Targeted learning with an undersmoothed LASSO propensity score model for large-scale covariate adjustment in health-care database studies.

Autor:	Wyss R; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, United States., van der Laan M; Division of Biostatistics, School of Public Health, University of California, Berkeley, Berkeley, CA 94720, United States., Gruber S; Putnam Data Sciences, LLC, Cambridge, MA 02139, United States., Shi X; Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, MI 48109, United States., Lee H; Office of Biostatistics, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD 20903, United States., Dutcher SK; Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD 20903, United States., Nelson JC; Kaiser Permanente Washington Health Research Institute, Seattle, WA 98101, United States., Toh S; Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, MA 02215, United States., Russo M; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, United States., Wang SV; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, United States., Desai RJ; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, United States., Lin KJ; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, United States.
Jazyk:	angličtina
Zdroj:	American journal of epidemiology [Am J Epidemiol] 2024 Nov 04; Vol. 193 (11), pp. 1632-1640.
DOI:	10.1093/aje/kwae023
Abstrakt:	Least absolute shrinkage and selection operator (LASSO) regression is widely used for large-scale propensity score (PS) estimation in health-care database studies. In these settings, previous work has shown that undersmoothing (overfitting) LASSO PS models can improve confounding control, but it can also cause problems of nonoverlap in covariate distributions. It remains unclear how to select the degree of undersmoothing when fitting large-scale LASSO PS models to improve confounding control while avoiding issues that can result from reduced covariate overlap. Here, we used simulations to evaluate the performance of using collaborative-controlled targeted learning to data-adaptively select the degree of undersmoothing when fitting large-scale PS models within both singly and doubly robust frameworks to reduce bias in causal estimators. Simulations showed that collaborative learning can data-adaptively select the degree of undersmoothing to reduce bias in estimated treatment effects. Results further showed that when fitting undersmoothed LASSO PS models, the use of cross-fitting was important for avoiding nonoverlap in covariate distributions and reducing bias in causal estimates. (Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2024.)
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu