An appropriate institutionalisation of health accounts is a vital step to ensure that data on health expenditure and financing of high quality are regularly produced. In this context, the governance of the production process, the technical aspects of the production process and the dissemination of the results including their use to inform policy making are crucial domains. This chapter teases out the best practices of institutionalising health accounts based on the experience in 13 OECD countries with a relatively long health accounts history. A key message is that there is no “one size fits all” approach. Many factors related to characteristics of the health system, the health data information landscape or administrative cultures and rules play a role in determining how health accounts should be best institutionalised in a country.
Best Practice in Institutionalising Health Accounts
Abstract
As seen in Chapter 2, nearly all OECD countries regularly report data on health spending and financing based on the categories and classifications set out in the SHA 2011 framework. Hence, the conclusion can be drawn that – to varying degrees – health accounts have in fact been implemented in those countries. Yet, the extent to which health accounts have been implemented and how the production of health accounts has been institutionalised differs across countries. To shed some light on this, the experiences of 13 OECD countries in institutionalisation of health accounts will be discussed in this chapter.
The terms implementation and institutionalisation have been used frequently in the previous chapters and might be thought of as interchangeable but in fact refer to quite different stages and aspects of introducing and sustaining practices or systems.
Implementation generally refers to the process of putting a new plan, policy, or idea into practice. It focuses on the initial adoption and application of a framework or methodology with an emphasis on the execution and operationalisation. In terms of steps, it often involves initial training, piloting, resource allocation, and overcoming initial barriers.
Institutionalisation, on the other hand, involves embedding that practice or system into the norms and structures of an organisation to ensure it becomes a stable and ongoing part of its operations. There is therefore a focus on sustainability and normalisation over the longer term. Resource needs often involve ensuring that there is sufficient structural and financial support to prevent any backsliding.
In summary, implementation is about starting something new, while institutionalisation is about making permanent and ingrained. Coming back to health accounting this means, that while many countries have implemented a health accounts framework at one point in time, not all were able to institutionalise this stream of work on a permanent basis.
The importance of adequate institutionalisation of health accounts has been discussed by a number of international organisations. In 2012, shortly after the release of the SHA 2011 framework, the World Bank published a strategic guide for the institutionalisation of health accounts (Maeda et al., 2012[1]), setting out some of the key factors that can lead to successful institutionalisation and illustrating the state of play based on a number of case studies with a particular focus on low- and middle-income countries. The report also proposed a framework for health accounts institutionalisation (Figure 3.1), where institutionalisation is defined as “... [the] routine government-led and country-owned production and utilisation of an essential set of policy-relevant health expenditure data using an internationally accepted health accounting framework”.
Source: Adapted from Maeda et al. (2012[1]), Creating Evidence for Better Health Financing Decisions – A Strategic Guide for the Institutionalization of National Health Accounts, http://hdl.handle.net/10986/13141.
The World Bank framework is circular in its structure applying a feedback mechanism between the use and the production of health accounts such that it focuses on the improvement of health financing data to support evidence‑based policy making in health systems (Maeda et al., 2012[1]). By providing detailed insights into health expenditures, the World Bank highlights how information from national health accounts (NHAs) can help policy makers understand health system objectives around resource allocation, efficiency, and equity in healthcare delivery. In the production phase of the cycle, the World Bank guide emphasises the importance of standardised, comprehensive, and accurate data based on internationally recognised guidelines, such as the System of Health Accounts (SHA 2011), which ensures consistency and comparability across countries. The World Bank framework also places at the centre the critical components of governance, capacity building to improve the technical skills, and sufficient funding at the national level – all required for robust and sustained data collection and analysis. To make health accounts more relevant for decision-making, the World Bank framework advocates for regular updates and timeliness by reducing the time lag between data collection and publication. This allows policy makers to respond to emerging challenges with up-to-date information.
Overall, a central tenet of the World Bank approach is to link NHA data and provide an input into analysing broader health sector reforms and planning processes. For example, data from NHAs can inform the design of universal health coverage (UHC) strategies, monitor financial protection against catastrophic health spending, or evaluate the cost-effectiveness of health interventions. The framework also places an emphasis on the need for close collaboration between ministries of health, finance, and statistical agencies, as well as engagement with international partners.
Given the still substantial differences in the degree of implementation of health accounts globally, WHO released their Framework for assessing maturity of health accounts institutionalisation (WHO, 2023[2]) Focusing on the domains of demand, governance and financing, institutional technical capacity, dissemination and use of data, the WHO framework identifies some key elements within each domain, with some of them being vital to the health accounts production process, while other enabling factors (Figure 3.2).
This framework served as the basis for WHO’s institutionalisation survey, carried out in 2024. Based on feedback from 97 countries, WHO clustered countries into four groups based on the level of institutionalisation of health accounts (WHO, forthcoming[3]). Applying a decision tree of characteristics related to health accounts production, the mutually exclusive grouping ranges from those countries that have not yet started producing health accounts or have not produced accounts in a long time (Group 0), through those who do not “own” the process – thus relying on international support for data collection (Group 1) and countries where production is not fully systematic but remains rather ad hoc (Group 2).
Finally, the Group 3 countries – those with the highest level of health accounts institutionalisation – have typically produced health accounts continuously over the last 6 years, the health accounts data collection and the staff carrying out the health accounts production are fully funded by the government and the health accounts production is systematic. The OECD countries participating in the survey were either allocated to Group 3 or Group 2 highlighting that health accounts institutionalisation across the OECD is perhaps unsurprisingly more advanced than in low- and middle‑income countries. Yet, across all country groups the survey revealed that the communication of health accounts results could be further strengthened to increase impact (WHO, forthcoming[3]).
Source: Adapted from WHO (2023[2]), Framework for assessing maturity of health accounts institutionalisation, https://iris.who.int/handle/10665/373992.
These two frameworks are very much linked in the sense that they identify essentially the same “building blocks” required for a comprehensive institutionalisation of health accounts. The World Bank highlights more the circular nature of this exercise, in the sense that successful production and dissemination of data leads to a widespread use of data in policy making and thus reinforcing demand for health accounts. The WHO teases out a bit more the different steps in the production process related to the governance and financing of this work and the technical production process. The analysis in this report builds on the WHO framework, with a focus on the domains of health accounts institutionalisation that are of particular relevance to OECD countries, related to the governance of health accounts production, some technical aspects of the production, and dissemination practices to ensure health accounts are used to their full potential.1
To better understand how health accounts have been institutionalised in practice, the OECD Secretariat carried out structured interviews with the national focal points of health accounts production in 13 OECD countries.2 This subset of countries represents the full variety of health systems in OECD countries and experience in the implementation and institutionalisation of health accounts. Sections 3.2 to 3.4 compare some key features of institutionalisation of health accounts across these countries. More detailed information can be seen in the corresponding country sheets in Annex C.
In the next three sections the communalities and differences across these experiences regarding the key institutionalisation dimensions Governance, Technical Production Process and Dissemination and Use of data are described, also in light of identifying some best practices.
3.2.1. Responsibility for Health Accounts production
Across OECD countries, there is a variety of arrangements through which health accounts are produced and how the production has been institutionalised. Table 3.2 provides an overview of the agencies (currently) responsible for health accounts production in the 13 OECD countries reviewed. Generally, the responsibility for production of health accounts can be clustered into four different types. Health accounts are produced by:
The Ministry of Health (Chile, Costa Rica, France and Spain), or an agency directly associated with a Department of Health (United States)
The National Statistical Institute (NSI), either in the Divisions of Health (as seen in Germany, the Netherlands and Switzerland) or other Divisions (Austria, the United Kingdom)
A Health agency established by the central government (or jointly with regional governments or other jurisdictions) with a relatively high degree of independence (Australia and Canada)
A consortium of various stakeholders, including, for example, research institutions (Korea – led by a university)
As can be seen from Table 3.2, there are substantial differences regarding the time when health accounts have been first implemented across the OECD: From 50 years ago or more in the United States and the Netherlands, to comparatively recently in Costa Rica or the United Kingdom. However, the year of implementation does not necessarily mark the starting year for current data availability. For example, while health accounts were only implemented around 2015 in the United Kingdom, a time series of comparable data for some key dimensions (HF and FS) is available back to 1997. In Switzerland, there have been several rounds of health accounts implementation. While partial data exists for the years 1960‑85, a first full implementation covers the period 1985‑95, with the second implementation including the years between 1995‑2010. In its current version, the Swiss health accounts can provide comparable data from 2010 onwards (Bundesamt für Statistik, 2024[4]). In Korea, the first NHA (a two‑dimensional table mapping financing source and function) including spending from private sources was constructed in the mid‑1990s. In 2003, the Korean Government requested Yonsei University to establish a NHA based on the original SHA manual published in 2000, and as a result, the first three‑dimensional NHA were constructed in 2004. The current time series covers data back to 1970. On the other hand, in some instances the initial health accounts implementations have been archived and may no longer be publicly available. This is, for example, the case in Germany where the first version of the NHA – covering a time series from 1970 to 1998 – is no longer maintained. The current version of the NHA is available back to the year 1992.
