Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
National Reference Metadata in Euro SDMX Metadata Structure (ESMS)
Compiling agency: Ministry of HealthDirectorate of research, evaluation, studies and statistics (Direction de la recherche, de l'évaluation, des études et des statistiques)
Ministry of Health Directorate of research, evaluation, studies and statistics (Direction de la recherche, de l'évaluation, des études et des statistiques)
1.2. Contact organisation unit
Sub-directorate of Health Statistics (Sous-direction Observation de la santé et de l'Assurance-maladie)
1.3. Contact name
Confidential because of GDPR
1.4. Contact person function
Confidential because of GDPR
1.5. Contact mail address
DREES
78-84 Rue Olivier de Serres
75015 Paris
FRANCE
1.6. Contact email address
Confidential because of GDPR
1.7. Contact phone number
Confidential because of GDPR
1.8. Contact fax number
Confidential because of GDPR
2.1. Metadata last certified
15 June 2026
2.2. Metadata last posted
14 July 2026
2.3. Metadata last update
11 June 2026
3.1. Data description
Statistics on healthcare non-expenditure provide information on healthcare human resources, healthcare facilities, and healthcare utilisation.
The people active in the healthcare sector (doctors, dentists, nurses, etc.) and their status (graduates, practising, migration of doctors and nurses, etc.);
The available healthcare technical resources and facilities (hospital beds, beds in residential care facilities, medical technology, etc.);
The health activities or patient contacts undertaken (hospital discharges, surgical procedures, ambulatory care data, etc.).
Annual national and regional data are provided in absolute numbers or as a rate of a relevant population.
Data are based mainly on administrative records (see section 18.1 ‘Source data’ for more information).
3.2. Classification system
For the collection data on healthcare non-expenditure, the classifications used in the System of Health Accounts (SHA) and its related set of International Classification for the Health Accounts (System of Health Accounts) are applied.
For Hospital discharges, the International Shortlist for Hospital Morbidity Tabulation (ISHMT) is used.
For Health Employment, the Directive 2005/36/EC of the European Parliament and of the Council of 7 September 2005 on the recognition of professional qualifications apply (Directive - 2005/36 - EN - EUR-Lex (europa.eu)).
Definitions of mandatory variables are laid down in Commission Regulation (EU) 2022/2294.
Where possible, the statistics are separated by sex (male/female), age group and NUTS2 region.
National deviations: see Annex at the bottom of the page.
3.3. Coverage - sector
Public Health
3.4. Statistical concepts and definitions
The healthcare non-expenditure statistics describe the public health sector from a non-monetary perspective. The statistics explain the number or rate of different healthcare resources, facilities and utilisations. A wide range of indicators are collected from a multitude of sources and therefore, details pertaining to individual variables are given in the Annex.
Definitions of mandatory variables are laid down in Commission Regulation (EU) 2022/2294.
The data collection process was organised so that :
the Ministry of Health has no access to information directly identifying the individuals, except those from health professionals directories publicly available ;
all transmissions of information are made according to protocols respecting the state of the art for security ;
the servers of the Ministry of Health are secured and protected by state of the art security protocols. Especially, the list of persons having access to microdata files is restricted to authorized persons.
8.1. Release calendar
Main indicators data are disseminated according to a release calendar.
The statistics commit to professional ethics of the national public statistics services, whose principles—codified and shared at the European level—are professional independence, commitment to quality, reliability, accessibility, respect for statistical confidentiality, impartiality, and objectivity.
Statistics are mostly derived from :
administrative data (full access to exhaustive data warehouses on health insurance reimbursements for resources and activities, social contributions and taxes for income and employment, health professional registers, etc.); these data benefit from regular control for administrative use; In most cases, statistics derived from these sources are treated as official statistics in France, either because they are produced within the Official Statistical Service (Insee and ministerial statistical departments – national statistical authorities / ONAs in the ESS) and therefore comply with the European Statistics Code of Practice, or because they have been formally accredited (“labelled”) by the French Authority for Official Statistics (Autorité de la statistique publique)
surveys labelled by the Official Statistics Quality Label Committee that verifies their technical quality, following the same standards and requirements as those implemented for European harmonised surveys
11.2. Quality management - assessment
The quality of the data meets the requirements for accuracy, timeliness and punctuality, comparability and consistency.
When evaluating statistical indicators, the quality of the obtained information is analyzed. The results of the calculation are compared with the results of the previous year. Outstanding values of indicators are identified, analyzed and explained. Significant deviations trigger a clarification loop with data providers; corrections/explanations are recorded in metadata to ensure traceability. Cross-source coherence checks are performed when possible, and Eurostat validation feedback is systematically reviewed and addressed.
