Frequently asked questions on health and social care financing

The state-allocated health and social care budget constitutes a lump sum that increases annually in accordance with four factors:

1. Growth in service need as forecast by THL’s SOME model
2. The wellbeing services county index produced by Statistics Finland
3. Changes in counties’ statutory responsibilities
4. Ex-post adjustment carried out by the Ministry of Finance

Since the funding is intended to be needs-based, the budget is increased annually in line with the growth in service need forecast by the SOME model. The model is based on the population’s age and gender distribution.

However, the health and social care budget is also adjusted for increases in health and social care costs that are not driven by growth in service need but are beyond the control of wellbeing services counties (including the City of Helsinki). Such increases may result, for example, from rising price levels. This increase is taken into account through the wellbeing services county index and the ex-post adjustment.

In the ex-post adjustment, the combined accrued costs of all wellbeing services counties are compared with the funding allocated to them, and the health and social care budget for the following year is decreased or increased by the difference.

The objective of the current funding model is to curb cost growth and to allocate funding regionally in an equitable manner. Unlike activity-based funding, in a needs-based model a wellbeing services county cannot increase its funding by increasing the number of treatment episodes or costs.

Funding is allocated equitably if the service need of each region can be determined with reasonable reliability and the region receives funding according to its need.

THL has examined the use of alternative need factors in need-adjustment model and already uses demographic and socioeconomic factors as part of the calculation of need coefficients. In the update proposal for the need-adjustment model published by THL in 2025, need coefficients are also presented in such a way that morbidity data is not used at all in determining service need. Such a model has been considered advantageous particularly in terms of incentives, as it does not encourage upcoding of diagnoses. 

However, individual-level morbidity data has a significant impact on regional need coefficients. For this reason, need coefficients calculated without morbidity data would disadvantage regions where morbidity is high relative to the age and gender distribution of the population. 

Based on current knowledge, register-based assessment of service needs is the only way to ensure funding that reflects service needs.

It does not. The need-adjustment affects only how health and social care budget from the state is distributed between wellbeing services counties, not the size of the budget.

The need-adjustment model does not systematically favour any region. However, due to the underlying logic of the model, a region’s funding may be lower than its potential level if the  diagnoses are recorded deficiently or transferred to THL’s registers incompletely in that region.

The population of a wellbeing services county is taken into account twice in the funding model: each resident increases both the funding allocated on the basis of service need and the funding allocated on the basis of population size. For this reason, the funding model favours the most populous wellbeing services counties.

Averaging refers to the practice whereby the service-need-based component of the calculated funding for wellbeing services counties (approximately 80% of total calculated funding) is determined based on the average of the need coefficients for the two most recent years – rather than only the most recent year’s need coefficient.

A change limiter, in turn, means that a maximum limit is set in advance for how much the funding of an individual wellbeing services county may decrease or increase. The difference between the procedures is that the change limiter affects only those regions that exceed the defined threshold, whereas averaging affects all regions equally.

The purpose of these procedures is to ensure that annual changes in observed morbidity do not lead to unreasonably large changes in the funding received by regions. Changes in observed morbidity may result particularly from changes in the data (e.g., improvements in registering diagnoses and transferring them to THL’s registers) and from changes in actual morbidity.

Averaging and the change limiter are particularly justified when changes in observed service need result from changes in the data and therefore do not reflect changes in actual service need.

If stability of funding is prioritised, averaging and the change limiter are useful, as annual changes are reduced. If, on the other hand, needs-based allocation is prioritised, they may be seen as a disadvantage. This is because the level of funding does not fully respond to changes in actual service need. In such cases, regions where actual service need decreases benefit, as their funding does not decrease as much as service need. Conversely, regions where actual service need increases lose out: funding does not increase as much as actual service need.

Service provision is taken into account in need-adjustment model because otherwise the regression coefficients of other factors would be biased. For example, a particular disease group could receive too large a coefficient if people with that disease tend to live more often in areas with a high level of service provision: their use of the public sector would be higher on average simply because of the superior access to care. Conversely, ignoring service provision may also produce regression coefficients that are too low.

