The French Emergency National Survey: A description of emergency departments and patients in France

Introduction Some major changes have occurred in emergency department (ED) organization since the early 2000s, such as the establishment of triage nurses and short-track systems. The objectives of this study were to describe the characteristics of French EDs organization and users, based on a nationwide cross-sectional survey. Methods The French Emergency Survey was a nationwide cross-sectional survey. All patients presenting to all EDs during a 24-hr period of June 2013 were included. Data collection concerned ED characteristics as well as patient characteristics. Results Among the 736 EDs in France, 734 were surveyed. Triage nurses and short-track systems were respectively implemented in 73% and 41% of general EDs. The median proportion of patients aged > 75 years was 14% and median hospitalisation rate was 20%. During the study period, 48,711 patients presented to one of the 734 EDs surveyed. Among them, 7% reported having no supplementary health or universal coverage (for people with lower incomes). Overall, 50% of adult patients had been seen by the triage nurse in less than 5 minutes, 74% had a time to first medical contact shorter than one hour and 55% had an ED length of stay shorter than 3 hours. Conclusion The French Emergency Survey is the first study to provide data on almost all EDs in France. It underlines how ED organization has been redesigned to face the increase in the annual census. French EDs appear to have a particular role for vulnerable people: age-related vulnerability and socio-economic vulnerability with an over-representation of patients without complementary health coverage.


Introduction
Some major changes have occurred in emergency department (ED) organization since the early 2000s, such as the establishment of triage nurses and short-track systems. The objectives of this study were to describe the characteristics of French EDs organization and users, based on a nationwide cross-sectional survey.

Methods
The French Emergency Survey was a nationwide cross-sectional survey. All patients presenting to all EDs during a 24-hr period of June 2013 were included. Data collection concerned ED characteristics as well as patient characteristics.

Results
Among the 736 EDs in France, 734 were surveyed. Triage nurses and short-track systems were respectively implemented in 73% and 41% of general EDs. The median proportion of patients aged > 75 years was 14% and median hospitalisation rate was 20%. During the study period, 48,711 patients presented to one of the 734 EDs surveyed. Among them, 7% reported having no supplementary health or universal coverage (for people with lower incomes). Overall, 50% of adult patients had been seen by the triage nurse in less than 5 minutes, 74% had a time to first medical contact shorter than one hour and 55% had an ED length of stay shorter than 3 hours. PLOS

Introduction
The main aim of emergency medicine (EM) is prevention, diagnosis, treatment and orientation for patients with a wide range of un-anticipated illnesses or injuries [1][2][3][4]. Emergency physicians (EPs) are often the first contact with the health care system for patients presenting an emergency medical condition, regardless of gender, age, insurance status and time of day [2][3][4][5]. In the French system, different outpatient and hospital providers collaborate to deliver the most appropriate level of care in emergency situations, mainly EPs and general practitioners (GPs) [6,7]. Several "out-of-hours" services exist to give patients access to a GP at night, on weekdays and all day on weekends or public holidays. A hospital-based department, the Service d'Aide Médicale Urgente (SAMU), offers a 24/7 telephone medical advice service, and whenever necessary, can send mobile intensive care units (MICUs), with EPs, in the pre-hospital field. Ultimately, patients seeking urgent care can visit the emergency department (ED) of any hospital they choose, at their own initiative or upon referral from one of the previous providers [6,7]. In the last 15 years, ED visits have been steadily increasing worldwide [5,[8][9][10][11][12]. French EDs share this situation with a 45% increase in ED visits during this period of time [7,8]. EDs became a common hospital admission route, as even if scheduled admissions appear to be stable, the number of patients admitted to the hospital via the ED (i.e unscheduled admissions) increased from 3.6 to 4.7 million/year from 2004 to 2011, thus reflecting that unscheduled admissions are becoming a larger part in total admissions [13]. The increase of non-urgent [14] and repeated visits [15][16][17][18] contribute of the increase in the number of total ED visit and ED overcrowding which is known to be associated with an increased morbi-mortality [19,20], increased frequency of medical errors [21] and reduced satisfaction of patients [22]. Almost 15 years ago, a major reorganization of emergency medicine was set in motion by the emergency care actors and the French government [23,24]. Several measures, as triage nurses or ED short-track systems, were introduced. The triage system consists of basic clinical assessment that aims to determine in how much time the patient should be seen and the amount of resources to be used. In France, there is no homogeneity in the use of triage scores but the most frequently used is the FRENCH triage scale [25]. Short-track systems set apart patients with non-urgent complaints and possible rapid discharge in a dedicated area, limiting fragmentation of care [26]. They both aim to decrease waiting times, ED length of stay, reduce ED overcrowding and increase patient and staff satisfaction [26]. However, we lack data to comprehensively describe emergency care after these major changes. This study aimed to examine and describe French ED organization and ED patients, based on a 1-day nationwide cross-sectional survey.