The reasons why countries decided to institutionalise the production of health accounts in the agency of their choice need to be seen in the country-specific context. Generally, there is a lot of continuity, and changes of responsibilities are rare. In Australia and Canada, the health accounts production moved to the respective health agencies in line with their establishment (1987 and 1994, respectively). Prior to that, health spending measurement was done in the Commonwealth Department of Health in the case of Australia (Goss, 2022[5]) and the Department of National Health and Welfare in the case of Canada. In Germany, health accounts production moved in the mid‑1990s within the Federal Statistical Office from the public budget division to the health division. The high level of continuity observed in the health accounts production responsibilities across countries suggests that, overall, governments are satisfied with their choice of where health accounts had been institutionalised and see little reason for change, but may also signal the absence of a viable alternative.
Table 3.2 suggests that, in each country, a range of agencies can potentially be responsible for health accounts production, and certain trade‑offs need to be considered by countries in their decision where health accounts should be institutionalised.
Embedding the production of health accounts in the National Statistical Institutes typically has the advantage of widespread availability and direct access to relevant data (in particular outside of the government sector) used in the construction of health accounts, and privileged access may in some cases also be enforceable via statistical legislation. It also allows for a close alignment between health accounts and national accounts which can be helpful since many concepts of health accounting originate from national accounting concepts. However, translating health accounts results into effective policies can be more challenging due to the (physical and/or cultural) distance between statistical experts and policy makers in ministries, and the lack of in-depth health sector and policy knowledge available in statistical agencies.
The advantage of producing health accounts within Health Ministries lies in their potential usage for policy making. Transforming results from health accounts analyses into policies should be generally much simpler if both activities occur under the same roof. However, not all ministries have necessarily the technical expertise to implement a statistical accounting system. One additional disadvantage may be going beyond the remit of the Ministry of Health to capture the whole economy perspective as well as having ready access to the broad range of data sources required to compile comprehensive health accounts.
One advantage of tasking technical health agencies with the implementation of health accounts can be their relative independence, which can possibly add to the legitimacy of the results and may be perceived as unbiased. Assigning this work to a (relatively independent) health agency may also be good option in countries where matters of health policy are generally more decentralised and where using an agency associated with the central government may not be perceived the most appropriate solution. However, there may be the question whether work on health accounts would fit into the mandate of existing health agencies and, whether these agencies would obtain easy access to data in a number of countries.
Using a consortium of stakeholders that collectively produce health accounts can have the advantage that it assembles a wide range of experts with a lot of technical expertise strengthening the overall quality of health accounts but a downside of such an arrangement may be the co‑ordination effort that such an arrangement can require and possibly unclear responsibilities. Generally, this type of arrangement might be seen as a more short-term measure to ensure proper implementation and ensure that the capacity and technical expertise is built up with the responsibilities eventually shifted to a more stable institutional structure.
This section compares and contrasts some key elements of the governance of health accounts production across the 13 OECD countries included in this review, such as the existence of a legal mandate, the organisation to obtain access to relevant data, the nature of collaboration across different stakeholders and the involvement of an advisory board in health accounts production.
3.2.2. Establishing a legal mandate
In many OECD countries a legal mandate for the production of health accounts exists. At the supranational level, for the 27 Member States of the European Union,3 since 2015, European legislation requires Member States to annually submit health expenditure and financing data based on the SHA 2011 framework to the Statistical Office of the European Union (Eurostat). The legal provision for this mandate is Commission Regulation (EU) 2021/19014 implementing Regulation (EC) 1338/20085 of the European Parliament and of the Council as regards statistics on healthcare expenditure and financing. Regulation (EC) 1338/2008 establishes a common framework for the systematic production of statistics on public health and health and safety at work, and broadly defines the domains and subjects on which data is collected (which includes healthcare expenditure and financing under the domain healthcare). Commission Regulation (EU) 2021/1901 defines in detail the health expenditure variables that need to be submitted to Eurostat (which includes variables of the SHA 2011 core framework dimensions health financing [HF], providers [HP], and functions [HC] and which are a subset of the two‑dimensional tables of the JHAQ template) as well as the metadata file and specifies the submission deadlines. It is important to understand in this context that Commission Regulations are directly applicable in EU Member States and become immediately enforceable without needing to be adopted and transposed into national law by its Member States.
Yet despite this legal basis, some EU Member States adopted their own additional national legislation to fortify the production of health accounts. In some cases, these predate the European legislation.
In France, for example, a “health accounts commission” was initially established in 1970 in a joint decree issued from the ministers of finance and economy, agriculture, public health and social security.6 The main purpose of the commission was to analyse employment and resource use in the health sector, including the annual production of accounts to track health spending and the use of the funds and the social categories that benefit from those funds. It also defines the composition of the said commission, including representatives of several ministries, public agencies, health providers and other stakeholders (see Section 3.4). In its current version, the decree stipulates that the secretariat of the commission is ensured by the Directorate of Research, Studies, Evaluation and Statistics (DREES), situated within the Ministry of Health.
In 2021, the Federal Parliament in Germany adopted a law (“Gesetz über die Statistiken zu Gesundheitsausgaben und ihrer Finanzierung, zu Krankheitskosten sowie zum Personal im Gesundheitswesen“) which stipulates the production of health expenditure accounts, costs-of-illness accounts, health labour accounts and a regional health labour monitoring.7 The law defines the scope of the different accounting systems, designates the Federal Statistical Office as the responsible agency for the production of these accounts and identifies the data holders that are required to submit input data for this work. The law authorises the Federal Ministry of Health to decree the frequency of the production of these accounts and the timelines by when data holders need to submit the requested input data.
In the United Kingdom, the above‑mentioned (EU) Commission Regulation led to the implementation and institutionalisation of health accounts within the Office of National Statistics (ONS), but this legal mandate has not been replaced by national legislation after the United Kingdom exited the European Union. Currently, the annual health accounts have the status of an official statistic, and their production continues without any legal mandate, mainly because of the reputational risk associated with a discontinuation of this work and the non-submission of health spending data to international organisations, as well as the perceived value of this stream of work. Eventually, health accounts could apply to be included in the list of “accredited official statistics”. This set of statistics are reviewed by the Office for Statistics Regulation (OSR), which certifies that statistics comply with the standards of trustworthiness, quality and value in the Code of Practice for Statistics. This label could potentially put the production of health accounts on a more solid foundation.
In other countries explicit legal mandates for the regular production of data on health expenditure and financing are missing but can be derived implicitly.
In Canada, the Canadian Institute for Health Information (CIHI) was created by the federal, provincial and territorial governments in 1994 as a non-profit organisation to improve healthcare, health system performance and population health across Canada. There is an implicit understanding that the regular collection and publication of health spending data for the country as a whole but also for the individual provinces and territories is key to fulfil this mandate.
The situation is similar in Australia. For the Australian Institute of Health and Welfare (AIHW), established in 1987 as an Australian Government statutory agency, the production of health accounts data is implicitly included in its mandate to inform health policy decisions in the country.
In the United States, the production of health accounts has been institutionalised in the Office of the Actuary at the Centers for Medicare and Medicaid Services (CMS), a federal agency with the Department of Health and Human Services. There is no direct legal mandate to produce health accounts, but the information derived from health accounts are vital inputs for CMS to accomplish its mission, such as to analyse the changes in the amount and costs of health services purchases, the economic causal factors at work in the health sector, the impact of policy changes and comparisons at an international level. Additionally, various legislative provisions require use of national health expenditure trends to adjust insurance benefit parameters and to evaluate CMS programmes.
There are other interesting arrangements for the production of health accounts:
While the submission of health spending data to Eurostat is obligatory for Austria under European legislation, the practical implementation of health accounts is based on a contractual fixed-term agreement between the Central Statistical Office (Statistik Austria) and the Federal Ministry of Health – in the sense that the Statistical Office is producing health accounts on behalf of the ministry. This arrangement predates the Commission Regulation, and the contract specifies the health accounts dimensions and categories that need to be produced. Every three years this contract is up for renewal which also presents the opportunity to discuss changes in the granularity of health accounts. Health accounts work is also explicitly mentioned in the “target management agreement” between the federation, states and social insurance schemes (see Section 3.4)
Among the countries interviewed, Korea seems to be the only one where a consortium of different stakeholders is collectively responsible for the production of health accounts. The consortium is led by a professor of a university (Yonsei University) and includes the Korean Institute for Health and Social Affairs (KIHASA), the National Health Insurance Service (NHIS) and the Health Insurance Review and Assessment Service (HIRA). Since 2007, National Health Accounts are designated as nationally approved statistics under the Korean Statistics Act. The government commissions annually the production of health accounts through a research service contract.
In Switzerland, Spain and the Netherlands, the production of health accounts is included in the respective Multi-Annual Statistics Programmes, which needs to be drawn up in accordance with the national statistical laws. In Switzerland, for example, the Federal Statistical Office is leading the development of this programme in consultation with other producers of statistics on the federal level, an expert group on federal statistics and statistical agencies on a subnational level and other expert groups. The current programme covers the period 2024‑27 (BFS, 2024[6]).
In Chile, health accounts production was initially promoted by an interinstitutional working group – constituted in 2004 via ministerial resolution – which included the Ministry of Health, the Health Superintendency, Central Bank, and the National Health Fund (FONASA) and was supported by the Economic Commission for Latin America and the Caribbean (ECLAC) and Pan American Health Organization (PAHO). Similarly, in Costa Rica, the production of health accounts was promoted by the National Health Accounts Council (CONACUSA) – constituted in 2015 via Executive Decree No. 39169 and composed of the Ministry of Health, the Costa Rican Social Security Fund (CCSS), the University of Costa Rica, the National Insurance Institute (INS), the National Institute of Statistics and Census (INEC), the Ministry of Finance, the General Insurance Authority (SUGESE), the Central Bank, and the National Union of Local Governments (UNGL), with PAHO participating as an observer.
