Comparing New-User Cohort Designs: The Example of Proton Pump Inhibitor Effectiveness in Idiopathic Pulmonary Fibrosis.

In conclusion, the marginal structural model and the conventional prevalent new-user design, both accounting for informative censoring, produced similar results. However, the prevalent new-user design variant based on never-users introduces selection bias and should be avoided. PMID: 33124647 [PubMed - as supplied by publisher]

https://www.ncbi.nlm.nih.gov/pubmed/33124647?dopt=Abstract

Source: Am J Epidemiol - October 30, 2020 Category: Epidemiology Authors: Tran T, Suissa S Tags: Am J Epidemiol Source Type: research

More News: Epidemiology | Proton Pump Inhibitors PPIs | Study