Introduction

Emergency departments (ED) are a key component of the healthcare system, as they are one of the very few services widely accessible (i.e., 24/7) to patients in need (Kromka & Simpson, 2019). ED use is also an important indicator of the quality of healthcare systems. Defined as 3 + visits to ED per year (Brennan et al., 2014; Gaulin et al., 2019), high ED use strongly contributes to ED overcrowding (Burns, 2017). Patients with mental disorders (MD), including substance-related disorders (SRD), are known to be among the most frequent ED users (Vandyk et al., 2013). A multicenter longitudinal study conducted in 2008–2010 evaluating ED use among 1.76 million adults in the US found that, compared to patients without MD, those with MD were 4.7 times more likely to be high ED users (Brennan et al., 2014). Other studies found that high ED users with MD accounted for between 38% to nearly half of ED visits for mental health reasons, this while only representing 8% to 17% of all ED users (Fleury et al., 2019a, 2019b, 2019c, 2019d; Schmidt et al., 2018a, 2018b). Patients with MD are often affected by co-occurring physical illnesses (Lin et al., 2015) and suicidal behaviors (Gili et al., 2019), which increases their risk of becoming high ED users. The high ED use of patients with MD may also be due to problems in access, continuity, or adequacy of outpatient care (Aagaard et al., 2014; Li et al., 2022). For about 45% of patients, ED is also the entry point to healthcare services, and may in fact be the only resource available to them (Fleury et al., 2019a; Kurdyak et al., 2021). But as it is one of the costliest healthcare service (Heyland & Johnson, 2017), ED should not replace outpatient care, particularly for such vulnerable patients. Some ED visits might be prevented if alternative care was made more available. A better understanding of barriers and facilitators to high ED use among patients with MD may contribute to recommendations that could help improve ED’s response to these patients’ needs.

While numerous quantitative investigations based on medical records have been published, relatively few qualitative studies provide reasons for the repeated use of ED by patients with MD (Aagaard et al., 2014; Digel Vandyk et al., 2018; Li et al., 2022; Poremski et al., 2020; Schmidt et al., 2018a, 2018b; Vandyk et al., 2019). These studies often report high ED use as being inevitable (Digel Vandyk et al., 2018), attributing it to a lack of knowledge regarding care alternatives (Parkman et al., 2017a), limited access to care (Digel Vandyk et al., 2018), or the incapacity of outpatient care to respond to the patients’ needs (Poremski et al., 2020). High ED use was also associated to patients being involuntarily admitted to ED due to suicidal behaviors, violence, intoxication, acts of public nuisance (McCormack et al., 2015; Poremski et al., 2020; Vandyk et al., 2019), or because they were referred there by a care provider or relative (Wise-Harris et al., 2017). High ED users were usually identified as patients with few support networks to help them cope with MD (Poremski et al., 2020; Wise-Harris et al., 2017). Patients reportedly used ED as a last resort, when feeling threatened by their health conditions (Parkman et al., 2017b; Vandyk et al., 2019; Wise-Harris et al., 2017). ED was often perceived as a safe and secure place for patients with multiple health issues and social problems, including unstable housing (Parkman et al., 2017a, 2017b; Poremski et al., 2020; Wise-Harris et al., 2017). Some patients also said they felt stigmatized by ED clinicians, or that they received poor care (Wise-Harris et al., 2017). In quantitative investigations, predictors of high ED use were mostly associated with being men (Fleury et al., 2019c; Schmidt, 2018), younger (Boyer et al., 2011; Sirotich et al., 2016), of lower socioeconomic status (Moulin et al., 2018), and diagnosed with schizophrenia (Kaltsidis et al., 2020; Vu et al., 2015), SRD (Moulin et al., 2018; Vu et al., 2015), or personality disorders (Chang et al., 2014; Richard-Lepouriel et al., 2015).

The few existing qualitative studies were based on small samples of high ED users (Digel Vandyk et al., 2018; Poremski et al., 2020), recruited patients from too few ED settings (Poremski et al., 2020; Schmidt et al., 2018a, 2018b), or focused on some specific MD diagnoses (e.g., personality disorders [(Vandyk et al., 2019)], alcohol-related disorders only [(Parkman et al., 2017b)]). Moreover, we found no previous studies that used a conceptual framework to study the reasons for high ED use from the perspective of patients with MD; such a framework would allow us to structure findings in a more comprehensive way and recommend more targeted service improvements. Previous literature suggests that healthcare system features, patient profiles, professional practice, and organizational characteristics within ED may reveal multiple reasons for high ED use. Using a framework for studying high ED use may also guide decision-makers by highlighting deployment strategies in outpatient care and professional practices, and informing us on patient behaviors, ultimately contributing to the development of personalized care and the reduction of unmet needs. Based on 299 patients with MD recruited in four large ED located in Quebec (Canada), this study aimed to explore reasons for high ED use, so we can formulate recommendations that could improve care for this vulnerable population.

