First symptoms and health care pathways in hospitalized patients with acute heart failure: ICPS2 survey. A report from the Heart Failure Working Group (GICC) of the French Society of Cardiology

Abstract Background Acute heart failure (AHF) is a common serious condition that contributes to about 5% of all emergency hospital admissions in Europe. Hypothesis To assess the type and chronology of the first AHF symptoms before hospitalization and to examine the French healthcare system pathways before, during and after hospitalization. Material and Methods A retrospective observational study including patients hospitalized for AHF Results 793 patients were included, 59.0% were men, 45.6% identified heart failure (HF) as the main cause of hospitalization; 36.0% were unaware of their HF. Mean age was 72.9 ± 14.5 years. The symptoms occurring the most before hospitalization were dyspnea (64.7%) and lower limb edema (27.7%). Prior to hospitalization, 47% had already experienced symptoms for 15 days; 32% of them for 2 months. Referral to hospital was made by the emergency medical assistance service (SAMU, 41.6%), a general practitioner (GP, 22.3%), a cardiologist (19.5%), or the patient (16.6%). The modality of referral depended more on symptom acuteness than on type of symptoms. A sudden onset of AHF symptoms led to making an emergency call or to spontaneously attending an emergency room (ER), whereas cardiologists were consulted when symptoms had already been present for over 15 days. Cardiologists referred more patients to cardiology departments and fewer patients to the ER than general practitioners or the SAMU. Conclusion This study described the French healthcare system pathways before, during and after hospitalization AHF. AHF clinic network should be developed to provide adequate care for all HF patients and create awareness regarding AHF symptoms.


| INTRODUCTION
Acute heart failure (AHF) is a public health problem that affects about 64.3 million individuals world-wide and 15 million in Europe. It is an important reason for emergency room (ER) visits. [1][2][3] In France, AHF affects over 1 million patients and causes numerous single and multiple hospitalizations. 4 In 2014, 165 000 patients were hospitalized for HF. 5 It is estimated that the number of these already frequent hospitalizations will increase even more in the future, with important consequences on public health costs and health system functioning, if no relevant measures are undertaken. 6 To date, for France, almost no information is available about the pathway of patients with AHF prior or during hospitalization and about problems that they may encounter, from first symptom appearance to patient care. 7 Therefore, we conducted a survey on hospitalized AHF patients to learn about health care pathways as experienced by the patients themselves, and about problems encountered prior to hospitalization.

| MATERIAL AND METHODS
Between April 1, 2018, and September 30, 2018, we conducted a national multi-center observational, transversal survey in France. The survey was performed according to the guidelines of the International Association for proper conduct in epidemiological research and conformed to local legal requirements for the conduct of this type of study. 8 In Europe, no ethics committee approval for this type of investigation is required. Patient participation was free and questionnaires were anonymous.
All members of the French Heart Failure Working Group of the French Society of Cardiology (GICC) were invited to participate in this survey. Data were collected using a questionnaire generated by cardiologists and public health physicians. Patients had to be at least 18 years of age and hospitalized or had to be previously hospitalized for a first episode of AHF as the main diagnosis for admission within the last 12 months. AHF was defined by the presence of new or worsening symptoms of HF (pulmonary edema, acute decompensated heart failure, or cardiogenic shock), requiring initiation or intensification of treatment for HF, according to the published standardized definitions. 9 Eligible subjects were identified using the Medical Information System Program (MISP) coding or during consultations. Notably, the MISP identified all patients hospitalized with AHF as the main diagnosis for admission using a dedicated specific MISP code during the period of the study. This identification was independent of the existence of follow-up by a cardiologist or not, and regardless of the recruitment center. Therefore, although the identification request by the MISP was performed by cardiologists from our working group, all eligible patients during the period could be included.
Patients who agreed to participate in the study completed the survey and mailed it back anonymously. Questions were about patient demographics, medical history, cardiac failure characteristics, types of symptom prior to hospitalization, pathway of care (before, during, and after hospitalization), current use of medication and of non-medical treatment, as well as how patients perceived their illness. All questions including in the survey were presented in the Appendix S1. Depending on the question, patients could provide one or several answers, thus prevalence could add up to more than 100%. The survey was generated in a way that patients were not geared towards a HF diagnosis.

