Related Experiment Videos
The predictive value of asthma medications to identify individuals with asthma--a study in German general practices
W Himmel1, E Hummers-Pradier, H Schümann
1Department of General Practice, University of Göttingen, Germany. whimmel@gwdg.de
Background:
The assessment of prescribing performance by aggregated measures mainly developed from automated databases is often helpful for general practitioners. For asthma treatment, the frequently applied ratio of anti-inflammatory to bronchodilator drugs may, however, be misleading if the specificity of a drug for the treatment of asthma, compared with other diseases, is unknown.
Aim:
To test the association of specific drugs with the diagnosis of asthma compared with other diagnoses.
Design Of Study:
Cross-sectional study analysing prescription data from a retrospective chart review.
Setting:
Eight general practices and one community respiratory practice in a town in Northern Germany.
Method:
All patients in the participating practices who received at least one of the 50 asthma drugs most frequently prescribed in Germany within the past 12 weeks were identified. Odds ratios (ORs) with 95% confidence intervals (ClI) were calculated to reveal any association between a specific drug and the diagnosis of asthma. The unit of analysis was the item prescribed.
Results:
Topical betamimetics (e.g salbutamol, fenoterol) were the most often prescribed asthma drugs in the general practices (52.1 ) and in the respiratory practice (57.6%). Inhaled steroids accounted for 15% and 13%; systemic steroids accounted for 10% and 13%, respectively. In the general practices, inhaled betamimetics had a moderate marker function for asthma (OR = 2.0; 95% CI = 1.14-3.58). A fixed oral combination drug of clenbuterol plus ambroxol was a marker drug against asthma (OR = 0.35; 95% CI = 0.20-0.61). In the respiratory practice, the diagnosis of asthma was strongly marked by fixed combinations of cromoglycate plus betamimetics (OR = 29.0; 95% CI = 6.86-122.24) and moderately by inhaled betamimetics (OR = 2.6; 95% CI = 1.28-5.14). In contrast, systemic steroids (OR = 0.24; 95% CI 0.10-0.57) and even inhaled steroids (OR = 0.46; 95% ClI= 0.22-0.96) proved to contradict the diagnosis of asthma.
Conclusion:
Only betamimetics were markers for asthma patients in both types of practices; inhaled steroids, however, were not. Combinations of cromoglycate were markers in the respiratory practice only. Limited specificity of drugs for a disease (e.g asthma) should be taken into account when analysing prescribing data that are not diagnosis linked.