Accurate HEDIS calculations reward Payer with Millions in Reimbursements

Healthcare Case Study

The Challenge

HEDIS metrics #1

  • A regional Healthcare Payer was having trouble determining accurate calculations for their Healthcare Effectiveness Data & Information Sets (HEDIS) metrics.
  • Based on the criteria set forth by the Affordable Care Act (ACA), majority of HEDIS metrics are the quality measures that determine Medicare/Medicaid reimbursements. For these metrics, having an accurate calculation could be the difference in payment of tens of millions of dollars.
  • As it stood, the organization’s year-over-year variances were inconsistent – +/- 15% in many cases. Their internal Data Services department lacked the resources to identify and explain these variances, and the vast size of the datasets made analysis extremely challenging.
  • Additionally, there appeared to be repeated shifts in utilization and demographic metrics. This was primarily due to a sharp drop in utilization cause by COVID-19 and an aging population. In essence, the Payer organization was dealing with a moving target.
  • The need to debug systems while uncovering the actual drivers of shifts in metrics is what led to the need for the HEDIS Benchmarking project which our team was brought in to lead.

The Conflict

HEDIS metrics #2

  • To get the project started, we were given extensive amounts of records which represented individual members both included and excluded from each calculation and tasked to identify inconsistencies.
  • When inconsistencies were found, we compared this against the national HEDIS standard and provided recommendation for altering the Extract, Transform & Load (ETL) process to correct the issue in either of the systems.
  • When a variance was identified, the task was modified to tie the changes to overarching demographic shifts. We discovered that this would often explain the variances.

The Solution

HEDIS metrics #3

  • By utilizing R software, we were able to accomplish tasks such as displaying summary statistics, grouping data, and even joining the results.
  • This allowed us to see each individual member and whether their inclusion status changed year-to-year and why.
  • By taking this approach, we we’re able to scale up that effort and go through the entire dataset in the time it would typically take to review just a handful of cases.
  • The R software proved to be the right tool for the job because of its great transparency. When we created a script, we’re essentially making a record of the steps performed in the analysis. These can then be referred to in order to review how steps performed.

The Result

HEDIS metrics #4

  • As a result of our work, we uncovered several issues with the current metric calculations.
  • For metrics that did not include errors, we were able to illustrate how the changes were related to demographic and usage shifts.
  • We were also able to look at the entire data set and found that the missing members had an eligibility date before a certain date. What we discovered from that information is that not enough history was being sent to correctly calculate the metric.
  • Now that we were able to understand the issue, we were able to send our findings to the ETL development team for corrective action.
  • Our findings helped pinpoint and address several ETL issues and could provide insight on how global trends were impacting their metrics.
  • With a consistent calculation now on hand, our clients could trust that they are audit compliant and maximize their Medicare/Medicaid reimbursements.