The Role of Massive Databases in the Post-market Clinical Follow-up

of Medical Devices

Marion Burland

and Thierry Chevallier

2,3,4

Department of Quality, Regulatory and Clinical Affairs, DEDIENNE Santé, Le Mas des Cavaliers,

217 rue Charles Nungesser, 34130 Mauguio, France

Department of Biostatistics, Epidemiology, Public Health and Innovation in Methodology (BESPIM),

CHU Nîmes, Place du Pr. Robert Debré, 30029 Nîmes, France

UMR 1302, Institute Desbrest of Epidemiology and Public Health, INSERM, Univ Montpellier, Montpellier, France

Tech4Health-FCRIN, France

Keywords: Massive Database, Real Life Study, Post-Market Clinical Follow up, Medical Device.

Abstract: With the application of new European regulations on medical devices in May 2021, the requirements for

clinical evaluation have been strongly reinforced. Post-marketing clinical follow-up is now a key activity for

manufacturers to keep their medical devices on the market. The use of material-epidemiology studies and

real-life databases has multiple strengths and advantages. However, the weaknesses and limitations identified

do not yet allow manufacturers (especially small and medium-sized companies) to fully utilize these tools for

post-market clinical follow-up. Yet certain technological and regulatory developments already implemented,

and to be implemented over time, suggest that these tools could play a crucial role in the clinical monitoring

of medical devices in the future. In order to better define the future use of real-life data in post-market clinical

follow-up activities, a comprehensive update of technological and regulatory surveillance is still required.

1 INTRODUCTION

Regulation (EU) 2017/745 on medical devices (other

than in vitro diagnostic medical devices) has been in

force since May 26, 2021 in all member states of the

European Union. It was adopted to establish a

rigorous, transparent, predictable and sustainable

regulatory framework for medical devices. This

framework must guarantee a high level of safety and

health protection while promoting innovation

(European_Parliament_and_European_Council, 2017).

Regulation (EU) 2017/745 constitutes a complete

overhaul of the regulations governing the rules for

placing medical devices on the market, making them

available and putting them into service. Adopting

these changes represents a real challenge for the

different protagonists of the sector, particularly for

manufacturers. The dimensioning of the clinical

evaluation and the obligation for the manufacturer to

ensure a post-market clinical follow-up integrated

into its post-market surveillance plan are among the

major changes brought by the new regulation.

(Nicolas Martelli, 2019; Beata Wilkinson, 2019; Alan

G. Fraser, 2020).

There are various methods to support post-market

clinical follow-up. However, some methods do not

always cover all the objectives of these activities. The

use of real-life data is therefore essential for the

clinical evaluation of medical devices.

Here again, various strategies for the use of real-

life data are available to medical device

manufacturers: while the implementation of clinical

investigations is one of the most suitable means of

generating clinical data to address targeted issues, the

multiplication of data warehouses and the use of the

latter could also make it possible to achieve these

objectives without necessarily involving human

beings.

The reflection carried out within the framework of

this work concerns the following issues: "What is the

current position of material-epidemiology and more

particularly of massive databases in the post-

marketing clinical follow-up of medical devices?"

and "What are the prospects of using these tools in

this field of activity?"

Burland, M. and Chevallier, T.

The Role of Massive Databases in the Post-market Clinical Follow-up of Medical Devices.

DOI: 10.5220/0010952600003123

In Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 1: BIODEVICES, pages 243-249

ISBN: 978-989-758-552-4; ISSN: 2184-4305

 2022 by SCITEPRESS – Science and Technology Publications, Lda. All r ights reserved

243

2 POST-MARKET CLINICAL

FOLLOW-UP

According to Annex XIV, Part B, Paragraph 5

of Regulation (EU) 2017/745, post-marketing

clinical follow-up "shall be understood as a

continuous process that updates the clinical

evaluation" and "shall be addressed in the

manufacturer’s post-market clinical plan"

(European_Parliament_and_European_Council,

2017). The methods and measures used in post-

market clinical follow-up activities are documented

in a post-market clinical follow-up plan.

