Performance measures address high-level patterns and outcomes of care comparing various dimensions of quality and cost across organizations and/or geographic areas. The primary purpose of performance measurement should be to identify opportunities to improve patient care. Performance measures may be used for improvement efforts, value-based payment programs, public reporting, patient decision-making, accountability and/or allocating resources toward identified gaps in community and population health needs.
The American Academy of Family Physicians (AAFP) advocates for the utilization of performance measures consistent with the criteria described below for evaluating and improving health, health care, and cost.
The following criteria shall be used by the AAFP to evaluate the need, quality and acceptability of a performance measure. We urge other stakeholder organizations to use these criteria as well.
Measure Alignment. If similar or competing measures exist, measures should be compared to select the best measure, to be used across programs, unless multiple measures are justified.
Potential for improvement. A significant gap should exist between optimal and current performance and/or a disparity in care across population groups should be demonstrated. The gap or disparity should be amenable to substantial improvement by means of feasible interventions.
Severity and prevalence. The severity of the condition and its prevalence in the population should be significant enough to justify targeting the condition for improvement.
Substantial impact. The measure should be patient-centered, and address a national health priority or hold the potential for substantial impact on the health status, health outcomes, and/or health care experience of individual patients or populations.
Relevant. The measure should be important to physicians and their patients, kept up to date to reflect current knowledge and science and be amenable to evaluation.
Improve value. Measures should have the potential to improve the value of health services for patients, plans, and purchasers of health care. For composite measures, the data required to compute the measure must be readily available, the components must be rationally related and weighted, and the composite must provide added value over the individual component measures, avoid all-or-none scoring, and not create undue burden.
Balanced. Performance measures aimed at controlling costs must not negatively impact quality of care, accurate and timely diagnosis and patient experience.
Outcome-Focused. The aim of the measure should be to improve health outcomes. When intermediate clinical outcomes or processes of care are assessed, the casual pathway to improve patient-oriented outcomes should be strong.
Evidence-Based. The measure should be grounded in science, explicit and reflect the degree of scientific certainty.
Reliability. The measure should be clearly defined, reproducible, and consistent across different practice settings.
Validity. The measure should be consistent with high quality evidence of efficacy and effectiveness, and accurately represents the concept of interest. There is face validity, indicated by obvious appropriateness or agreement by experts. The measure also has construct validity established through research and statistical analysis. The measure should be tested and validated at the level of care to which it is applied.
Risk adjusted. The measure should be risk adjusted, if possible and appropriate, to account for factors beyond a physician or health system's control, such as differences in practice settings, patient preferences, co-morbid conditions, cultural factors, and social determinants of health.
Risk adjustment must not mask disparities in quality or equity of care resulting from implicit or explicit bias due to race, ethnicity, socioeconomic status or gender. It is important to retain accountability for developing systems and processes that strive for continuous quality improvement.
Use of Artificial Intelligence (AI). If AI is used in performance measurement, it should abide by the AAFP's policy, "Ethical Application of Artificial Intelligence in Family Medicine."
Improvement attainable. It should be possible for physicians to improve their performance.
Precisely defined and specified. The measure specifications should include:
Clear attribution. The measure must include a defined, fair and accurate method for establishing accountability for specific patients.
Reasonable resource requirements and cost. The measure should not impose an inappropriate financial or resource burden on those collecting and reporting the data. Any costs of collecting the data and affecting improvements should be justified by impact on patient-oriented outcomes.
Ease of data collection and reporting. Data for measures should be generated during routine care delivery and discretely captured. The AAFP's "Performance Measurement in Value-Based Payment Models for Primary Care (Position Paper) provides a more detailed discussion of the AAFP's position on performance measurement, particularly as it relates to value-based care.
(1998) (December 2024 BOD)