In the majority of OECD countries there exists a legal basis for the production of health accounts (either explicitly or implicitly). Such a mandate can facilitate the long-term budgeting and allocation of sufficient funding and staff to ensure a continuity of health accounts production and a sustainable future for the accounting system. Yet even without such a mandate, OECD countries have created more ad hoc arrangements that have so far guaranteed a continuous production of health accounts – but their funding situation may be more precarious when budgets become tight or government policies shift.
3.2.3. Co‑operation with data holders or other stakeholders
The production of health accounts relies on the access and use of many different data sources which requires a clear plan and organisation how to co‑operate with different stakeholders and data producers. In the initial implementation phase of health accounts, it is advisable to first map out the entire health data landscape of appropriate public and private data providers in a country that could be potentially beneficial, before assessing which health datasets would be the most adequate ones to use. This assessment should refer to “primary sources”, which generally record monetary transactions but can also include non-financial “secondary sources” (collecting, for example, utilisation data), which may be potentially useful for the creation of “allocation keys”, in case health spending needs to be distributed across different health providers, functions or financing schemes. The eventual choice of data sources will depend on many factors, including the quality of the data and its comprehensiveness, the public availability of data, its timeliness, or the costs of data (for those that are not freely available). The choice of the most appropriate data sources used in health accounts is also driven by the methodologies adopted in countries – that is, “how” health accounts have been technically implemented; whether they are primarily relying on nation-wide data sources collected via financing schemes, or rather focus on data sources gathered from the provider-side, or combine several regional data sources to a national total, or a combination of these approaches.8 Once the choice of data sources to produce health accounts is made, a strategy needs to be developed how the routine production of health accounts can be maintained in the medium to long-run but also how to potentially adapt health accounts if new, previously unused, data sources become available.
The practical arrangements through which health accounts teams co‑operate with data producers and other stakeholders vary widely across the countries included in this review due to differences in some key characteristics of the health systems, the health data infrastructure but also the administrative and statistical rules and cultures. Yet, all have some strategic co‑operation plans with data holders in place. Generally, there appear to be some differences in the co‑operation between data producers from the public and private sector, which may relate to all data, be it on financing schemes, providers or services.
Since across OECD countries, health spending is financed predominantly via public schemes (either based on residence‑based entitlements or different forms of compulsory health insurance) having accurate data on public spending on health is paramount for high-quality health accounts. Generally, obtaining data from the public sector is typically straight-forward for health accounts teams. In many cases, the data needed are already publicly available at the required level of detail and are produced by various public agencies based on existing regulation. This can apply to budgets of specific central or regional ministries, or municipalities; or to financial accounts of social health insurance funds, or other statistics on public finances. In other instances, the necessary data is not freely available but can be obtained on demand for the specific purpose of health accounting. Many experts interviewed highlighted that “goodwill arrangements” or “gentlemen’s agreement” with public data holders are an important avenue how key data for health accounts production is procured. Typically, these arrangements foresee that data can be obtained for the specific purpose of health accounts production but there may be limitations for use beyond this purpose. Establishing and maintaining good relationships with data providers is generally considered an important component of health accounts production.
Health accounts experts also rely in many instances on freely available data from the private sector, which can relate to private spending or revenues from private health providers or any other useful data from private actors in the health system. Beyond this, “goodwill agreements” for the sake of obtaining unpublished data are also critical for data from the private sector in many OECD countries. Compared to actors from the public sector, it may however require more time to convince private actors that this type of voluntary collaboration in the production of health accounts can be mutually beneficial.
Beyond these informal arrangements, OECD health accounts teams also rely in some instances on formalised co‑operations:
The ONS in the United Kingdom relies on a number of “service level agreements” to obtain access to important data. For example, for data on private sector providers and provision of healthcare, ONS has a long-standing “data partnership agreement” with an independent business intelligence organisation operating across the healthcare and social care sectors. In France, data acquisition is mainly based on good-will although there is a more formal arrangement with the Banque de France for some information on private insurance spending.9 In the Netherlands, a formalised agreement exists with Health Insurers to access their claims data.
Since the production of health accounts is carried out by a consortium of different actors in Korea, this co‑operation has of course been formalised with clearly defined tasks for all consortium partners involved. When it comes to external data holders, a co‑operation with Private Health Insurance agency has been established.
Where the responsibility for health accounts production lies with National Statistical Institutes, these countries have potentially additional legal levers to obtain data via the various existing statistical laws. Broad legal frameworks can often provide a legal mandate for statistical offices to collect data of high economic or societal importance. The Netherlands and Switzerland have mentioned that this legal tool is helpful. For example, Article 23 of the Swiss Federal Law on Compulsory Health Care assigns the Federal Statistical Office the role to compile the necessary statistical data to assess the functioning and effectiveness of this law, by collecting the required data from insurers, service providers and the population – who are obliged to provide this information free of charge.10
In Germany, since 2023, the extent of the required co‑operation by different data holders (including private entities) has been stipulated in a decree issued by the Ministry of Health.11 The requested data and the timelines by which they have to be submitted to the Federal Statistical Office are very detailed. This includes, for example, reporting requirements by the umbrella organisation of Private Health Insurance companies (PKV) to provide overall spending data by private health insurers (and co-payments by insured) and a breakdown by type of service, the Federal Association of Dentists associated with the Social Health Insurance Scheme (KZBV) for cost-sharing for dentures, or the Institute of the Evaluation Committee (IdB) for billing data of SHI-affiliated physicians and psychotherapists.
In Spain, Article 82 of the General Health Law12 stipulates that for the purpose of determining the total amount of funding allocated to healthcare, the autonomous communities shall promptly submit their approved budgets to the Ministry of Health and Consumer Affairs and provide information on their execution as well as their final settlement.
In a number of cases, health accounts teams also purchase some vital data from private providers that is not available free of charge. This is for example the case in the United States or in Australia where data on pharmaceutical spending is acquired from IQVIA.
Managing the relationship with data producers and other stakeholders is an important task of health accounts productions. These relationships should be built on trust and should become easier to manage over time. It is difficult to say whether data acquisition should rely more on formal than informal arrangements – this is pretty much country-specific. Of course, in case some informal arrangements have shown to be problematic (e.g. data not submitted on time), health accounts teams should explore options for a formal co‑operation. These, however, are generally associated with some administrative work. In some instances, especially when dealing with private data companies formalised data agreements are the more natural option.
3.2.4. Role of advisory group in health accounts production
In many countries interviewed for this report, a group of experts regularly advises on issues related to the production of health accounts. Yet, how these groups are set up and the scope of their advisory role differs across countries. Moreover, a general distinction also needs to be made into advisory groups that were established in the implementation phase of health accounts and those groups that advise during the routine health production, as well as expert groups that are more technical in nature and those discussing expenditure data in a broader policy context.
In Canada, two separate groups have been set up to advise on health accounts production and dissemination. The National Health Expenditure Expert Advisory Group meets annually with the health accounts team and discusses updates to the latest health spending trends and highlights issues to monitor in the future. The group also advises on possible methodological improvements, new products, and analytical work on health spending more broadly (CIHI, 2024[7]). It consists of a dozen experts, with a mix of researchers, and representatives of federal and provincial bodies, including health ministries and statistical agencies. In the past, the group has, for example, advised on the development of new econometric models to evaluate data quality, the use of new data sources on home and community care spending, and the development of a report on the drivers for hospital, physician and drug spending (CIHI, 2011[8]). In addition, a technical group composed of members of the health accounts team at CIHI and data providers in federal and provincial agencies meets twice a year to discuss issues related to the data and production process.
In Australia, a Health Expenditure Advisory Committee oversees the health accounts production at AIHW and meets 2‑3 times per year. Its Terms of Reference include advising the AIHW on data sources, analysis methods and presenting of health spending estimates; on the integration of AIHW expenditure data collection with other (sub)national collections and with international frameworks; and on longer term directions of the health expenditure work (AIHW, n.d.[9]). So far, data providers and users from the public sector are represented in this committee but membership may eventually be extended to non-government organisations and researchers. Several agencies from the Australian Government (for example, the Department of Health and Aged Care, the Department of Veterans’ Affairs, the Treasury, the Australian Bureau of Statistics and Medicare) as well as health departments from states and territories are currently represented. Additionally, a technical advisory group – a subgroup of the committee – is consulted for specific methodological issues.
France set up a health accounts commission as early as 1970 by decree to (i) establish accounts to study the various arrangements through which health expenditure is channeled, the services that are paid for and the social categories that benefit; and (ii) to assess the impact of the healthcare system on economic development. Participation is formalised, as the decree clearly stipulates the membership. Currently, the commission has around 60 members, headed by a president (nominated per decree for a period of three years), a vice‑president (Director of DREES), 17 members representing the public administration, such as the Ministries of Health, Economy, Budget, Agriculture but also agencies associated with the Ministry of Health, such as Technical Agency for Hospital Information (ATIH), 27 members representing a variety of health professionals, health providers, the pharmaceutical industry, social and private health insurers, patient representatives and other agencies, as well as 14 members with other background such as researchers (who are nominated per decree for a period of three years). The commission generally meets once a year to discuss the results and methodological issues – the latest findings are typically discussed in the context of wider health policy issues (see Section 3.4).
