When discussing the changing healthcare landscape, I find the late Clayton Christensen’s examples of the evolution in healthcare – from intuitive to empirical to precision – a helpful shorthand. Empirical methods like randomized controlled trials and peer-reviewed research helped build upon the foundation laid by early physicians who used trial and error as well as community knowledge to treat patients. However, Christensen finds that in the age of information, tailored analytics and massive amounts of data are rapidly becoming more accessible. This change can be found across healthcare and for Xsolis, we have brought the precision mindset to utilization management in a sea change we term Precision Utilization Management (Precision UM).
Xsolis defines Precision UM as “using Xsolis analytics to automate inpatient determinations.” Simply put, it’s a process of focusing clinicians on difficult cases and automating administration out of the mix. With Precision UM, we are centering our efforts around two key themes:
- A significant portion of status determinations can be automated because of the amount of data available to make accurate predictions
- As machine learning and predictive analytics continue to refine our predictions, the portion of cases available for Precision UM increases
These two themes flesh out our product delivery roadmap for Precision UM and aim to support the clinical determinations of utilization management staff while reducing administrative burden.
Authorize Clear Status Determinations
Xsolis uses artificial intelligence and machine learning to identify cases with an extremely high inpatient likelihood and assigns each case a numerical score (the higher the score, the more likely inpatient) known as the Care Level Score™ (CLS™) (which ranges from 0-157).
For Medicare cases: After data analysis and clinical validation, an organization can use Xsolis analytics for inpatient determinations due to the high degree of accuracy with the Care Level Score™ (CLS™) . This process adheres to the Conditions of Participation.
For commercial/MA cases: Based on analysis provided by Xsolis, hospitals alongside participating health plan partners can pinpoint appropriate inpatient authorizations by Care Level Score™ analysis (any cases above the high probability CLS™ threshold are automatically statused as inpatients).
Advance and Optimize Based on Millions of Cases
As Xsolis takes in millions of case files and makes millions of predictions, our impact for your organization increases. This “learning” behind our machine learning has resulted in our team being able to assess more cases more accurately.
The percentage of cases that can be automated varies by the threshold you select and your Care Level Score™ analysis, but organizations typically find accuracy in excess of 98% at Care Level Score™ of 131 and above. For an average health system, the number of cases above that Care Level Score™ can free up your nurses and staff to review and assess cases and patients that aren’t as clinically clear.
While this number will never reach zero – nurses with clinical experience and expertise are a crucial part of the utilization review process – we expect that the volume of cases we can impact will continue to grow, ensuring that those nurses and physicians are able to focus only on the unclear cases, reducing the time spent on administrative work across the board.