When specifying measures, measure developers should consider whether a similar measure exists for the same condition, process of care, outcome, or care setting. Measure developers should consider harmonization for every measure under development or maintenance throughout the Measure Lifecycle and harmonize measures unless there is a compelling reason for not doing so (e.g., significant risk variation by age, comorbidity, race) that would justify a separate measure. Harmonization standardizes similar measures when their differences do not make them scientifically stronger or more valuable. Harmonization should not result in inferior measures, but in scientifically strong, clinically valuable, evidence-based measures, which are important to persons/families/caregivers. Quality measures should be based on the best way to capture and specify the measure based on the current scientific information and guidelines. Do not assume a CMS consensus-based entity (CBE)-endorsed measure is better than a new measure.

Measure harmonization is important because it reduces duplication and overlap across quality measures. Measure duplication is undesirable because it may result in unnecessary data collection burden and make the processes of measure selection and interpretation less straightforward. The CMS consensus-based entity (CBE) requires consideration of measure harmonization as part of its endorsement processes.

When developing specifications, measure developers should consider various aspects of the measure for potential harmonization. Harmonization often requires close inspection of specification details of the related measures. Harmonizing measure specifications during measure development is more efficient than harmonizing a fully developed and specified measure. The earlier in the process that measure developers identify related or competing measures, the sooner there is resolution to problematic issues.

Harmonization may include comparison and reconciliation of

• Age ranges
• Measurement period
• Allowable values for medical conditions or procedures (e.g., codes, code systems, code lists, descriptions)
• Allowable conditions for inclusion in the denominator (e.g., codes, code systems, code lists, descriptions)
• Denominator or numerator exclusion categories, whether the exclusion is from the denominator or numerator, and whether optional or required
• Calculation algorithm
• Risk adjustment and/or stratification methods

Examples of when the Measure Developer Should Consider Harmonization

• Diabetic Foot & Ankle Care, Peripheral Neuropathy (CMIT Member ID 00199) – Neurological Evaluation (Steward: American Podiatric Medical Association) is a process measure reporting the frequency of peripheral neuropathy evaluations by measured entities; the proposed measure addresses peripheral neuropathy outcomes.
• An existing diabetes measure includes individuals aged 18 to 75. A new process of care measure is based on new clinical practice guidelines that recommending a specific treatment only for individuals aged 65 years and older.
• An existing diabetes measure includes individuals aged 18 to 75. CMS has requested measures for beneficiaries aged 75 years and older.
• Influenza immunization measures exist for many care settings, but the new measure is for a new care setting.
• Readmission rates exist for several conditions, but the new measure is for a different condition.
• A set of new hospital measures may be able to use data elements already in use for existing hospital measures.

If harmonization of the measure can occur with one or more existing measures, then the measure developer should use existing definitions for those attributes. Other resources (e.g., CMS Measures Inventory Tool [CMIT], Measure Authoring Development Integrated Environment [MADiE], and the Electronic Clinical Quality Improvement [eCQI] Resource Center) contain specifications and shared libraries to help identify opportunities for further harmonization. If the measure developer decides not to harmonize measures, they must document the reasons and include any literature used to support this decision.

Some reasons not to harmonize include

• The science, such as clinical practice guidelines, behind the new measure does not support using the same variable(s) found in the existing measure.
• The measures’ intentions vary across programs/payors, which requires the measures to be distinct.
• The measures have differing denominator populations at significantly different risk (i.e., the denominators are risk stratified).