Réseau REDSIAM
Value of a national administrative database to guide public decisions: From the système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) to the système national des données de santé (SNDS) in FranceL’utilité d’une base médico-administrative nationale pour guider la décision publique : du système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) vers le système national des données de santé (SNDS) en France

https://doi.org/10.1016/j.respe.2017.05.004Get rights and content

Abstract

In 1999, French legislators asked health insurance funds to develop a système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) [national health insurance information system] in order to more precisely determine and evaluate health care utilization and health care expenditure of beneficiaries. These data, based on almost 66 million inhabitants in 2015, have already been the subject of numerous international publications on various topics: prevalence and incidence of diseases, patient care pathways, health status and health care utilization of specific populations, real-life use of drugs, assessment of adverse effects of drugs or other health care procedures, monitoring of national health insurance expenditure, etc. SNIIRAM comprises individual information on the sociodemographic and medical characteristics of beneficiaries and all hospital care and office medicine reimbursements, coded according to various systems. Access to data is controlled by permissions dependent on the type of data requested or used, their temporality and the researcher's status. In general, data can be analyzed by accredited agencies over a period covering the last three years plus the current year, and specific requests can be submitted to extract data over longer periods. A 1/97th random sample of SNIIRAM, the échantillon généraliste des bénéficiaires (EGB), representative of the national population of health insurance beneficiaries, was composed in 2005 to allow 20-year follow-up with facilitated access for medical research. The EGB is an open cohort, which includes new beneficiaries and newborn infants. SNIIRAM has continued to grow and extend to become, in 2016, the cornerstone of the future système national des données de santé (SNDS) [national health data system], which will gradually integrate new information (causes of death, social and medical data and complementary health insurance). In parallel, the modalities of data access and protection systems have also evolved. This article describes the SNIIRAM data warehouse and its transformation into SNDS, the data collected, the tools developed in order to facilitate data analysis, the limitations encountered, and changing access permissions.

Résumé

En France, le législateur a souhaité en 1999 que les régimes d’Assurance Maladie développent un système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) afin de mieux connaître et évaluer le recours aux soins et les dépenses de santé des assurés. Ces données sur près de 66 millions d’habitants en 2015 ont déjà donné lieu à de nombreuses publications internationales s’intéressant à différentes thématiques : prévalence et incidence de pathologies, parcours de soins de patients, état de santé et recours aux soins de populations spécifiques, usage des médicaments en vie réelle, mesure d’effet indésirables de médicaments ou d’autres procédures de soins, suivi des dépenses de l’Assurance Maladie, etc. Le SNIIRAM contient des informations individuelles sur des caractéristiques sociodémographiques et médicales des assurés et l’ensemble des remboursements de soins hospitaliers et de médecine de ville, codés selon différents référentiels. L’accès aux données est soumis à des autorisations qui sont fonction de la nature des données demandées ou utilisées, de leur temporalité et du statut du demandeur. De façon générale, les données peuvent être analysées par les organismes habilités sur la période couvrant les trois dernières années plus l’année en cours, et des demandes spécifiques peuvent être réalisées pour extraire des données portant sur des historiques plus longs. Un échantillon généraliste de bénéficiaires (EGB) au 1/97e, représentatif de la population protégée au niveau national, a été constitué en 2005 pour permettre un suivi sur 20 ans avec un accès facilité pour la recherche médicale. Il s’agit d’une cohorte ouverte, qui inclut les nouveaux affiliés et nouveau-nés. Le SNIIRAM a continué à évoluer mais aussi à s’étendre pour devenir, en 2016, le socle du futur système national des données de santé (SNDS) qui intègrera progressivement de nouvelles informations (causes de décès, données médicosociales et des assurances complémentaires santé). Conjointement, les modalités d’accès aux données et les systèmes de protection évoluent aussi. Cet article a pour but de décrire le SNIIRAM et l’évolution vers le SNDS, les données recueillies, les outils développés afin d’en faciliter l’analyse, les limites rencontrées, les autorisations d’accès et leurs évolutions.

Introduction

Health insurance administrative databases are increasingly frequently and effectively used to guide public decisions in many countries [1], [2], [3], [4], [5], [6], [7]. In addition to monitoring the various types of medical expenditures, these databases can also be used to conduct epidemiological studies on the health status of populations, their health care utilisation and health care expenditure, evaluate medical practices or health system experimentations. They can also be used for international comparisons [8], [9], [10], [11], [12].

