AB019. Longitudinal asthma management profiles: visualisation of patient histories using multiple data sources

Background: Electronic medical records (EMR) offer valuable information for research and clinical case management, but are currently underused. Barriers to EMR use include the limited information on medication use and health outcomes provided by single data sources, the challenge of linking multiple sources, and the lack of methods to integrate all information to reconstitute patients’ complete medical trajectories. The ASTRO-LAB cohort study, assessing the safety of long-acting β-agonists (LABA) in asthma, collected data from direct patient follow-up and healthcare databases, and thus allowed a more comprehensive exploration of medication use patterns, asthma control and exacerbations over time. To develop longitudinal asthma management patient profiles for the ASTROLAB cohort by integrating data on prescription and dispensation events, and patient-reported medication exposure and occurrence of severe asthma exacerbations (SAEx).

Methods: Children and adults with persistent asthma (aged 6–40) were included in France and the United Kingdom (UK) if on a stable therapy pattern of either LABA without inhaled corticosteroids (ICs), ICs without LABAs, LABAs and ICs in separate canisters, or fixed-dose combination (FDC) for ≥6 of 12 baseline months. The main study outcome was occurrence of asthma-related courses of oral corticosteroids, unscheduled medical contacts or death (SAEx). Patients were followed for ≤24 months via 4-monthly computer-assisted telephone interviews (CATIs) inquiring on recent asthma control, medication use, and SAEx, and via monthly text messages. Linkage was performed with prescription data (UK primary care records, physician reports in France) and dispensing data (French health insurance records). Prescription and exposure patterns were described and data visualization plots of longitudinal medication management profiles during follow-up developed for each patient.

Results: At inclusion, of 1,051 patients [(48.3% women, mean (SD) age =21.6 (10.7)], 2.9% were prescribed LABA without ICs, 29.8% ICs without LABA, 10.0% LABA and ICs in separate canisters and 57.4% LABA/ICs fixed-dose combination. During baseline CATI, 15% patients reported non-exposure. Among patients with ≥12 months of follow-up, 36% had ≥1 SAEx (range, 0–18). For individual patients, longitudinal plots allowed identification of the regularity and concordance in prescription and dispensation events, and possible relations with patient-reported drug exposure and SAEx occurrence. Several profiles will be presented to illustrate possible research and clinical applications (e.g., drug safety, tailoring care to patient history).

Conclusions: Prescription, dispensation and patient-reported medication use may give different estimations of drug exposure. Combining routinely-collected data with easily-recordable self-reported data allows a comprehensive overview of asthma management over time. Further improvements of data visualization and applications in clinical practice and research are worth exploring, and would facilitate EMR valorisation.

Keywords: Electronic medical records (EMR); asthma; long-acting β-agonists (LABA); inhaled corticosteroids