Study design and selection of participants
The French Emergency Survey (FES) was a nationwide cross-sectional survey with a two-level design aiming to describe hospital-based emergency care in France through ED organization and ED patients. It was developed by the Directorate for Research, Studies, Evaluation and Statistics of the French Health and Social Affairs Ministry (DREES) with the help of the French Society of Emergency Medicine (SFMU), and the SAMU French ED association (SUdF).
The study took place on June 11, 2013, in all emergency departments in France (as defined by the decree 2006-577 [24]). Any patients presenting to an ED in France during the study period (24-hr, June 11 th , 2013 from 8:00 am to June 12 th , 2013 8:00 am) were included. June 11 was a weekday (Tuesday) and the month of June has a high rate of visits [27], which allowed inclusion of a large number of patients that could be representative of ED patients.

Survey development
The research tool was developed by the study steering committee based on a previous survey of a sample of EDs [28], a literature review and a qualitative study as previously described [28]. Briefly, the steering committee searched MEDLINE via PubMed, RefDoc, ameli.fr, and grey literature (including scientific society websites and scientific meetings proceedings) for all articles and reports of emergency medicine and ED organization in France. In total, 36 studies and reports were identified. Then a qualitative study was conducted among the main institutional actors involved in France's emergency care. All interviews were performed from December 23, 2011 to March 8, 2012. A total of 33 institutions were approached and their representatives interviewed. These institutions were the French Health and Social Affairs Ministry, Regional Health Agencies, French National Health Insurance Fund for Salaried Employees, Emergencies Regional Observatories (ORU), National Medical Council, federations of health care facilities, university and general hospitals, EP and GP unions and nursing home physicians. Patient representatives were also interviewed. After that preliminary phase, the research tool was tested with 23 EDs in February 2013 and modified according to their feedback.

Survey description
A two-level design was used for the survey. The first part was an ED-centered questionnaire, aimed at describing the surveyed hospital and ED. It was to be completed once by each ED administrator. The second part was the patient questionnaire, which was to be completed by the emergency physician for each patient who presented to any of the surveyed EDs during the study period.

ED questionnaire
The ED questionnaire comprised 105 questions organised as follows: 1) department identification (e.g., name, localization, adult or paediatric ED); 2) local organization of emergency care (e.g., existence of a short-track system, existence of direct access to specialized care [cardiology, neurology, gynaecology, geriatrics and ophthalmology]; 3) human resources (e.g., number of doctors, residents, nurses and other healthcare workers); 4) department organization (e.g., presence of a triage nurse, social workers, paediatrician, psychiatrists and/or inpatient geriatric consultation); 5) collaboration with the other hospital departments (e.g., resuscitation procedures, surgical, medical and geriatric consult, CT and MRI accessibility and for how many hours per day, availability of hospital rooms, presence of a bed manager, existence of a local major incidents management plan [as recommended by the Circular No. DHOS/CGR/2006/401 of September 14, 2006 [29]]); and 6) emergency attendance the day of the study.

