David Suarez1, Roger Borràs, Xavier Basagaña. 1. Epidemiology and Assessment Unit, Fundació Parc Tauli, Universitat Autònoma de Barcelona, Sabadell, Spain. david.suarez.lamas@gmail.com
Abstract
BACKGROUND: Marginal structural models were developed to address time-varying confounding in nonrandomized exposure effect studies. It is unclear how estimates from marginal structural models and conventional models might differ in real settings. METHODS: We systematically reviewed the literature on marginal structural models since 2000. RESULTS: Data to compare marginal structural models and conventional models were obtained from 65 papers reporting 164 exposure-outcome associations. In 58 (40%), estimates differed by at least 20%, and in 18 (11%), the 2 techniques resulted in estimates with opposite interpretations. In 88 papers, marginal structural models were used to analyze real data; only 53 (60%) papers reported the use of stabilized inverse-probability weights and only 28 (32%) reported that they verified that the mean of the stabilized inverse-probability weights was close to 1.0. CONCLUSIONS: We found important differences in results from marginal structural models and from conventional models in real studies. Furthermore, reporting of marginal structural models can be improved.
BACKGROUND: Marginal structural models were developed to address time-varying confounding in nonrandomized exposure effect studies. It is unclear how estimates from marginal structural models and conventional models might differ in real settings. METHODS: We systematically reviewed the literature on marginal structural models since 2000. RESULTS: Data to compare marginal structural models and conventional models were obtained from 65 papers reporting 164 exposure-outcome associations. In 58 (40%), estimates differed by at least 20%, and in 18 (11%), the 2 techniques resulted in estimates with opposite interpretations. In 88 papers, marginal structural models were used to analyze real data; only 53 (60%) papers reported the use of stabilized inverse-probability weights and only 28 (32%) reported that they verified that the mean of the stabilized inverse-probability weights was close to 1.0. CONCLUSIONS: We found important differences in results from marginal structural models and from conventional models in real studies. Furthermore, reporting of marginal structural models can be improved.
Authors: Milena A Gianfrancesco; Laura Trupin; Stephen Shiboski; Mark van der Laan; Jonathan Graf; John Imboden; Jinoos Yazdany; Gabriela Schmajuk Journal: J Rheumatol Date: 2018-12-01 Impact factor: 4.666
Authors: Catherine E de Keyser; Maarten J G Leening; Silvana A Romio; J Wouter Jukema; Albert Hofman; M Arfan Ikram; Oscar H Franco; Theo Stijnen; Bruno H Stricker Journal: Eur J Epidemiol Date: 2014-09-12 Impact factor: 8.082
Authors: Jessie K Edwards; Stephen R Cole; Daniel Westreich; Heidi Crane; Joseph J Eron; W Christopher Mathews; Richard Moore; Stephen L Boswell; Catherine R Lesko; Michael J Mugavero Journal: Epidemiology Date: 2015-09 Impact factor: 4.822