The services provided by digital clinics must be separable from the data, so that the use and users of digital clinics can be studied in general. Existing national register data enable research on digital clinics. In this article for researchers, we present a solution based on utilizing the service provider's service unit (SOTE-OIDs).
The majority of wellbeing services counties have already adopted or are in the process of adopting a digital clinic or a digital health and social services centre, i.e. a digital and often chat-based alternative to attendance services in primary health care. Remote phone calls to health centres have long been common. If you, as a researcher, are interested in doing register research on digital clinic services in wellbeing services counties, the key question is how to separate digital clinic operations from phone calls to health centres in the register of outpatient care in primary health care (Avohilmo).
The connection mode variable is currently not a satisfactory solution in public outpatient care
In open hilmo, remote interaction has traditionally been examined with the help of the connection method variable (Hilmo-Contact method). The classification used in recent years has made it possible to divide remote transactions into two: 1) real-time remote transactions (R52) and 2) remote transactions without real-time contact (R56). Real-time remote transactions include calls, video calls and chats. Remote communication without real-time contact includes, for example, letters and asynchronous messages, such as messages sent in Apotti's Maissa service. The digital clinic is mainly a chat service, while the telephone is a key channel for contacting the attending services of health centres. As both represent real-time remote transactions, it is not possible to distinguish digital clinic operations from calls from attendance by using the connection mode variable.
Another challenge related to the mode of communication variable is its variable quality. When studying Avohilmo entries in public digital clinics, we have found that there are variations between wellbeing services counties in the way digital clinic activities are recorded. In some areas, real-time remote interaction is emphasized, while in other areas, consultations between professionals or the management of the customer's case are emphasized. In at least one well-being area, digital clinic events have also been recorded to a significant extent as appointments, although the appointment is clearly defined as an attendance in the Hilmo Guide. The figures are separate estimates of the need for treatment, i.e. Avohilmo events with information describing the assessment of the need for treatment but no information describing the visit - in this case, the connection method is completely missing.
In private healthcare, the connection mode variable has been useful in distinguishing between remote interaction and attendance
In private outpatient care and occupational health care, the mode of contact variable has been clearer, as these sectors do not have an assessment of the need for care corresponding to primary public health care and the interaction between professionals (e.g. nurse-doctor-pair or team model) is not as common as in public outpatient care. Thus, real-time remote transactions and appointments are the most commonly used values for the connection mode variable. However, one problem remains: digital clinic operations cannot be separated from video calls or calls by means of the connection mode variable.
The classification of the connection method variable was updated in the summer of 2025, and in the future it may be possible to separate digital clinic services using the connection method variable
The Hilmo - Connection method classification is used in Avohilmo, and it was updated in summer 2025 so that the classification is now more fine from the point of view of remote transactions, including the following categories: 1) remote transactions in real-time via voice (R50), 2) remote transactions in real-time via video (R53), 3) remote transactions in real-time via chat (R54), 4) remote transactions in real-time, more precisely defined (R47), 5) transactions by letter (R55), 6) digital transactions without real-time contact (R57) and 7) digital transactions without contact with a professional (R59). From the point of view of digital clinic research, this is a good thing: After all, the vast majority of digital clinic contacts are chat, while the vast majority of remote interactions with attending services are calls (voice).
However, getting the updated Hilmo - Connection method classification for the daily use of professionals in wellbeing services counties may take time, and this required not only guidance, but also changes to patient information systems. At the time of writing (29.9.2025), the new remote access classifications are not yet widely used. The old real-time remote transaction code (R52) will have 12.4 million transactions in 2025, while the new codes (R50, R53, R54, R47, R55, R57 and R59) will have a total of only 0.1 million transactions (source: THL; High quality report).
The best solution here and now: service provider's service unit (SOTE-OID) can be used to separate digital clinics in many wellbeing services areas
The mandatory field in Avohilmo data is the service unit of the service provider, which is based on the units in the SOTE organisation register (so-called SOTE-OID identification number). From the point of view of digital clinic research, it is fortunate that the majority of wellbeing services counties have created one or more SOTE-OID identifiers for digital clinic operations, which are fairly easily recognisable from the SOTE organisation register. Figure 1. For example, we understand that in the Lifecare patient information system, the place of performance field is linked to the SOTE organisation register, i.e. if the wellbeing services county establishes a separate place of performance for its new digital clinic, a new unit should also be included in the SOTE organisation register to cover the service production of the new digital clinic.