Patient questionnaire
For every patient who presented to a participating ED on that day, the patient section of the questionnaire was completed by the patient or the accompanying person under the supervision of the EP. The patient section included up to 123 questions. Sociodemographic questions included age, gender, level of education (none / less than high school degree / high school degree / more than high school degree), employment status as defined by the French national institute for statistical and economic studies (active [employed / unemployed] / inactive [retired / student / other inactive]), health care coverage (none / public health insurance / state medical assistance) and complementary health insurance coverage [none / private / universal complementary health care coverage (CMU-C)]. In France, a large part of the population enrols in a private complementary health insurance to cover co-payments required by public health insurance. People with less than an income threshold can benefit from a free complementary health insurance called the CMU-C. When the patient was a child, the sociodemographics reported were those of the accompanying family member. All patients were also asked about 1) the mode of arrival; 2) the genesis of the emergency consult (prior care procedures undertaken before the ED visit, self-reported reasons for the visit [isolated medical reason, difficulties in access to care, need for institutionalization etc.]); and 3) usual use of healthcare system and existence of a GP referent. Moreover, each step of care provided in the ED was recorded by the physician: initial complaint; final diagnosis; waiting times; biological, radiological or therapeutic interventions; and output mode (discharge, hospitalization etc.). Physicians were also asked to assess the adequacy of the emergency consult for each patient. Initial complaints were classified by using the SFMU thesaurus and the final diagnoses were matched according to their International Classification of Diseases (ICD-10) codes.

Data collection
One study referent in each ED transmitted data from completed questionnaires to a dedicated secure website. All data were anonymous. An external provider helped with website conception and data collection. All steps of data collection were supervised by the DREES. Reminders were sent to non-responding EDs, both through the external service provider and the SFMU.

Data quality
The DREES verified the completeness of the database. Internal consistency was, first, controlled by focusing on 3 major variables: age, gender and radiological prescriptions so as to detect discrepancies between age and type of ED (e.g., adult patients seen in a pediatric ED), gender and pathologies (e.g., male with a strictly female diagnosis) and type of radiological exam and type of ED (e.g., unconventional imaging procedure [MRI or tomodensitometry] in a structure that does not have such imaging available). For age and gender, less than 0.001% discrepancy was noted. For radiological exams, a 2% discrepancy was noted. This finding probably also reflects particular organizations, for example a partnership with a nearby imaging center. Then, data quality was assessed by comparing data from the FES and the Oscour 1 network. The Oscour 1 network was established in 2004 with support of the General Directorate of Health Care Provision, the SFMU and regional partners including the ORU. In 2013, 414 emergency structures participated in the monitoring network, covering nearly 65% of ED visits in France [27]. Data from EDs involved in the Oscour 1 network are routinely obtained from patient medical records and the emergency passage summary, including sociodemographic, administrative and medical data. The Oscour 1 network and FES data were consistent for ED visits recorded on June 11, 2013.

Ethics
This study was declared to be of public interest by the CNIS (Conseil National d'Information Statistique) and was integrated into the public statistical program (Visa no. 2013X080SA and publication in the official Journal of French Republic September 17, 2013). It was also approved by the CNIL (Commission Nationale de l'Informatique et des Libertés, French law no. 78-17) (identification no. 1663413). According to French law, written informed consent was not required for this type of study. Patients were informed by staff and a short handout and posters were in the waiting area; 0.3% refused to participate.

Statistical analysis
Descriptive analyses were performed. They included the characteristics of the EDs and then the characteristics of ED users. EDs users were classified in two ways: first according to the type of ED visited, then according to their age. Continuous variables, which were non-normally distributed, are reported as median (Q1-Q3). Categorical variables are reported as numbers (%). All statistical analyses were performed with SAS (SAS/STAT Package 2002-2003, SAS Institute Inc., Cary, NC, USA).

Characteristics of ED users and care
Among the 52,018 expected questionnaires (based on the number of ED visits to each ED during the study period, as described in the ED section of the survey) 48,711 patient questionnaires were collected (94%). They were completed by the accompanying person/family member (in most cases, the accompanying person was the parent) in 25% of cases. For 2% of patients, data collection was impossible because the patient was not able to communicate and no accompanying person was available.  1. The most frequent were traumatic injuries (30%), cardiovascular symptoms (10%) and gastroenterological symptoms (10%). Decreased general condition and loss of autonomy represented 2% (n = 557) and less than 1% (n = 58) of chief complaints, respectively. The most frequent cardiovascular and gastroenterological symptoms were chest pain (44%) and abdominal pain (73%). Overall, 25% of patients (n = 8597) had been seen in a short track system. Time to first medical contact after triage was shorter than one hour for 74% of patients (n = 25,883). ED length of stay (LOS) was shorter than 3 hours for 55% of patients (n = 19,165).

