Prior Evidence of Reliability & Validity for Measure Data Elements
Measure developers may use prior evidence of reliability of data elements for the data type specified in the measure (e.g., hospital claims) as evidence for those data elements. Prior evidence could include published or unpublished testing
- Including the same data elements
- Using the same data type (e.g., claims, chart abstraction)
- Conducted on a sample (i.e., representative, adequate numbers, and randomly selected).
Use of prior evidence of validity of data elements is acceptable if the measure under development uses the same data elements and data type and obtains a representative sample of sufficient size.
There is no requirement for separate reliability testing of data elements if the measure developer conducted validity testing on the data elements. If using patient scores from an instrument/scale in the measure under development, measure developers can use testing and documentation of the reliability of the scale as evidence of data element reliability. If the measure developer did not conduct validity testing of the data elements, they should use prior evidence of reliability of data elements.