Selecting Quality Measures

The NQMC Web site lets you search for quality measures based on various attributes found in the Template of Measure Attributes. NQMC also provides a feature that allows users to compare two or more measure summaries in a side-by-side comparison. The attributes displayed in the comparison feature (see the Measure Comparison Template) were chosen to address the measure selection considerations discussed below. Note that this guidance is tailored specifically to measures classified in the Clinical Quality domains: process, access, outcome, structure and patient experience. However, many of the same considerations will apply to measures classified into other domains.

Measure Selection

here are many issues to consider when selecting a quality measure. The first step is to identify the measurement purpose and intended use of the measure. Purposes of quality measures include:

• Quality improvement
• Accountability
• Research

Measures may also be used for a combination of these purposes (see Using Quality Measures for more details on the uses of a measure). It is crucial to determine your assessment and improvement priorities in order to select measures that will serve those priorities. In doing so, it is also important to consider and balance the different perspectives and expectations among different stakeholders. For example, a health plan administrator may have to consider priorities of the businesses that purchase the health plan for their employees along with the priorities of the physicians who are members of their network. Businesses may be more concerned with the value or efficiency of health care services whereas physicians may be more concerned with outcomes of care for their patients. The challenge may be to select a set of measures that balance these varying expectations and perspectives. A starting point is to make clear what the varying stakeholders "value and regard as an essential mission of health care. Areas of agreement among these perspectives ought to define the central focus for quality measurement." (1)

Once the goal or intended use of the measure has been determined, the following areas should be considered in selecting an appropriate measure for the desired purpose(s):

• Does the measure possess the desirable attributes of a measure as outlined by the conceptual areas of importance, scientific soundness, and feasibility of a measure? (Refer to NQMC's Desirable Measure Attributes for more detail.)
• What data sources are available? What is the feasibility and expense of collecting additional data? The three most common data sources used are administrative databases (often referred to as "claims" data), medical records including electronic health records, and survey data. Administrative data are usually readily available in electronic form and relatively inexpensive to use whereas medical records data, though richer in detail, are expensive to obtain and require personnel and time-intensive data abstraction methods.

  Refer to the NQMC Template of Measure Attributes Data Source field.

• Does the measure apply to the desired setting of care and to the providers who give care that the user wishes to assess?

  Refer to the NQMC Template of Measure Attributes Measurement Setting field and Professionals Involved in Delivery of Services field.

• Does the measure belong to a domain of measurement that will produce relevant data? For example, an organization wishing to focus on the perceptions of patients should use Patient Experience measures since the information is collected directly from the patient. Data from these measures will provide insight into patients' satisfaction and perception of the care they receive. Likewise, an organization wishing to assess quality issues that affect how health care is delivered should select measures that assess processes of care (i.e., Process measures).

  Refer to the NQMC Template of Measure Attributes Primary Measure Domain field.

• Have considerations been made for comparisons? When selecting a quality measure, it is important to determine an appropriate comparison in order to make reliable assessments of quality. For example, an organization that wishes to assess their performance in relation to prescriptive standards or goals, such as Healthy People 2010, should select measures for which prescriptive standards of comparison are available.

  Refer to the NQMC Template of Measure Attributes Standard of Comparison field.

• Some comparisons, such as national benchmarks, may require the additional consideration of risk or case-mix adjustment of factors that contribute to differences in results but are not related to quality of care. Alternatively, comparisons can be made by stratifying results to examine potential disparities in care among different subgroups.

  Refer to the NQMC Template of Measure Attributes Allowance for Patient or Population Factors field.

Applying Quality Measures

Quality measurement can be used to drive performance improvement. The quality improvement process is often iterative, and therefore measurement may need to be repeated over time. The first measurement can be used to identify problems and to establish baseline results. Subsequent measurements allow an evaluation of the impact of quality improvement efforts and make it possible to monitor and sustain improvement.

Once users identify appropriate quality measures for implementation, complete measure specifications should be obtained from the submitting organization. The measure summaries provided in NQMC are intended to provide an abstraction of the important attributes of a measure incorporating the information provided by the submitting organization(s). NQMC identifies contact information for requesting measure documentation in the Measure Availability field.

Interpreting Quality Measures

Quality measures often require aggregating data about the health care given to many patients and expressing the results as a rate, ratio, frequency distribution, or score for average performance. The measure result is often composed of a number and unit of measure. The number provides the magnitude, and the unit provides a context for interpreting that number. Refer to the NQMC Template of Measure Attributes Scoring field and Interpretation of Score field

The result of a quality measure can be interpreted in comparison to a standard if one is available. Different types of comparisons include: external comparison to similar providers at a single point in time, external comparison to similar providers over time, internal comparison over time (comparing scores before and after quality improvement efforts), and prescriptive standard (e.g., goals set by Healthy People 2010). Well-established standards of care exist for some areas of health care treatment and services. In these cases, it is possible to conclude that a quality problem does or does not exist. Where there is no standard of care, results can be meaningful if compared to set goals such as those set by the organization implementing the measure or to national goals, such as Healthy People 2010. Refer to the NQMC Template of Measure Attributes Standard of Comparison field.

1. McGlynn EA. Six challenges in measuring the quality of health care. Health Aff (Millwood). 1997 May-Jun;16(3):7-21.