Pharmacy & Therapeutics- August 2008 - (Page 463) Monitoring Asthma Control continued from page 456 Schatz et al. In another study by Schatz et al., claims data were used to evaluate the relationship of potential asthma quality-of-care markers to subsequent emergency hospital care.20 Controller medications and beta agonists were found to be possibly better predictors of emergency hospital care than quality-of-care markers for asthma (as determined by HEDIS measures, such as use of any controller medication). In the total sample of approximately 110,000 patients, the use of one or more controller drugs (odds ratio [OR], 1.35) and four or more controller drugs (OR, 1.98) was associated with an increased risk of emergency hospital care. In contrast, a controller/total asthma medication ratio of more than 0.5 (OR, 0.73) and the dispensing of fewer than six beta-agonist canisters (OR, 0.30) were associated with a decreased risk. On the basis of these findings, the authors concluded that a medication ratio of greater than 0.5 functioned as the best quality-ofcare marker for identifying those patients with persistent asthma. Leone et al. This study also assessed medication use as a predictor of asthma.21 The authors observed that the level of asthma severity, as determined by the intensity of asthma treatment, indicated acute exacerbations requiring hospital admission. The model used in the study might be helpful in targeting subgroups of populations for intensive intervention programs, but an important limitation to this study was the inability of the model to estimate the specificity and sensitivity of the predictive parameter. COMMENT: Although retrospective reviews regarding the types of prescribed therapy and prescription fill rates may be helpful in identifying asthma patients at risk for exacerbations, the sensitivity of this information must be confirmed in larger studies. Ideally, analyzing the prescription history of asthma patients, in addition to medical claims data with and without the inclusion of PRO information, would be beneficial in determining the usefulness of retrospective prescription data. Further complicating this equation is the difficulty in maintaining long-term asthma control because of the unpredictability of disease exacerbations. This difficulty highlights the need for constant monitoring for disease control, as previously mentioned, as well as documenting the need for adequate and consistent use of controller therapy. Stempel et al. In this study, 53% of patients who exhibited asthma control in the first year had a period of uncontrolled asthma in the second and third years.22 At the end of three years, only 27% of patients were considered to be continuously in control of their asthma. Likewise, despite the extensive use of short-acting beta-agonists, oral corticosteroids, and ED visits, fewer than 50% of patients who had uncontrolled asthma had filled prescriptions for any type of controller medication. The use of controller medications over time was greater in the patients who had controlled asthma, but this rate declined to 46% at the end of the study. COMMENT: The addition of pharmacy data to medical claims data can improve the accuracy of predictive modeling, but it does not achieve the degree of precision found with disease conditions having a more predictable response to therapy. care in asthma. Similar to models that employ medical and pharmacy claims data, HEDIS criteria have shown a lack of effectiveness in identifying patients with uncontrolled asthma and in stratifying asthma severity and risk for health care utilization by health plan members. Cababa et al. These authors compared the assessment of asthma severity based on criteria from HEDIS and from the National Heart, Lung, and Blood Institute (NHLBI).23 In this analysis of 896 pediatric patients, using the NHLBI criteria, the authors identified 656 (73%) as having had persistent asthma, based on HEDIS criteria, compared with 338 patients (38%) who did not. Although the HEDIS criteria for persistent asthma were fairly sensitive (0.89), they were not very specific (0.70). For children not consistently using daily controller medications (n = 346), sensitivity was even lower (0.45) but specificity was similar (0.68). As a result, the authors recommended that these findings be interpreted with caution. Berger et al. The findings of Cabana and colleagues were subsequently confirmed in another study that assessed HEDIS measures for the appropriate use of asthma medications among 49,637 health plan members.24 Employing both pharmacy and medical claims, including outpatient, hospitalization, and ED visits, the Berger investigators used HEDIS criteria to identify patients with persistent asthma and their subsequent risk of ED visits and hospitalizations based on controller medication use and adherence to therapy. The investigators found that 45.9% of patients with persistent asthma were not using any type of controller medication, whereas 35.7% were using one class of long-term controller medications and 18.4% were using more than one class of these controller medications. For the following year, however, more than 25% of these patients did not need any asthma medication. Adherence to medication regimens played an important role in predicting the risk for an ED visit or hospitalization: • Patients with low adherence to their controller medication had the highest risk for an ED visit or hospitalization (OR, 1.72 and 2.23, respectively). Low use was defined as needing a single controller medication with less than a 120-day supply. • Patients with moderate or high adherence had the lowest risk of ED visit or hospitalization. Moderate use was defined as needing a single controller with more than a 120day supply but less than a 180-day supply. High use was defined as needing a single controller with a 180-day supply or more. The authors stated that current HEDIS measures of the appropriate use of asthma medications tend to cause patients to be mislabeled as having persistent asthma when in fact they might have intermittent asthma. Although HEDIS measures may show some predictive value for populations, they are not adequate as tools for predicting future disease activity in individuals. COMMENT: The analysis of raw claims data, even when coupled with a chart review that employs HEDIS measures, does not capture reductions in a patient’s functional status resulting from clinical disease. Important quality-of-life problems, such Effectiveness of HEDIS Measures HEDIS measures are commonly used to evaluate quality Vol. 33 No. 8 • August 2008 • P&T® 463
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.