12.1. Relevance - User Needs
Main users of national healthcare data are researchers, policy makers, civil society, and the media. The data are shared with WHO, OECD and Eurostat.
However, user needs and the relevance of official statistics are regularly assessed through structured stakeholder dialogue.
For statistical surveys produced within the French Official Statistical Service, the opportunity, scope and burden of surveys are routinely discussed with the CNIS (National Council for Statistical Information), which brings together representatives of administrations, social partners, researchers, professional organisations and civil society. These exchanges take place notably during the (re-)labelling/authorisation process for surveys, and may lead to adjustments in questionnaires, definitions, coverage or periodicity, and to improvements in dissemination.
For administrative data sources, information systems and data warehouses are regularly enhanced to address emerging policy and research needs. User requirements expressed by researchers and institutional users are channelled through the Official Statistical Service and dedicated expert groups/working arrangements with data owners, which can result in new variables, improved classifications, better linkage possibilities, and strengthened documentation/metadata.
12.3. Completeness
All the requirements laid down in Commission Regulation (EU) 2022/2294 are fulfilled.
Exceptions concern timeliness for a limited number of series: (i) medical, dentistry and pharmacy graduates, for which France benefited from a derogation (Implementing Decision (EU) 2022/2306) up to N+21 months for reference years 2021–2023 and similar delays may also affect reference year 2024 due to source-data consolidation issues; and (ii) “Number of practising nurses”, currently provided with a one-year delay due to data availability constraints linked to the ADELI-to-RPPS transition (from 2022) and the ongoing reconstruction of the series using alternative sources.
13.1. Accuracy - overall
The overall accuracy of healthcare non-expenditure data depends on the accuracy of the data used to compile them. Almost all indicators are compiled from exhaustive administrative data (registers/claims/management systems) or from surveys produced within the Official Statistical Service, which follow established quality standards. For administrative sources, accuracy is supported by high coverage and routine validation checks; for surveys, accuracy is monitored through standard survey quality processes and consistency checks over time.
13.2. Sampling error
Sampling error is negligible for most variables, as they are derived from exhaustive administrative sources (no sampling). Where indicators are based on sample surveys, sampling error is managed through standard survey design and estimation practices (e.g. appropriate sample design, use of sampling weights and calibration).
13.3. Non-sampling error
For administrative sources, potential non-sampling errors mainly relate to coverage/definition issues, reporting or coding errors, and changes in administrative rules or information systems; these are mitigated through consistency controls and time-series/outlier analysis, and are documented in metadata when they affect comparability.
For survey-based indicators, non-sampling errors may include non-response and measurement errors; they are addressed through standard procedures (follow-ups, editing, weighting adjustments/calibration and, when relevant, imputation), and through coherence checks over time and across related variables.
14.1. Timeliness
Majority of final data for the year N is provided to international organisations at N+14 months. France has been however granted a derogation by the Implementing decision - 2022/2306 as concerns the timeliness of medical, dentistry and pharmacy graduates, up to N+21 months for reference years 2021-2023. Although this derogation formally ends with reference year 2023, similar delays may also affect reference year 2024 due to consolidation issues in the source data transmitted by the Ministry in charge of Higher Education. In addition, the indicator “Number of practising nurses” is currently provided with a one-year delay due to data availability constraints and the ongoing reconstruction of the series following the ADELI-to-RPPS transition.
14.2. Punctuality
Punctuality is generally very good. Nearly all datasets and metadata are transmitted within the legal deadlines, as the compilation mainly relies on regularly produced administrative and management information systems with stable annual cycles (e.g. health insurance claims and reimbursement databases, hospital discharge and emergency care information systems, annual hospital activity and capacity reporting, national registers of health professionals, and national registers of healthcare facilities). In addition, some indicators come from official surveys with established fieldwork and release calendars (e.g. surveys on training schools for health professions, population health interview surveys).
Punctuality is supported by internal production schedules aligned with the European transmission calendar and by routine validation prior to delivery. Delays remain limited to a small number of series, mainly when upstream source data are not consolidated in time (e.g. administrative sources on higher education graduates) or when major source transitions require additional processing to ensure reliability (e.g. reconstruction of the practising nurses series). In such cases, delays are documented in metadata.
15.1. Comparability - geographical
Data covers metropolitan France and overseas departments and regions (DROM). Statistics are broadly comparable across territories when based on nationwide administrative sources; however, some geographical comparability and coverage limitations remain for specific sources and variables:
DROM are excluded from EHIS coverage, but since 2019 France has set up a complementary survey to extend its coverage to the DROM.
