In this joint research project with the Association of Finnish Local and Regional Authorities and Aalto University, we analysed Finnish Regional Social and health care systems based on national registry data. Our purpose was to predict changes in long-term care states of elderly people. The data was gathered from several Finnish national-level registries. Gathering comprehensive national-level data, and creating possibilities for its research use, creates possibilities for regional social and health care officials to develop better local health policies. We utilized iterative statistical analysis methods, which were based on co-creation workshops with regional social and health-care authorities.
The first stage of the process was to gather permission for the research use of data from the Finnish social and health care registries. This phase required comprehensive knowledge of the contents of each registry, which was achieved through information provided by the authorities and direct communication with registry keepers. This phase took a relatively long time, including writing a clear description of the project plan, use, storage and disposal of data and the personnel handling the data. Also, the specification of the datasets and required information took several weeks.
The second stage was obtaining the data, through the authority that pseudonymised the data so personal identification numbers of patients could not be retrieved. The timeframe for this phase was several weeks, after getting the permission for the data from each authority.
The results from the initial statistical analyses were discussed in a workshop with Finnish municipal social and health care leaders. Based on previous analyses and discussions in the workshop, we finally created predictive models for the whole Finnish regional social and health care system. The one-day workshops had to be prepared some months beforehand.
Even though the national data can be retrieved relatively easily from registries, there are several challenges in combining these datasets together: categorisation of data is usually done based on legislative framework, which is usually different for each registry. Getting access to data requires knowledge of the nature of information contained in each registry, which is quite well documented by the registry authorities. Achieving the datasets requires clear and open communication with registry keepers.
The social and health care officials of different regional social and health care systems with whom the results of the analysis were discussed were positive and helpful towards the co-creation process. This was helped by focusing on the interpretability of the analyses: the information had to be contextualized so that the relevancy of the analyses was clear for the officials.
A key success factor was the Finnish comprehensive social and health care data gathered in national-level registry, that has information on all social and health care use of Finnish citizens. Also, other national-level registries have comprehensive data, such as changes in marital statuses and places of residence of individuals. The data is accessed by going through the permission and data achievement procedure.
However, due to the personal and intimate nature of social and health care information, the procedure for achieving the data is time-consuming and might take a considerable amount of time. This is a potential barrier for using social and health care data in co-creation of new service innovations.
Finnish social and health care data is quite comprehensive and well registered on international level. However, accessing the data needs a rigorous process ensuring both the non-compromisation of individual confidentiality and secure handling of the data. The data usually is gathered and structured around a specific legislative framework and designed for certain use: this has to be acknowledged when designing a research setting using these data.
- Get familiar with the legal framework of each registry, as the format of information is based on these.
- Contextualize the findings of analyses for the target audience, so that the information and policy recommendations can easily be implemented in practice.
- Assume the compatibility of different datasets: as the data have been gathered to serve different purposes, they might be on different levels and with varying levels of detail.