In Korea, the “Health Accounts Forum” was initially established in 2011 as an irregular meeting of experts in the area of health accounting but became a regular event in 2022. At this meeting, the various members of the consortium share their work, and experts in health and LTC expenditures discuss current issues and provide technical advice. This forum also contributes to the dissemination of health account results.
In the United Kingdom, ONS has established a “Health Accounts Steering Group” to advise on health accounts developments, consisting of the ONS health accounts team, representatives of the Departments of Health of the four administrations, NHS England, independent experts and OECD to provide an international perspective. At the moment, this steering group meets on an annual basis but there were more frequent interactions of the Steering Group during the health accounts implementation phase. The annual meeting is typically organised a few weeks before the release of the latest estimates and allows for an in-depth discussion of the first preliminary results, their plausibility and any methodological advancements. Steering group members are also invited to provide comments on the analytical piece accompanying the data release.
In Austria, a “scientific SHA advisory board” was established during the initial implementation of SHA in 2005 and includes representatives of ministries, research institutes, chambers of health professionals, universities, state representatives, representatives of social insurance organisations and the state health funds. Initially, members were appointed by the Ministry of Health, but access is less formalised now. The role of the board has changed over time. While in the early years of the initial implementation of health accounts, it provided important feedback on the methodology, it has now become a forum to discuss results. Board members receive a substantial report (~120 pages) on the recent results and the health accounts work carried out in the previous year. In addition to this “user side” advisory board, every few years an external expert group reviews the health accounts as per standard procedure within Statistik Austria for all statistical products. The main focus of this “peer review” lies on the discussion of the methodology applied in the construction of health accounts and calculation methods (Statistik Austria, 2020[10]).
While the United States has no formalised advisory board, CMS actively reaches out for feedback from users, data providers and policy makers when they revise their historical data (every five years). They also obtain feedback outside of this cycle from the Department of Health and Human Services.
In Chile, the above‑mentioned interinstitutional committee spearheaded the implementation of health satellite accounts. Following a period of inactivity, there is willingness to restart the committee’s involvement in the production of health accounts.
In Costa Rica, CONACUSA initially led the technical and methodological implementation of health accounts during development phase. Following a period of inactivity, it has now resumed activities serving primarily as a policy-oriented steering group.
In the Netherlands and Spain, there are a no systematic groups to advise on health accounts production but experts have been consulted ad hoc on specific issues in the past.
Overall, the establishment of a group of external experts who regularly advise on health accounts production from a methodological perspective but who may also provide guidance on dissemination and the policy context in which health spending data needs to be understood is welcomed. The existence of a group provides an instrument to help assure a high level of quality of the health accounts and may strengthen the legitimacy of the results. Of course, the establishment of such groups come at a cost since the health accounts teams would generally need to organise these meetings and prepare some background material.
The technical implementation of health accounts is influenced by numerous factors, such as the methodology applied in the construction of health accounts, the dimensions and categories calculated within health accounts, whether national versions of health accounts exist in a country in addition to SHA 2011‑based accounting, the complexity of the health system, the corresponding health information system, the availability and access to data but also the agency where health accounts production is institutionalised. As an accounting framework, health accounts have to incorporate a large amount of information included in different data sources, such as administrative records, financial accounts, household budget surveys, business statistics and many other.
3.3.1. Annual production cycle
The regular production of health accounts requires a number of critical steps, starting with the acquisition of data and assuring their quality, the processing of the data and the compilation of the various health accounts tables, the validation of the plausibility of the results, an analysis of the results (potentially including also the drafting of reports), and finally, the dissemination of the various health accounts outputs using a range of different tools. After this, the final step in the production cycle would be to dedicate sufficient time to reflect on the applied processes and methods used and review those (if necessary), and scan for new appropriate data sources that could potentially improve the health expenditure data in the future.
All countries reviewed for this report publish health accounts on an annual basis and thus work on an annual production cycle, while they all appear to have clearly established processes in place. As discussed in Chapter 2, some of the countries produce health accounts exclusively based on SHA 2011 (e.g. Austria, Korea, the United Kingdom), whereas other countries produce national versions of health accounts which then serve as the basis to submit SHA 2011‑based health accounts table to the international organisations (e.g. Australia, Canada, France, the United States), in some instances implementing additional dimensions, such as a regional breakdown. Finally, in a third group of countries, the national and international versions are more or less produced simultaneously (e.g. Germany, the Netherlands, Switzerland). This general approach determines the country-specific timeline in the production of health accounts.
In countries where health accounts are exclusively based on SHA 2011 accounting, the production cycle is partially influenced by the schedule of the international data collection on health accounts data (JHAQ).13 This means that, for example, data acquisition in countries such as Austria or the United Kingdom starts around five to six months before the planned release of data, which is either just before (in the case of Austria) or just after (in the case of the United Kingdom) the end of the OECD March deadline for data submission. In Austria, the remaining part of the year is spent on the calculation of spending estimates for year t‑1, a review of the methodology, the drafting of reports for the ministry and the advisory board, as well as carrying out health accounts-related analyses as part of their mandate in the health target monitoring (See Section 3.4). In Costa Rica, data acquisition begins nine months before the OECD deadline and data release follows the full completion of the international data validation process.
Australia, Canada and the United States work on a different time schedule for their data releases. In all three countries, national health accounts data are disseminated around November-December each year. In Australia and Canada, the first data requests are sent 10‑11 months before the eventual release date (around 6 months in the case of the United States). After the release of the data, the following months are used to map and convert national health expenditure into the SHA 2011‑based data templates and submit those to the international organisations, as well as for other work, including more detailed national analyses.
In Germany and Switzerland, the publication of results stemming from the national health accounts versions is done around the same time as when the JHAQ tables are submitted to the international organisations. This requires these countries to dual-code the identified transactions in their databases – once according to the national classifications and categories and then according to SHA 2011 classifications.
3.3.2. Technical implementation and data processing
The methodologies applied in the construction of health accounts differ widely across OECD countries. In very general terms, one common methodological approach chosen by many countries is to focus predominantly on collecting health spending data from the financing side and then allocate the various transactions to healthcare services and providers. Others rely more on revenue or income data from health providers as their main sources. However, in most instances, integrative approaches are used combining data sources from both the financing and provider perspectives. There are many different factors that can affect the methodological choice of a country, but generally, health accounts teams give preference to the approach that is most suited to the health data infrastructure in their country, taking into account availability and quality of the data. This observation reinforces the need to clearly map out the health information flows between the various actors (financing and provision) in a country’s healthcare system in an initial implementation step.
As the methodologies applied in the OECD countries included in this report are so country-specific, this is mirrored by the technical processes used in health accounts production.
Generally, health accounts teams process most of the “raw” input data that they obtain from the various data holders themselves. They can either access this data themselves if available on data portals, download it on websites, or receive submissions from data holders directly (e.g. in an electronic format via email). This raw data can vary substantially in its granularity which may have an impact on the resource needs to process it. For example, in some instances countries use patient-level claims data from insurers in the construction of health accounts, which have substantial file sizes (e.g. France). In others, the raw data refers to aggregated budget results, public spending statistics or consolidated financing results of health insurers (e.g. Austria, Germany, the United Kingdom). These files are, by their nature, smaller by comparison. In some instances, the raw data has already been adjusted to some extent by data holders, but generally some processing or adjustment is required on the side of the health accounts team before it can be used in the health accounts calculations. For example, CIHI needs to standardise the different information received on health spending by Canadian provinces as reporting standards may differ. In Korea, standardised data output/submission formats are regularly negotiated between the institution handling the raw data and the lead institution responsible for data collection and final entry into the health account master table.
In some instances, part of the data processing is carried out by other actors. In the United Kingdom, for example, the allocation of total spending by government financing schemes (HF11) to health providers and services is done by the health departments of the four administrations with the ONS providing the control total and eventually combining and validating the four submitted tables. In Spain, public health spending by the 17 autonomous regions is collected by the respective health departments on the regional level before being submitted to the centralised health accounts team and processed both for the Spanish version of public health spending aggregates (EGSP) and the SHA 2011 based submissions. As mentioned, in Korea a consortium of stakeholders is collectively responsible for the health accounts production which makes the technical implementation and co‑operation a bit more complex (Box 3.1).
For specific transactions, the health accounts team may need to resort to some specialised analysis carried out by external partners that will be plugged into the health accounts calculation models. For example, in the Netherlands, data on patient cost-sharing by type of services are calculated by a business intelligence agency (Vektis) based on insurance claims. Germany, for example, receives estimates of out-of-pocket spending on dentures directly by the umbrella organisation of SHI-affiliated dentists.
Health accounts are produced by a consortium led by a professor from Yonsei University (the lead institution), together with the Korean Institute for Health and Social Affairs (KIHASA), the National Health Insurance Service (NHIS) and the Health Insurance Review and Assessment Service (HIRA) based on annual contracts with the Ministry of Health and Welfare.
Work starts in July with the establishment of an annual workplan and a co‑operation plan between the members of the consortium, also reflecting on the previous year’s production cycle and potential for improvement. When it comes to data acquisition, the lead institution acquires and processes all data that are outside of the direct responsibility of other consortium partners. Hence, the lead institution collects, for example, in co‑operation with KIHASA all official statistics or public expenditure data from 250 public nationwide health centres.
Each consortium member processes the data it is responsible for in its own database but need to submit its output in the format defined by the lead institution. The lead institution will then perform a validity check of these inputs into the final calculations before processing the data further and the production of the final health accounts tables. Those are submitted to the International Health Accounts Team by end of April and could potentially be revised depending on the outcome of the validation exercise.
