REG study: Real-life, longitudinal ICS adherence patterns in a UK asthma population

Abstract

Background: Inhaled corticosteroid (ICS) adherence is often suboptimal. Adherence can be estimated using electronic medical records, but studies seldom consider changes over time or distinguish between ICS persistence (duration of use) and implementation (used versus prescribed). Aim: To better understand real-life ICS adherence patterns and how they vary over time within and between patients. Methods: Records were extracted from the Optimum Patient Care Research Database for patients initiating ICS at an index date (ID). Eligible patients (≥6yrs) had ≥3-yrs9 continuous data (1 baseline; 2 outcome); asthma diagnosed ≥1yr before ID, & ≥2 separate ICS and/or SABA prescriptions in each outcome yr. ICS persistence & implementation were estimated over consecutive 6-month intervals. Treatment episodes (persistent use) were calculated in each interval and implementation (MPR) for each episode then examined longitudinally. Results: 13,922 patients received 112,984 ICS prescriptions over 2yrs. For the subgroup of patients with ≥1 ICS prescription in all intervals (52.1%): median(IQR) treatment episode duration was 148(104)days; mean(±SD) MPR was high in the 0-6mnth interval (89.02±0.24%) and decreased on average 0.89±0.08% over the successive intervals, and substantial intra- & interpersonal MPR variance was found (64% & 36%, respectively). Conclusions: Defining ICS adherence as implementation during treatment episodes accounts for non-persistence and results in higher MPRs than typically reported. Consideration must be given to the appropriate definition of adherence when designing adherence studies. The variation in MPR detected over time may allow causal modelling of adherence determinants and effects.