LOS Outlier & Excess Days calculations

Clinical Analytics LOS (Length of Stay) Outlier calculation

In order to assess the effect of hospital care on patient outcomes and utilization, it is essential to take into account the differences between looking at all encounters, just inliers, and just outliers. Clinical Analytics flags each encounter considered a length of stay outlier.

NOTE: Clinical Analytics defines an outlier as "any patient with a length of stay greater than two standard deviations from the Nationwide All-Payer Geometric Mean LOS for the MS-DRG that the encounter is grouped into."

For example, the Nationwide All-Payer Geometric Mean LOS for MS-DRG 291 (heart failure and shock with MCC) is 6.2035 days. The standard deviation for it is 5.0423. Therefore, an encounter would be considered an outlier if their LOS is greater than or equal to 16.2881.

Including/Excluding outliers

In Clinical Analytics, you have the capability of including or excluding outliers in two places.

  • When creating your profile, you can include inliers, outliers or both.
    • Add a filter on "LOS Outlier" or LOS Outlier 10 Days"

  • In Details Section, you can view your data based upon ‘Length of Stay Outlier’ as Both, Outliers or Inliers.

Outliers and Length of Stay Benchmark values

Note that all of our benchmarks are based upon a percentile, rather than a mean, so outlier cases have minimal impact on the median and other percentile values. We first calculate the median LOS for each of hospital in the peer group, then rank those hospitals by median LOS. That ordered list is then used to create the percentile values you see in the tool. For example, the median LOS of the hospital at the 75th percentile is the benchmark value you will see in Clinical Anaytics for the 75th percentile of that peer group. Since the individual hospital value is the median patient LOS for that hospital, outlier values do have as much of an impact on the benchmark value as using a mean LOS approach.

Excess Days calculation

The Excess Days (All Patients) measure is an aggregation of the Excess Days value calculated for each encounter.

Excess Days for an encounter = (encounter LOS) - (Nationwide All Payer median LOS for that encounter's MS-DRG when the encounter was loaded into Clinical Analytics)

Excess Days for an encounter will equal 0 if:

  • the encounter's LOS is less than the benchmark
  • the discharge disposition is Missing/Invalid/Other

NOTE: The benchmark LOS for this measure is determined when the encounter is loaded into Clinical Analytics, so you may notice minor discrepancies between the Excess Days value and the LOS Opportunity if the Nationwide All Payer year you're using for LOS is different from the year of benchmark data used when that encounter was loaded.

If an encounter has Excess Days (Excess Days >0), it will be flagged by the Excess Days (flag) measure and will be included in the Excess Days (patients with Days) measure.

There are 3 Excess Days measures:

  • Excess Days (All Patients) measure composite displayed is: (Sum of all Excess Days for all encounters) divided by (number of encounters)
  • Excess Days (Patients with Days) measure composite displayed is: (Sum of all Excess Days for all encounters with Excess Days) divided by (number of encounters with Excess Days)
  • Excess Days (flag) measure composite displayed is: (number of encounters flagged with Excess Days) / (number of encounters)

Excess Costs/Charges measures

Once Clinical Analytics determines the number of excess days, the costs (and charges) incurred on those excess days are summed to populate the excess costs/charges measures.