Methodology

To create the ratings, CHPI collects claims data from Anthem Blue Cross, Blue Shield of California, UnitedHealthcare, and Medicare. The data, which covers 10 million Californians, indicates the treatments and services that patients received during the measurement period. CHPI maps each patient to the physician that the patient saw the most during the measurement period and runs a set of nationally-endorsed quality measures on the data. Take, for example, the Testing Blood Sugar for People with Diabetes measure. CHPI measures the percentage of a physician’s diabetic patients, aged 18-75, that had the recommended hemoglobin A1c (HbA1c) testing during the measurement period. In some instances, CHPI combines related individual measure scores into a “composite” score. In this example, the Diabetes Care Composite is a simple average of the three diabetes measures, Testing Blood Sugar for People with Diabetes, Testing Kidney Function for People with Diabetes and Treating High Blood Pressure for People with Diabetes. CHPI only reports scores for physicians that have enough data for a reliable result. If a physician does not have enough data for a measure, a score is not reported for that physician.

CHPI converts these scores into star ratings to show how one physician’s scores compare to those of similar physicians. Physicians in the highest 25% receive four stars, and physicians in the lowest 25% receive one star. Star ratings are calculated separately for each measure, so a physician can receive four stars in one measure and one, two, or three stars in another measure.

In addition to rating physicians, CHPI also rates practice sites. CHPI defines a “practice site” as a specific address (e.g. “123 PINE ST STE 200”) at which two or more physicians of the same specialty see patients. The practice site score is a summary of all of the physician scores at the practice site. On this website, not all of the physicians whose scores are included in the practice site score are listed – only the physicians with reliable individual results are listed.

For more information, please see the Rating Methods.