Elamur Research⌕
☰
home / papers / 2026 / Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT

Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT

Paper · 2026 · Adam Sutton, Niko Moller-Grell, Thomas Searle et al.

Summary

Converting MIMIC-IV into the OMOP common data model usually depends on enterprise tools such as BigQuery and on complex orchestration. This puts standardised MIMIC data out of reach for many smaller research groups. The authors present dbt_mimic_omop, an open-source ETL pipeline built on dbt-Core and DuckDB. It runs entirely on local, consumer-grade hardware and maps MIMIC-IV v3.1 to OMOP CDM v5.4.

The pipeline follows a bronze/silver/gold medallion design:

  • Bronze: raw files and OHDSI Athena vocabularies are ingested.
  • Silver: data are cleaned and source codes are mapped to standard concepts.
  • Gold: the final OMOP tables are built.

Beyond the structured data, the pipeline loads MIMIC-IV-Note texts into the note table. It also stores MedCAT-generated SNOMED CT annotations from CogStack Dashboards in note_nlp. Chest radiographs from MIMIC-CXR go into the Medical Imaging extension tables.

Compared with an earlier PostgreSQL conversion, the result is more complete and maps concepts better. All three MIMIC modalities become available through one schema. This supports clinical NLP work such as entity linking, cohort selection, and RAG evaluation.

Key points

  • The pipeline produces over 343 million standardised clinical events for 364,627 patients. It finishes in under three hours on a laptop with 16 GB RAM and 8 CPU cores, and is split into 69 modular dbt models.
  • Concept coverage improves over the PostgreSQL solution of Legrand et al. (2025):
    • measurement concepts rise from 0% to about 95%;
    • unit concepts reach 98.6%;
    • drug concepts rise from 86.1% to 96.6%;
    • race reaches 52.23%;
    • ethnicity stays unmapped because suitable standard vocabulary translations are missing.
  • The note_nlp table holds roughly 195.6M MedCAT annotations coded in SNOMED CT. Because generating them is costly, the table is also offered as a ready download.
  • MIMIC-CXR fills two imaging tables:
    • image_occurrence holds 372,209 images, of which 88.8% are linked to visits through temporal matching;
    • image_feature holds 656,697 CheXpert findings, with 100% concept mapping.
  • The code is on GitHub as CogStack/dbt_mimic_omop. Planned extensions include other MIMIC versions and further public datasets.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓King's College LondonLondon · United Kingdom42 works · 1,617 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksHealth Professions · 69 worksHealth Information Management · 50 works

Cite

Cite as Adam Sutton, Niko Moller-Grell, Thomas Searle and Richard Dobson. 2026. Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT. BioNLP 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.bionlp-1.80

@inproceedings{sutton2026fast,
  title = {{Fast, Accurate, and Local Conversion of MIMIC-IV to OMOP with DBT}},
  author = {Adam Sutton, Niko Moller-Grell, Thomas Searle and Richard Dobson},
  booktitle = {BioNLP 2026},
  year = {2026},
  doi = {10.18653/v1/2026.bionlp-1.80},
  url = {https://doi.org/10.18653/v1/2026.bionlp-1.80}
}

