Finland has embarked on a transformation in which randomised experiments are becoming mainstream not only in medicine. The large-scale trial of digital reception in Ostrobothnia, which was launched in the spring of 2025 and is significant both in Finland and internationally, is part of this transformation. The close cooperation between the wellbeing services area and the research group already during the planning phase of the introduction of the platform has ensured a high-quality planned and implemented experiment, which enables the reliable evaluation of the effectiveness of the nurse's chat appointment.

Randomized experiments coming mainstream in Finland

Randomized controlled experimental designs (RCTs) are the gold standard for medical evidence and their use has been regulated since the 1970s, for example before new medicines are introduced on the market. However, despite their apparent usefulness, they have been little used to assess the impact of changes in the service system. Even major changes to the healthcare service system have been made and will be made continuously without a credible comparison, in which case it is difficult, if not impossible, to say afterwards what kind of effects the change has had on patients. 

One reason for this is that it may be impossible to assess the impact of some changes, even with large resources. However, a significant number of the changes are those whose evaluation would be clearly small and cost-effective in relation to the benefits of the evaluation.

However, the direction has changed and experiments are gradually becoming a part of impact-based decision-making. Some good examples are the ones that started in 2017. basic income experiment Started in 2021 Experiment with two-year pre-primary education. The experiments are also gradually landing in the social and health care service system, where Sanmark et al. (https://clinicaltrials.gov/study/NCT06836258) study the effects of AI-assisted recording on professional time and patient satisfaction, and Sääksvuori et al. (2022) investigate the impact of reminder letters on the vaccine coverage of seasonal influenza vaccines, using RCT designs. Finland's comprehensive national registers make the evaluation of the SOTE service system particularly attractive through experiments, as information on the effects of an intervention or change can be extracted from the registers at a low cost (THL, 2017).

Together with the Ostrobothnia Welfare Region, SoteDataLab has contributed to driving forward this transformation of impact data based on randomised comparisons with the large-scale, randomised domestic level launched in April 2025. experiment. In the experiment, half of the residents of Ostrobothnia had access to a new digital health care reception for the duration of the nine-month pilot before the service expanded to the entire population after the trial period. The pilot will provide nationally and internationally significant information on how the digital option affects the use of primary health care services, access to care and primary health care resources.

The experiment has been carried out cost-effectively in a situation where the digital reception of medical care was in any case being introduced in the wellbeing services area. With minor changes in the timing of implementation and utilizing Finland's internationally high-quality and versatile register data, a test was created that enables a reliable evaluation of the effectiveness of the new service. At the same time, useful descriptive information on the use of the new service is produced, for example, on what kind of people have found the service.

Why RCT?

The assessment of changes in the service system is often based on observational data, such as before-after comparisons, where it is often difficult to distinguish causal effects or relies on strong assumptions. In a randomized experimental design, the study group is randomly drawn into two control groups, the experimental group and the control group. Due to randomisation, the experimental and control groups do not, on average, differ according to the background characteristics of the groups, which means that the differences between the groups can be interpreted as the effects of intervention. A randomised experimental design is often the most reliable way to distinguish the effect of an intervention from all other possible effects and confounding factors. Therefore, RCTs are considered to be the gold standard for measuring the effectiveness of interventions.

Why Ostrobothnia?

The digital reception experiment was made possible in Ostrobothnia by a combination of three key factors: the right timing, a common goal to produce reliable research evidence on the impact of digital health services and effective cooperation. At the same time, the wellbeing services area was developing its digital services platform, allowing for a randomised experimental design to be taken into account in the design and procurement of the platform. In any case, the new platform would have been deployed, but elements allowing for randomisation were needed in connection with the deployment.

In practice, a function was needed to allow the platform to allow only the persons of the trial group to access the digital reception of medical care on the basis of strong authentication, while the comparison group would receive a polite message that the service is currently being piloted only for a limited population. In addition, it was important for the wellbeing services county to create a unique SOTE-OID identifier for the digital clinic, for which all digital appointments are recorded in the patient information system. This ensured that all digital receptions can be identified from the Avohilmo register (Read more about utilizing SOTE-OID).

The beginning of the collaboration between the research group and the wellbeing services county already at the planning stage was a prerequisite for ensuring the prerequisites for a randomized comparison design.

From the point of view of starting the cooperation, it was important that the research group and the wellbeing services county recognised the clear need for an impact study on digital health services. Digital health services have grown rapidly since the COVID-19 pandemic. For its part, the research team identified the need for impact data, which led to the establishment of the SoteDataLab project. In Ostrobothnia, praise must be given to the research-positive Director of Strategy and Development Suvi Einola, who promoted the implementation of the experiment in the wellbeing services area. In addition, the management of the wellbeing services county identified a need for research into digital services, which enabled the county to fully support the experiment.

The implementation of the ambitious experiment required preparation in close cooperation: the research group was responsible for the scientific planning of the experiment and the implementation of randomisation, while the wellbeing services county was responsible for the operational implementation and practical planning. The research team conducted the experiment Pre-Scientific Registration, the necessary power calculations and plans, and the randomisation of the test group on the basis of data extracted from the Population Information System of the DVV. In terms of the well-being area, the practical implementation of the experiment has been promoted by a professional team (special thanks to Petra Fager and Sofi Nyman), with whom the research team has been easy to work closely.

The result is a large-scale randomized experiment that meets high scientific criteria.

What's next?

Successful cooperation with the research team and the Ostrobothnian Welfare Region provides a solid basis for further experimentation – in Ostrobothnia or elsewhere in Finland. The SoteDataLab research group is part of a new consortium (Generative AI and digital solutions: Improving Effectiveness and Productivity in Health Services (GAINS)), which was recently awarded funding from the Strategic Research Council (SRC) of the Academy of Finland. The follow-up plans include assessing the effectiveness of the new AI tools in both social and health care.

Authors of the blog:

Alex Kivimäki, Tapio Haaga, Oskari Nokso-Koivisto