The influenza surveillance system
The Sentinelles Network
The Sentinelles network, coordinated by the Pierre Louis Institute of Epidemiology and Public Health (UMR-S1136) of Inserm and Pierre and Marie Curie University, is composed of general practitioners and pediatricians. Throughout the year, based on cases reported by general practitioners—who represent approximately 0.5% of all general practitioners—the network estimates the weekly number of patients seeking care for influenza-like illness (ILI), expressed as incidence (number of visits) or incidence rate (number of visits per 100,000 inhabitants) at the national and regional levels. The case definition for influenza-like illness is: fever above 39°C, with sudden onset accompanied by muscle aches and respiratory symptoms. General practitioners and pediatricians collect nasopharyngeal swabs, which are sent primarily to the National Reference Center (NRC) for Respiratory Infection Viruses (including influenza) and to the virology laboratory at the University of Corsica for confirmation of the influenza diagnosis. Samples for confirming the diagnosis of influenza are collected from a representative sample of patients presenting with influenza-like illness during the season. These samples enable monitoring of the weekly trend in the proportion of samples testing positive for influenza, identification of the circulating influenza virus types and subtypes, and, at the end of the season, assessment of the incidence of consultations for laboratory-confirmed influenza, as well as contribution to national and European estimates of vaccine efficacy.
SOS Médecins
The SOS Médecins associations provide on-call medical care, making it possible to obtain data on consultations conducted 24/7, including on holidays and during school breaks, within the geographic area they cover. This data is transmitted daily to Santé publique France via a single encrypted file, which enables the estimation of the daily and weekly number of patients seeking care for influenza-like illness. The case definition used by SOS Médecins is: sudden onset of fever exceeding 38.5°C, accompanied by muscle aches and respiratory symptoms. In 2019–2020, 62 SOS Médecins associations participated in this surveillance. The weekly proportion of patients diagnosed with influenza or influenza-like illness among all procedures coded by SOS Médecins is calculated for each region and at the national level.
Emergency department visits and hospitalizations
Since June 2004, Santé publique France has been collecting data daily from emergency departments across France through the Oscour® network (Organization for Coordinated Emergency Surveillance).
For the 2017–2018 season, more than 600 emergency departments participated in influenza surveillance, covering over 93% of emergency department visits in metropolitan France. Analysis of this data allows for the observation of weekly variations in the number of patients visiting emergency departments or hospitalized with a diagnosis of influenza (coded according to ICD-10). This hospital surveillance is broken down by age group and by geographic level (national and regional). The indicator used is the proportion of hospitalizations with a diagnosis of influenza/influenza-like illness among all hospitalizations, for all ages combined or for the specific age group under consideration. This indicator, combined with the proportion of hospitalizations following visits for influenza/influenza-like illness, serves as an indicator of the severity of the epidemic.
Severe cases hospitalized in intensive care
The new joint surveillance of severe cases of influenza and COVID-19 began on October 5, 2020 (week 40 of 2020) for an indefinite period initially. The surveillance period will be reviewed regularly based on changes in epidemiological trends and knowledge regarding the interaction between SARS-CoV-2 and so-called “winter” respiratory viruses, particularly influenza viruses.
Data collection relies on the network of sentinel intensive care units established in 2018, initially for the surveillance of severe influenza cases. The surveillance of severe COVID-19 cases has relied on this same network, supplemented by approximately twenty intensive care units in metropolitan France that are not part of the network.
The objectives of this surveillance system are:
General objective:
To track the epidemiological trends of severe cases of influenza and COVID-19 admitted to intensive care as the “winter season” approaches, in the context of the COVID-19 pandemic
Specific objectives:
To document the specific contribution of influenza viruses and SARS-CoV-2 among intensive care admissions;
To describe and document the comorbidities and characteristics of patients admitted to intensive care for influenza or SARS-CoV-2 infection.
Regional influenza and COVID-19 reporting forms for 2022-2023 available for download:
Unlike in mainland France, influenza surveillance is conducted year-round in the French overseas departments and regions (DROM) and overseas communities (COM) due to the different patterns of influenza virus circulation in tropical climates.
