Weng W, Gray B, Lipner RS. — American Board of Internal Medicine
Presented: AcademyHealth Annual Research Meeting, June 2011
Research Objective: To understand the relative importance of patient satisfaction and the patient-centered medical home's practice infrastructure measures to the clinical quality of diabetes care.
Study Design: The ABIM Diabetes Practice Improvement Module (PIM) was used to collect medical record data for 10 clinical (intermediate-outcomes and processes) measures. Through counting the number of measures met, a composite score was created ranging from 0 to 10 for each patient. Physicians' patient-centered medical home measure was defined using NCQA's Physician Practice Connections - Readiness Survey (PPC–RS) scale, which focuses on practice infrastructure. The patient satisfaction indicator was calculated as the percentage of patients answering “excellent” to overall diabetes care from the patient survey component of the PIM.
A negative binomial model was used to examine the association of PPC-RS scale and patient satisfaction with the composite quality score. A logit model was utilized to examine the association of PPC-RS scale and patient satisfaction with A1C at goal (<7.0% or <8.0% based on patients), LDL at goal (<100 mg/dl) and blood pressure at goal (<130/80) respectively. All models' standard errors were adjusted for clustering of patients within physicians, and included patients' age, gender, race/ethnicity, factors limiting self-care, physicians' subspecialty indicator, percentage of patients by primary insurance coverage as covariates. To account for the possibility that better patient satisfaction was caused by better practice infrastructure, we first excluded the patient satisfaction indicator to examine the association between the PPC-RS scale and the quality measure (short-model). We then included both the PPC-RS scale and patient satisfaction indicator in our regression to examine the association of patient satisfaction with the quality measure (full-model).
Population Studied: 21,920 patient chart reviews from 944 physicians who completed the Diabetes PIM during January 2008 to March 2010. Each physician also completed a PPC-RS practice infrastructure survey and received on average 23 patient surveys on diabetes care.
Principal Findings: Coefficients of PPC-RS scale and patient satisfaction were significant across all models (P-value < 0.05) except patient satisfaction coefficient in A1C full-model. Comparing short and full models, the PPC-RS scale coefficients in the full model were only slightly smaller than those in short model (e.g., PPC-RS coefficients were 0.45 and 0.42 in LDL short and full models, respectively). Simulations using estimated coefficients indicated that reaching the maximum PPC-RS scale or having all physicians’ patient satisfaction as “excellent” would, on average, increase the composite quality score by 11.3% and 9.5%, respectively, from current levels; increase by 8.3% and 9.2% patients with blood pressure at goal (currently 40.6%); and increase by 4.0% and 5.8% patients with LDL at goal (currently 63.4%).
Conclusions: In terms of potential for quality improvement, both the practice structure measure, as captured by PPC-RS scale, and the patient satisfaction indicator had equally strong associations with the quality of diabetes care. It is important to consider both practice infrastructure and patient satisfaction when considering qualification as a patient-centered medical home.
Implications for Policy, Delivery or Practice: An assessment tool determining qualification as a patient-centered medical home that solely relies on either practice structure or patient satisfaction of care will likely miss important determinants of diabetes care.
For more information about this presentation, please contact Research@abim.org.