Other surveys exclude the coverage of Mayotte
Administrative data coverage (health insurance reimbursements, payrolls) is also quite incomplete for Mayotte.
These limitations are documented in the metadata in annex.
15.2. Comparability - over time
On the whole, the data are comparable over time. Potential breaks in series (e.g. changes in definitions, coverage, classifications or source systems) are monitored through annual time-series checks and are documented variable by variable in the Annex at the bottom of the page.
15.3. Coherence - cross domain
Cross-domain coherence is generally good, as many indicators rely on the same national administrative sources and reference registers, ensuring consistent concepts and coverage across related domains. The main area where differences may be observed across domains concerns health workforce indicators, where figures may come from different sources depending on the concept measured (e.g. professional registers or claims/payroll sources for headcounts of some practising professionals versus establishment reporting for hospital employment in full-time equivalents). These differences reflect conceptual scope and source-specific measurement rather than inconsistencies. Relevant limitations and source choices are documented in national documentation (e.g. : Évolution des effectifs salariés hospitaliers depuis 15 ans - Méthodologie de construction d’agrégats nationaux | Direction de la recherche, des études, de l'évaluation et des statistiques or “Démographie des infirmières et des aides-soignantes” )
15.4. Coherence - internal
Internal coherence is ensured by using the definitions and classifications recommended in the Joint Questionnaire and by applying routine consistency checks (e.g. totals vs breakdowns, consistency across classifications). Any inconsistencies identified during production or validation are investigated and, when necessary, corrected; otherwise they are documented in the metadata/Annex.
In addition, revisions follow the general principles of official statistics (transparency, traceability and documentation of changes). For the indicators reported in this questionnaire, revisions are not part of a regular revision cycle. When revisions occur, they mainly reflect explicit methodological changes (changes in definition, concept, coverage, classification or source), rather than late reporting or routine data consolidation. Such changes are systematically documented in the Annex (date, series concerned and nature of the change).
17.2. Data revision - practice
Revisions are relatively limited. When a change affects a published series, the date, variables concerned and reason for the revision (e.g. definition/coverage/source change) are explicitly documented in the Annex at the bottom of the page (and, where relevant, in national documentation).
18.1. Source data
The data reported in this questionnaire are mainly derived from administrative data processes (national registers and information systems covering healthcare facilities, hospital activity, health insurance claims/reimbursements and health professional registers) and from official surveys (population health interview surveys and surveys on health training schools). For most variables, a single primary source is used. For a limited number of indicators, a multisource process is applied in order to improve coverage or align with the required concept (e.g. combining claims/insurance data with payroll/employment information for specific workforce indicators). Source data are extracted, validated (completeness, plausibility and consistency checks), aggregated to the required breakdowns and transmitted according to the Joint Questionnaire definitions. A variable-by-variable description of the source, coverage and processing is provided in the Annex at the bottom of the page.
18.2. Frequency of data collection
Annual
18.3. Data collection
Health non-expenditure statistics are compiled from multiple administrative and survey sources. Administrative data are primarily collected for operational/management purposes by the responsible institutions and extracted annually for statistical production; data are transmitted to the statistical producer through established channels and schedules. Survey-based data are collected through standard official survey operations (fieldwork, editing and weighting procedures) following their dedicated calendars. For the Joint Questionnaire, data collection consists of coordinating yearly extracts, performing validation checks (completeness, plausibility and consistency), and compiling the required breakdowns prior to transmission. Detailed information on the collection mode and procedures is provided for each variable/source in section 10.6 and in the Annex.
18.4. Data validation
Prior to submission, a set of national validation checks is performed. This includes completeness checks, format/range checks, and internal consistency checks (e.g. totals versus breakdowns, consistency across classifications and time periods). Time-series plausibility checks and outlier detection are carried out by subject-matter experts; major year-to-year changes are investigated and documented (source change, coverage/definition change, or genuine trend). When errors or inconsistencies are identified, data are corrected and the final dataset is validated before transmission. In addition, the built-in checks in the Excel questionnaire and the validation feedback from Eurostat/EDAMIS are reviewed and addressed prior to final submission.
18.5. Data compilation
Aggregate figures are compiled according to the definitions and guidelines of the Joint Questionnaire. Most indicators are derived from exhaustive administrative sources and are transmitted as aggregates; therefore, no design weights are applied and imputation is generally not used.