The sheer volume of data collected as part of health accounts requires the use of appropriate statistical software and data management tools to process the data and calculate results. There is a variety of IT tools employed across the 13 OECD health accounts teams, reflecting differences in the health data infrastructure in countries, corporate IT policies in the agencies responsible for the production of health accounts but possibly also the IT competences and preferences of health accounts team members. It is important that data processing is flexible enough to accommodate and incorporate data submissions from all data holders received in different formats. Most countries appear to use a mixed approach by using various IT tools for the different steps in the process from extraction, adjustment, calculation through to dissemination.
That said, a good number of countries including Austria, Canada, Korea, Spain, the United Kingdom and the United States rely predominantly on spreadsheets such as MS-Excel and link a vast number of tables. In the United Kingdom, some quality checks are done using Python, Austria uses SAS for some tasks and the United States also relies on Eviews and SAS for some production steps.
For very large datasets, particularly in the early data processing and cleaning steps, the use of spreadsheets may however become impractical. This can be the reason why in Chile, France, the Netherlands and Switzerland, health accounts are generally processed in R, an open-source and freely available programming language, through the different production steps.
Some countries predominantly rely on other software: Germany uses MS-Access for most of the processing including the final calculations, while auxiliary calculations (such as to construct allocation keys for services and providers) may be done in MS-Excel. In that country, using MS-Access also allows a for a seamless linkage to the Cost-of-illness database where data is processed using the same software. The approach is similar in Chile which uses MS-Access to process results from intermediate databases containing classifications of the HF and HP dimensions which are then linked to MS-Excel to generate HC distributions. Australia predominantly uses SAS to access the important AIHW databases such as MBS, as well as for health accounts calculations and the generation of output tables. Costa Rica uses the WHO’s Health Accounts Production Tool (HAPT), which is fed with the various data sources following adjustments made using MS Excel.
3.3.3. How to assure data quality?
Quality assurance is a vital part of the health accounts production process. This refers on the one hand to the input data – that is, the data that is used in the calculation of health expenditure figures – as well as the health accounts output – the health accounts results.
All countries reviewed in this report check the plausibility of their input data but the extent of how systematic and comprehensive this is can differ according to the general policy adopted but also the type of data source. There is generally sufficient confidence in data from public finance budgets or insurance funds since this data should have been audited before any possible publication. For other data sources, health accounts team may return to original data holders to have data verified before processing in case it looks unusual. Health accounts regularly undergo a rigorous series of quality tests of the results before any publication and dissemination. These can refer to checks for completeness and plausibility, be limited to the latest year of publication or the entire time series. In this context, the quality assurance tool and the validation work of the International Health Accounts Team (IHAT) has frequently been mentioned by experts as an important support tool (Box 3.2)
The International Health Accounts Team (IHAT) consists of health accounts experts of OECD, Eurostat and WHO. They jointly manage the annual Joint Health Accounts Questionnaire (JHAQ) collection on health expenditure and financing. The JHAQ data request is sent to the health accounts focal points in all 38 OECD countries and beyond and is the key instrument to populate the international databases on health expenditure and financing.
The JHAQ data request consists of an Excel-based template (one file for each year submitted) which includes six two‑dimensional worksheets, combining the various SHA 2011 dimensions, for example spending on health services by financing schemes (HCxHF). Macros are embedded in these excel files to flag to country respondents if data is inconsistent within one worksheet, across different worksheets or in case negative values are entered in any table or an atypical combination of two dimensions is reported (e.g. inpatient long-term care provided by diagnostic service providers).
When countries officially submit their JHAQ data, IHAT will analyse this data carefully. In addition to checking each submission for consistency, IHAT monitors the plausibility of health spending trends over time, on an aggregate level but also for key health spending categories. Typically, several rounds of consultation between IHAT and country respondents are needed before the data is finally validated by IHAT – this frequently also involves the correction and resubmission of some data tables.
Many countries also rely on their different types of advisory bodies that exist for quality assurance (see Section 3.2). The United Kingdom, for example, seeks feedback from their health accounts steering group before the release of new data in what is termed a “curiosity session”. The ad hoc Austrian statistical peer-review also has a clear quality assurance focus. Australia publishes annually a “data quality statement” with the release of new data. In Canada, preliminary results on forecasted data for year t‑1 and actuals (year t‑2 and revised data) are shared with provinces before publication to get feedback on the validity of these results.
In addition, many countries interviewed for this report also publish some methodological information accompanying their data (See Section 3.4). Since one of the key objectives of health accounts is improving the transparency of financing flows in the health system, health accounts themselves also need to be transparent. It also adds to the validity of data and can be considered as a quality assurance instrument – since it allows stakeholder to potentially point to accounting issues.
3.3.4. Ensuring sufficient capacity and knowledge transmission
The production of health accounts generally relies on a core team of experts with diverse backgrounds, such as in economics, statistics, accounting, IT and health policy. In the OECD countries interviewed for this report, the size of the core team varied from 1 to 12 people, with an average of 3 to 5. Differences in the size of the team can reflect on the one hand differences in the technical implementation of health accounts, such as the extent of data use, the degree of internal processing and the sophistication of the methodology applied, and on the other hand the scope of the implementation (including the range of different outputs produced), as well as the complexity of the health system. The size of the core team does not cover other staff located in other agencies that may contribute to the production of health accounts to varying extents. In many countries, core members of the health accounts team are not solely allocated to health accounts activities but are also contributing to other projects outside of the peak time of health accounts production which makes it challenging to directly compare the resources allocated to health accounts production.
In Spain, for example, the health accounts work in the Ministry of Health is carried out by only one person. However, some of the initial data collection and processing necessary for national health accounts is already carried out by the 17 autonomous regions.14 This is similar in the United Kingdom, where the allocation of government spending to providers and functions is done by entities of the four administrations (England, Northern Ireland, Scotland, Wales) with the ONS ensuring that this is aligned with UK-wide calculated control total.
On the other hand, a dozen people in the United States are part of the health accounts team. This comparatively large number can partly be explained by the fact that the National Health Expenditure Accounts is constructed on a much more detailed level than in most other countries and above what is required for international comparisons, including, for example, a break-down of health spending for all 50 states.
In Korea, each of the consortium members have 3‑4 staff working on health accounts.
In the majority of OECD countries interviewed, there has only been limited change in the overall size of health accounts teams but also a relative consistency among core member of the teams in recent years, which greatly contributes to a consistent production of health expenditure and financing data of high quality. That being said, most countries realise that some sort of documentation detailing the processes by which health accounts are produced is critical to enable a smooth transition without disruptions in health accounts production if there is staff turnover.
How these internal documentations look like varies in practice but generally, countries do not record all aspects of the health accounts work in a single handbook. One exception is Costa Rica, where a manual describing the production of health accounts is currently under development. It includes a detailed flowchart of the process and the various data sources and institutions involved with the purpose of ensuring continuity in case of staff turnover. The other countries rather seem to rely on the collection of separate documents or files which can be readily updated and revised as required. These are not intended to be published but the general nature of these documents would cover:
A documentation of processes – referring to the sequencing (and timelines) of different steps that need to be taken from data acquisition to disseminating the final results, including a list (or database) of data sources used with the data custodians, as well as a description and information on availability, timeliness, frequency of this data.
Comments on some key methodological decisions – referring to a summary of the internal justification on the reasoning on some particular decisions regarding the choice of data sources or calculation methods applied and an evaluation of the alternatives.
Comments on calculation and estimation methods – referring, for example, to notes on why certain calculations have been modified (this is included either in log-files, as comments on excel files or commented on in codes)
In addition, the methodological documents that many countries produce for external purposes also serve as vital background information for incoming new staff. Ideally, a hand-over period should be factored in when there is a change of personnel to ensure that proper training and knowledge transfer can take place between experienced staff and those fresh to the work. This is clearly of greater importance when the production of health accounts is reliant on a single person or a very small team. Circumstances may mean that an orderly hand-over is not always possible, stressing again the need for comprehensive and up-to-date documentation.
In summary, regarding the technical aspects of institutionalising the routine production of health accounts, there are a number of important aspects to reiterate. First, there should be clear operational procedures in place that set out the annual production cycle with a division of responsibilities among health accounts team members. This requires establishing a coherent production timeline from the initial acquisition of the various input data and their validation, the processing of data and the calculation of the health accounts, through to the final review of the results and analyses, and then publication as well as dissemination of the accounts and associated products. Of course, sufficient staff and adequate IT infrastructure is needed to accomplish this cycle. For a smooth regular production of health accounts, processes should be standardised as much as possible. This means, for example, data requests sent to data holders should ideally request the same submission format with minimal changes between years, and some continuity in the adjustment and calculation methodologies. Of course, reviewing and improving data sources, methods and processes (if necessary) for future production rounds based on past experience is a vital component of best practice in the institutionalisation of health accounts. Quality assurance at all steps in the production cycle, from validating input data to checking the plausibility of health accounts results is also key. Finally, comprehensive documentation of the different stages in the production process, the methodology used, and the various calculation steps can be very useful to sustain the process by facilitating staff handover and the training of new staff.
Health accounts should not be considered an end in themselves. While their primary purpose allows for the systematic reporting and analysis of health spending and financing across different financing sources, providers, and functions, the ultimate goal is to use the insights from health accounts together with other information to inform health policy, improve resource allocation, and ensure equitable and efficient use of health funds. To that aim, the dissemination of health accounts should ensure that their findings are accessible, understandable, and actionable for policy makers, researchers, and the general public.
