This article was produced and financed by Norwegian centre for E-health research

It is very difficult for researchers to obtain data from electronic health records. However, a researcher has now developed a computer program that gives researchers ability to analyse the data, while ensuring that the patient's privacy is protected. (Illustration photo: Colourbox)

Analysing health records without exposing sensitive information

With the aid of a computer program, researchers can now analyse information in the patient's electronic health record without compromising privacy.

Norwegian centre for E-health research

The Norwegian Centre for E-health Research was created 1st January 2016, with the objective of contributing to a common national ICT solution for health and care services.

Two thirds of all Norwegians saw their regular general practitioner at least once in 2015. As a result, they leave a significant amount of data about symptoms and diseases that may be relevant for research. Many common health problems can be combated by research on health data collected at local health services such as general practitioner offices and nursing home care.

Yet, there is little research done on this data, as it is extremely difficult to gain access to this data for privacy reasons. In Norway, only three to four studies are done on this data each year. Whereas in Scotland, which has approximately the same number of inhabitants and a similar national health service structure, around 60 studies are carried out a year.

However, this may change.

Can guarantee privacy

Kassaye Yitbarek Yigzaw currently employed at the Norwegian Centre for E-health Research has created a method that provides researchers with the information they need, while protecting the patients' and healthcare providers' privacy.

Yigzaw has developed a computer program that analyses the records distributed across health institutions, such as general practitioners, while hiding which patients and health professionals have contributed information for the analysis results.

The computer program makes it difficult to learn private information about patients and general practitioners. Therefore, the regular general practitioners can make patients’ records available for research, while guarantee their privacy and confidentiality, explains Yigzaw.

The patient can opt out

From manual and extremely time-consuming data collection with many security holes, we will now have a process that can actually provide real-time statistical data, and it complies with all standards of security.

"We have only used standards that are safety-approved for the use of sensitive information," explains the researcher.

Now he is doing a postdoctoral research on the same subject, in order to further develop the functionality of the software. His research is part of a larger project called Snow, which will further develop digital services for regular general practitioners, while the information they have recorded will be easier to use in clinical research, quality work in health institutions and public health research.

The precondition is, as always, that the patients must approve the use of their own record information for research purposes. It is naturally possible to opt out. In this case, the regular general practitioner can easily remove the relevant record from the data analysis.

Until now, Yigzaw has only run the software in a test environment, and he hopes to have the opportunity to try it out in practise quite soon.

"There are around 1,500 medical offices in the country. If they all use this computer program, we can produce continually updated statistics in the areas in which we want to research," he says.

Easier to carry out surveys

Kassaye has also developed and tested a method that can change the way we collect and analyse digital surveys. The method ensures that the participants' record their answers on their computing devices and their responses are collected and analysed in such a way no respondent’s responses are revealed.

The data collection and analysis are also significantly improved, while the privacy of the participants is now better protected.

"Both methods have applications in many areas. I believe that these tools could also be useful for analysing business data distributed across multiple institutions without sharing sensitive information,"says the researcher.


Kassaye, Y.Y. Towards Practical Privacy-Preserving Distributed Statistical Computation of Health Data. Doctoral Thesis at UiT (2016)

External links

Powered by Labrador CMS