It's crucial to understand and quantify what’s driving CMI trends, especially when CMI has declined. With a few clicks, our application quantifies the impact of the DRG mix and the severity capture on CMI (we can do this at all levels of your organization! how cool!).
The DRG mix is less controllable via severity documentation but is typically the key driver of month-to-month fluctuations in your CMI. It is largely determined by the principal diagnoses and operating room procedures of the patients who happen to get admitted.
The medical vs. surgical DRG share is looked at to figure out if either or both are responsible for the CMI trend.
Such limited insight does not really identify the root cause. If you determine that the surgical share went down, it only begs the questions of which types of surgical DRGs, and of what the impact on the overall CMI trend was.
We identify whether each DRG is high or low value relative to your unique patient population. Then, the impact of each DRG can be quantified based on the change in its share during a given time period.
ClinIntell’s CMI Analysis provides immediate quantified insight into the impact of each component on a CMI trend by assessing performance relative to your unique patient population.
Severity capture is impacted by the quality of documentation and coding of CCs and MCCs (severity documentation).
Seeks to assess severity documentation performance by calculating capture rates at the Med/Surg or specific DRG level.
Capture rates will naturally change with the patient mix just like CMI. For example, certain Orthopedic DRG groups simply have lower capture rates. An increase in share of such DRGs can be hard to distinguish from lower severity performance. Furthermore, capture rates do not provide insight into the value of the CC or MCC, which differs for each DRG group.
ClinIntell's Severity CMI metric isolates the severity capture component of CMI but also controls for your unique patient population. Since it is not a percentage, the impact on CMI can be directly quantified.