HEDIS® Measure Development Process

Developing a measure is a multi-step process. It involves identifying the clinical area to evaluate; conducting an extensive literature review; developing the measure with the appropriate MAP and other panels; vetting it with various stakeholders; and performing a field-test that looks at feasibility, reliability and validity. 

View development process

Desirable Attributes of HEDIS

These attributes were designed to assess measures for comparison among health care systems, not measures for quality improvement (although some may also be valid for the assessment of quality improvement). A “health care system” may be an MCO, a PPO, a physician group practice, a hospital or other institutional setting that provides acute and nonacute care.

NCQA staff and MAP members recognize that no measure will be perfect with respect to each of these attributes, and not all attributes will be equally important for all measures. For example, a measure that requires information not currently available in plans’ information systems may be considered for inclusion in HEDIS if it has other desirable attributes, captures meaningful information on quality of care delivered and will stimulate the improvement of information systems. Similarly, measures relating to access, satisfaction and informed health care choices may require consideration of criteria in addition to those listed.

Whether considering the value of a new measure or the continuing worth of an existing one, we must define what makes a measure desirable. These are some of the questions we ask when developing a measure.

  1. Relevance: Is the information captured important to different groups (e.g., individual consumers, purchasers, physicians, health care systems)? Does the measure evaluate or capture meaningful information about an important feature of health care that will inform health care choices? Will the measure stimulate internal efforts toward quality improvement?
    1. Meaningfulness: What is the significance of the measure to the different groups concerned with health care? Is the measure easily interpreted? Are the results meaningful to target audiences? 
    2. Health importance: What is the prevalence and overall impact of the condition in the U.S. population? What are the significant health care aspects that the measure will address?
    3. Financial importance: What are the financial implications resulting from the actions evaluated by the measure? Does the measure relate to activities that have high financial impact?
    4. Cost effectiveness: What is the cost benefit of implementing the change in the health care system? Does the measure encourage the use of cost-effective activities or discourage the use of activities that have low cost-effectiveness?
    5. Strategic importance: What are the policy implications of implementing the measure? Does it encourage activities that use resources efficiently to maximize health?
    6. Controllability: What impact does the organization have on the condition or disease? What impact does the plan have on the measure?
    7. Variance among systems: Will there be wide variations across systems?
    8. Potential for improvement: How much room is available for plans to improve performance?
  2. Scientific Soundness: Perhaps in no other industry is scientific soundness as important as in health care. While no manufacturer or service provider is allowed to be negligent,
    when was the last time a bank was required to demonstrate that a change was scientifically proven to be better for consumers? And has that change been scientifically studied?
    1. Clinical evidence: What is the strength of the evidence supporting the measure? What guidelines have been published for the condition? What do guidelines say about aspects of the measure? Is there evidence that documents the link between clinical processes and outcomes that the measure addresses?
    2. Reproducible: Does the measure produce the same results when repeated in the same population and setting?
    3. Valid: Does the measure make sense logically and clinically?
    4. Accurate: Does the measure precisely evaluate what is actually happening?
    5. Risk adjustment: Is it appropriate to stratify the measure by age or some other variable?
    6. Comparability of data sources: If different systems use different data sources for a measure, are accuracy, reproducibility and validity affected?
  3. Feasibility: Does the measure have reasonable specifications? If the measure is relevant but does not have reasonable specifications, can the specifications be made feasible?
    1. Precise specification: Does the measure have clear specifications for data sources and methods for data collection and reporting?
    2. Reasonable cost: Does the measure impose an inappropriate burden on health care systems?
    3. Confidentiality: Does the data collection violate accepted standards of member confidentiality?
    4. Logistical feasibility: Is the required data that needs available?
    5. Auditability: Is the measure susceptible to manipulation or “gaming” that would be undetectable in an audit?

For in-depth information about the quality measurement process, please refer to the HEDIS Volume 1: Narrative publication.