Article Text
Statistics from Altmetric.com
One problem with large studies that focus on a limited number of patient relevant outcomes endorsed by Tattersfield and Harrison1 is that the mean response to a given drug may hide a wide variability in individual response. This is particularly likely to occur in patients with symptomatic asthma despite treatment with low/moderate doses of inhaled corticosteroids. Studies have shown marked heterogeneity of airway inflammation and disordered airway function in these patients2-4 and there are wide differences in the effects of the candidate drugs on these measures.5 Smaller crossover studies comparing the effects of different treatments in patients who have been extensively characterised are needed to establish whether important heterogeneity of response occurs. Such studies may uncover easily identifiable patient characteristics that predict a response to an individual drug.
authors' reply Patients with asthma clearly vary in their response to drugs and it is possible that smaller detailed studies will allow the identification of patients more likely to respond to a particular drug. For such a test to be useful in practice, however, it would need to be easily accessible, reliable, and have a very high predictive value. History suggests that it is often more practicable for the patient to undergo a trial of a drug, rather than to undergo an indirect assessment which is unlikely to be 100% predictive of the response and hence may result in some patients not receiving a drug from which they would benefit.