References

25 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. All of Us Research Program Investigators. 2019. The “all of us” research program. New England Journal of Medicine, 381(7):668–676. find
  2. Jeffrey S Brown, Aaron B Mendelsohn, Young Hee Nam, and 1 others. 2022. The us food and drug administration sentinel system: a national resource for a learning health system. Journal of the American Medical Informatics Association, 29(12):2191–2200. find
  3. Kenneth Clark, Bruce Vendt, Kirk Smith, John Freymann, Justin Kirby, Paul Koppel, Stephen Moore, Stanley Phillips, David Maffitt, Michael Pringle, Lawrence Tarbox, and Fred Prior. 2013. The cancer imaging archive (TCIA): Maintaining and operating a public information repository. Journal of Digital I… find
  4. DrivenData. 2026. Snomed ct entity linking benchmark. https://www.drivendata.org/benchmarks/310/benchmark-snomed-ct/. Accessed: 2026-02-25. annotation toolkit. Artificial intelligence in medicine, 117:102083. ↗
  5. Melissa Finster, Markus Wenzel, and Elham Taghizadeh. 2025. Common data models and data standards for tabular health data: a systematic review. BMC Medical Informatics and Decision Making, 25(1):422. find
  6. Paul Legrand, Kawsar Noor, Satyam Bhagwanani, and Richard Dobson. 2025. An open-source text-to-sql pipeline for omop-formatted electronic health records. Technical report, King’s College London. Available from https://github.com/red1dwarf/ text-to-omop/tree/main/ETL (accessed 2026-02-20). ↗
  7. Maryam Garza, Guilherme Del Fiol, Jessica Tenenbaum, Anita Walden, and Meredith Nahm Zozus. 2016. Evaluating common data models for use with a longitudinal community registry. Journal of biomedical informatics, 64:333–341. find
  8. Jeremy Irvin, Pranav Rajpurkar, Michael Ko, Yifan Yu, Silviana Ciurea-Ilcus, Christopher Chute, Henrik Marklund, Behzad Haghgoo, Robyn Ball, Katie Shpanskaya, Jayne Seekins, David A Mong, Safwan S Halabi, Jesse K Sandberg, Ricky Jones, David B Larson, Curtis P Langlotz, Bhavik N Patel, Matthew P Lun… find
  9. Alistair Johnson, Tom Pollard, Nathaniel Greenbaum, Matthew Lungren, Chih-ying Deng, Yifan Peng, Zhiyong Lu, Roger Mark, Seth Berkowitz, and Steven Horng. 2019. MIMIC-CXR: A large publicly available database of labeled chest radiographs. find
  10. Niko Möller-Grell, Linglong Qian, and Shihao Shenzhang. 2026. Multimodal ards detection through agentic medgemma workflows. https://www.kaggle.com/competitions/ med-gemma-impact-challenge/writeups/ new-writeup-1770377682785. Kaggle Med-Gemma Impact Challenge write-up. Accessed: 2026-02-25. ↗
  11. Benjamin Moody, Sicheng Hao, Brian Gow, Tom Pollard, Wei Zong, and Roger Mark. 2022. MIMIC-IV Waveform Database. PhysioNet. Version 0.1.0. find
  12. Observational Health Data Sciences and Informatics. 2025. ATHENA – OHDSI standardized vocabularies. find
  13. Observational Health Data Sciences and Informatics. 2026. OMOP Common Data Model. Technical report, OHDSI. Accessed 2026-02-20. find
  14. Alistair Johnson, Tom Pollard, Steven Horng, Leo Anthony Celi, and Roger Mark. 2023a. MIMIC-IV-Note: Deidentified free-text clinical notes. find
  15. Alistair EW Johnson, Lucas Bulgarelli, Lu Shen, Alvin Gayles, Ayad Shammout, Steven Horng, Tom J Pollard, Sicheng Hao, Benjamin Moody, Brian Gow, and 1 others. 2023b. Mimic-iv, a freely accessible electronic health record dataset. Scientific data, 10(1):1.
  16. Alistair EW Johnson, Lucas Bulgarelli, Lu Shen, Alvin Gayles, Ayad Shammout, Steven Horng, Tom J Pollard, Sicheng Hao, Benjamin Moody, Brian Gow, and 1 others. 2023c. Mimic-iv, a freely accessible electronic health record dataset. Scientific data, 10(1):1.
  17. Alistair EW Johnson, Tom J Pollard, Lu Shen, Li-wei H Lehman, Mengling Feng, Mohammad Ghassemi, Benjamin Moody, Peter Szolovits, Leo Anthony Celi, and Roger G Mark. 2016. Mimic-iii, a freely accessible critical care database. Scientific data, 3(160035).
  18. Vipina K Keloth, Juan M Banda, Michael Gurley, Paul M Heider, Georgina Kennedy, Hongfang Liu, Feifan Liu, Timothy Miller, Karthik Natarajan, Olga V Patterson, and 1 others. 2023. Representing and utilizing clinical textual data for real world studies: An ohdsi approach. Journal of Biomedical Informa… find
  19. Zeljko Kraljevic, Thomas Searle, Anthony Shek, Lukasz Roguski, Kawsar Noor, Daniel Bean, Aurelie Mascio, Leilei Zhu, Amos A Folarin, Angus Roberts, and 1 others. 2021. Multi-domain clinical natural language processing with medcat: the medical concept Nicolas Paris, Antoine Lamer, and Adrien Parrot.… find
  20. PCORnet. 2025. Pcornet common data model v7.0 specification. Technical report, Patient-Centered Outcomes Research Network (PCORnet). find
  21. Tom J Pollard, Alistair EW Johnson, Jesse D Raffa, Leo A Celi, Roger G Mark, and Omar Badawi. 2018. The eicu collaborative research database, a freely available multi-center database for critical care research. Scientific data, 5(1):180178. find
  22. Mark Raasveldt and Hannes Mühleisen. 2019. Duckdb: an embeddable analytical database. In Proceedings of the 2019 international conference on management of data, pages 1981–1984. find
  23. Observational Health Data Sciences and Informatics. 2020. Ohdsi/mimic: Etl for transforming mimic-iv to the omop common data model. Technical report, OHDSI. GitHub repository. Accessed 2026-02-20. find
  24. SNOMED International. 2025. SNOMED CT. International Health Terminology Standards Development Organisation (IHTSDO) / SNOMED International, International Edition. find
  25. Adam Sutton, Vlad Dinu, Thomas Searle, and Richard Dobson. 2026. Cogstack dashboards. Accessed: 2026-02-20. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/sutton-2026-fast-accurate-local-conversion-mimic/