In our own work, we rely heavily on SOTE-OID identifiers, when the aim is to separate the service production of a digital clinic from the service production of health centres. At the moment, it can be said that the service of digital clinics in wellbeing services counties consists mainly of chat conversations but also video receptions, while in attendance services, i.e. health stations, real-time remote interaction mainly means phone calls. Thus, the combination of the SOTE-OID unit and the connection mode variable makes it possible to distinguish visits to a digital clinic from the service production of health centres (visits and calls) and, moreover, the connection mode variable makes it possible to distinguish between visits to health centres and calls to health centres.
Unfortunately, it is not possible to identify digital clinics in all wellbeing services areas using the service provider's service unit (SOTE-OID).
As stated above, the majority of wellbeing services counties have created a separate and easily recognisable service provider unit for digital clinic operations. This is not the case for all welfare areas. In our view, the underlying idea may be that digital service activities are not separate entities (digiclinic) but digital is seen as a tool that is used by everyone, rather than as a separate function. For example, on the digital platform of the Kymenlaakso wellbeing services county, the service provider is either Kaiku24, if it is a first-line digital contact, or a health centre, if the patient is treated by a wellbeing services county professional. In this case, you can try to separate digital contacts using the connection method variable, but there are many challenges, as we illustrated at the beginning of the article.
We will now list some areas where we understand that there is no separate SOTE-OID for digital clinic operations (indicative, there may be errors): Jämsä, Kristiinankaupunki, Kymenlaakso Welfare Area, Kuusiokunta (Alavus, Kuortane, Soini, Ähtäri), Meri-Lappi (Kemi, Tornio, Keminmaa), Mänttä-Vilppula, Parkano, Siikalatva and Siuntio. We understand that there was no separate SOTE-OID in Varkaus and Joroinen (digiklinikka acquired from Mehiläinen, ended on 4/2023) and Ylä-Savo (service ended at the end of 2024). In Etelä-Savo, a separate SOTE-OID referring to digital clinic operations has only been in place since the beginning of 2023, even though the digital clinic was already piloted in the area from 2021 to 2022. Päijät-Häme has had a separate SOTE-OID referring to digital clinic operations only from the beginning of 2023, even though the digital clinic was introduced in part of the region already in 2021. We have also realised that, as a rule, no separate SOTE-OID has been created for digital clinic operations in the outsourcing of individual health stations.
It should also be mentioned that in the case of Apotti municipalities (Helsinki, Vantaa, Kerava and Kauniainen), asynchronous messages in countries do not, in our view, have their own health centre-specific health and social services OID units (these messages are answered from health centres).
Identification of the service provider's service unit (SOTE-OID) in research use may require consent from the wellbeing services county
The licences for using the data used in our SoteDataLab project are granted by THL and Statistics Finland, and these data are used in Statistics Finland's Fiona remote access environment. In principle, Statistics Finland protects by pseudonymising all direct identification data before exporting the data to a remote user environment, because according to the Statistics Act, the data may not be published or disclosed for research purposes in such a way that the subject of the statistics (e.g. person, company, organisation, etc.) is revealed. As a result, Statistics Finland would in principle also pseudonymise the service unit identifiers of SOTE-OID.
However, the use of SOTE-OID units in research and the publication of data at unit level is possible, subject to the consent of the statistical unit. In our contacts with Statistics Finland, we created a solution model: the wellbeing services county may give its written consent that the direct identification data of the service provider units in its SOTE organisation register can be identified from the national register materials used by the project in remote Fiona use. Consent is limited to this research project, and the purpose of use is described in the research plan for the project submitted as an attachment to the licence applications.
Consents in Fiona use may not be required either if (a) the study allows the averages of several wellbeing services counties to be examined instead of a single wellbeing services county, or (b) the service use of residents of a particular wellbeing services county by public digital clinics is examined instead of a specific service provider (when digital clinics are included throughout the country). Outside Fiona, researchers could first compile a table of three variables from the SOTE organisation register: SOTE-OID, sector and self-created indicator of whether the service provider unit is a public digital clinic or not. Statistics Finland could export this table for Fiona remote use and protect the SOTE-OID tags with the same encryption key as Avohilmo data. Consent may also not be required if (c) the survey is based on Findata's (and Statistics Finland's) data permit.
Category: Bloggers
Tapio Haaga, Meeri Seppä, Tanja Saxell, Which Kortelainen, Oskari Nokso-Koivisto, Lauri Sääksvuori and Alex Kivimäki.
We thank Tuulikki Vehko, Heli Suho, Merja Korajoki, Tuukka Holster and Sonja Lumet who commented on the draft text; from Statistics Finland's Reetta Salokan on promoting the above-mentioned consent-based solution; welfare areas that have given their consent and representatives of service providers (lots. Katja Rääpysjärvi, Jukka Karjalainen and Kaisa Kujansivu), who have kindly answered our questions about the SOTE-OID units in the regions.