These databases present a number of advantages. They comprise very large population samples and can sometimes even cover the entire population, ensuring high statistical power, without biases related to the representativity of a sample, thereby allowing more detailed analyses of subgroups according to age or specific geographical territories, for example. Use of these databases is much less expensive that conducting specific surveys in populations or health care institutions, by providing rapid access to data collected in a standardized format [13], [14], [15], [16], [17], [18].

In France, the système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) [national health insurance information system] was gradually developed from 1999 onwards. It initially provided essentially aggregate data to various authorities, but was subsequently transformed into a functional tool with the availability of individual data in 2006. The SNIIRAM data warehouse comprises the characteristics and medical information of the beneficiaries of the various national health insurance schemes, as well as the in-hospital or office medicine health care reimbursed to this population. Cash payments related to sickness, disability or death are also recorded. After a detailed review of data protection and the possible risks of re-identification of persons as a result of data sharing and the growing use of these data by numerous organisations, the 2016 health system modernisation act used SNIIRAM as the cornerstone to create the système national des données de santé (SNDS) [national health data system] (http://www.snds.gouv.fr/) [19], [20], [21], [22]. Broadening of the scope of SNIIRAM and launching of SNDS will allow increased use of these data for public health purposes.

This article describes the SNIIRAM data warehouse and its transformation into SNDS, the data that it contains, the tools that have been developed to facilitate analyses, the potential limitations encountered during analyses and interpretations, and the changing modalities of data access.

Section snippets

From creation of SNIIRAM to the birth of SNDS

The Social Security system, created in France in 1945, is now composed of several schemes constructed around various occupational sectors (Fig. 1) [23], [24]. Of the 66 million inhabitants in France at the end of 2015, the general scheme covers salaried employees of the private sector and their dependents (i.e. about 76% of the population living in France), as well as people covered by sections locales mutualistes (SLM) [local mutualist sections], essentially civil servants, employees of

Beneficiaries are identified in SNIIRAM by an pseudonymised identifier

Constitution of the SNIIRAM data warehouse is based on reliable identification of individuals by the numéro d’inscription au répertoire (NIR) [social security number] derived from the répertoire national d’identification des personnes physiques (RNIPP) [national repertory for the identification of individuals], based on registry office data, including legal immigrants. SNIIRAM therefore allows registry office certification for social security bodies and the fiscal administration and management

SNIIRAM échantillon généraliste des bénéficiaires (EGB)

In 2005, a decree concerning SNIIRAM implemented the EGB, which is managed by CNAMTS and has been the subject of a specific approval from CNIL. It is derived from a 1/97th random sample based on the confidential number of the control key of each individual's NIR (beneficiaries or dependents) ranging between 1 and 97. A preliminary study based on randomisation of the RNIPP ensured that the distribution by five-year age-groups and by sex of the population living in France was independent of

External data that can be linked to SNIIRAM

An administrative database that will be linked to SNIIRAM, allowing identification of older people living in établissements d’hébergement pour personnes âgées dépendantes (EHPAD) [nursing homes] not included in the regulatory framework of SNIIRAM. This Resid-Ehpad system, governed by a special decree, allows nursing homes attached to a pivotal general scheme health fund to transmit the lists of their residents, together with dates of admission and discharge, to national health insurance. The

Use of SNIIRAM for epidemiological and economic purposes: advantages and limitations

SNIIRAM data therefore provide a wealth of information that can be used for public health purposes to guide decision-making. However, these data were not initially collected for research purposes and they can therefore be subject to random or systematic measurement errors, which can have major consequences when defining study populations, exposures, events and covariables. However, in the case of outpatient care, errors should not concern administrative data, which are entered by computer, as

Valorization in the form of scientific publications based on national health insurance databases

Between 2007 and 2016, more than 400 scientific publications (210 SNIIRAM/EGB) based on national health Insurance data (Table 5) were identified in the MEDLINE database (search and Figure in Supplementary material – S1 and S2). From 2009, the annual number of publications has linearly increased to reached 80 in 2016. Seventy-eight percent of publications were in English (Supplementary material, Table S3). However, a lack of homogeneity in the English terms used to describe the French national

Access to SNDS

The data access permission policy in the context of upgrade to SNDS has recently been defined by decree [27]. Modification of the French personal data protection in the field of health studies act (fusion of chapters IX and X) and access request circuits are defined in a second decree [74].

Conclusion

France now has an extensive medical and administrative information system, which has largely contributed to public health information and as an aid to decision-making, and has given rise to a large number of international publications. However, it is difficult to find these publications by using simple search terms. The use of acronyms (SNDS or SNIIRAM) is word-consuming when they are first described in full. Nevertheless, the data source must be indicated in the abstract. An appropriate

Disclosure of interest

The authors declare that they have no competing interest. They are employed by CNAMTS, the SNIIRAM administrator.

