Artikel
Data-Driven Surgery: IMI-EDC – Paving the Way for AI Analysis
Suche in Medline nach
Autoren
-
Thomas M. Pausch
- Allgemein-, Viszeral- und Transplantationschirurgie, Universitätsklinikum Heidelberg, Heidelberg, Germany; Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Max Blumenstock
- Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Ulf Hinz
- Allgemein-, Viszeral- und Transplantationschirurgie, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Christian Niklas
- Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Pascal Fuchs
- Allgemein-, Viszeral- und Transplantationschirurgie, Universitätsklinikum Heidelberg, Heidelberg, Germany; Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Cornelia Lyu
- Allgemein-, Viszeral- und Transplantationschirurgie, Universitätsklinikum Heidelberg, Heidelberg, Germany; Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Matthias Ganzinger
- Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Nelly Zental
- Klinik für Anästhesiologie, Universitätsklinikum Heidelberg, Heidelberg, Germany; Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Tobias Dittrich
- Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany; Klinik für Hämatolologie, Onkologie und Rheumatologie, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Martin Loos
- Allgemein-, Viszeral- und Transplantationschirurgie, Universitätsklinikum Heidelberg, Heidelberg, Germany
-
Martin Dugas
- Institut für Medizinische Informatik, Universitätsklinikum Heidelberg, Heidelberg, Germany
| Veröffentlicht: | 6. September 2024 |
|---|
Gliederung
Text
Background: Modern healthcare relies on sophisticated data analytics, with Electronic Data Capture (EDC) and Electronic Health Records (EHR) systems at the forefront [1]. Furthermore, artificial intelligence (AI) promises fundamental improvements in surgical data science [2]. But most healthcare organizations lack the data infrastructure required to collect the data needed to optimally train algorithms. And simply adding AI applications to a fragmented system will not create sustainable change [3]. We present the pilot study of the EDC/EHR-system we develop, called IMI-EDC (IMI= Institute of Medical Informatics) [4], implemented in the European Pancreas Centre at the Heidelberg University Hospital. IMI-EDC aims to seamlessly integrate clinical data for upcoming AI analyses.
Objective: We assessed IMI-EDC's implementation in pancreatic surgery, focusing on quantity and quality of captured data to enable AI research. We compared patient comfort, data quality, and workflow between IMI-EDC and the previous paper-based system of health records.
Methods: IMI-EDC was used in a period from November 2022 to April 2024 to capture medical history and quality of life survey submissions by patients via tablets. Users' feedback evaluated comfort and satisfaction with change of processes. We compared the first 100 IMI-EDC submissions with 100 paper submissions for data completeness, correctness, and clarity. Data transfer times to a scientific SQL database were compared between eight IMI-EDC and seven paper submissions.
Results: IMI-EDC seamlessly integrated into the clinical workflow. During a pilot phase of 4 months we collected single-source data sets of 495 outpatient visits, followed by 2595 data sets during process automatization within the next 12 months. In addition, the system was extended to the inpatient area including clinical data capture by doctors and nurses plus usage in related disciplines of anesthesiology and radiology. Tablet using patients demonstrated higher comfort and submitted more complete documents (98/100) compared to paper users (66/100). Data transfer to the scientific database was significantly faster with IMI-EDC (4.2 mins) than with paper (9.9 mins). IMI-EDC in combination with data interfaces from our EHR systems proved the further evolution of our data management into a Next Generation Database. This includes fulfillment of FAIR-principles of data science.
Conclusions: IMI-EDC's success in ensuring data completeness and efficiency in automated data capture and clinical data transfer positions it as a vital tool for upcoming AI utilization. It enables single-source “triple-flow” data science into clinical surgery, research, and quality control. Its scalability within the surgical hospital ecosystem offers the promise of streamlined data preparation for advanced AI research.
The authors declare that they have no competing interests.
The authors declare that a positive ethics committee vote has been obtained.
References
- 1.
- Dugas M, Blumenstock M, Dittrich T, Eisenmann U, Feder SC, Fritz-Kebede F, Kessler LJ, Klass M, Knaup P, Lehmann CU, Merzweiler A, Niklas C, Pausch TM, Zental N, Ganzinger M. Next-generation study databases require FAIR, EHR-integrated, and scalable Electronic Data Capture for medical documentation and decision support. npj Digit Med. 2024;7(1):10. DOI: 10.1038/s41746-023-00994-6
- 2.
- Piccialli F, Somma V Di, Giampaolo F, Cuomo S, Fortino G. A survey on deep learning in medicine: Why, how and when? Inf Fusion. 2021;66(July 2020):111–137. DOI: 10.1016/j.inffus.2020.09.006
- 3.
- Panch T, Mattie H, Celi LA. The “inconvenient truth” about AI in healthcare. npj Digit Med. 2019;2(1):4–6. DOI: 10.1038/s41746-019-0155-4.
- 4.
- Ganzinger M, Blumenstock M, Fürstberger A, Greulich L, Kestler HA, Marschollek M, Niklas C, Schneider T, Spreckelsen C, Tute E, Varghese J, Dugas M. Federated electronic data capture (fEDC): Architecture and prototype. J Biomed Inform. 2023;138(January):104280.
