{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,1,16]],"date-time":"2026-01-16T13:00:10Z","timestamp":1768568410084,"version":"3.49.0"},"posted":{"date-parts":[[2025,9,28]]},"group-title":"Epidemiology","reference-count":24,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,9,28]],"date-time":"2025-09-28T00:00:00Z","timestamp":1759017600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.medrxiv.org\/about\/FAQ#license"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,9,28]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>To develop privacy-enhancing statistical methods for estimation of binary disease risk model association parameters across multiple electronic health record (EHR) sites with heterogeneous selection mechanisms, without sharing raw individual-level data. We illustrate their utility through a cross-biobank analysis of smoking and 97 cancer subtypes using data from the NIH All of Us (AOU) and the Michigan Genomics Initiative (MGI).<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Materials and Methods<\/jats:title>\n                  <jats:p>\n                    Large-scale biobanks often follow heterogeneous recruitment strategies and store data in separate cloud-based platforms, making centralized algorithms infeasible. To address this, we propose two decentralized sequential estimators namely, Sequential Pseudo-likelihood (SPL) and Sequential Augmented Inverse Probability Weighting (SAIPW) that leverage external population-level information to adjust for selection bias, with valid variance estimation. SAIPW additionally protects against misspecification of the selection model using flexible machine learning based auxiliary outcome models. We compare SPL and SAIPW with the existing Sequential Unweighted (SUW) estimator and with centralized and meta learning extensions of IPW and AIPW in simulations under both correctly specified and misspecified selection mechanisms. We apply the methods to harmonized data from MGI (\n                    <jats:italic>n<\/jats:italic>\n                    = 50,935) and AOU (\n                    <jats:italic>n<\/jats:italic>\n                    = 241,563) to estimate smoking-cancer associations.\n                  <\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>In simulations, SUW exhibited substantial bias and poor coverage. SPL and SAIPW yielded unbiased estimates with valid coverage probabilities under correct model specification, with SAIPW remaining robust under selection model misspecification. Both approaches showed no notable efficiency loss relative to centralized methods. Meta-learning methods were efficient for large sites but failed in settings with small cohort sizes and rare outcome prevalence. In real-data analysis, strong associations were consistently identified between smoking and cancers of the lung, bladder, and larynx, aligning with established epidemiological evidence.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>Our framework enables valid, privacy-enhancing inference across EHR cohorts with heterogeneous selection, supporting scalable, decentralized research using real-world data.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2025.09.26.25336642","type":"posted-content","created":{"date-parts":[[2025,9,28]],"date-time":"2025-09-28T18:55:11Z","timestamp":1759085711000},"source":"Crossref","is-referenced-by-count":0,"title":["Privacy-Enhancing Sequential Learning under Heterogeneous Selection Bias in Multi-Site EHR Data"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0967-6755","authenticated-orcid":false,"given":"Ritoban","family":"Kundu","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8566-9552","authenticated-orcid":false,"given":"Xu","family":"Shi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kumar Kshitij","family":"Patel","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8005-7327","authenticated-orcid":false,"given":"Lucila","family":"Ohno-Machado","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3659-1514","authenticated-orcid":false,"given":"Maxwell","family":"Salvatore","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Peter X.K.","family":"Song","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Bhramar","family":"Mukherjee","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2025093010000703000_2025.09.26.25336642v1.1","doi-asserted-by":"publisher","DOI":"10.1377\/hlthaff.2016.1651"},{"key":"2025093010000703000_2025.09.26.25336642v1.2","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMsr1809937"},{"issue":"1","key":"2025093010000703000_2025.09.26.25336642v1.3","doi-asserted-by":"crossref","first-page":"214","DOI":"10.1111\/biom.13400","article-title":"Statistical inference for association studies using electronic health records: handling both selection bias and outcome misclassification","volume":"78","year":"2022","journal-title":"Biometrics"},{"key":"2025093010000703000_2025.09.26.25336642v1.4","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMsr1110507"},{"key":"2025093010000703000_2025.09.26.25336642v1.5","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMp1006114"},{"key":"2025093010000703000_2025.09.26.25336642v1.6","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.2019.1677241"},{"key":"2025093010000703000_2025.09.26.25336642v1.7","doi-asserted-by":"crossref","unstructured":"V. Chernozhukov , D. Chetverikov , M. Demirer , E. Duflo , C. Hansen , W. Newey , and J. Robins . Double\/debiased machine learning for treatment and structural parameters, 2018.","DOI":"10.3386\/w23564"},{"key":"2025093010000703000_2025.09.26.25336642v1.8","first-page":"30","article-title":"Odal: