A secure cloud for medical data
FAU researchers develop cloud architecture for use in healthcare
When patients are treated in hospitals, large amounts of data are created. Until now, it has not been possible to analyse data from free text which could be used to draw conclusions about the efficacy of medications and treatments. Over the past three years, medical IT experts at FAU have developed cloud architecture which makes this possible. In this cloud, patient data is anonymised and processed for research purposes in line with data protection laws. This could make a significant contribution to improving patient care. FAU researchers are now starting work on the follow-up project ‘Klinische Datenintelligenz’ (clinical data intelligence) which will also incorporate medical imaging results and data from gene analyses for medical research.
How can patient data be distributed centrally and made available to researchers while ensuring that patient anonymity is guaranteed? Researchers from FAU’s Chair of Medical Informatics considered this question in the project ‘cloud4health’. Together with Averbis GmbH, Fraunhofer SCAI, TMF and Rhön-Klinikum AG, they have developed a cloud which meets the special requirements which are specific to healthcare.
With the help of cloud4health, researchers can search free texts – such as medical findings or case histories – which contain particularly valuable information for various elements. For example, patients who have been diagnosed with tumours can be filtered out of the collection of data or searches can be made for critical effects of medications that are mentioned in free texts – the cloud makes it possible to analyse the data according to content. It also gives researchers access to data for large patient populations which can be used to analyse a wide range of questions in research, development and health economics in a way which is compatible with data protection laws. Until now, this was only possible for structural data – data which is presented in the same form for each patient.
To carry out the project, the researchers began by working closely with data protection officers and data protection experts in regional government to develop a data protection and security concept which was able to guarantee the protection of sensitive medical data to meet the special requirements found in healthcare. In the following stage, the researchers combined text analysis technologies and cloud computing techniques in scenarios where they could be used for medical and economic purposes and evaluated them.
Using the cloud4health infrastructure, it is possible to recognise the unwanted side effects of medications which have recently entered the market, for example. With the help of the cloud-based analyses, researchers can find out more quickly whether there are any side effects – and if there are, which ones. The data that is gathered in this way allows conclusions to be drawn about suitable dosages, which in turn enables improved treatment.
Follow-up project incorporates medical imaging
The cloud4health projects has set the direction for future technologies which can be used for further big data analyses, as it includes data such as laboratory results and free texts. However, this does not provide a complete picture of patients’ situations. Imaging techniques, such as X-ray, CT or MRT, and gene analyses also contribute to a better understanding of what goes on inside the body. To perform a complete evaluation of patient data, it is therefore logical to also collect and process data from these procedures. The challenge here is being able to process such diverse data in a way which makes useful medical evaluations possible. In addition, the amounts of data produced are so enormous that large memory and computing capacities must be available in order to be able to carry out evaluations which usually take several days in just a few minutes in the future.
To construct such a comprehensive pool of data in line with data protection laws and make it usable for various applications in patient care and research is the goal of the new Klinische Datenintelligenz project which will run for three years. The Chair of Medical Informatics has been collaborating on this project with Siemens AG, Universitätsklinikum Erlangen, the German Research Center for Artificial Intelligence (DFKI), the Institute of Women’s Health (ifg), the Fraunhofer Institute for Integrated Circuits IIS, Charité Berlin and Averbis GmbH since October 2014. The project has been awarded 3.5 million euros in funding by the Federal Ministry of Economic Affairs and Energy as part of the Smart Data programme.
The researchers presented their project results in a public workshop at the lecture theatre at the Medical Valley Center at 10.30 a.m. on Thursday 20 November.
For more information, please see http://www.imi.med.uni-erlangen.de.
Further information:
Dr. Martin Sedlmayr
Phone: +49 9131 8526755
martin.sedlmayr@fau.de