Post-market clinical follow-up activities

"proactively collect and evaluate clinical data from

the use in or on humans of a CE marked medical

device, placed on the market or put into service within

its intended purpose" (MDCG, MDCG 2020-7- Post-

market clinical follow-up (PMCF) Plan Template,

2020). The results of these activities should help to

achieve the following objectives:

i) confirm the safety and performance of the

medical device throughout its intended

lifetime and cover the limitations identified in

the clinical evaluation report;

ii) detect unknown adverse effects, monitor

them, and identify possible contraindications;

iii) identify and analyze emerging risks based on

the evidence;

iv) ensure the acceptability of the benefit/risk

ratio;

v) identify any possible misuse or off-label use of

the medical device

(Josep Pane, 2018; https://www.qualitiso.com/scac-

suivi-clinique-apres-commercialisation, 2021).

The limits of clinical evaluation can only be

identified when the context of the medical device’s

clinical use and its associated performances have

been precisely identified. Indeed, identifying the

limitations corresponds to an analysis of the

sufficiency of clinical evidence performed through

clinical evaluation. Upon conclusion of the clinical

evaluation report, the manufacturer should be able to

answer the following questions:

i) Are there any unsubstantiated or partially

unsubstantiated claims?

ii) Is the benefit/risk ratio acceptable over the

lifetime of the medical device?

iii) What are the complications and other residual

risks associated with the use of the medical

device?

The implementation of post-marketing

surveillance activities and more particularly post-

marketing clinical follow-up activities depends on the

answers to these questions and must make it possible,

in the more or less long term, to cover the limits

identified, otherwise the CE marking of the medical

device could be jeopardized.

There are several ways of conducting post-market

clinical follow-up activities. General methods and

procedures (analysis of the scientific literature,

feedback from congresses, etc.) have the advantage of

being economical for the manufacturer, but generally

do not make it possible to overcome all the limitations

identified. Implementing specific methods and

procedures for each medical device is therefore very

often necessary to compensate for the lack of clinical

evidence. Each method has its own advantages and

disadvantages. This makes the post-market clinical

follow-up strategy and planning even more complex.

3 REAL-LIFE DATA FOR

POST-MARKET CLINICAL

FOLLOW-UP

Material-epidemiology can be defined as follows:

"like pharmacoepidemiology, as a discipline that

applies epidemiological methods and/or reasoning

to evaluate, generally on large populations and

over long periods of time, the effectiveness, risk

and use of medical devices in real life"

(Equipe_pédagogique_DU, 2020-2021). Material-

epidemiological studies are observational studies that

make it possible to approach the reality in the field

without disrupting the usual behaviors of prescribing,

fitting and using the medical device by collecting

real-life data. They can be prospective, retrospective

or ambispective.

The collection of real-life data and the use of

massive databases are now recognized in the

European regulation on medical devices just as in

Article 108 of Regulation (EU) 2017/745 "The

Commission and the Member States shall take all

appropriate measures to encourage the establishment

of registers and databanks for specific types of

devices setting common principles to collec

comparable information. Such registers and

databanks shall contribute to the independent

evaluation of the long-term safety and performance

of devices, or the traceability of implantable

devices, or all of such characteristics."

(European_Parliament_and_European_Council, 2017).

This therefore suggests an acceleration in the use of

ClinMed 2022 - Special Session on Dealing with the Change in European Regulations for Medical Devices

244

this type of study in the clinical follow-up activities

of medical devices.

3.1 Real Life Data Warehouses

Bégaud et al. define "real-life data," as "data that are

without intervention in the usual patient management

arrangements and are not collected in an

experimental context (which, notably, is the case with

randomized controlled trials), but which are

generated during the routine care of a patient, and

which therefore reflect a priori current practice. Such

data can come from multiple sources : they can be

extracted from computerized patient records, or

constitute a by-product of the information used for

healthcare reimbursement; they can be collected

specifically [...], or to constitute registries or cohorts,

or more punctually as part of ad hoc studies; they can

also come from the web, social networks, connected

objects, etc." This definition is included in the more

general definition of "health data" proposed in

Article 4 of the General Data Protection Regulation

(GDPR) and supplemented in Recital 35 of the GDPR

(Ministère_des_affaires_sociales_et_de_la_santé,

2016; Bernard Bégaud, 2017;

Club_de_la_Sécurité_de_l'information_Français,

2019).

As a result, there are a multitude of data sources

available to provide insight into questions related to

the efficacy, safety, and use of medical devices.

