Published on March 30, 2019

SNF Utilization Algorithm Developed by Mercy ACO Published in AJAC

Mercy Accountable Care, LLC, Data and Analytics team members Mark E. Lewis, MPH, Sr. for Data and Analytics, and Avery M. Day, MPH, Jr. Data Analyst, created a custom risk-adjusted algorithm to evaluate current and potential candidates for skilled nursing facility partnerships at Mercy Health System.

An article detailing their efforts and results “ACO Use of Case Mix Index to Comprehensively Evaluate Post Acute Care Partners” was published in The American Journal of Accountable Care (AJAC)’s March 2019 issue.

AJAC Abstract

Accountable care organizations (ACOs) continually strive to achieve the triple aim: an enhanced patient experience and improved population health at a decreased cost. One area of opportunity identified by a Philadelphia-area ACO (PAACO) to achieve this goal was to decrease skilled nursing facility (SNF) utilization and cost. The PAACO’s 2017 attributed population had utilized more than 130 facilities despite a group of preferred SNFs having been previously determined.

There is currently no standard risk-adjusted algorithm for rating SNFs. The PAACO thus created a risk-adjusted algorithm that would rank the facilities based on a weighting system of key indicators of quality. Points were awarded or detracted based on expected amount thresholds; the sum of the points was multiplied by the case mix index, which served as a normalizing factor. The results revealed that the group of preferred SNFs was performing worse than all other facilities included in the analysis.

Areas for opportunity were identified for the preferred group, and as a true community partner, the PAACO will work with these facilities to improve their ranking. This approach could not only decrease costs for the PAACO, but also increase the CMS star rating for these facilities.

Team Effort

“First and foremost, I need to say this was a team effort and acknowledge my co-author, Avery, for all of her major contributions. There was a glaring need in the overall evaluative process, and after a bit of investigation, it became apparent that we would need to create our own solution,” explains Mark. “I also thought this would be something fun for the team to work on, which as an additional perk ended up proving to be a great learning experience.

We went through countless iterations, versions, debates of inclusion/exclusion until we arrived on a suitable solution. You might be surprised that with the right environment and application of logic, you might actually ‘enjoy’ math.

We’ve since been fine-tuning and exploring additional utility of the theory. This underlying theory is really what’s published, the basics of which can be applied broadly. Is it perfect, no, but I wanted people to know it could be done and it’s a major step in the right direction, with the end result aimed at better care, health, and well-being of the patient. Whatever we do or work on, we always strive to benefit the patient, and the populations we serve.”