Framework for approaching federated causal inference based on real-world observational data sources

Abstract Context Causal inference methods, generally requiring detailed individual information, allow us to evaluate public health interventions in real-world observational settings. When comparing interventions across borders and for this purpose leveraging distributed sensitive individual-level data, a federated approach can be aided. However, reusing such distributed data while maintaining privacy and interoperability, is a challenging endeavor. Methodological framework In the context of the BY-COVID project, a methodological framework to approach federated causal inference was published in October 2023. It is presented as a set of guidelines in the form several consecutive steps, starting with defining a research question, to building a causal model, translating the causal model into data requirements captured within a Common Data Model (CDM), generating synthetic data complying with the requirements captured in the CDM, and developing interoperable analysis scripts which can be distributed for deployment in different federated nodes. The proposed framework was built through implementing and combining existing methodologies and taking into account the principles of legal, organizational, semantic and technical interoperability. Objectives This presentation will provide an overview of the proposed framework, and its demonstration by performing an international comparison of the real-world effectiveness of SARS-CoV-2 primary vaccination campaigns in preventing infections at different federated sites. Challenges encountered (e.g. difficulties in complying with data requirements by linking heterogeneous local data sources) and lessons learned in the course of the demonstrator will be highlighted. Implications We aim to contribute to a public health research field where the rapid assessment of emerging population-health research questions (potentially feeding into evidence-based policy-making) is achievable, in that way contributing to European pandemic preparedness.

Authors

Institutions

Publication Details

Journal
European Journal of Public Health
Published
2024-10-28
DOI
https://doi.org/10.1093/eurpub/ckae144.682
Primary Topic
Bayesian Modeling and Causal Inference
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Framework for approaching federated causal inference based on real-world observational data sources

Nienke Schutte, Nina Van Goethem, Francisco Estupiñán‐Romero, Enrique Bernal et al.
European Journal of Public Health
Bayesian Modeling and Causal Inference
article

Framework for approaching federated causal inference based on real-world observational data sources

Nienke Schutte, Nina Van Goethem, Francisco Estupiñán‐Romero, Enrique Bernal, Marjan Meurisse
article en

Abstract

Abstract Context Causal inference methods, generally requiring detailed individual information, allow us to evaluate public health interventions in real-world observational settings. When comparing interventions across borders and for this purpose leveraging distributed sensitive individual-level data, a federated approach can be aided. However, reusing such distributed data while maintaining privacy and interoperability, is a challenging endeavor. Methodological framework In the context of the BY-COVID project, a methodological framework to approach federated causal inference was published in October 2023. It is presented as a set of guidelines in the form several consecutive steps, starting with defining a research question, to building a causal model, translating the causal model into data requirements captured within a Common Data Model (CDM), generating synthetic data complying with the requirements captured in the CDM, and developing interoperable analysis scripts which can be distributed for deployment in different federated nodes. The proposed framework was built through implementing and combining existing methodologies and taking into account the principles of legal, organizational, semantic and technical interoperability. Objectives This presentation will provide an overview of the proposed framework, and its demonstration by performing an international comparison of the real-world effectiveness of SARS-CoV-2 primary vaccination campaigns in preventing infections at different federated sites. Challenges encountered (e.g. difficulties in complying with data requirements by linking heterogeneous local data sources) and lessons learned in the course of the demonstrator will be highlighted. Implications We aim to contribute to a public health research field where the rapid assessment of emerging population-health research questions (potentially feeding into evidence-based policy-making) is achievable, in that way contributing to European pandemic preparedness.

European Journal of Public HealthVol. 34(Supplement_3)
Sciensano (Belgium) (BE), Instituto Aragonés de Ciencias de la Salud (ES), UCLouvain (BE)
Good health and well-being
Openalex Percentile: Top 83%
Bayesian Modeling and Causal Inference
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.