High-dimensional multiple imputation for partially observed confounders including natural language processing–derived auxiliary covariates
American Journal of Epidemiology · 2026
DOI: 10.1093/aje/kwaf017 ↗Kurz gefasst
Fachartikel in der Zeitschrift American Journal of Epidemiology, 2026. Verfasst von Janick Weberpals und 19 weiteren Autor:innen. Beteiligte Einrichtung: Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, Uni…. Hauptthemen laut MEDLINE: Natural Language Processing, Acute Kidney Injury, Models, Statistical. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 48 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NHLBI NIH HHS, NLM NIH HHS, NIH HHS.
Abstract (Originalsprache)
Multiple imputation (MI) models can be improved with auxiliary covariates (ACs), but their performance in high-dimensional data remains unclear. We aimed to develop and compare high-dimensional MI (HDMI) methods using structured and natural language processing (NLP)–derived AC in studies with partially observed confounders. We conducted a plasmode simulation with acute kidney injury as outcome and simulated 100 cohorts with a null treatment effect, incorporating creatinine labs, atrial fibrillation (AFib), and other investigator-derived confounders in the outcome generation. Missingness was imposed on creatinine based on creatinine itself and AFib. Different HDMI candidate ACs were created using structured and NLP-derived features, and we mimicked scenarios where AFib was unobserved by omitting it from all analyses. Using the least absolute shrinkage and selection operator, we selected HDMI covariates for MI and propensity score models. The treatment effect was estimated after propensity score matching in MI datasets, and HDMI methods were compared to baseline imputation and complete case analysis. High-dimensional MI using claims data showed the lowest bias (0.072). Combining claims and sentence embeddings led to an improvement in the efficiency with a root mean square error (RMSE) of 0.173 and 94% coverage. Natural language processing–derived AC alone did not outperform baseline MI. High-dimensional MI approaches may decrease bias in studies where confounder missingness depends on unobserved factors.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States.
- Biostatistics Division, Kaiser Permanente Washington Health Research Institute, Seattle, WA, United States.
- Division of General Internal Medicine and Primary Care, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States.
- Department of Population Health Sciences, Duke University School of Medicine, Durham, NC, United States.
- Office of Biostatistics, Center for Drug Evaluation and Research, US Food and Drug Administration, Silver Spring, MD, United States.
- Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, MA, United States.
ORCID-Profile: Janick Weberpals, Pamela A Shaw, Joseph M Plasek, Sudha R Raman, Bradley G Hammill, John G Connolly, Fang Tian, Wei Liu, José J Hernández-Muñoz, Sebastian Schneeweiss, Rishi J Desai
Förderung
- NHLBI NIH HHS
- NLM NIH HHS
- NIH HHS
- Master Agreement
- FDA HHS
- NIH
- Task Order
Zitieren
APA
Weberpals, J., Shaw, P. A., Lin, K. J., Wyss, R., Plasek, J. M., Zhou, L., Ngan, K., DeRamus, T., Raman, S. R., Hammill, B. G., Lee, H., Toh, S., Connolly, J. G., Dandreo, K. J., Tian, F., Liu, W., Li, J., Hernández-Muñoz, J. J., Schneeweiss, S., & Desai, R. J. (2026). High-dimensional multiple imputation for partially observed confounders including natural language processing–derived auxiliary covariates. American Journal of Epidemiology. https://doi.org/10.1093/aje/kwaf017
Vancouver
Weberpals J, Shaw PA, Lin KJ, Wyss R, Plasek JM, Zhou L, et al. High-dimensional multiple imputation for partially observed confounders including natural language processing–derived auxiliary covariates. American Journal of Epidemiology. 2026. doi:10.1093/aje/kwaf017
BibTeX
@article{Weberpals2026,
author = {Weberpals, Janick and Shaw, Pamela A and Lin, Kueiyu Joshua and Wyss, Richard and Plasek, Joseph M and Zhou, Li and Ngan, Kerry and DeRamus, Thomas and Raman, Sudha R and Hammill, Bradley G and Lee, Hana and Toh, Sengwee and Connolly, John G and Dandreo, Kimberly J and Tian, Fang and Liu, Wei and Li, Jie and Hernández-Muñoz, José J and Schneeweiss, Sebastian and Desai, Rishi J},
title = {High-dimensional multiple imputation for partially observed confounders including natural language processing–derived auxiliary covariates},
journal = {American Journal of Epidemiology},
year = {2026},
doi = {10.1093/aje/kwaf017}
}
EUVIMED