Assessing the real-world impact of interstitial lung disease in patients with rheumatoid arthritis on healthcare resources and treatment patterns through natural language processing

Objectives: Interstitial lung disease (ILD) is a serious pulmonary complication and a major cause of morbidity and mortality in several connective tissue diseases, notably rheumatoid arthritis (RA). We developed a study to compare the healthcare resource utilisation in patients with RAILD and RA-non-ILD and establish the impact of ILD diagnosis in the treatment of patients with RA using natural language processing (NLP) of electronic health records (EHRs). Methods: This observational, retrospective, multicentre study analysed unstructured clinical data from EHRs of patients with RA with and without ILD (RAILD and RA-non-ILD) from 9 hospitals in Spain, spanning January 2014 to December 2019. NLP was used to extract and standardise data into Systematized Nomenclature of Medicine - Clinical Terms (SNOMED-CT) terminology. The study aimed to compare healthcare resource utilisation and treatment patterns between the 2 groups, focusing on changes in treatment following ILD diagnosis in patients with RAILD. Results: Among 13,958 patients with RA, 712 (5.1%) were identified as having RA-ILD, with 433 diagnosed during the study period. Patients with RAILD had significantly higher healthcare resource utilisation than patients with RA-non-ILD, particularly in pneumology (60.5% vs 9.0%), internal medicine (37.4% vs 19.9%), and emergency departments (57.3% vs 39.1%). Hospitalisation rates were also higher in patients with RAILD (64.5% vs 31.9%). Prior to ILD diagnosis, tumour necrosis factor inhibitors were the most commonly prescribed treatments, whereas after diagnosis, drugs with different mechanisms, such as rituximab and abatacept, were more frequently initiated. Conclusions: The development of ILD in patients with RA is associated with increased healthcare resource utilisation and different treatment patterns. NLP on EHRs is a valuable tool for offering insights into disease management and healthcare burden in patients with RAILD.