To illustrate the wide variability of real-life data

sources and processing possibilities, two typologies

of real-life data repositories that can be used in

manufacturers' post-market clinical follow-up

strategy are presented below. It should be noted that

these examples are not intended to be an exhaustive

presentation of all the solutions available to

manufacturers.

3.1.1 Practice Registries

Practice registers are integrated with the aim of

evaluating, monitoring and improving practices.

They are databases made up of standardized data

(specifically entered to feed the registry), resulting

from professional practices, most often relating to a

specific theme. The collection and analysis of these

data are widespread within professional

organizations, learned societies or networks. The

setting up of a practices register is orchestrated by a

professional structure made up of peers who run the

i) the theme of the register;

ii) the design of the register (this includes

compliance with the regulations applicable to

the collection and processing of the data

collected and guarantees of confidentiality of

the data relating to the patients and health

professionals involved);

iii) the quality of the data collected and the

methodology for entering the data;

iv) analysing and exploiting the data collected

(CNIL; HAS, 2014;

Group_IMDRF_Patient_Registries_Working, 2016).

Certain fields, such as thoracic and vascular surgery,

orthopaedics and interventional radiology, are among

those for which practice registers are frequently

implemented (CHRU_Tours, 2021; Besse, 2020;

Berghmans, 2020). However, as long as they comply

with the regulations and methodology described

above, these tools can be deployed in many fields of

application, with the operation and constraints

specific to each register.

Exploiting the real-life data provided by practice

registries is part of the routine activities carried out by

manufacturers in the context of their post-market

clinical follow-up activities. Indeed, the periodic

reports published within the framework of the

bibliographic monitoring carried out by

manufacturers, are a means of updating the state of

knowledge on medical devices and the pathologies

under evaluation. They generally feed the state of the

art with, for example, data relative to the clinical

conditions of use of medical devices (target

population, indications, type of medical device or

assembly preferred, etc.). In addition, some practice

registries now offer services that allow a

manufacturer to access aggregate data analysis

reports for medical devices for which it is the legal

manufacturer. That way, the manufacturer can use

practice registry data as a source of clinical data

specific to its medical devices.

However, the independent and often voluntary

nature of this type of approach does not always allow

for comparability of data between different registries

(items filled in, methods of inclusion, granularity of

information provided not always sufficient for the

manufacturer to accurately identify the medical

device used, etc). In addition, the data format and

access restrictions generally do not allow the

manufacturer to remove all uncertainties regarding

the medical device of interest (aggregated data,

access to manufacturer data only). Finally, the reports

submitted reflect the results of device use as of the

date of the report. This limits longitudinal follow-up.

The Role of Massive Databases in the Post-market Clinical Follow-up of Medical Devices

245

3.1.2 Health Data Warehouses

The main purpose of health data warehouses is to

concentrate and guarantee long-term access to

existing massive data relating to the medical care of

patients, socio-demographic data, data from previous

research, practice registers, etc. These data are

exploited for research, studies or evaluations in the

field of health. Massive data warehouses often make

it possible to bring together data initially stored in

different heterogeneous databases. Centralizing this

information helps to:

i) consolidate it,

ii) guarantee its coherence and quality,

iii) consult it in a transversal way,

iv) identify it and also

v) use it more easily by quickly constituting

cohorts.

Unlike a targeted research project, study or

evaluation (the aim of which is to respond to a

specific objective limited in time), a health data

warehouse corresponds to the constitution of a large

database for which, in the long term, the data

controller can envisage processing the data in several

ways within the framework of different research

projects. However, it should be noted that health data

warehouses can only be created for the sake of public

interest (CNIL; https://www.cnil.fr/professionnel,

2021; https://www.has-sante.fr/, 2021).

As examples,

i) The National Health Data System, is a large-

scale real-life data warehouse implemented to

analyze and improve population health

(SNDS) (https://www.snds.gouv.

fr/SNDS/Accueil, 2021);

ii) The implementation of hospital warehouses is

increasingly frequent in order to concentrate,

in a single data warehouse, a set of real-life

data collected over a limited territory;

iii) The constitution of personal data warehouses

for the purposes of research, study or

evaluation in the health field by specialized

companies is also frequent.

