Asthma quality-of-care measures using administrative data: identifying the optimal denominator

Abstract

INTRODUCTION The Healthcare Effectiveness Data and Information Set (HEDIS) was developed by the National Committee for Quality Assurance and has been applied to many diseases, including asthma.1 The National Committee for Quality Assurance attempts to create and facilitate HEDIS measures that reflect “quality care” that then may be used to identify plans offering such quality care and to compare results among plans. The HEDIS asthma measure identifies patients with “persistent asthma” based on inpatient, outpatient, and pharmacy utilization data and assesses the proportion of such patients dispensed at least 1 controller medication. Problems with this measure include that this administrative definition of persistent asthma may not adequately reflect persistent asthma as defined clinically2,3 and that patients who meet these criteria for appropriate treatment may actually be at increased risk for subsequent emergency hospital care.4,5 Since 2005, 2 years of continuous enrollment and qualification as “persistent asthma” have been required to increase the specificity of the HEDIS denominator.1 However, health plans may prefer a measure that can be applied to members with a single year of continuous enrollment so as to maximize the number of patients who can be evaluated. A Joint Task Force of the American Academy of Allergy, Asthma and Immunology and the American College of Allergy, Asthma and Immunology was convened to study this subject. Working with pharmaceutical company collaborators, a study5 was performed in 4 different commercial insurance data sets to try to identify an improved asthma quality-of-care measure using administrative data. The strategy was to test the relationship of different numerator-denominator combinations of medical and pharmacy claims in year 1 to asthma exacerbations in the subsequent year. A measure associated with reduced exacerbations would presumably be a better measure of quality care than would a measure not associated with reduced exacerbations. The main findings of this previous study5 were as follows: 1. Based on the lowest number of patients, the highest prevalence of controller use, and the highest prevalence of exacerbations in patients not receiving controller medications, the diagnosis plus 4 medications denominator was the most specific of the 1-year denominators tested for persistent asthma. 2. Compared with other numerators tested, the ratio measure (ratio of controllers to total medications 0.5) was most consistently associated with improved outcomes in all age groups with the more specific denominators. However, the previous Task Force report5 did not evaluate close alternative denominators (eg, 1 encounter with an asthma diagnosis plus 3 medications or 2 encounters plus 2 medication dispensings). The purpose of the present study was to test several additional denominators to define the optimal 1-year denominator to be used with the ratio measure. The main characteristic of the denominator being sought was optimal discrimination of the ratio in predicting acute asthma episodes. Other aspects of the denominators that were evaluated included the size of the denominator and the proportion of patients with medication ratios of 0.5 or greater.