In the French Antilles (Guadeloupe, Saint Martin, Saint Barthélemy, Martinique), influenza surveillance is conducted by sentinel networks of volunteer general practitioners (15–20% of all general practitioners), who report weekly to the Regional Health Agencies (ARS) the number of patients seen for influenza-like illness. The data is then analyzed by Santé publique France Antilles. In addition, the number of visits for influenza-like illness is also tracked in Martinique by the SOS Médecins association, as well as the number of pediatric emergency room visits and hospitalizations for influenza-like illness. In Guadeloupe, Saint Martin, and Saint Barthélemy, the number of emergency room visits and hospitalizations for influenza-like illness is available weekly via the Oscour®Network. Surveillance of severe influenza cases requiring hospitalization is also conducted from October to April in the intensive care units of hospitals in Martinique and Guadeloupe. Virological surveillance is carried out using samples collected by a group of volunteer general practitioners between October and April each year and by hospital physicians for patients diagnosed in the hospital. Virological analyses are performed by university hospitals (CHUs) and the CNR (Pasteur Institute of French Guiana). Results are centralized by the Antilles regional unit, which provides regular feedback on the influenza situation to its partners through epidemiological bulletins.
In French Guiana, influenza surveillance relies on a sentinel network of general practitioners, the Cayenne Medical Guard, and the Decentralized Prevention and Care Centers (CDPS). All data on consultations for influenza-like illness are transmitted to Santé publique France Guyane, which analyzes and interprets them. Surveillance of emergency department visits and hospitalizations for influenza and influenza-like illness is also conducted via the Oscour® network (hospitals in Kourou and Cayenne), and severe cases of influenza admitted to the intensive care unit in Cayenne are reported to Santé publique France Guyane. In addition, virological surveillance complements syndromic surveillance and is conducted through samples collected in private practice and at CDPS (at least 2 samples per week corresponding to the first 2 individuals meeting the case definition for influenza-like illness) or within the department’s hospitals. The analyses are then performed by the virology laboratory at the Cayenne Hospital and by the CNR (Pasteur Institute of French Guiana). The results are centralized by Santé publique France Guyane, which provides regular updates on the influenza situation in French Guiana to its partners through epidemiological bulletins.
In Réunion, influenza surveillance relies on a network of sentinel physicians coordinated by Santé publique France Indian Ocean. Fifty general practitioners and two private pediatricians spread across the island report weekly on the number of consultations conducted, as well as the number of consultations for influenza-like illness. The number of cases is estimated by extrapolating data from sentinel physicians to consultation reimbursement data provided by the General Social Security Fund (CGSS). During epidemic periods, sentinel physicians are required to collect one to two random samples per week from patients presenting with influenza-like illness. Sentinel physicians are also encouraged to collect one to two nasopharyngeal swabs per week from patients who have had flu-like symptoms for less than 3 days. During inter-epidemic periods, physicians are also required to collect one to two swabs per week if they encounter flu-like cases meeting the sampling criteria. These samples are collected throughout the year and help characterize the viruses circulating on the island. Virological analyses (strain typing via RT-PCR) are performed by the virology laboratory at the Centre Hospito-Universitaire Nord de Saint-Denis (CHU-Nord) and the CNR (South).
In Mayotte, surveillance has been conducted since June 2009 by a network of approximately twenty sentinel physicians—including hospital-based, private practice, and association-affiliated doctors, as well as civilian and military personnel—distributed across the island, who report the weekly number of patients seeking care for flu-like symptoms. Sentinel physicians are also encouraged to collect one to two nasopharyngeal swabs per week from patients who have had flu-like symptoms for less than 3 days. Virological analyses (strain typing by RT-PCR) are conducted year-round by the virology laboratory at the Mayotte Hospital Center and the CNR (South). Analysis of epidemiological and virological data is carried out by the Indian Ocean regional unit, which provides feedback to its partners through epidemiological updates.