Depending on the specificities of each data

warehouse, access (direct or indirect) to

manufacturers of medical devices is not always

allowed. However, when this is possible, these tools

offer a multitude of processing possibilities within the

limits of each health data warehouse (data

compartmentalized within the health data warehouse,

often significant and uncontrolled implementation

time, inconsistency of data in case of absence of

consolidation and monitoring process, relatively high

cost of data exploitation for manufacturers, etc).

3.2 Current Strengths and Limitations

of using Real-Life Data in the

Activities of Post-market Clinical

Follow-up

The use of massive databases on real-life data for

post-marketing studies appears to be a solution that

has the advantages over clinical investigations of

consuming fewer financial and human resources,

being implementating quickly, and having access to

large panels.

However, due to limitations in the use of these

data identified for medical device manufacturers; the

place of real-life data in post-market clinical follow-

up activities is not always optimized. Among the

main weaknesses of these tools we may note:

i) The compartmentalization within the

framework of the health data warehouse and

the independence of these approaches, which

do not always provide answers to the

deficiencies identified in the clinical

evaluation files;

ii) The very restricted and limited access to

medical device manufacturers;

iii) The operating cost of these tools (depending

on the methods of accessing the data in each

data warehouse), which remains relatively

high despite their economic nature;

iv) The difficulty of exploiting and comparing

data processing due to the diversity of data

sources, the cross-referencing of structured

and unstructured data, and the non-

standardization of certain data such as the

designation of medical devices (reduced

interoperability).

3.3 Prospects of using Real-Life Data

in Post-market Clinical Follow-up

In order to consider the prospects of using real-life

data warehouses and the future role of these tools in

the context of post-marketing clinical follow-up

activities, a bibliographic research was conducted.

This consisted of identifying, among the

technological, organizational and regulatory

developments identified in the literature, various

avenues for overcoming the limitations of use

ClinMed 2022 - Special Session on Dealing with the Change in European Regulations for Medical Devices

246

previously identified. It should be noted that the

possibilities of improvement presented in this section

are not intended to be exhaustive.

3.3.1 The Health Data Hub: A Single

Platform to Facilitate Access to Data

from Various Sources

The Health Data Hub was officially created on a

national scale, by the law of July 24, 2019 on the

organization and transformation of the healthcare

system. Its creation is one of the highlights of the

French strategy for artificial intelligence. The Health

Data Hub aims to create a dynamic ecosystem for the

innovative exploitation of health data. The objective

of this project is to facilitate the sharing of and access

to health data from a wide variety of sources by

creating a unique patform to promote research. This

platform must be able to facilitate the reconciliation

of health data from various sources (National Health

Data System, various real-life data warehouses,

registries, cohorts, learned societies with clinical

databases including connected objects, health

surveys, prevention data, school medicine,

occupational medicine, etc.) and their exploitation

from a regulatory and technical viewpoint. It must act

as a trusted third party between data producers and

users and be accompanied by a service offer that

includes support procedures, matching operations

between datasets, support for data collection and

consolidation, and the provision of human and

technical resources to exploit them (Villani, 2018;

Marc Cuggia, 2019; Chloé Picavez, 2019).

In light of these ambitions, the Health Data Hub

should become a true enabler for the use of real-life

data by medical device manufacturers.

3.3.2 UDI: Structuration and

Standardisation of Medical Device

Identification

Following the example of the FDA, which has made

the Unique Device Identifier (UDI) mandatory since

2013, Regulation (EU) 2017/745 now requires the

implementation of the UDI for all devices governed

by European regulations (except for custom-made

medical devices and devices under investigation) in

order to improve patient safety and optimize their

care pathway. The unique identification number for

medical devices is an alphanumeric code containing

standardized information to identify each medical

device placed on the market (with a part related to the

identification of the manufacturer and the model of

the medical device (identical for all medical devices

with a common designation:UDI-DI) and a variable

part related to the production unit of the medical

device (UDI-PI)). In order to ensure traceability of

medical devices, the unique identification numbers

are recorded and stored in a common European

database accessible to all member states: EUDAMED

(European database on medical devices) (Elisabetta

Bianchini, 2019; Dorothée Camus, 2019).

With the standardization of medical device

identification, UDI presents a real opportunity for the

use of real-life data warehouses by medical device

manufacturers.

3.3.3 Digitization of Patient Monitoring for

the Benefit of Patient Intervention in

Evaluating the Quality of Care

The development of online platforms is booming.