Although supply factors are taken into account in the need-adjustment model, funding for wellbeing services counties is not intended to be allocated on the basis of how easy it is for the population to access care. For this reason, need factors are neutralised before regression coefficients are used to predict each person’s service use for the following year. This means that the value of each person’s individual need factor is replaced with the national average.

If a wellbeing services county hires more personnel, this (i) has no direct effect on need coefficients. However, it (ii) has an indirect effect on need coefficients, as it changes the regression coefficients of other need factors, such as disease groups.

It is impossible to determine whether an increase in personnel leads to a higher or lower need coefficient in the region in question. In any case, the effect is moderate, as it affects the coefficients only indirectly through the regression coefficients of need factors.

The need coefficients affecting funding for 2025 and 2026 are based on the need model update published in 2022. The need-adjustment model report published by THL in 2025 includes a proposal to update the model. The decision to update the model is made by Parliament, and the proposal for the update is prepared by the Ministry of Finance on the basis of THL’s proposal.

THL continuously seeks to improve reporting related to funding calculations and the transparency of data.

For example, in December 2025, we published a new reporting tool that allows users to examine diagnosis data used in health and social care funding calculations at national level and by wellbeing services county. The tool contains data on the number of cases of different diseases and their development based on different data sources for the years 2021–2024 in table and chart form.

There are differences between information system providers in how the technical instructions for data transfer have been interpreted. Similarly, the interpretation of THL’s diagnosis recording guidelines has been inconsistent between providers. These differences in interpretation already existed before the health and social care reform.

In the calculation of funding allocations, it is essential that they are based on data that are as accurate as possible. In addition, the data must be sufficiently comprehensive.

It should be noted that perfectly accurate data do not exist. To identify service needs, information on the population’s morbidity is required, and demographic data alone does not describe need sufficiently comprehensively.

The averaging of need coefficients proposed by THL aims to reduce year-to-year variation in coefficients and thereby increase the predictability of funding.

Averaging means that the need-based funding is determined by the average of need-coefficients of the two most recent years. Averaging reduces variation in need coefficients arising both from morbidity and other need factors, as well as from changes in data production.

This temporary solution smooths variation until challenges in morbidity data are resolved.

The calculation of service need must be based on available register data, as no better data is available. The aim is to use the most recent data that are comprehensively available.

Although register data are not perfect, their quality is continuously improving. Wellbeing services counties and other actors are increasingly transferring data to THL’s registers in a more comprehensive and consistent manner, which improves the accuracy of need calculations.

However, it should be noted that counties must produce data that is comparable to other counties. Comparable data allows its extensive use in need-adjustment calculations. For the data to be comparable with other data used, they must be recorded in registers in accordance with national guidelines, not through ad hoc and separate data extractions.

THL continuously engages in dialogue with system providers so that data transfers comply with THL’s guidelines. In addition, several automated validation mechanisms have been built into THL’s system, making it possible to detect anomalies in data transfers.

A fundamental problem is that neither THL nor wellbeing services counties always have visibility into how system providers transfer data, and therefore do not have the means to detect all possible errors.

The error was caused by the fact that the need coefficients for 2023 used in the funding allocations were based on incomplete data. THL researchers identified that one of the datasets used in the calculation was outdated. This dataset had been extracted already in June 2024 and did not include, for example, supplementary data deliveries related to the Apotti system made in December 2024.

The error was corrected immediately, and the preliminary need coefficients for 2026 were updated using the corrected data without delay. The Ministry of Finance published the updated funding calculation based on the corrected data on 26 June 2025. Data and calculations for previous years have been reviewed, and no similar error has been identified in them.

To prevent recurrence, THL has strengthened personnel resources for the calculation of funding needs. In the future, more people will review the submitted data and the calculation process. In addition, THL has improved validation procedures associated with the data management.

It does not increase. Diagnoses do not affect the national level of health and social care budget, but only its distribution between wellbeing services counties.

Physicians are required to record diagnoses and related entries on medical, not financial, grounds. 

Physicians must follow the guidelines for recording, in particular the Finnish handbook for recording diseases and the general guidelines, for recording patient information. They must record the diagnosis they have determined for the patient during an inpatient care episode or an outpatient visit, and if there are multiple diagnoses, select the primary one to be recorded as the main diagnosis.

It does not. In need-adjustment process, costs are assigned to each individual primarily based on the disease categories to which they belong on the basis of recorded diagnoses and other codes included in those categories.

