The review of measure denominator exclusions, numerator exclusions, and denominator exceptions should be based on the testing data and should include, at a minimum

• Evidence of sufficient frequency of occurrence of denominator exclusions, numerator exclusions, and denominator exceptions
• Evidence the data elements (e.g., codes) used to identify denominator exclusions, numerator exclusions, and denominator exceptions are valid

Review may also include evidence there is distortion of the measure results without the denominator exclusions, numerator exclusions, and denominator exceptions.

Example

• Evidence that a denominator exclusion distorts a measure may include variability of a denominator exclusion across comparison groups and sensitivity analyses of the measure score with and without the denominator exclusion.

The measure may require additional review when the measure developer allows patient preference or other individual clinical judgment based on unique patient conditions as a denominator exception category. The measure developer should analyze whether the denominator exception will make a major change to measure results. The measure developer must also consider whether patient preference represents a clinical denominator exception to eligibility or whether provider intervention can influence patient preference. Measure developers should always evaluate these measures with and without the denominator exception and include the proportion of denominator exceptions for any group-level tabulations.