They now involve the patient, who must enter

information related to his or her quality of life or more

specific dimensions such as physical functioning,

satisfaction, the relationship with care providers, etc.

This evolution is part of the process of continuous

improvement of practices. Indeed, it is becoming

essential for the patient to participate in the evaluation

of the quality of care in real life. By using PROMs

(Patient-Reported Outcomes Measures) and PREMs

(Patient-Reported Experience Measures), patients

can describe their feelings and their experience in real

time and in detail and prevent adverse events from

occurring (Lisa S. Rotenstein, 2017; Rie Fujisawa,

2018).

The use of these new sources of real-life data is a

real opportunity for medical device manufacturers,

healthcare professionals and patients. Indeed, the

exchange of information should make it possible to

improve the manufacturer-practitioner-patient

relationship and to adapt the therapeutic strategy in an

individualized manner and/or on a global scale.

3.3.4 Portable Devices for Monitoring the

Health and Fitness of Subjects

The use of portable medical devices is now an integral

part of the monitoring and treatment of certain

chronic diseases such as diabetes or certain cardiac

pathologies. The use of portable fitness-tracking

devices is also becoming more common. For

example, consumers are equipping themselves with

connected scales or watches in an effort to improve

their health and fitness. Wearable devices that include

connected bracelets and watches, sensors or any other

medical device collect information through consumer

and patient declarations and also passively. This

passive, automated collection of information from

The Role of Massive Databases in the Post-market Clinical Follow-up of Medical Devices

247

sensors is done directly with interfaces connected to

databases that concentrate information from various

sources and of various types (Catherine Dinh-Le,

2019).

The data-processing possibilities offered by the

use of portable devices for health and fitness

monitoring are numerous. So exploiting these new

data sources could become widespread for the

evaluation of the performance and safety of medical

devices. This is a real opportunity for medical device

manufacturers. However, the acceptance of these

tools by patients, healthcare professionals and

competent authorities; the respect of regulations for

data access and exploitation (ethical, legal, etc.); the

standardization, processing and development of

predictive analysis models for the exploitation of data

are still obstacles to the democratization of the use of

these tools.

4 CONCLUSION

The entry into force of the new medical device

regulations has required manufacturers to review

their clinical evaluation processes. Among the major

changes brought about by the overhaul of the

European medical device regulatory framework are

the dimensioning of clinical evaluation, requirements

for post-market surveillance and post-market clinical

follow-up. Post-market clinical follow-up is now a

key activity for manufacturers to keep their medical

devices on the market.

Recognized in European regulations for the first

time, the collection of real-life data and the use of

massive databases are methodologies of interest for

post-market clinical follow-up activities. Although

the use of these tools has multiple strengths and

benefits, the associated weaknesses and limitations do

not yet allow medical device manufacturers to fully

exploit them. Consequently, material-epidemiology

studies and massive databases of real-life data are

currently only complementary to other post-market

clinical follow-up activities because they rarely meet

all the targeted objectives.

In view of the multiplicity of solutions developed

or being developed to combine data sources, facilitate

access to medical device manufacturers, standardize

data and feed data warehouses with new sources of

information, the use of real-life data warehouses is

expected to soon become a key part of post-market

clinical follow-up activities.

In order to better define the future role of these

tools in the clinical evaluation of medical devices, an

update of the technological and regulatory

surveillance should be considered in order to

exhaustively identify developments to facilitate their

use. The obstacles to the use of real-life massive

databases in post-marketing clinical follow-up

activities must also be identified. The ethical aspects

(pseudonymized and anonymized data) of massive

databases, wich are a major topic, should be

addressed in this update.

Lastly, this article is based on a purely industrial

vision (small to medium-sized companies responsible

for the marketing of medical devices). One way of

working would be to widen this reflection to a more

global vision including the viewpoints of the different

actors involved in material-epidemiology studies and

the exploitation of real-life data.

ACKNOWLEDGEMENTS

We would like to thank the entire teaching staff of the

University Diploma in “Methodology in the clinical

evaluation of medical devices” for their continued

pedagogical support in the writing of this work.

REFERENCES

Alan G. Fraser, R. A. (2020). Implementing the new

European Regulations on medical devices - clinical

responsibilities for evidence-based practice: a report

from Regulatory Affairs Committee of the European

Society of Cardiology. European Heart Journal.