3.4.1. Standard dissemination tools
All countries participating in the survey disseminate high level and/or detailed health spending data on an annual basis using a variety of different outlets, but the degree of this dissemination activity differs. In some instances, health accounts teams use the full suite of communication instruments available, to cater for different audiences and user groups:
There are two main instruments by which detailed health spending data are made available. Some countries create extensive data tables that can be downloaded by interested users. For example, the United States and Canada create publicly available zip-files for this purpose which include dozens of readymade spreadsheet files; Switzerland allows users to download all data back to 1960 in a single px-file (a standard format for statistics files used now by many statistical offices) and releases data via “data cubes” [STAT-TAB or.Stat]; Australia combines this information in a single excel-file. Others release their data via “data cubes” in various data platforms where users can interact directly and tailor to their specific needs. This is, for example, the procedure in the Netherlands, Germany or Austria.
In many instances, the publication of new health accounts data is accompanied by a short press release (1‑2 pages), issued by the agency where health accounts are institutionalised and targeted towards the media to raise awareness of health accounts and the latest results. This will often then be picked up and covered in both the mainstream and more specialist media. This is the case, for example, in Austria, Germany, the Netherlands or Switzerland with press releases occurring at least once a year. The media strategy of Destatis (Germany) is to release new data including estimates for year t‑1 around World Health Day (7 April) with a second release in June on the financing framework. In Austria, a first release occurs in February (covering data for year t‑2), with a second release in June with preliminary estimates for year t‑1.
A number of countries accompany the release of new data with more substantive analyses of around 5‑10 pages including visualisations, frequently in a html format. The “statistical bulletin” of the ONS in the United Kingdom (ONS, 2024[11]) or the “snapshot” of CIHI in Canada (CIHI, 2024[12]) serve as examples for this type of communication tool.
In a few countries, detailed health accounts analyses are disseminated, and results put into broader health policy context. These types of analytical reports cater more for informed users and experts. In France, the latest health accounts results are released in a substantial publication in September each year as part of its well-established dissemination strategy (Box 3.3). The “Health expenditure Australia” report issued by the AIHW is a 120‑page long publication and includes the key results but also contains very detailed background information on concepts used and definitions (AIHW, 2024[13]). In Korea, a National Health Accounts report is published annually, including figures consistent with the OECD health database.
In some cases, detailed methodological information is also made available in separate documents, accompanying the analytical release. This is, for example the case in Canada, Spain and the United States. In Switzerland, a note on methodological changes is made available. Related to this, Australia publishes annually a “data quality statement” with the release of new data. It covers sections on the institutional environment, timeliness, accessibility, interpretability, relevance, accuracy and coherence (AIHW, 2024[14]).
In person media briefings or press conferences are also used in some countries to promote health accounts results but are not necessarily set out as annual events. In the United States, for example, media briefings occur annually but since the pandemic as virtual events. In Germany, press conference for health accounts were organised occasionally over the past decades but this is, in general, exceptional. The launches were led by the President of the German Statistical Office.
The Direction de la Recherche, des Études, de l’Évaluation et des Statistiques (DREES) of the Ministry of Health in France has long been responsible for health accounts production. The most recent expenditure data including a revised time series is generally published each September of the following year, alongside a very detailed analysis of the health spending data and placing it in the context of important ongoing policy discussions. The 2024 edition of the French health account publication ran to nearly 300 pages (DREES, 2024[15]).
The timing of the release is important, not only dictated by data availability but chosen to have a tangible impact on health policy discussions: notably, at this end of September, the Social Security Financing Bill (Projet de loi de financement de la sécurité sociale (PLFSS)) for the following year is adopted in the Council of Ministers before being discussed in parliament. The Comptes de la Santé annual report is predominantly based on the French-specific version of NHA and covers detailed analysis of the trends for each individual healthcare service and financing arrangement (both in current and constant prices). It also includes substantial comparison of French health spending in an international context for which the SHA 2011 framework is used (the differences between the French NHA and the French SHA data are also explained at length in the report). Some methodological information on the health expenditure calculations is also covered in the annexes to the report. Finally, the publication always includes some pertinent and complementary health policy analysis which go beyond health spending. In the 2024 edition, for example, their analysis explored the implications of a policy reform to reduce co-payments for specific medical goods and the evolution of the compensation for sick leave.
The publication is released on the website of the Ministry of Health to align with the annual meeting of the health accounts commission (Commission des comptes de la santé). This meeting is a half-day event, frequently opened by the minister, and attended by members of the commission and invited experts (around 50‑80 people). In addition to presenting and discussing the findings of the last health accounts data, the agenda includes complementary items covering timely health policy topics (e.g. dental care, mental health, long-term care). The OECD regularly attends the annual Commission meetings providing an international perspective on some of the results and topics discussed.
Alternative means to communicate and publicise data and other material have become increasingly popular in recent years. Infographics are tools to visualise complex content in an easily accessible and understandable way. In Chile, France, the United States and Canada, the creation of online infographic content has become part of the communication strategy of the health accounts teams (Figure 3.3). Another, relatively modern, form of communication are social media channels. These are used, for example, in Canada, France and the United Kingdom in their media strategies to announce key health spending data.
In two countries, the health accounts team regularly publish health spending analyses and results in peer-reviewed scientific journals addressing the research community, which can provide some additional legitimacy to the data and the methodologies applied. In the United States, the latest analyses of results and trends (for years including year t‑1) are typically published the following December in Health Affairs. In the 2024 article, for example, one of the key messages of the health accounts team of CMS was that the share of GDP allocated to health in 2023 was similar to the one seen in 2019 (Martin et al., 2025[16]). Since health accounts production in Korea is led by a professor from Yonsei University, there is already a strong link to the academic world, and Professor Jeong has published policy analyses based on health accounts results in a number of Korean and international journals, including in Health Policy (Jeong, 2005[17]) and Health Affairs (Jeong, 2011[18]), demonstrating the applicability of health accounts.
Finally, health accounts results are frequently also disseminated in other publications which cover “health” more broadly, but which may have wider outreach than specific publications related to health financing. In Austria, for example, health spending data is included in the health statistics yearbook issued annually by the Statistical Office (Statistik Austria, 2024[19]). In Australia, health expenditure data is included in the biennial publication “Australia’s health” (AIHW, 2024[20]). In Spain, both SHA 2011‑based health accounts as well as health spending according to the national version are annually published in reports on the national health system (Ministerio de Sanidad, 2024[21]). The report includes, for example, a breakdown of private health expenditure by region, by sex and by educational attainment.
Source: Canadian Institute for Health Information (2024[22]) National Health Expenditure Database www.cihi.ca/en/national-health-expenditure-trends-2024-infographics#where
3.4.2. Informing policy and the use of health accounts results
A key objective of a systematic production of health accounts is that results and analyses are used to inform policy making. For this to happen, health accounts results and analyses need to be communicated to senior staff in ministries, ministers, as well as other decision-makers and key stakeholders in a way that can influence debates and policy discussions.
The first step in this process is to ensure that there is sufficient awareness and buy-in from senior officials at the time of release of the latest estimates. In countries where health accounts have not been institutionalised within ministries of health, the responsible authority will typically allow prior access to press notifications and reports under embargo to health ministries shortly before the official releases. In other cases, ministers are automatically informed about new health accounts data releases.
In the United Kingdom, staff within the Department of Health and Social Care produce a briefing note for the minister on the day health accounts data are release by ONS. In France, the annual meeting of the health accounts committee where the new annual report is discussed is typically opened by the minister, who will have been suitably briefed on the results beforehand (Box 3.3)
For the most part, the health accounts teams interviewed for this report indicated that they are in regular interactions with officials in health and other ministries or (staff of) parliamentarians – groups of people who can potentially contribute to translating health accounts information into policy. This communication can be successful since high-level health accounts results are also regularly cited in parliamentary debates. Other examples of where health accounts can make a notable impact relate to discussions with auditing courts, associations of health professionals, health providers or insurers – actors who all play an important role in influencing public opinion and can thus also influence policy making to a certain degree. Some examples, where health accounts contributed to shaping specific policies or have the potential to do so include:
In Korea, for example, the information from the health accounts has in the past served to inform pharmaceutical pricing policies and coverage policies under the National Health Insurance and provided evidence in the discussion on the sustainability of financing of public spending (Jeong, 2024[23]). The annual 10‑year forecast of National Health Expenditure carried out by CMS in the United States can also be seen in this context.
The timing of the release of the health spending report in France is chosen so the latest data can inform debates on the social security budget (Box 3.3). The most recent report also looked into the impact on policy initiatives to reduce co-payments. In Chile, health spending data is also currently informing discussions on how to address high out-of-pocket spending. In the Netherlands, international comparisons on long-term care spending have attracted some attention in recent years, as health accounts data identifies the country as the highest long-term care spender. In Spain, public health spending data (EGSP) data was used as the basis for an analysis of health spending presented at a conference of the Presidents of the 17 autonomous regions.
There is a particularly interesting practical application of health accounts in Austria, where the data is used to monitor the cost containment path and national and state spending targets (Box 3.4).
In some cases, health spending data influences national plans and strategies, such as in the case of Costa Rica, where the high levels of curative care spending highlighted the need for a greater focus on prevention in the National Health Policy 2023‑33, as well as in the National Plan for the Quality of Health Service Delivery 2024‑30, which included statistics on hospital and out-of-pocket spending.
