Amsterdam
Medical Data Science

About

The Data Sharing Initiative by the European Society of Intensive Care Medicine (ESICM) and its Data Science Section aims to promote responsible intensive care patient data sharing across Europe. This will facilitate the development of big data based algorithms using artifical intelligence techniques such as machine learning. The ultimate goal is to use these algorithms at the bedside to improve care and outcome for future critically ill patients.

Collabaration with the Dutch Society for Intensive Care Medicine (NVIC), its Research Network (RCCnet) and Amsterdam UMC has led to the production of AmsterdamUMCdb. This is the first freely accessible intensive care database from within the European Union. It contains de-identified health data related to tens of thousands of European intensive care unit admissions, including demographics, vital signs, laboratory tests and medications.

Press release

A common press release by Amsterdam UMC, the European Society of Intensive Care Medicine (ESICM) and the Dutch Society of Intensive Care Medicine was issued on Sunday, November 17th, 2019 at 20.00 CET in multiple languages, co-inciding with coverage on Dutch national TV news.

The English version of this joint press release may be found here.
De Nederlandse versie van dit gezamenlijke persbericht vindt u hier.
Die deutsche Version dieser gemeinsamen Pressemitteilung finden Sie hier.
La version française de ce communiqué de presse commun est disponible ici.
La versión en español de este comunicado de prensa conjunto se puede encontrar aquí.
此新闻稿的中文版本在这里。

This document contains the press release in all available languages.
This document shows a list of available contact persons from supporting organisations.

Database content and technical aspects

AmsterdamUMCdb is a relational database consisting of 7 tables. The current version is v1.0. It is distributed as tab separated files and will be made available at the EASY repository from the Dutch Archiving and Networked Services (DANS). A detailed description of AmsterdamUMCdb and its data is available at our GitHub repository.

Requesting Access

The database, although de-identified, still contains detailed information regarding the clinical care of patients, so must be treated with appropriate care and respect and cannot be shared without permission. This is outlined in the End User Licence. Those seeking to use AmsterdamUMCdb must complete the following formal steps:

1. Complete a required training course.

Valid training courses include the Data or Specimens Only Research (DSOR) course from CITI, the Basic Course for Clinical Investigators (BROK) from NFU or an equivalent course. The DSOR course may be taken free of charge and is also needed to gain access to the MIMIC and eICU intensive care databases from the USA. Further information on taking the DSOR course may be found here. More information on the BROK course may be found here.

2. Fill out and sign the combined Access and End User License form.

You may download the form here. Please note that this license covers use for lawful, non-commercial, scientific research purposes only. However, it is our specific goals to facilitate everyone wanting to contribute to better health for future patients. Therefore, please contact us should you wish to discuss other lawful use cases.

3. Submit your application

Send your form and proof of training course completion to access@amsterdammedicaldatascience.nl. Please allow up to five business days to approve your application. Incomplete applications will not be processed. Upon approval, AmsterdamUMCdb will be made available for download through EASY at DANS: the Netherlands institute for permanent access to digital research resources which may be found here.

Legal, Privacy and Ethical Considerations

AmsterdamUMC fully complies with all applicable European and Dutch national Laws including the General Data Protection Regulation (GDPR). A Data Privacy Impact Analysis has been conducted.

Independent experts led by Prof. Sijbrands from Erasmus MC and the Netherlands Federation of University Medical Centers (NFU) reviewed AmsterdamUMCdb to report on its re-identification risk in the context of the GDPR. They concluded that, taking into account all means reasonably likely to be used, re-identification is not reasonably likely for AmsterdamUMCdb which can therefore be considered as anonymous information in the context of the GDPR.

An independent ethics review was conducted by the group of clincial ethicist Dr. Erwin Kompanje, renowned for their expertise on ethics related to intensive care medicine. They concluded that the use of de-identified medical data in this context is ethically sound given the minimal risk and the ethical principles of duty of easy rescue and charitable contribution and, therefore, that the advantages of responsible data sharing far outweighh potential disadvantages.

National Engagement

The Dutch Society of Intensive Care Medicine (NVIC), and in particular its Research Collaboration in Critical Care Netherlands network (RCCCnet), strongly support this ESICM initiative. Many centers have publicly expressed future intent to share data related to patients treated in their intensive care departments. These centers include UMC Utrecht, Maastricht UMC+, Radboudumc Nijmegen, OLVG Amsterdam, and Erasmus MC Rotterdam.

Patient Involvement

Both the Dutch ICU patient organisation IC Connect and the Dutch Foundation of Family and Patient Centered Care (FCIC) approve AmsterdamUMCdb and responsible data sharing for the benefit of future patients given the privacy protection measures that have been taken.

Acknowledgements

Amsterdam Medical Data Sciences wishes to acknowledge the project leaders for AmsterdamUMCdb, Paul Elbers, MD, PhD and Patrick Thoral, MD and all intensive care professionals from the department of Intensive Care Medicine, led by Prof. dr. Armand Girbes, MD, PhD, EDIC and Hans van der Spoel, MD, EDIC. We also wish to thank our clincial IT specialists Jan Peppink, Ronald Driessen and Dagmar Ouweneel for their invaluable contributions to the production of AmsterdamUMCdb. In addition, we are grateful to Physionet who produced the MIMIC and eICU databases that have been a great source of inspiration.