A one-shot distributed algorithm to perform logistic regressions on electronic health records data from multiple clinical sites","volume":"24","year":"2019","journal-title":"In Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing"},{"issue":"3","key":"2025093010000703000_2025.09.26.25336642v1.9","doi-asserted-by":"crossref","first-page":"376","DOI":"10.1093\/jamia\/ocz199","article-title":"Learning from electronic health records across multiple sites: A communication-efficient and privacy-preserving distributed algorithm","volume":"27","year":"2020","journal-title":"Journal of the American Medical Informatics Association"},{"issue":"1","key":"2025093010000703000_2025.09.26.25336642v1.10","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1093\/biomet\/asab007","article-title":"Heterogeneity-aware and communication-efficient distributed statistical inference","volume":"109","year":"2022","journal-title":"Biometrika"},{"key":"2025093010000703000_2025.09.26.25336642v1.11","doi-asserted-by":"publisher","DOI":"10.1136\/eb-2018-102911"},{"key":"2025093010000703000_2025.09.26.25336642v1.12","doi-asserted-by":"crossref","unstructured":"M. Hu , X. Shi , and P. X.-K. Song . Collaborative inference for treatment effect with distributed data-sharing management in multicenter studies. Statistics in Medicine, 2024.","DOI":"10.1002\/sim.10068"},{"key":"2025093010000703000_2025.09.26.25336642v1.13","unstructured":"M. I. Jordan , J. D. Lee , and Y. Yang . Communication-efficient distributed statistical inference. Journal of the American Statistical Association, 2019."},{"issue":"3","key":"2025093010000703000_2025.09.26.25336642v1.14","doi-asserted-by":"crossref","first-page":"606","DOI":"10.1093\/jrsssa\/qnae039","article-title":"A framework for understanding selection bias in real-world healthcare data","volume":"187","year":"2024","journal-title":"Journal of the Royal Statistical Society Series A: Statistics in Society"},{"key":"2025093010000703000_2025.09.26.25336642v1.15","article-title":"A doubly robust method to counteract outcome-dependent selection bias in multi-cohort ehr studies","year":"2024","journal-title":"arXiv preprint"},{"key":"2025093010000703000_2025.09.26.25336642v1.16","doi-asserted-by":"publisher","DOI":"10.1093\/biomet\/asq006"},{"issue":"5","key":"2025093010000703000_2025.09.26.25336642v1.17","doi-asserted-by":"crossref","first-page":"1370","DOI":"10.1080\/03610926.2024.2335543","article-title":"Superpopulation model inference for non probability samples under informative sampling with high-dimensional data","volume":"54","year":"2025","journal-title":"Communications in Statistics-Theory and Methods"},{"issue":"1","key":"2025093010000703000_2025.09.26.25336642v1.18","doi-asserted-by":"crossref","first-page":"69","DOI":"10.1111\/rssb.12352","article-title":"Renewable estimation and incremental inference in generalized linear models with streaming data sets","volume":"82","year":"2020","journal-title":"Journal of the Royal Statistical Society Series B: Statistical Methodology"},{"key":"2025093010000703000_2025.09.26.25336642v1.19","doi-asserted-by":"crossref","unstructured":"L. Luo , M. Risk , and X. Shi . Online causal inference with application to near real-time post-market vaccine safety surveillance. Statistics in Medicine, 2024.","DOI":"10.1002\/sim.10095"},{"issue":"1","key":"2025093010000703000_2025.09.26.25336642v1.20","doi-asserted-by":"crossref","first-page":"205","DOI":"10.1111\/rssa.12275","article-title":"A re-evaluation of fixed effect (s) meta-analysis","volume":"181","year":"2018","journal-title":"Journal of the Royal Statistical Society Series A: Statistics in Society"},{"key":"2025093010000703000_2025.09.26.25336642v1.21","doi-asserted-by":"crossref","unstructured":"M. Salvatore , R. Kundu , X. Shi , C. R. Friese , S. Lee , L. G. Fritsche , A. M. Mondul , D. Hanauer , C. L. Pearce , and B. Mukherjee . To weight or not to weight? the effect of selection bias in 3 large electronic health record-linked biobanks and recommendations for practice. Journal of the American Medical Informatics Association, page ocae098, 2024.","DOI":"10.1093\/jamia\/ocae098"},{"issue":"3","key":"2025093010000703000_2025.09.26.25336642v1.22","first-page":"40","article-title":"meta: An r package for meta-analysis","volume":"7","year":"2007","journal-title":"R news"},{"key":"2025093010000703000_2025.09.26.25336642v1.23","doi-asserted-by":"crossref","unstructured":"M. Zawistowski , L. G. Fritsche , A. Pandit , B. Vanderwerff , S. Patil , E. M. Schmidt , P. Vande-Haar , C. J. Willer , C. M. Brummett , S. Kheterpal , et al. The michigan genomics initiative: a biobank linking genotypes and electronic clinical records in michigan medicine patients. Cell Genomics, 3(2), 2023.","DOI":"10.1016\/j.xgen.2023.100257"},{"key":"2025093010000703000_2025.09.26.25336642v1.24","unstructured":"W. Zhou , M. Kanai , K.-H. H. Wu , H. Rasheed , K. Tsuo , J. B. Hirbo , Y. Wang , A. Bhat-tacharya , H. Zhao , S. Namba , et al. Global biobank meta-analysis initiative: Powering genetic discovery across human disease. Cell genomics, 2(10), 2022."}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.09.26.25336642","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:47:06Z","timestamp":1768510026000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.09.26.25336642"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,28]]},"references-count":24,"URL":"https:\/\/doi.org\/10.1101\/2025.09.26.25336642","relation":{},"subject":[],"published":{"date-parts":[[2025,9,28]]},"subtype":"preprint"}}