Data Capture collects available data from hospital databases or research cohorts.
In addition, clinical data are de-personalised to protect patient privacy.
It should also track consent and trigger the appropriate mechanisms to remove data from the platform in case the patient has removed her consent.
The MIP has been designed to support a very wide range of methods to gather data, however some datasets are easier to include in the MIP if they meet some criteria.
It can be seen as safer and more secure for a hospital to take care of the de-personalisation of its data, especially as this procedure may expose the hospital to ethical and legal questions.
We encourage hospitals to provide de-personalised data to MIP, and we can also help to put in place a technical solution, FedEHR© Anonymizer©, provided by our partner Gnubila.
MIP uses internally in its Data Factory several databases based on the I2B2 schema. If your data has been prepared for data sharing and follows the I2B2 standard, this effort will fasten the inclusion of your data in the platform.
MIP can connect to and query any PACS system compatible with the DICOM standard.
To better protect your clinical PACS, we recommend that you put in place a research PACS that will contain only the selection of medical images available for research.
Provided some development or configuration on our side, we can support the following methods of integrating data into MIP.
We can ingest data extract provided as a set of files.
For MRI data, files can be in the DICOM or NIFTI formats. We support in particular the BIDS organisation of data
For EHR data, files can be in CVS, TSV, JSON or XML formats.
We can customise the data capture to use online research databases with a Web Service API.