Research has shown that over half of adults who visit a physician do not receive necessary health services. Physician-based quality-of-care measurement initiatives, such as pay-for-performance, have become increasingly common, emphasizing the importance of critically examining methods used to assess quality of care. This research proposal examines diabetes disease-specific quality of care measures and the feasibility of applying advanced psychometric modeling techniques (i.e., Item Response Theory (IRT)) to measure variation between physicians in quality of care and patient outcomes. This new methodology could increase the accuracy and reliability of current scoring procedures, and could also have larger implications for the generalizability and interpretability of provider-level quality measurement. The main objective of this research is to test the feasibility of using an IRT model to determine patient-level diabetes quality-of-care scores, and then use those scores to assess provider quality.
The specific aims of this study are to (1) define a measurement model for quality of diabetes care;(2) compare the current methods of patient-level quality-of-care measurement to a new psychometric modeling technique for a diabetic population;and (3) determine whether a psychometric modeling technique can identify physician specific variations in patient outcomes better than current composite scores. The research design will consist of a cross-sectional and longitudinal secondary data analysis of patient-level primary care process of-care data, specifically focusing on diabetes care management. Data will be abstracted from the Veterans'Health Administration electronic health record system. It will examine patient-level diabetes process-of-care data collected over 2 years. By developing new methods of measuring quality of care, this study has the potential to work toward the Agency for Healthcare Research and Quality's strategic goal of strengthening quality measurement and improvement in health care. In addition to the healthcare-specific applications of quality measurement, assessments of quality-of-care patterns in chronically ill populations have benefits in the larger context of public health, since they can improve chronic disease interventions, as well as patient and provider education initiatives.