Discussion
To the best of our knowledge, the French Emergency Survey is the first study that provides data aimed to portray the characteristics of all EDs in a European Country with such exhaustivity [30]. A previous study performed in 2002 among a sample of 150 French EDs (and considered as representative of all French EDs) reported an annual census of 23,000 [31]. In our study, this number was 26,500 corresponding to an increase of 15% in 10 years. The triage nurse concept was introduced in 1991 in French EDs [32]. The decree No 2006-577 regarding ED organization [24] recommends that when the activity of the emergency structure allows it, the team should also include a triage nurse. In 2013, the SFMU published triage recommendations [33] calling for a triage nurse in all EDs with more than 5 patients per

ED visit in the last 7 days N (%)
Yes in the same ED 1500 (4) Yes in another ED 426 (1)

Onset of complaint N (%)
The day of ED visit 16690 (48) The day before ED visit 4572 (13) More than one day before ED visit 10041 (29) Missing data 3222 (9)

Output mode N (%)
Return home 24893 (71) hour (corresponding to 43,800 ED visits per year). In our study, every ED with more than 45,000 ED visits per year had a triage nurse. Numerous international studies have been published about the use of triage scales [34,35]. But few data are available regarding the rate of triage implementation in developed countries. A study reported in 2011 that 97% of Swedish EDs used triage scales compared to 50% in the early 2000's [36]. To mitigate overcrowding, triage, as well as other interventions that aim at improving patient flow, has been extensively evaluated. Short tracks have been shown to be effective in reducing low priority patients waiting time and length of stay, without negatively affecting the times of patients with higher priority [26,[37][38][39][40][41]. In our study, 41% of general EDs had a short-track system which concerned 25% of adult patients. Among patients older than 15 years, 7% had universal supplementary health coverage (CMU-C) and 7% no supplementary health insurance coverage. In 2010, according to the French Institute for Research and Information in Health Economics [42][43][44], 94.7% of the French population had supplementary health insurance coverage: 89.0% private supplementary health insurance coverage and 5.7% CMU-C coverage. Only 4.2% had no supplementary health insurance coverage. Our results show that ED patients were less likely to have a supplementary health insurance or CMU-C coverage, as compared to the overall French population. Numerous studies have demonstrated that no or poor supplementary health insurance coverage is associated with both difficulties in health care access and frequency of ED visits [45][46][47][48]. Thus, over-representation of people with no supplementary health insurance coverage or CMU-c might reflect the social vulnerability of this population.
Finally, median hospitalization rate was 20%. We found approximately the same hospitalization rate as the previous study performed in 2003 [31]. As in 2003, the median hospitalization rate was higher in public hospitals than private hospitals [31].

Onset of complaint N (%)
The same day of ED visit 7299 (57) One day before ED visit 2036 (16) More than one day before ED visit 2916 (23) Missing data 645 (5)

Study limitations
The main limitation is that data regarding patients were recorded in real time (i.e., during patient care), which can suggest reporting errors. Second, this study has limitations that are common to this type of study design (declarative surveys) such as reporting bias (including social desirability bias) and representativeness of the sample. This is especially true if the patient-questionnaire was completed with the assistance of a proxy (in our study 25% of patients). But it is important to note that in most cases, the proxy was a family member. Thus, we can assume that the proxy is sufficiently close to the individual to correctly answer, particularly concerning socio-demographic characteristics. In addition, the design of the survey allowed for responses to the ED questionnaire provided by an ED administrator, which reduced the risk of missing data and reporting bias. Concerning representativeness of the sample, as 99.7% of French EDs provided data with a response rate of 94% for The French Emergency National Survey: Emergency departments and patients patient questionnaire, we can assume that our study sample is representative of French EDs and patients. And finally, data referred to French EDs, which raises the question of the generalizability of the results. In 2002, the French Health and Social Affairs Ministry compared the organization of emergency systems in 10 European countries [49] and reported that major differences existed for pre-hospital care but that EDs were similar.

Conclusion
The French Emergency Survey provides data regarding French EDs and their organization. It also describes precisely ED patients' characteristics, details of ED care and length of stay. In addition, it underlines how ED organization has been redesigned to face the increase in the number of ED visits, such as the establishment of a triage nurse or of a fast-track process. Moreover, French EDs appear to have a particular role for vulnerable people, age-related vulnerability as one in five patients are more than 75 years old, but also socio-economic vulnerability, with an over-representation of patients with no complementary health coverage. These major considerations might guide future studies on the subject of ED organization and care.