+ Site Statistics
References:
54,258,434
Abstracts:
29,560,870
PMIDs:
28,072,757
+ Search Articles
+ Subscribe to Site Feeds
Most Shared
PDF Full Text
+ PDF Full Text
Request PDF Full Text
+ Follow Us
Follow on Facebook
Follow on Twitter
Follow on LinkedIn
+ Translate
+ Recently Requested

Evaluating large-scale propensity score performance through real-world and synthetic data experiments



Evaluating large-scale propensity score performance through real-world and synthetic data experiments



International Journal of Epidemiology 47(6): 2005-2014



Propensity score adjustment is a popular approach for confounding control in observational studies. Reliable frameworks are needed to determine relative propensity score performance in large-scale studies, and to establish optimal propensity score model selection methods. We detail a propensity score evaluation framework that includes synthetic and real-world data experiments. Our synthetic experimental design extends the 'plasmode' framework and simulates survival data under known effect sizes, and our real-world experiments use a set of negative control outcomes with presumed null effect sizes. In reproductions of two published cohort studies, we compare two propensity score estimation methods that contrast in their model selection approach: L1-regularized regression that conducts a penalized likelihood regression, and the 'high-dimensional propensity score' (hdPS) that employs a univariate covariate screen. We evaluate methods on a range of outcome-dependent and outcome-independent metrics. L1-regularization propensity score methods achieve superior model fit, covariate balance and negative control bias reduction compared with the hdPS. Simulation results are mixed and fluctuate with simulation parameters, revealing a limitation of simulation under the proportional hazards framework. Including regularization with the hdPS reduces commonly reported non-convergence issues but has little effect on propensity score performance. L1-regularization incorporates all covariates simultaneously into the propensity score model and offers propensity score performance superior to the hdPS marginal screen.

(PDF emailed within 0-6 h: $19.90)

Accession: 062919361

Download citation: RISBibTeXText

PMID: 29939268

DOI: 10.1093/ije/dyy120


Related references

Usage Of Propensity Score, Instrumental Variable, Or Machine Learning For Real World Data Analysis. Value in Health 18(3): A11-A12, 2015

The paradox between randomized controlled trials and propensity score-matched real-world data. Moving from dissonance to dialog?. Journal of Thoracic and Cardiovascular Surgery: -, 2018

A comparison of overall survival with 40 and 50mg/m 2 pegylated liposomal doxorubicin treatment in patients with recurrent epithelial ovarian cancer: Propensity score-matched analysis of real-world data. Gynecologic Oncology 143(2): 246-251, 2016

Effect of tiotropium on lung function decline in early-stage of chronic obstructive pulmonary disease patients: propensity score-matched analysis of real-world data. International Journal of Chronic Obstructive Pulmonary Disease 10(): 2185-2192, 2016

Two studies in one: A propensity-score-matched comparison of fingolimod versus interferons and glatiramer acetate using real-world data from the independent German studies, PANGAEA and PEARL. Plos One 12(5): E0173353-E0173353, 2017

Impact of postdilatation on performance of bioresorbable vascular scaffolds in patients with acute coronary syndrome compared with everolimus-eluting stents: A propensity score-matched analysis from a multicenter "real-world" registry. Cardiology Journal 23(4): 374-383, 2016

Differences Between Conditional and Marginal Propensity Score Estimates: A Real-World Application. Journal of the American College of Cardiology 70(1): 117-117, 2017

Get the most from your data: a propensity score model comparison on real-life data. International Journal of General Medicine 9(): 123-131, 2016

Reverse engineering large-scale genetic networks: synthetic versus real data. Journal of Genetics 89(1): 73-80, 2010

Improved tomographic reconstruction of large-scale real-world data by filter optimization. Advanced Structural and Chemical Imaging 2(1): 17-17, 2016

Preparing Laboratory and Real-World EEG Data for Large-Scale Analysis: A Containerized Approach. Frontiers in Neuroinformatics 10: 7-7, 2016

Fingolimod vs dimethyl fumarate in multiple sclerosis: A real-world propensity score-matched study. Neurology: -, 2018

Real-World Evidence and Real-World Data for Evaluating Drug Safety and Effectiveness. JAMA: -, 2018

The FARSEEING real-world fall repository: a large-scale collaborative database to collect and share sensor signals from real-world falls. European Review of Aging and Physical Activity 13(): 8-8, 2016

Ergonomic criteria in large-scale engineering design--II. Evaluating and applying requirements in the real world of design. Applied Ergonomics 30(3): 207-221, 1999