Fig. 1

Flow diagram presenting the processes used for enrolling patients, grouping for target trial emulation, and propensity score matching. (A) Flow diagram of patient enrollment. (B) Flow diagram of grouping patients with idiopathic pulmonary fibrosis for target trial emulation with a new-user design for antifibrotics. Patients who had not used antifibrotics between the date of idiopathic pulmonary fibrosis (IPF) diagnosis and the date of censoring or death were categorized as unexposed patients, whereas those who started using antifibrotics on the date of IPF diagnosis were categorized as antifibrotic-exposed patients. The intention-to-treat analysis included the unexposed patients and all patients who were assigned to first-line antifibrotic treatment (nintedanib or pirfenidone). Multivariate Fine–Gray models were used in the intention-to-treat analyses of the effects of antifibrotics on mortality. (C) Flow diagram of propensity score matching. In the target trial emulation, propensity score-matched comparisons were made between antifibrotic-exposed and unexposed patients. Propensity scores were calculated using a logistic regression model adjusted for age, sex, cerebrovascular disease, dementia, acquired immunodeficiency syndrome/human immunodeficiency virus, myocardial infarction, renal disease, congestive heart failure, peripheral vascular disease, chronic pulmonary disease, peptic ulcer, liver disease, diabetes mellitus, hemiplegia or paraplegia, venous thromboembolic disease, pulmonary hypertension, long-term oxygen use, and corticosteroid use at baseline