References (74)

  • P. Ricordeau et al.

    The prevalence and cost of diabetes in metropolitan France: what trends between 1998 and 2000?

    Diabetes Metab

    (2003)
  • A. Aguade et al.

    Estimating the burden of diabetes to the French national health insurance

    Value Health

    (2015)
  • P. Ricci et al.

    Reimbursed health expenditures during the last year of life, in France, in the year 2008

    Rev Epidemiol Sante Publique

    (2013)
  • J.P. Fagot et al.

    HPV vaccination in France: uptake, costs and issues for the National Health Insurance

    Vaccine

    (2011)
  • E.H. Morrato et al.

    Using population-based routine data for evidence-based health policy decisions: lessons from three examples of setting and evaluating national health policy in Australia, the UK and the USA

    J Public Health

    (2007)
  • T. Kimura et al.

    Pharmacovigilance systems and databases in Korea, Japan, and Taiwan

    Pharmacoepidemiol Drug Saf

    (2011)
  • R. Platt et al.

    The US Food and Drug Administration's Mini-Sentinel program: status and direction

    Pharmacoepidemiol Drug Saf

    (2012)
  • N. Gavrielov-Yusim et al.

    Use of administrative medical databases in population-based research

    J Epidemiol Community Health

    (2014)
  • P.J. Wallace et al.

    Optum Labs: building a novel node in the learning health care system

    Health Aff (Millwood)

    (2014)
  • D. Milea et al.

    A review of accessibility of administrative healthcare databases in the Asia-Pacific region

    J Mark Access Health Policy

    (2015)
  • M. Andersen et al.

    The Asian Pharmacoepidemiology Network (AsPEN): promoting multinational collaboration for pharmacoepidemiologic research in Asia

    Pharmacoepidemiol Drug Saf

    (2013)
  • R.A. Charlton et al.

    Healthcare databases in Europe for studying medicine use and safety during pregnancy

    Pharmacoepidemiol Drug Saf

    (2014)
  • F.R. Estupiñán Romero et al.

    Handbook on methodology: ECHO information system quality report

    (2014)
  • M.T. Bazelier et al.

    Data management and data analysis techniques in pharmacoepidemiological studies using a pre-planned multi-database approach: a systematic literature review

    Pharmacoepidemiol Drug Saf

    (2015)
  • G. Hripcsak et al.

    Characterizing treatment pathways at scale using the OHDSI network

    Proc Natl Acad Sci U S A

    (2016)
  • I. Pigeot et al.

    Establishment of a pharmacoepidemiological database in Germany: methodological potential, scientific value and practical limitations

    Pharmacoepidemiol Drug Saf

    (2008)
  • J. Raftery et al.

    Potential use of routine databases in health technology assessment

    Health Technol Assess

    (2005)
  • G.F. Riley

    Administrative and claims records as sources of health care cost data

    Med Care

    (2009)
  • L.C. Thygesen et al.

    When the entire population is the sample: strengths and limitations in register-based epidemiology

    Eur J Epidemiol

    (2014)
  • M. Goldberg et al.

    Mieux utiliser les bases de données nationales pour la santé publique et la recherche

    (2012)
  • P.L. Bras et al.

    Rapport sur la gouvernance et l’utilisation des données de santé

    (2013)
  • Direction de la recherche, des études, de l’évaluation et des statistiques

    Données de santé : anonymat et risque de ré-identification

    DREES Solidarite Sante

    (2015)
  • Loi no 2016-41 du 26 janvier 2016 de modernisation de notre système de santé

    JORF

    (2016)
  • K. Chevreul et al.

    France: health system review

    Health Syst Transit

    (2010)
  • CNAMTS

    Améliorer la qualité du système de santé et maîtriser les dépenses. Propositions de l’Assurance Maladie pour 2017. Rapport au ministre chargé de la Sécurité sociale et au Parlement sur l’évolution des charges et des produits de l’Assurance Maladie au titre de 2017 (loi du 13 août 2004)

    (2016)
  • G. Trouessin et al.

    FOIN: a nominative information occultation function

    Stud Health Technol Inform

    (1997)
  • Décret no 2016-1871 du modifiant le décret no 2005-1309 du 20 octobre 2005 modifié pris pour l’application de la loi no 79-17 du 6 janvier 1978 relative à l’informatique, aux fichiers et aux libertés

    JORF

    (2016)
  • Cited by (0)

    View full text