[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-290488-105":59,"doc-detail-290488-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","machine-learning-in-causal-inference-for-epidemiology-methods","Machine learning in causal inference for epidemiology - Methods","","Causal inference often relies on parametric models to estimate the effect of interest, but valid inference depends on correct model specification—a difficult requirement, especially with high-dimensional data. Integrating machine learning can reduce bias from model misspecification because ML avoids imposing a fixed functional form. Directly plugging ML predictions into effect formulas, however, can create plug-in bias. To address this and retain useful asymptotic properties, estimators combining ML prediction with traditional inference are reviewed for epidemiologists: TMLE, AIPW, and DML.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/machine-learning-in-causal-inference-for-epidemiology-methods/290488/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/machine-learning-in-causal-inference-for-epidemiology-methods/290488.png","ImageObject",300,407,{"name":92,"@type":93},"Aria","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-29","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",14,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"Why are parametric models challenging in high-dimensional epidemiological settings?","Question",{"text":113,"@type":114},"They depend on correct model specification, which becomes difficult as dimensionality increases, increasing the risk of biased effect estimates.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"How does machine learning help causal inference, and what is the plug-in bias risk?",{"text":118,"@type":114},"ML can reduce bias from model misspecification because it does not require specifying a functional form. But using ML predictions directly in a pre-defined effect formula can introduce plug-in bias.",{"name":120,"@type":111,"acceptedAnswer":121},"Which three doubly-robust estimators are presented for addressing plug-in bias?",{"text":122,"@type":114},"The article reviews Targeted Maximum Likelihood Estimation (TMLE), Augmented Inverse Probability Weighting (AIPW), and Double/Debiased Machine Learning (DML).","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},290488,1789641814,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":41,"preview_page_count":4},2336464648322,"https://ap-avatar.wpscdn.com/avatar/2200025388227c56fec?_k=1778556882303663488","European Journal of Epidemiology (2024) 39:1097–1108  \n[https://doi.org/10.1007/s10654-024-01](https://doi.org/10.1007/s10654-024-01)173-x  \nMETHODS  \nMachine learning in causal inference for epidemiology  \nChiara Moccia1 · Giovenale Moirano1 · Maja Popovic1 · Costanza Pizzi1 · Piero Fariselli2 · Lorenzo Richiardi1 · Claus Thorn Ekstrøm3 · Milena Maule1  \nReceived: 7 December 2023 / Accepted: 21 October 2024 / Published online: 13 November 2024 © The Author(s) 2024  \nAbstract  \nIn causal inference, parametric models are usually employed to address causal questions estimating the effect of interest. However, parametric models rely on the correct model specification assumption that, if not met, leads to biased effect estimates. Correct model specification is challenging, especially in high-dimensional settings. Incorporating Machine Learning (ML) into causal analyses may reduce the bias arising from model misspecification, since ML methods do not require the specification of a functional form of the relationship between variables. However, when ML predictions are directly plugged in a predefined formula of the effect of interest, there is the risk of introducing a “plug-in bias” in the effect measure. To overcome this problem and to achieve useful asymptotic properties, new estimators that combine the predictive potential of ML and the ability of traditional statistical methods to make inference about population parameters have been proposed. For epidemiologists interested in taking advantage of ML for causal inference investigations, we provide an overview of three estimators that represent the current state-of-art, namely Targeted Maximum Likelihood Estimation (TMLE), Augmented Inverse Probability Weighting (AIPW) and Double/Debiased Machine Learning (DML) .  \nKeywords Machine learning · Causal inference · Targeted learning · Doubly-robustness  \nIntroduction  \nThe advent of advanced technologies and data collection methods has led to an increase in the complexity of modern epidemiological studies, compelling researchers to work with high-dimensional data more frequently. In parallel, the adoption of Machine Learning (ML) techniques has risen, thanks to their ability to learn patterns and relationships from the data, without explicitly programming for every condition.  \nUntil now, ML algorithms in epidemiology have been mostly used to perform prediction tasks, for example in  \n􀀍 Chiara Moccia chiara.moccia@unito.it  \n1 Cancer Epidemiology Unit, Department of Medical Sciences, University of Turin and CPO Piedmont, Via Santena 7, Turin 10126, Italy  \n2 Department of Medical Sciences, University of Turin, Turin, Italy  \n3 Section of Biostatistics, Department of Public Health, University of Copenhagen, Copenhagen, Denmark  \ndisease diagnosis, patient prognosis, or treatment response [1–3] . ML algorithms excel at learning complex patterns from data, allowing analysts to generate accurate predictions based on the available information. The increasing use of ML in epidemiological research has sparked interest in the context of causal inference, where the goal is to draw causal conclusions on a relationship of interest. In this context, researchers aim to define a causal estimand, representing the quantity they seek to estimate, and then establish the assumptions necessary to express it in terms of observed data through a process known as identification. Thereafter, the focus shifts to estimation and inference tasks. A major risk to causal inference when using observational data is the presence of confounding. Common confounding adjustment techniques include multivariable regression models, propensity score methods, and g-methods [4]. All these approaches typically employ parametric models. However, parametric models rely on correct model specification, which can be particularly challenging in the context of high-dimensional data. For example, in genetic epidemiology, researchers often deal with datasets containing information on thou","cbCaiq5Ma3DdGdJd","https://ap.wps.com/l/cbCaiq5Ma3DdGdJd","pdf",1192413,12,"English","# Methods\n## Machine learning and model misspecification\n## Plug-in bias and doubly-robust estimators\n## Targeted Maximum Likelihood Estimation (TMLE)\n## Augmented Inverse Probability Weighting (AIPW)\n## Double/Debiased Machine Learning (DML)\n# Introduction\n## High-dimensional epidemiological data\n## Confounding and causal estimands","[{\"question\":\"Why are parametric models challenging in high-dimensional epidemiological settings?\",\"answer\":\"They depend on correct model specification, which becomes difficult as dimensionality increases, increasing the risk of biased effect estimates.\"},{\"question\":\"How does machine learning help causal inference, and what is the plug-in bias risk?\",\"answer\":\"ML can reduce bias from model misspecification because it does not require specifying a functional form. But using ML predictions directly in a pre-defined effect formula can introduce plug-in bias.\"},{\"question\":\"Which three doubly-robust estimators are presented for addressing plug-in bias?\",\"answer\":\"The article reviews Targeted Maximum Likelihood Estimation (TMLE), Augmented Inverse Probability Weighting (AIPW), and Double/Debiased Machine Learning (DML).\"}]","Machine learning in causal inference for epidemiology - Methods | PDF"]