Self-Reported Overall Adherence and Correct Inhalation Technique Discordance in Chronic Obstructive Pulmonary Disease Population.

Background: Adherence to inhaled medication constitutes a major problem in patients with chronic obstructive pulmonary disease (COPD) globally. However, large studies evaluating adherence in its entirety and capturing a large variety of potentially associated factors are still lacking. Objective: To study both elementary types of adherence to chronic inhaled COPD medication in "real-life" COPD patients and to assess relationships with a wide-ranging spectrum of clinical parameters. Methods: Data from the Czech Multicentre Research Database (CMRD) of COPD, an observational prospective study, were used. Overall adherence (OA) was evaluated with Morisky Medication Adherence Scale (©MMAS-4) and adherence to an application technique (A-ApplT) with the Five Steps Assessment. Mann-Whitney U test, Spearman's correlation, and logistic regression were used to explore relationships between variables. Results: Data of 546 participants (69.6% of all patients from the CMRD) were analyzed. Two-thirds self-reported optimal OA, but only less than one-third demonstrated A-ApplT without any error. OA did not correlate with A-ApplT. Next, better OA was associated with higher education, a higher number of inhalers, a lower rate of exacerbations, poorer lung function, higher degree of upper respiratory tract symptoms (SNOT-22), absence of depressive symptoms, ex-smoking status, regular mouthwash after inhaled corticosteroids (ICS), and flu vaccination. By contrast, better A-ApplT was associated with a lower number of inhalers, better lung function, and regular mouthwash after ICS. Independent predictors of nonoptimal OA included lower degree of education, absence of flu vaccination, anemia, depression, and peptic ulcer history, whereas independent predictors of lower A-ApplT were lower education, absence of regular mouthwash after ICS, and higher COPD Assessment Test score. Conclusions: Parameters associated with OA and A-ApplT differ, and those associated with both adherence domains are sometimes associated inversely. Based on this finding, we understand these as two separate constructs with an overlap.