In many OECD countries, an important use of SHA 2011‑based health accounts is international benchmarking. Important comparison indicators used in country-specific publications in this context are per capita health spending, health spending as a share of GDP, or health spending growth in real terms, the public-private financing split and the share of out-of-pocket spending.
As mentioned in Chapter 2, a number of countries go beyond the standard application of health accounts to inform on country-specific policy needs:
In Australia, there is great interest in monitoring how the health system performs to meet the healthcare needs of their indigenous peoples. In this context, indicators stemming from the national health accounts are used to populate the Aboriginal and Torres Strait Islander Health Performance Framework (AIHW, 2023[24]). In this assessment, health spending of the indigenous population is compared with the non-indigenous population, and broken down by type of service, funding and state or territory (Figure 3.4).
A number of countries, including Australia, Canada, Chile, Spain and the United States monitor health spending at a subnational level, as an indicator to measure equitable access to healthcare across regions. In the United States, for example, comparing personal health spending across states highlights that the per capita spending in the state of Utah is only half that seen in New York state or D.C. (Figure 3.5). The health accounts team also has analysed these results to explain the observed differences, and found that, over time, personal income per capita and the percentages of the population enrolled in Medicare and Medicaid were the key variables explaining the variation in state health spending (CMS, 2022[25]).
The Austrian health system is characterised by a complex mix of responsibilities between the federal government, the state‑level governments and the social health insurances funds when it comes to the organisation and financing of the system. Article 15a of the Constitutional law enables the federation and states to conclude agreements with each other on matters within their respective spheres of competence.
Based on this legal provision, in 2013, the federation, states and social health insurance funds concluded such an agreement pertaining to the target-based health governance for the planning, organisation and financing of healthcare (“Zielsteuerung Gesundheit”).
Defining financial targets for the states and social health insurance funds is an important element of this Target-based Governance Agreement; in its current version, financial targets are set until 2028. The agreed spending growth path in the Target-based Governance Agreement refers to public and also private spending based on SHA (excluding long-term care) and detailed SHA (unpublished) data and analyses are used to monitor whether targets are met. This refers to, for example, spending on public hospitals by federal state, social health insurance spending by legal entity, a detailed breakdown of federal healthcare expenditure, and a detailed breakdown of expenditure for hospitals on a 2nd digit level of HP.
Total health funding for Indigenous and non-Indigenous Australians, per person by area of funding, 2019‑20, in AUD
Source: AIHW (2023[24]), Aboriginal and Torres Strait Islander Health Performance Framework – 3.21 Expenditure on Aboriginal and Torres Strait Islander health compared to need, www.indigenoushpf.gov.au/measures/3-21-expenditure-on-aboriginal-and-torres-strait-i#keymessages.
Total all payers per capita state estimates by state of residence, 2020, personal healthcare, in USD
Source: CMS (n.d.[26]), National Health Expenditure Data: Health Expenditures by State of Residence, 1991‑2020, www.cms.gov/data-research/statistics-trends-and-reports/national-health-expenditure-data/state-residence (accessed on 11 January 2025).
Another important additional extension of health accounts work is allocating health spending by disease, age and gender. This helps identify the economic burden of disease of a health system and can give indications for priority setting and resource allocation. Spending by disease is regularly reported in Australia, France, Korea or the Netherlands. However, this stream of work is not necessarily institutionalised within the same team (or even the same agency) as the health accounts implementation and may therefore use different health expenditure boundaries. This is not the case in Germany, where disease accounts are produced on a regular basis and aligned with the German national health accounts framework Table 3.3). Here, in a first step, health spending is produced based on the three core dimensions (financing schemes, services and providers). In a second step, using complex allocating models, for each health provider, health spending is broken by diseases (based on ICD‑10), gender and age. In Costa Rica, spending by disease category is based on the DIS classification used as part of the WHO HAPT tool (Figure 3.6). The DIS classification is itself linked to the International Classification of Diseases (ICD). Data is available for years 2018‑20, with spending from the 2021‑22 period currently being estimated. The aim is to consolidate the production of spending by disease estimates as part of the annual cycle.
Overall, while in many OECD countries the use of health accounts to provide evidence to design health policy could be further strengthened, it needs to be remembered that alternative data sources to health accounts may exist that can perform this role more appropriately. For example, for an assessment of the financial sustainability of a particular national health insurance fund, a country may rely on the detailed financial accounts of this fund rather than the more aggregated results drawn from health accounts.
Health expenditure in medical practices (HP.3.1) in Germany, all age groups, 2020, in million EUR
Note: Disease classification is based on ICD‑10.
Source: Gesundheitsberichterstattung des Bundes (n.d.[27]), Krankheitskosten in Mio. € für Deutschland. Gliederungsmerkmale: Jahre, Geschlecht, ICD‑10, Einrichtung, (accessed on 11 January 2025).
Health expenditure on NCDs in Costa Rica, 2020
Note: Disease classification is based on DIS classification (WHO HAPT).
Source: Adapted from Ministry of Health, Costa Rica (2025[28]) Cuentas de salud, www.ministeriodesalud.go.cr/index.php/biblioteca/material-educativo/material-publicado/indicadores-en-salud/indicadores-de-servicios-de-salud/indicadores-economicos-en-salud/cuentas-de-salud.
3.4.3. Interacting with the public and creating a positive feedback loop
In addition to delivering evidence and information crucial to health policy decision making or the monitoring of reforms, health accounts results should also be readily accessible to academia and the public at large. Health accounts should provide data to support scientific research and a better understanding of the financing flows in the health systems. Broad public access to this data is therefore an important element to promote transparency and accountability.
Researchers and the general public, however, do not need to remain passive consumers of this data. In some countries, health accounts teams have opened up channels for the public to interact with those responsible for compiling the health accounts. This type of interaction can reinforce the validity or legitimacy of the accounting results and in some instances may also improve its quality.
Across countries interviewed for this report, any press communication or websites promoting health accounts results is accompanied with contact details -such as an (generic) email address or phone numbers of team members – for questions and overall comments. This feedback mechanism is considered valuable since it can point to misunderstandings in the press material which may be improved in future press communication (e.g. for example using simpler language and avoiding technical jargon). The research community or the general public may also identify potential weaknesses in health accounting systems.
In addition, the existing advisory boards in many countries (as described in Section 3.2) can also serve as a more formalised feedback mechanism. Yet, there are also other formalised feedback mechanism. In Canada, for example, the National Health Expenditure work undergoes a regular product review every 4‑5 years to assess whether the products are still relevant. There is a similar procedure in place in the United States.
In summary, the regular and timely dissemination of health expenditure and financing data is of crucial importance to fulfil the objectives of health accounts – such as increasing the transparency of financing flows in the health systems, enhancing the accountability of the various actors in the health system, and providing evidence to inform health policy. Good dissemination practices include the development of products for various audiences with different level of expertise. Describing some high-level results in a press release can be sufficient for the general public to raise awareness but publishing more in-depth analytical pieces – potentially also putting health expenditure data in the wider health policy context – is desirable to increase the impact or the data. Getting buy-in and the active participation of senior decision makers or major stakeholders in the dissemination process can also be a useful way to signal that health accounts data are an import source to inform policy making. Providing access to all health accounts results (on a website) to interested users and publishing methodological notes is standard practice to increase the transparency of the data.
While most OECD countries already engage in many of these activities, some teams may be able to potentially do more to advocate the use of health accounts data in the policy-making process, but it is understood that this can be a challenge. However, having regular interactions and developing a working relationship with staff in ministries and other stakeholders may help to raise awareness of the potential of health accounts to contribute to health policy discussions.
Finally, experts from countries participating in the structured interviews informing this report were invited to share their perspectives on the crucial factors for a successful institutionalisation of health accounts. Some common threads can be discerned, but country-specific factors can also play a role. Experts frequently mentioned the following key aspects:
Focus on the production of high-quality data to add value: Experts pointed to a possible positive feedback loop between the production of insightful, robust health spending data and its perceived value, and by consequence its use to inform evidence‑based policy making. In other words, the use of health accounts can be an incentive to continue producing and further improve data. In this context, the timeliness of data and its perceived impartiality are also important elements to strengthen its worth. Moreover, dissemination should go beyond simply providing key figures by also putting the results into the wider policy context.
Liaising with stakeholders and policy makers is key: Garnering the attention from key stakeholders and policy makers was considered as another decisive factor for a successful institutionalisation, since these groups can be instrumental in reinforcing the positive feedback loop described above. If stakeholders are involved to some degree in the production process, it is more likely that they will “buy-in” and acknowledge the importance and legitimacy of health accounts and promote their use.
Establishing and maintaining a fruitful co‑operation with data providers: Many experts emphasised that building a close co‑operation with data providers, both public and private, as well as a developing a deep understanding of the data landscape, plays a major role in producing health accounts of high-quality in a timely and consistent manner. While collaborating with data holders should be uncomplicated and straight-forward in most cases, more time may be needed to establish a (formal or informal) working arrangement in some instances, where data providers may need to be convinced that co‑operating may be beneficial for both sides.
Some other key aspects for a successful institutionalisation were also mentioned. These include, for example ensuring the continuity of health accounts production. This can be supported by having a legal mandate for producing health accounts, which should facilitate data exchanges with data providers and other stakeholders. Yet, other, related, factors are also important to sustain the process in the long term. This includes allocating sufficient resources to the health accounts work, as well as establishing dedicated units responsible for the health accounts production (within the designated agency) with the appropriate mix of expertise.
