanced analytics are seeping their way into many areas of healthcare, one that remains untapped is in accurate forecasts of staffing needs. Using historical census data, predictive analytics can help improve staffing problems by accurately aligning staff to meet patient demand weeks in advance of a shift.
Using time series analysis, predictive models are created and validated and continually refined based on what actually happened to adjust to projections going forward. Within 60 days in advance of a shift, the prediction can get within one staff member of what is actually needed 96 percent of the time.
But data isn’t flawless, and algorithms are not magic. It requires a clear-eyed view to filter out emotional responses to the data to avoid errors.
This requires a good amount of trust. For the cautious, control-prone individuals working in healthcare, it’s often a big leap to trust staffing predictions when they feel that no one knows their hospital or department better than them.
Predictive analytics is a tool to be used in combination with extensive knowledge of staffing strategies; it will not solve all of an organization’s problems alone. You need experts to routinely monitor the predictive model and functional leaders within the healthcare organization to make sure the model is being applied as intended.
In Moneyball, the application of sabermetrics didn’t replace the need for scouts, coaches, and good, old-fashioned effort. It is simply a tool for recruiters to use to gain a more objective view of players. And most didn’t accept the new statistical method right away. It took time – proof that the statistics were reliable.
Organizations may be looking to achieve their own Moneyball effect, but they should be knowledgeable about how the process actually works to avoid unrealistic expectations. Predictive analytics is only effective if key stakeholders feed the model reliable data and buy into using the information that it produces. People must trust the predictive model in order to see success. Finding the right partner to provide guidance on predictive analytics and staffing strategies is fundamental to hitting the home run.
The Predictive Analytics in Healthcare 2016 survey was developed by AMN Healthcare and Avantas. It found that accurate forecasting of future patient demand and workforce needs would be very valuable in solving nurse scheduling and staffing problems.