Article
Mathematical & Computational Biology
Siyun Yang, Elizabeth Lorenzi, Georgia Papadogeorgou, Daniel M. Wojdyla, Fan Li, Laine E. Thomas
Summary: This article introduces analytical methods and visualization tools for causal subgroup analysis, including subgroup weighted average treatment effect and overlap weighting method. The proposed methods aim to achieve balance within subgroups and to address the bias-variance tradeoff in SGA. The Connect-S plot is designed for visualizing subgroup covariate balance.
STATISTICS IN MEDICINE
(2021)
Article
Health Care Sciences & Services
Cecile Payet, Stephanie Polazzi, Jean-Francois Obadia, Xavier Armoiry, Jose Labarere, Muriel Rabilloud, Antoine Duclos
Summary: The study evaluated the performance of high-dimensional propensity scores (hdPSs) in surgical comparative effectiveness studies, finding them to be more accurate and balanced compared to traditional propensity scores (PSs). Results showed that hdPS estimates were more consistent with results seen in randomized controlled trials, indicating hdPSs as a promising alternative for controlling indication bias in comparative studies of surgical procedures.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2021)
Article
Health Care Sciences & Services
Siyun Yang, Ruiwen Zhou, Fan Li, Laine E. Thomas
Summary: This study investigates the propensity score weighting method for causal subgroup survival analysis. Two causal estimands are introduced, and corresponding propensity score weighting estimators are provided. The logistic model with subgroup-covariate interactions selected by least absolute shrinkage and selection operator consistently outperforms other propensity score models. Additionally, overlap weighting generally outperforms inverse probability weighting in terms of balance, bias, and variance.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2023)
Article
Health Care Sciences & Services
Nicolas H. Thurin, Jeremy Jove, Regis Lassalle, Magali Rouyer, Stephanie Lamarque, Pauline Bosco-Levy, Corentin Segalas, Sebastian Schneeweiss, Patrick Blin, Cecile Droz-Perroteau
Summary: This study examines how specific medical procedures may affect treatment effect estimation in propensity score-adjusted comparative studies and proposes a solution. The analysis shows that excluding the immediate pre-exposure time can reduce the risk of including potential instrumental variables and bias in the study.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2023)
Article
Chemistry, Multidisciplinary
Teng Ma, Su Yu
Summary: There are biases in recommendation systems, and not considering these biases leads to unfairness to users, items, and platforms. This study addresses the issue of selection bias by utilizing the propensity score. A selection bias propensity score estimation method (SPE) is developed, which accurately estimates the user's choice tendency by considering user and item information. The SPE method is combined with traditional recommendation algorithms, and experiments on a public dataset demonstrate its effectiveness in reducing selection bias and improving the performance of the recommendation model.
APPLIED SCIENCES-BASEL
(2023)
Review
Rheumatology
Ibrahim Almaghlouth, Eleanor Pullenayegum, Dafna D. Gladman, Murray B. Urowitz, Sindhu R. Johnson
Summary: Observational studies provide insights into rheumatic conditions, risk factors, and treatment effects, but may have confounding bias. Propensity score methods can help achieve study group balance and reduce confounding effects in rare disease research.
JOURNAL OF RHEUMATOLOGY
(2021)
Article
Environmental Sciences
Jayanath Ananda, Gamithri Gayana Karunasena, David Pearson
Summary: Household food management behavior changed considerably during the COVID-19 pandemic in Australia. The study found that Australian households reduced food waste by 9% on average during the pandemic and made positive changes in their food behavior, such as using grocery lists and discount purchases.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2023)
Article
Health Care Sciences & Services
Yingrui Yang, Molin Wang
Summary: In epidemiology, incorporating propensity score in the Cox regression model can effectively control for confounding, but determining exposure effect using propensity score remains challenging in situations with moderate to substantial error in exposure measurement. This paper proposes an estimating equation method to correct bias caused by exposure misclassification, providing more accurate estimation of exposure-outcome associations. Simulation studies are conducted to evaluate the performance of the proposed estimators in various settings, with an application to correct bias in estimating the association of PM2.5 levels with lung cancer mortality.
STATISTICAL METHODS IN MEDICAL RESEARCH
(2021)
Article
Orthopedics
Dirk R. Larson, Isabella Zaniletti, David G. Lewallen, Daniel J. Berry, Hilal Maradit Kremers
Summary: Many arthroplasty research studies are based on nonrandomized, retrospective, registry-based cohorts, which may introduce treatment selection bias and confounding. Propensity scores can help balance cohort characteristics and minimize potential bias and confounding in nonrandomized studies. This article explains how propensity scores are created and provides examples of their application in the analysis of nonrandomized studies.
JOURNAL OF ARTHROPLASTY
(2023)
Article
Public, Environmental & Occupational Health
Marissa J. Seamans, Hwanhee Hong, Benjamin Ackerman, Ian Schmid, Elizabeth A. Stuart
Summary: This report examines the generalizability of subgroup effects and investigates bias caused by unmeasured heterogeneity of subgroup effects across sample and population through deriving bias and conducting Monte Carlo simulation. Understanding the generalizability of subgroup effects may lead to increased use of these methods for making externally valid inferences of treatment effects using a study sample.
Article
Oncology
Yifan Li, Sujiao Liang
Summary: This study analyzed the prognostic value of including S-1 in chemotherapy regimens for stage II and III gastric cancer patients. The results showed that including S-1 was beneficial for stage II patients, improving 5-year overall survival and progression-free survival. However, for stage III patients, including S-1 was associated with inferior overall survival and increased risk of recurrence and distant metastases.
Article
Public, Environmental & Occupational Health
Remi Lenain, Julie Boucquemont, Karen Leffondre, Cecile Couchoud, Mathilde Lassalle, Marc Hazzan, Yohann Foucher
Summary: This study used observational data to simulate a clinical trial and evaluate the benefits of kidney transplantation using time-dependent propensity scores. The results showed a significant gain in life expectancy for kidney transplant recipients compared to those on dialysis, supporting transplantation as the best treatment for end-stage renal disease. This simple method could also be considered for estimating the effects of time-dependent exposures.
Article
Health Care Sciences & Services
Florian Chatelet, Benjamin Verillaud, Sylvie Chevret
Summary: This article discusses the implementation of propensity-score (PS) modeling in finding treatment-by-subset interaction for a right-censored outcome based on observational data. The study compares two main strategies for PS estimation and suggests the use of interaction terms to improve the accuracy of treatment effect estimation. The simulation study shows that the across subsets strategy is preferred in small samples, with the inclusion of interaction terms.
BMC MEDICAL RESEARCH METHODOLOGY
(2023)
Article
Public, Environmental & Occupational Health
Til Sturmer, Michael Webster-Clark, Jennifer L. Lund, Richard Wyss, Alan R. Ellis, Mark Lunt, Kenneth J. Rothman, Robert J. Glynn
Summary: This study extends previous simulations on the performance of propensity score (PS) weighting and trimming methods under different settings including unmeasured confounding. The results suggest that unmeasured confounding can lead to biased estimates, and trimming methods such as Sturmer and Walker can help reduce this bias in treatment effect estimation.
AMERICAN JOURNAL OF EPIDEMIOLOGY
(2021)
Article
Mathematical & Computational Biology
Mitchell M. Conover, Kenneth J. Rothman, Til Sturmer, Alan R. Ellis, Charles Poole, Michele Jonsson Funk
Summary: The study found that differential misclassification of a strong predictor of exposure can lead to biased and imprecise estimations with IPTW, but using asymmetric trimming can reduce bias and result in more precise estimations.
STATISTICS IN MEDICINE
(2021)