Beata Wilkinson, R. v. (2019). The Medical Device

Regulation of the European Union Intensifies Focus on

Clinical Benefits of Devices. Therapeutic Innovation

and Regulatory Science.

Berghmans, T. (2020). Epithor. Revue des Maladies

Respiratoires.

Bernard Bégaud, D. P. (2017). Les données de vie réelle, un

enjeu majeur pour la qualité des soins et la régulation

du système de santé - L'exemple du médicament.

Besse, J. L. (2020). Rapport 2020 sur le registre national

des Prothèses totales de cheville de l’AFCP.

Catherine Dinh-Le, R. C. (2019). Wearable Health

Technology and Electronic Health Record Integration:

Scoping Review and Future Directions. JMIR

MHEALTH AND UHEALTH.

Chloé Picavez, E. S. (2019). Focus données de santé :

Recherches – HDH-Entrepôts de données de santé

Jurisprudence de la CNIL 2019.

CHRU_Tours. (2021, janvier). Communiqué de presse -

L’entrepôt de données cliniques du CHRU, le CDC,

outil majeur pour les études liées au COVID-19 et pour

la recherche en général.

Club_de_la_Sécurité_de_l'information_Français. (2019).

Le traitement des données de santé. Clusif.

ClinMed 2022 - Special Session on Dealing with the Change in European Regulations for Medical Devices

248

CNIL. (s.d.). Référentiel Relatif aux traitements de données

à caractère personnel mis en oeuvre à des fins de

création d'entrepôts dans le domaine de la santé.

Dorothée Camus, D. T.-E.-B. (2019). 2018New European

medical device regulation: How theFrench ecosystem

should seize the opportunity ofthe EUDAMED and the

UDI system, while overcomingthe constraints thereof.

Therapie .

Elisabetta Bianchini, M. F. (2019). Unique device

identification and traceability for medical software: A

major challenge for manufacturers in an ever-evolving

marketplace. Journal of Biomedical Informatics.

Equipe_pédagogique_DU. (2020-2021). Supports de cours

DU méthodologies en évaluation clinique des DM.

Group_IMDRF_Patient_Registries_Working. (2016).

Principles of International System of Registries Linked

to Other Data Sources and Tool.

HAS. (2014). Développement professionnel continu (DPC)

- Fiche méthode - Le registre des pratiques.

https://www.cnil.fr/professionnel. (2021, 06). Récupéré sur

CNIL .

https://www.has-sante.fr/. (2021, 06). Récupéré sur HAS.

https://www.qualitiso.com/scac-suivi-clinique-apres-

commercialisation. (2021, 06). Récupéré sur Qualitiso.

https://www.snds.gouv.fr/SNDS/Accueil. (2021, 06).

Récupéré sur SNDS.

Josep Pane, R. D. (2018). EU postmarket surveillance plans

for medical devices. Pharmacoepidemiology and Drug

Safety.

Lisa S. Rotenstein, R. S. (2017). Making Patients and

Doctors Happier — The Potential of Patient-Reported

Outcomes. The New England Journal of Medicine.

Marc Cuggia, S. C. (2019). The French Health Data Hub

and the German Medical Informatics Initiatives: Two

National Projects to Promote Data Sharing in

Healthcare. IMIA Yearbook of Medical Informatics.

MDCG. (2020, April). MDCG 2020-7- Post-market clinical

follow-up (PMCF) Plan Template.

Ministère_des_affaires_sociales_et_de_la_santé. (2016).

Colloque Big Data en santé : Quels usages ? Quelles

solutions ?

Nicolas Martelli, D. E. (2019). New European Regulation

for Medical Devices: What Is Changing?

Cardiovascular and Interventional Radiological

Society of Europe.

Parlement_Européen_et_du_Conseil_Européen. (2017,

Avril 5). Règlement (UE) 2017/745 du Parlement

Européeen et du Conseil du 5 avril 2017 relatif aux

dispositifs médicaux.

Rie Fujisawa, N. K. (2018). Measuring Patient Experiences

(PREMS): Progress Made By The OECD And Its

Member Countries Between 2006 And 2016. OECD

Health Working Papers No. 102.

Villani, C. (2018). Donner un sens à l'intelligence

artificielle - Pour une stratégie nationale et

européenne.

The Role of Massive Databases in the Post-market Clinical Follow-up of Medical Devices

249