Since the 2015–2016 season, each region has defined its own epidemic period on a weekly basis during the winter season by approving or rejecting the proposals generated by the SurSaUD® and Sentinelles Data Analysis Module (MASS) [1] developed by Santé publique France. This tool establishes statistical thresholds for each region based on regional historical data from three epidemiological data sources (the Sentinelles network, Oscour®, and SOS Médecins) and using three different modeling methods. The models use a 5-year rolling window of historical data. The results are presented as alerts. Each week, MASS classifies regions into 3 phases:
no-alert phase: <40% of alerts activated
pre- (or post-)epidemic phase: ≥40% and <100% of alerts activated
epidemic phase: 100% of alarms triggered
The regional units of Santé publique France validate or reject this statistical analysis, also taking into account other complementary surveillance data, such as virological data. This summary is presented in map form in the national weekly bulletin.
[1] Pelat C, Bonmarin I, Ruello M, Fouillet A, Caserio-Schönemann C, Levy-Bruhl D, et al. Improving regional influenza surveillance through a combination of automated outbreak detection methods: the 2015/16 season in France. Euro Surveill Bull Eur Sur Mal Transm Eur Commun Dis Bull. Oct 2017;22(32).
Syndromic surveillance: deaths from all causes
This surveillance is based on vital statistics data (death certificates) submitted to the National Institute of Statistics and Economic Studies (INSEE). Derived from a sample of 3,000 municipalities, these data cover nearly 80% of national mortality. The estimate of the number of deaths across the entire country is therefore calculated by multiplying the number of deaths observed in this sample of 3,000 municipalities by 1.25. However, these data do not include information on the medical causes of death. They therefore allow for the detection of excess mortality, but not for attributing it to a specific event, even one of national scope.
Currently, INSEE is the most effective source for monitoring mortality. However, there is a delay in reporting the number of deaths from this network, due to the legal deadline for reporting a death to the civil registry (24 hours on business days) and the time it takes for the civil registry office to enter the information. Thus, deaths occurring on day D are available for analysis only starting on day D+10 on average. To account for this delay, the analysis of mortality for a full week can therefore only be conducted two weeks after the last day of the week in question. At the two-week mark, mortality trends (increase, stability, decrease) can be assessed, but a reliable quantitative assessment of this trend cannot yet be provided, as not all deaths have been reported. In the context of a large-scale public health event, based on 10 years of experience using this surveillance system, it is estimated that a minimum of three weeks is required to provide a reliable quantitative assessment of the number of deaths that occurred during the observed period.
Specific surveillance: deaths directly attributable to influenza
Specific surveillance of influenza-related mortality makes it possible to:
determine the severity of the epidemic by examining the proportion of deaths among severe cases on the one hand and among cases in ARI clusters on the other, comparing these proportions to those of previous years;
detect changes in the distribution of epidemiological characteristics of individuals whose deaths are attributed to influenza in order to adapt control measures as quickly as possible.
Surveillance of influenza-related mortality is based on:
A statistical model developed by Santé publique France and used since the 2016–2017 season, which estimates the number of deaths directly and indirectly attributable to influenza during the epidemic period, based on all-cause mortality data, epidemiological and virological data on the circulation of influenza viruses and respiratory syncytial virus (RSV), as well as meteorological data (temperature and humidity). These data are available for the age groups 0–64 years, 65–74 years, and 75 years and older;
The number of deaths among severe influenza cases admitted to intensive care;
The number of deaths among clustered cases of acute respiratory failure (ARF) in long-term care facilities
The analysis of causes of death conducted by the CépiDc based on handwritten death certificates allows for tracking the number of deaths directly attributed to influenza; however, the two-year delay in obtaining this information prevents its use for monitoring or assessing the epidemic. Electronic certification of medical causes of death, implemented since 2007, allows for the analysis of causes of death with a shorter delay. Electronic certification accounted for 10% of deaths in 2015.
Interpretation of all-cause mortality
It is the difference, over a given period, between the number of expected deaths and the number of observed deaths that allows us to estimate the potential impact of a health event. This is referred to as “excess mortality.” The expected number of deaths over a given period is obtained using a statistical model that accounts for historical data (over 6 years), including the general trend and seasonal fluctuations. When the observed number of deaths exceeds the statistical threshold defined by the model, excess mortality is considered to have occurred. Therefore, excess mortality attributable to an event should not be confused with the total number of deaths that occurred during the period in which that event took place.
