Danish register data undoubtedly hold great potential for high-quality research and the possibilities to attract funding for such research. Before initiating a new research project on register data, a number of challenges should be acknowledged and overcome, and CIRRAU has successfully guided many AU-affiliated researchers through these challenges.
A note of caution: Inexperienced researchers often expect that the vast amount of information recorded on the Danish population enables fast and easy publishable, high quality scientific papers. This is partly true, but researchers should also acknowledge the administrative work and the processing time involved in obtaining approval for their research projects as well as the data-administrative challenges in generating a consistent longitudinal data set for the population of interest.
CIRRAU offers in-depth experience with data-administrative challenges for most health-related registers and many of Statistics Denmark’s population-based registers with the amount of knowledge increasing steadily.
Read more below about the different challenges and how CIRRAU can help you overcome them.
APPROVAL
Before initiating a research project utilizing individual level information on any Dane, the research project must be approved by relevant Danish authorities. The needed approvals depend on the data sources requested and where the data are analysed. While some data may be accessed at the individual Danish research institution, other data must be accessed through Statistics Denmark’s research platform.
Irrespective of the chosen solution, individual level data must be used exclusively for the project for which approval was granted, and individual level data must be treated in accordance with Danish legislation.
APPLICATION
A typical application for a research project should describe and motivate all individual level data needed; both databases and variable names.
Typically, approvals are needed from the Danish Data Protection Agency, Sundhedsdatastyrelsen and Statistics Denmark, and this process typically takes 6-12 months.
HOW CAN CIRRAU HELP?
Before initiation of the project and access to the individual level data, it may be especially challenging to choose and motivate all variables needed for a specific research project.
CIRRAU has in-depth knowledge and experience regarding these challenges and has supported many researchers to successful approvals for research projects.
Once a research project has been granted and the researcher has access to the individual level data, other challenges may arise.
DATA FOR ADMINISTRATIVE PURPOSES - NOT RESEARCH
All individual-level data were collected for and by administrative systems. Typically there are administrative changes during the study period of interest and data were set up with no intention of using data for other purposes, such as research, later.
Depending on the data sources of interest, one should not underestimate the workload involved in dealing with and accounting for all changes that influence collection of data in order to generate a consistent longitudinal dataset.
DEALING WITH MISSING AND INCONSISTENT INFORMATION
The Danish Civil Registration System contains information on permanent residence including emigrations, immigrations and disappearances. This is recorded in 9 different data sources with (although minor) missing and inconsistent information, where persons may be registered to reside at a specific address in Denmark while also being registered as living abroad.
The residential information available through CIRRAU contains consistent longitudinal information on residence for all Danish residents who lived at least one day in Denmark since 1968, accounting for these issues.
The municipal reform in 2007 reduced the number of municipalities from 276 to 96, and CIRRAU data make it possible to “re-code” municipalities across the reform.
In addition, CIRRAU also offers access to geocoded information nearly complete for the total Danish population from 1978 onwards. These geocode data take into account discontinued, abandoned and newly established addresses.
FROM 100 DATA SETS TO 3 DATABASES
The National Patient Register contains an administrative database containing patient-level information, a database describing diagnoses given at somatic in- and outpatient departments, and procedures performed at somatic in- and outpatients departments.
This information is contained in one database per year, so far yielding 35 administrative databases, 35 diagnostic databases, and 35+ databases on procedures performed. In addition, there are some changes in what types of information are included, as well as inconsistencies due to varying variable lengths between years.
The National Patient Register available through CIRRAU has been transformed from these 100+ data sets into one administrative database covering the whole period, one diagnostic database covering the whole period, and one database on procedures covering the whole period. All the minor inconsistencies in information between years have been corrected. Also, information from DUSAS is included in these 3 databases.