Each person can be included in each disease category only once. For example, the first record of specific diagnosis code belonging to the diabetes category observed for a person during a calendar year places them in the diabetes category, but the second, third, fourth, etc. record in the same category no longer affects funding in any way.

It does not. A new record of a diagnosis can affect your region’s funding only if the person receiving the diagnosis has not already had other codes belonging to the same disease category recorded during that year. This is because one person can belong to each disease category only once.

In addition, funding is not allocated directly on the basis of regression coefficients, and the coefficients do not affect the total level of health and social care funding. Instead, regression coefficients are used to calculate predicted service use for each individual.

The need coefficients of wellbeing services counties are calculated as the ratio of the average predicted service use of the population living in the area to the national average predicted service use. The need coefficients ultimately determine, to a large extent, how health and social care budget is distributed between regions.

Because regression coefficients affect need coefficients only through service use predictions, and need coefficients are always calculated relative to the situation in all other regions, it is impossible to determine with any quick or simple calculation how much one additional record of a diagnosis in a single disease category affects the funding of a single region. To perform such a calculation, it would be necessary to know how the diagnosis changes the predicted service need of the region and of the whole country, and then calculate the need coefficient on that basis.

When extracting morbidity data included in the disease classification of the need-adjustment model, diagnosis data from outpatient visits and inpatient care episodes in the care registers (Terveys-Hilmo, Avohilmo and Sosiaalihilmo) are taken into account (primary, secondary and long-term diagnoses, as well as cause-of-injury and symptom diagnoses).

ICPC-2 entries are also taken into account for outpatient data. In primary health care, data are taken into account only from physicians’ outpatient visits.

Yes. In addition to the care registers, morbidity data is also taken into account from several other data sources maintained by THL, the Social Insurance Institution (Kela) and the Finnish Centre for Pensions (ETK).

This may be due to several factors: 

In need-adjustment process, even a single record of diagnosis belonging to, for example, the diabetes category in any data source during a calendar year is sufficient to classify the individual to the disease group.

In addition to the care registers, data from other registers are also taken into account (see previous question), which regions are not able to include in their own analyses.

In general, all diagnoses recorded in THL’s care registers, including long-term diagnoses, affect the funding received by regions.

However, the incorrect recording of long-term ICD-10 diagnoses as  ICPC-2 codes identified in spring 2025 has not affected funding allocations. This is because, in the need-adjustment formulas, diagnosis data extracted from both the visit reason field and the long-term diagnosis field are treated as equivalent, meaning that the information can be extracted from either field. In addition, the same diagnosis for the same person affects the need-calculations only once, and repeated occurrences of the diagnosis have no further impact.

The care record guidelines specify how long-term diagnoses must be submitted to the care registers. The guidelines also apply to other data submitted to THL. THL’s quality reports allow the number and share of long-term diagnoses in events to be examined, for example by region and by system provider.

Diagnoses recorded in the private sector for outpatient care services are taken into account in the definition of morbidity.

In particular, the number of private sector data submissions to Avohilmo has increased significantly in recent years, but it has so far been difficult to determine whether the submitted data also include services purchased from the private sector by wellbeing services counties (and previously by municipalities) through service vouchers or other outsourcing arrangements.

As morbidity data has in any case already included diagnoses recorded by the private sector physicians through other data sources (for example sickness allowance data), private sector data have not so far been excluded.

However, the need-adjustment model is continuously being developed. In the update proposal published by THL in 2025, it is proposed that diagnoses recorded in all Avohilmo service categories, including those recorded in private healthcare, occupational health services and the Finnish Student Health Service, be taken into account in determining individuals’ morbidity.

The reason for this is that, under the law, a wellbeing services county is responsible for health and social care services for the whole population residing in its catchment area, even if some of them do not actually use public services.

This choice may increase funding in regions where private healthcare or occupational health services are widely used. For this reason, it is proposed that models predicting the need for publicly funded services should also include variables describing the use of other sectors, measuring how much an individual has used services other than publicly funded ones in monetary terms.

This, in turn, reduces funding in regions where private healthcare or occupational health services are widely used.

Because sufficiently comprehensive data on their use have not been available from THL’s care registers.