Overall Recommendations

Improving Performance Measures

• Performance measures should be constructed so that the credit for achieving the measure is commensurate with the likelihood of benefit to the patient, consistent with the Institute of Medicine definitions of quality. The most credit should be given for achieving goals or clinical actions with large potential benefits in downstream outcomes for the patient (e.g., based on life expectancy, comorbidity, etc.).
• Performance measures should be constructed so that the eligible population reflects the population(s) that will receive the benefit.
• Performance measures should motivate improvements in quality while minimizing problems with patient safety and unintended consequences.
• Performance measures should be improved through the use of clinically detailed data, and the limitations of measures that use only utilization data should be disclosed.
• Performance measures should incorporate, when possible, considerations of patient preferences and patient choice.
• Performance measures should incorporate patient assessments of quality.
• When operationalizing performance measures, differences in quality across special population groups should be assessed so that appropriate quality improvement interventions can be implemented, as necessary.

Return to Contents

Recommendations for Future Research

Participants identified many avenues of possible future research and questions to be tested. The main themes included:

• Test alternate specifications for technical quality measures for their potential to improve risk factor control and motivate or minimize unintended consequences. Measures to be tested should include tightly linked clinical action measures, continuous weighted measures (including those incorporating QALY metrics), longitudinal measures, and dichotomous stringent control measures.
• Work with health plans and health care organizations to develop and test methods to systematically capture detailed clinical data (e.g., pharmacy, laboratory) for incorporation into technical quality measures; test the reliability of measures constructed with these data.
• Test alternate specifications of patient-reported quality measures for their potential to improve risk factor control and motivate or minimize unintended consequences.
• Develop and implement measures that incorporate patient preferences and clinical factors as components of quality assessments. In particular, test methods that incorporate factors such as medication intolerance, frailty, life-expectancy, and comorbidities into risk-factor control assessments of patient preferences and goals.

Return to Contents

Next Steps

Research ideas and potential research projects to test some of the above recommendations are being discussed with the National Committee for Quality Assurance, Kaiser Permanente of Northern California, and the Department of Veterans Affairs, Veterans Health Administration.

Return to Contents
Proceed to Next Section