For example, for a given week, if the observed number of deaths is 1,000 while the expected number of deaths is 700, the excess mortality will be 300, or +43% compared to the expected number of deaths (-> (300/700)x100). Furthermore, given that these data are calculated based on a sample covering only 80% of mortality in France, we can estimate excess mortality on a national scale by multiplying by 1.25 (-> 300 × 1.25 = 375). In this example, we would therefore report an observed excess mortality of 43%, with 375 excess deaths. However, excess mortality occurring during a flu epidemic is not 100% attributable to the flu. Other factors can influence mortality in winter and contribute to explaining excess mortality (other seasonal respiratory viruses, temperatures, potential intercurrent conditions). Research conducted by Santé publique France indicates that, for the 2014–2015 influenza epidemic in metropolitan France, approximately three-quarters of the excess mortality was linked to influenza. However, the proportion of excess mortality attributable to influenza varies from year to year.
The goal of virological surveillance is to detect and isolate circulating influenza viruses at an early stage and to determine their antigenic characteristics and susceptibility to antiviral drugs. Influenza virus testing is performed by direct detection using molecular biology techniques (Reverse Transcriptase-Polymerase Chain Reaction, RT-PCR) or immunological techniques (immunofluorescence, ELISA) and by culture. Viral characterization (subtyping for influenza A, lineage determination for influenza B, and antigenic characterization) is performed using a hemagglutination inhibition assay. Screening for mutations associated with antigenicity, antiviral susceptibility, or virulence is performed by sequencing several dozen strains for each type, subtype, or lineage of influenza viruses.
Virological surveillance in France is coordinated by the National Reference Center (CNR) for Respiratory Infection Viruses, which comprises a coordinating laboratory and two associated laboratories:
the Pasteur Institute in Paris (North);
the Hospices Civils de Lyon (South, Indian Ocean);
the Pasteur Institute of French Guiana (Antilles-French Guiana).
Virological analysis is performed:
in urban areas: primarily by the CNR (mainland France) and the virology laboratory at the University of Corte (Corsica), using nasopharyngeal swabs collected by general practitioners and pediatricians in the Sentinelles network;
in hospitals: by the CNR and the hospital laboratories of the Rénal network (National Network of Hospital Laboratories).
Santé publique France conducts international surveillance to detect and characterize health threats that could affect France. This surveillance is based primarily on the analysis of information gathered from specific international surveillance tools on institutional websites (WHO, ECDC). This surveillance covers seasonal influenza, influenza of zoonotic origin (avian and swine), and the emergence of new influenza viruses.
Since the 2017–18 season, Santé publique France has been responsible for estimating vaccination coverage among high-risk individuals targeted by French vaccination recommendations. Data on influenza vaccine reimbursements from the National Health Data System (SNDS) are used to estimate this vaccination coverage.
Data collected for surveillance purposes are analyzed weekly and summarized in a weekly national epidemiological bulletin published every Wednesday during the surveillance period.
In each region, the regional unit of Santé publique France regularly publishes feedback documents intended for local partners and members of regional public health networks. These documents are available in the “publications” tab on the page dedicated to each region or via the search bar (“influenza”) on the Santé publique France website:
Since January 2012, the GrippeNet.fr participatory surveillance study has been coordinated by the Sentinelles Network. This European surveillance and research project aims to collect epidemiological data on flu-like symptoms directly from the public via the Internet and anonymously.
Unlike traditional surveillance methods, GrippeNet.fr collects its data directly from the general public residing in France, without going through hospitals or doctors.
In relation to
thematic dossier
The First 1,000 Days
The period from conception through the first two years of life is critical for a child’s development and the health of the adult they will become. That is why it is essential to prioritize the...
thematic dossier
Aging Well
People over the age of 60 could account for one-third of the population by 2050. The priority today is no longer to extend life expectancy but to improve the quality of life for older adults.
Vaccination
thematic dossier
Contagious diseases most often affect children at a very young age. Because children are particularly vulnerable, they are a priority target for vaccination programs.