Finally, experts mentioned various methodological and technical aspects that facilitate a successful health accounts institutionalisation. This refers, for example, to establishing efficient operations by collecting standardised input data and automating processes, and striving for a continuous improvement in sources and methodology applied. In the initial implementation phase, following an iterative approach in constructing health accounts, focusing first on the “low-hanging fruits” before progressively filling gaps when feasible was a recommended strategy. Documenting the production process – for both internal and external purposes – was also emphasised as an important aspect in the institutionalisation of health accounts.
References
[20] AIHW (2024), Australia’s health 2024, Australian Institute of Health and Welfare, https://www.aihw.gov.au/getmedia/6b19e493-0ebe-420f-a9a3-e48b26aace9f/aihw-aus-249-ib.pdf?v=20240715154739&inline=true.
[13] AIHW (2024), Health expenditure Australia 2022–23 - web report, Australian Institute of Health and Welfare, https://www.aihw.gov.au/getmedia/78531cd2-900a-43b6-97b1-8072b441de6c/health-expenditure-australia-2022-23.pdf?v=20241120164118&inline=true (accessed on 10 January 2025).
[14] AIHW (2024), Health expenditure database 2022–23; Quality Statement, Australian Institute of Health and Welfare, https://meteor.aihw.gov.au/content/799593 (accessed on 11 January 2025).
[24] AIHW (2023), Aboriginal and Torres Strait Islander Health Performance Framework - 3.21 Expenditure on Aboriginal and Torres Strait Islander health compared to need, Australian Institute of Health and Welfare, https://www.indigenoushpf.gov.au/measures/3-21-expenditure-on-aboriginal-and-torres-strait-i#keymessages (accessed on 10 January 2025).
[9] AIHW (n.d.), Health Expenditure Advisory Committee, Australian Institute of Health and Welfare, https://www.aihw.gov.au/about-us/committees/health-expenditure-advisory-committee (accessed on 7 January 2025).
[6] BFS (2024), Statistisches Mehrjahresprogramm des Bundes 2024-2027, Bundesamt für Statistik, https://www.bfs.admin.ch/bfs/de/home/bfs/oeffentliche-statistik/mehrjahresprogramm.assetdetail.30127368.html.
[4] Bundesamt für Statistik (2024), Kosten und Finanzierung des Gesundheitswesens - Steckbrief, https://www.bfs.admin.ch/bfs/de/home/statistiken/gesundheit/erhebungen/cou.assetdetail.31786686.html (accessed on 4 January 2025).
[22] Canadian Institute for Health Information (2024), National Health Expenditure Database, https://www.cihi.ca/en/national-health-expenditure-trends-2024-infographics#where.
[7] CIHI (2024), National Health Expenditure Trends 2024 - Methodology Notes, Canadian Institute for Health Information, https://www.cihi.ca/sites/default/files/document/nhex-trends-2024-meth-notes-en.pdf.
[12] CIHI (2024), National health expenditure trends, 2024 — Snapshot, Canadian Institute for Health Information, https://www.cihi.ca/en/national-health-expenditure-trends-2024-snapshot (accessed on 10 January 2025).
[8] CIHI (2011), Health Care Cost Drivers: The Facts, Canadian Institute for Health Information, https://secure.cihi.ca/free_products/health_care_cost_drivers_the_facts_en.pdf.
[25] CMS (2022), Econometric Analysis of Health Expenditures: Methodology and Model Specification, https://www.cms.gov/files/document/econometric-analysis-state-health-expenditures-methodology-and-model-specification.pdf (accessed on 11 January 2025).
[26] CMS (n.d.), National Health Expenditure Data: Health Expenditures by State of Residence, 1991-2020, https://www.cms.gov/data-research/statistics-trends-and-reports/national-health-expenditure-data/state-residence (accessed on 11 January 2025).
[15] DREES (2024), Les dépenses de santé en 2023 - Résultats des comptes de la santé, DREES, https://drees.solidarites-sante.gouv.fr/sites/default/files/2024-11/CNS24.pdf.
[27] Gesundheitsberichterstattung des Bundes (n.d.), rankheitskosten in Mio. € für Deutschland. Gliederungsmerkmale: Jahre, Geschlecht, ICD-10, Einrichtung, https://www.gbe-bund.de/gbe/!isgbe.indikatoren_set_page?p_uid=gast&p_aid=1461769&p_sprache=D&p_help=2&p_indnr=64&p_ansnr=82111364&p_version=2&p_thema_id=26890&p_thema_id2=1&p_thema_id3=&p_thema_id4=&D.000=3742&D.003=1000004&D.734=4504 (accessed on 11 January 2025).
[5] Goss, J. (2022), “Health Expenditure Data, Analysis and Policy Relevance in Australia, 1967 to 2020”, International Journal of Environmental Research and Public Health, Vol. 19/4, https://doi.org/10.3390/ijerph19042143.
[23] Jeong, H. (2024), Health Accounts for Public Use: Korean case.
[18] Jeong, H. (2011), “Korea’s National Health Insurance—Lessons From The Past Three Decades”, Health Affairs, Vol. 30/1, pp. 136-144, https://doi.org/10.1377/hlthaff.2008.0816.
[17] Jeong, H. (2005), “Health care reform and change in public–private mix of financing: a Korean case”, Health Policy, Vol. 74/2, pp. 133-145, https://doi.org/10.1016/j.healthpol.2004.12.017.
[1] Maeda, A. et al. (2012), Creating Evidence for Better Health Financing Decisions - A Strategic Guide for the Institutionalization of National Health Accounts, World Bank, http://hdl.handle.net/10986/13141.
[16] Martin, A. et al. (2025), “National Health Expenditures In 2023: Faster Growth As Insurance Coverage And Utilization Increased”, Health Affairs, Vol. 44/1, https://doi.org/10.1377/hlthaff.2024.01375.
[21] Ministerio de Sanidad (2024), Informe Anual del Sistema Nacional de Salud 2023, https://www.sanidad.gob.es/estadEstudios/estadisticas/sisInfSanSNS/tablasEstadisticas/InfAnualSNS2023/INFORME_ANUAL_2023.pdf (accessed on 11 January 2025).
[28] Ministry of Health, Costa Rica (2025), Cuentas de Salud, https://www.ministeriodesalud.go.cr/index.php/biblioteca/material-educativo/material-publicado/indicadores-en-salud/indicadores-de-servicios-de-salud/indicadores-economicos-en-salud/cuentas-de-salud.
[11] ONS (2024), “Healthcare expenditure, UK Health Accounts: 2022 and 2023”, Statistical Bulletin, https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/healthcaresystem/bulletins/ukhealthaccounts/2022and2023 (accessed on 10 January 2025).
[29] PAHO (2023), Best Health Accounting Practices Using SHA 2011, Pan American Health Organization, https://iris.paho.org/handle/10665.2/57137.
[19] Statistik Austria (2024), Jahrbuch des Gesundheitsstatistik 2022, Statistik Austria, https://www.statistik.at/services/tools/services/publikationen/detail/1808.
[10] Statistik Austria (2020), Standard-Dokumentation Metainformationen zu Gesundheitsausgaben nach System of Health Accounts für Österreich, https://www.statistik.at/fileadmin/shared/QM/Standarddokumentationen/VW/std_v_gesundheitsausgaben.pdf (accessed on 11 January 2025).
[2] WHO (2023), Framework for assessing maturity of health accounts institutionalization, World Health Organization, https://iris.who.int/handle/10665/373992.
[3] WHO (forthcoming), Assessment of the maturity of the institutionalization of health accounts: Results of a Global Survey, World Health Organization.
← 1. A recent report published by the Pan-American Health Organization provides guidance on the technical aspect of institutionalisation of Health Accounts (PAHO, 2023[29])
← 2. Australia, Austria, Canada, Chile, Costa Rica, France, Germany, Korea, the Netherlands, Spain, Switzerland, the United Kingdom, the United States.
← 3. Out of the 38 OECD countries, 22 are also Member States of the European Union. Another three EU Member States (Bulgaria, Croatia, and Romania) are currently in the process of accession to the OECD.
← 8. A full analysis of the various methodologies of how health accounts have been implemented in OECD countries is beyond the scope of this report. Some countries provide detailed methodological papers on this. For all OECD countries, the key data sources used in the calculation of health expenditure are included in the metadata files annually submitted via the JHAQ data collection. They can be accessed through the OECD data platform http://data-explorer.oecd.org/s/1o4.
← 9. The agreement with the Banque de France has a legal basis in the French Social Security Code. Article L862‑7 of the Social Security Code requires the administration to produce an annual report on complementary health organisations. It is on this basis that the Ministry of Health has signed an agreement with the Banque de France.
← 11. Verordnung zur Durchführung der Erhebungen nach dem Gesundheitsausgaben- und -personalstatistikgesetz www.recht.bund.de/bgbl/1/2023/369/regelungstext.pdf?__blob=publicationFile&v=2.
← 12. General Health Law 14/1986 (further modified by Laws 21/2001 and 62/2003).
← 13. In the JHAQ collection, OECD countries are expected to submit SHA 2011‑based health expenditure data for year t‑2 by end of March in year t, or shortly thereafter. Preliminary estimates for year t‑1 can be submitted after the deadline.
← 14. The health system in Spain is decentralised and the responsibility of financing and healthcare delivery generally lies with the 17 autonomous regions.