Full-time equivalents (FTE) are produced only for hospital employment indicators, using the annual hospital survey on healthcare facilities (SAE): for salaried staff, establishments report annual paid FTE (average over the year), which are then aggregated to the required categories; for self-employed medical staff, FTE are not directly available and are estimated from headcounts using conventional conversion rules (as documented in the Annex).
Statistics on healthcare non-expenditure provide information on healthcare human resources, healthcare facilities, and healthcare utilisation.
The people active in the healthcare sector (doctors, dentists, nurses, etc.) and their status (graduates, practising, migration of doctors and nurses, etc.);
The available healthcare technical resources and facilities (hospital beds, beds in residential care facilities, medical technology, etc.);
The health activities or patient contacts undertaken (hospital discharges, surgical procedures, ambulatory care data, etc.).
Annual national and regional data are provided in absolute numbers or as a rate of a relevant population.
Data are based mainly on administrative records (see section 18.1 ‘Source data’ for more information).
11 June 2026
The healthcare non-expenditure statistics describe the public health sector from a non-monetary perspective. The statistics explain the number or rate of different healthcare resources, facilities and utilisations. A wide range of indicators are collected from a multitude of sources and therefore, details pertaining to individual variables are given in the Annex.
Definitions of mandatory variables are laid down in Commission Regulation (EU) 2022/2294.
National changes of statistical concepts and national definitions deviating from Regulation (EU) 2022/2294: see Annex at the bottom of the page.
Registered health professionals or health care facility categories.
All health care staff
All available beds or equipment in hospitals or in nursing and residential care facilities
All discharges or procedures performed in all hospitals
Complete national territory
Calendar year; depending on the data set this can be annual average data or data as reported by 31st December.
The overall accuracy of healthcare non-expenditure data depends on the accuracy of the data used to compile them. Almost all indicators are compiled from exhaustive administrative data (registers/claims/management systems) or from surveys produced within the Official Statistical Service, which follow established quality standards. For administrative sources, accuracy is supported by high coverage and routine validation checks; for surveys, accuracy is monitored through standard survey quality processes and consistency checks over time.
The data are published in absolute numbers or rate per inhabitants.
Aggregate figures are compiled according to the definitions and guidelines of the Joint Questionnaire. Most indicators are derived from exhaustive administrative sources and are transmitted as aggregates; therefore, no design weights are applied and imputation is generally not used.
Full-time equivalents (FTE) are produced only for hospital employment indicators, using the annual hospital survey on healthcare facilities (SAE): for salaried staff, establishments report annual paid FTE (average over the year), which are then aggregated to the required categories; for self-employed medical staff, FTE are not directly available and are estimated from headcounts using conventional conversion rules (as documented in the Annex).
The data reported in this questionnaire are mainly derived from administrative data processes (national registers and information systems covering healthcare facilities, hospital activity, health insurance claims/reimbursements and health professional registers) and from official surveys (population health interview surveys and surveys on health training schools). For most variables, a single primary source is used. For a limited number of indicators, a multisource process is applied in order to improve coverage or align with the required concept (e.g. combining claims/insurance data with payroll/employment information for specific workforce indicators). Source data are extracted, validated (completeness, plausibility and consistency checks), aggregated to the required breakdowns and transmitted according to the Joint Questionnaire definitions. A variable-by-variable description of the source, coverage and processing is provided in the Annex at the bottom of the page.
Annual
Majority of final data for the year N is provided to international organisations at N+14 months. France has been however granted a derogation by the Implementing decision - 2022/2306 as concerns the timeliness of medical, dentistry and pharmacy graduates, up to N+21 months for reference years 2021-2023. Although this derogation formally ends with reference year 2023, similar delays may also affect reference year 2024 due to consolidation issues in the source data transmitted by the Ministry in charge of Higher Education. In addition, the indicator “Number of practising nurses” is currently provided with a one-year delay due to data availability constraints and the ongoing reconstruction of the series following the ADELI-to-RPPS transition.
Data covers metropolitan France and overseas departments and regions (DROM). Statistics are broadly comparable across territories when based on nationwide administrative sources; however, some geographical comparability and coverage limitations remain for specific sources and variables:
DROM are excluded from EHIS coverage, but since 2019 France has set up a complementary survey to extend its coverage to the DROM.
Other surveys exclude the coverage of Mayotte
Administrative data coverage (health insurance reimbursements, payrolls) is also quite incomplete for Mayotte.
These limitations are documented in the metadata in annex.
On the whole, the data are comparable over time. Potential breaks in series (e.g. changes in definitions, coverage, classifications or source systems) are monitored through annual time-series checks and are documented variable by variable in the Annex at the bottom of the page.