| THEORY/CALCULATION
Only data from patients who provided full information about their age and gender were included in the analyses. Continuous variables were expressed as mean ± standard deviation (SD) and categorical variables as frequency with percentage. Statistical analyses were performed using the R version 3.6.3 [R Foundation] for categorical and numerical variables. The t test was used and a p value <0.05 was considered significant.

| RESULTS
Forty centers participated in the study; 7000 questionnaires were sent to patients; of those, 1044 (15%) were returned. Among the 1044 returned, 251 (24%) were discarded (patient deceased, no information provided or more than 50% of the information missing, gender or age missing). Overall, 793 questionnaires were considered suitable for the statistical analysis.

| Patient characteristics
Patient characteristics are shown in Table

| Patient care pathway
Details about patient care pathway, from the symptom-appearing phase to the post-hospital phase, are provided in Figure 1 (Table 2). Prior to hospitalization, symptoms had already been present for over 15 days in 47.4% of the patients, and for over 2 months in 32.3% of them, and 39.6% reported having seen their physician the month prior to admission.

| Hospital referral and admission
Appendix S1 provides detailed information about patient referral to the hospital, according to type of symptom and time of appearance.
The emergency medical assistance service (SAMU) was the most common type of referral (41.6%). Only 19.5% of the patients were referred by their cardiologist. SAMU-referred patients were older and more likely to suffer from acute coronary symptoms and chest pain, as well as dyspnea. Patients referred by a general physician (GP) or a cardiologist were more likely to suffer from lower limb edema. Other symptoms and recent weight gain in particular were less frequently reported by patients. Most patients referred by their GP or cardiologist had symptoms for over 15  while those admitted to a non-intensive care unit were more likely to have edemas. Appendix S1 specifies the unit of hospitalization according to symptoms. Other prior medical problems, risk factors or comorbidities did not appear to influence the type of hospital admission.

| Care after admission
In total, 69.2% of the patients admitted to ER were transferred to a non-intensive care unit. Those remaining were transferred to a department other than cardiology (Table 4).

| Post-hospitalization management
After being discharged, 76.2% of the patients returned back home.
Only a minority was transferred to a cardiac rehabilitation ward (12.4%), to a convalescent home (6.4%), or went to live with a relative (2.1%). The youngest patients entered the cardiac rehabilitation program while the oldest ones were admitted to a convalescent home. AHF diagnosis can be challenging because symptoms vary at presentation, and many different factors can cause an episode of acute heart failure. Thus, rapid identification of patients with AHF is the first step in providing effective care. In our study, referral was more influenced by the type and severity of recent symptoms rather than by symptom types only. Only chest pain, but not dyspnea of any severity, prompted rapid contact with emergency services. This suggests that in France, the ER, the SAMU and the GP are currently the first medical contact for most AHF patients prior to hospital admission. 15 However, neither a GP nor a regular cardiologist played a major role in admitting the patient to hospital, and this was even more common when patients were not city residents. In the OFICA study, only about 10% of the patients hospitalized for AHF had consulted a cardiologist previously, and about half of the patients had not been referred by any of their regular practicians. 16 30 We admit that this study has some limitations. First, the design of this retrospective study was based on the use of a survey, with a risk of selection bias. In addition, the use of a patient survey causes a risk of bias introduced by the survey instrument itself. Patients were mainly selected by cardiologists working in AHF clinics. Moreover, the patient return rate was low. However, we considered this rate adequate, considering the patient population mean age. Even though almost half of the patients in this study had consulted a physician during the month prior to hospitalization, we did not ask whether a AHF diagnosis had been made during this consultation.

| Patient disease perception and knowledge
Moreover, the majority of patients did not provide a LVEF value and data on medications could not be analyzed.
In conclusion, over the past decades, with a decreasing number of cardiologists in France, non-cardiologists have difficulties in admitting their patients to specialized cardiology units. Both patients and the public need to be better informed about AHF symptoms. GPs, cardiologists, and hospitals need to network more efficiently to decrease ER admissions and, therefore, to improve patient prognosis.