The Relationship Between Real-World Inhaled Corticosteroid Adherence and Asthma Outcomes: A Multilevel Approach

BACKGROUND: Low inhaled corticosteroid (ICS) adherence is associated with increased asthma burden. This relationship is likely bidirectional, and may vary across adherence stages (initiation, implementation, and persistence). Studies rarely examine reciprocal influences. OBJECTIVE: To investigate the relationship between ICS implementation and asthma-related outcomes over 2 years, considering bidirectionality and temporal sequence. METHODS: Primary care records (1987-2012) from the Optimum Patient Care Research Database, United Kingdom, were used. Eligible patients were 6 years or older and had 3 or more years of continuous registration starting 1 year before ICS initiation (index date), physician-diagnosed asthma, 2 or more ICS and/or short-acting β-agonist prescriptions each follow-up year, and no long-acting β-agonists, leukotriene receptor antagonists, or maintenance oral corticosteroids in the preceding year. ICS implementation (percentage of days covered) and risk domain asthma control (RDAC; no asthma-related hospitalizations, emergency visits, or outpatient visits and no oral corticosteroid or antibiotic prescriptions with evidence of respiratory review) were estimated for each prescription interval (period between 2 successive prescriptions). Multilevel analyses modeled bidirectional relationships between ICS implementation and RDAC (and its components), controlling for sociodemographic and clinical characteristics. RESULTS: In prescription data from 10,472 patients, ICS implementation in the preceding interval did not predict RDAC, but was weakly positively associated with simultaneous RDAC. Being male, non-current smoker, without chronic obstructive pulmonary disease diagnosis, and with fewer than 4 comorbidities significantly increased odds of RDAC. Asthma-related antibiotics and outpatient visits in the same interval and short-acting β-agonist overuse in the preceding and same interval predicted lower ICS implementation. CONCLUSIONS: Patients may adapt their ICS use to their current needs without this impacting later RDAC