Methods

Study Context

In Quebec, most health and social services, including medical and some psychosocial services, are public (Ministère de la Santé et des Services Sociaux, 2017). Primary mental health care relies on medical clinics staffed by general practitioners and community healthcare centers that mainly provide psychosocial services. As complements to the mental healthcare system, helplines, crisis centers, suicide prevention centers, and detox centers are mostly run by community-based organizations, with counseling services provided by psychologists working mainly in the private system. Specialized MD care is provided by the psychiatric departments of general or psychiatric hospitals, or in addiction treatment centers for patients with SRD (Ministère de la santé et des services sociaux (MSSS), 2022). Patients access public services for MD in community healthcare centers or psychiatric specialized care, through one-stop services provided by community centers. However, receiving the appropriate outpatient services can be a lengthy process (Champagne et al., 2018; Fleury et al., 2016; Ministère de la Santé et des Services sociaux (MSSS), 2021; Vérificateur général du Québec, 2023).

Study Setting and Data Collection

Data for this exploratory qualitative study were drawn from a larger, mixed-methods study on high ED users with MD involving four large ED of various types located in Quebec’s healthcare networks, serving roughly two million inhabitants – about one fourth of the province’s population. The first ED was in a university-affiliated psychiatric hospital. The second was in a psychiatric hospital merged with a general hospital; most patients were admitted in the latter before being transferred to the psychiatric hospital ED, which was located nearby. The third was a psychiatric ED in a university-affiliated general hospital, whose patients were also referred to the psychiatric ED by the general ED, with both being in the same location. The fourth was a general ED located in a general hospital that included mental health clinicians.

Data were collected between March 1, 2021 and May 13, 2022. Participants had to be at least 18 years old, able to communicate in French or English, and sufficiently functional to complete a structured interview. If a participant was not sufficiently functional (i.e., too intoxicated or disorganized), the interview was postponed. Participants were required to grant the research team access to their medical records; they had to be diagnosed with a MD, including SRD, and be high ED users, defined as 3 + visits to ED due to mental health conditions within the 12 months preceding their interview (Fleury et al., 2019d; Gaulin et al., 2019). Recruitment was conducted randomly based on patient medical records – though taking into account the time of year when ED was used. More specifically, when selecting patients from each network, ED clinicians made sure that the sample was balanced by including patients who made ED visits at different periods of the year (e.g., winter, spring). The coordinates of the first 450 patients willing to take part in the study were transmitted to the research team, as they were deemed potential participants. The research coordinator then contacted each of them to validate their willingness to participate in the study, and scheduled a one-hour structured telephone interview featuring closed- and open-ended questions. With a total answering time of about 20 min, these open-ended questions had been validated by a steering committee of ED managers and clinicians. They probed reasons for high ED use, and what patients liked or disliked the most about ED use and outpatient services, including professional practice issues that could explain their high ED use. The steering committee was set up to help with study design and recruitment. Two interviewers conducted the interviews using LimeSurvey, while being closely monitored by the research team. The qualitative responses were recorded, then transcribed verbatim.

The interviews included a survey covering a broad range of sociodemographic and clinical characteristics that were used to articulate patient profiles: sex, age group, education, civil status, occupation (e.g., unemployed), household income ($Can/year), type of housing (e.g., supervised), suicidal behaviors (attempt or ideation), and service use patterns (e.g., having a case manager). Patients were also asked if they had consulted outpatient services before deciding to go to an ED. These data were self-reported and mostly based on multiple choice questions – or, in some cases, standardized scales.

Medical records were also used to identify patient health diagnoses, based on the ED database (BDCU) and inpatient/hospitalization database (MED-ECHO), as framed by the International Classification of Diseases, Tenth Revision, for Canada (ICD-10-CA) (codes detailed in Appendix 1). Referring to exclusive groups, principal MD included serious MD (schizophrenia spectrum and other psychotic disorders, bipolar disorders, personality disorders), common MD (e.g., anxiety, depressive and adjustment disorders), and SRD (alcohol and drug use, including induced use, intoxication, or withdrawal), in that order. As SRD tend to be underdiagnosed in medical records (Huynh et al., 2021), two standardized scales were included in the interviews and merged with results from medical records: the Alcohol Use Disorders Test (AUDT) (Bohn et al., 1995) and the Drug Abuse Screening Test-20 (Skinner, 1982). Chronic physical illnesses were identified based on the Charlson and Elixhauser Comorbidity indexes (Simard et al., 2018). High ED use was measured based on the ED database (BDCU). Data from medical records were merged with survey responses. Data from the survey and medical records were collected for the 12-month period preceding each patient’s interview. Patients provided consent to participate in the study prior to completing their interviews and received a modest financial compensation. The study followed the Consolidated Criteria for Reporting Qualitative Studies (COREQ) (Tong et al., 2007). The multi-site protocol was approved by the ethics board of the Douglas Mental Health University Institute.

Analyses

Descriptive analyses were produced for the quantitative data, while content analysis (Braun & Clarke, 2006) was performed on the qualitative data. The latter followed 6 steps: (1) becoming familiar with the data; (2) generating initial codes and an analysis grid; (3) combining codes into themes; (4) reviewing themes and verbatims for analysis consistency and completeness; (5) presenting and describing themes clearly, with relevant quotes; (6) interpreting the data. A conceptual framework (Fig. 1) based on previous implementation models (Damschroder et al., 2009; Fixsen et al., 2005; Greenhalgh et al., 2004) guided the content analysis, featuring the main barriers and facilitators to the implementation and deployment of innovations in healthcare services. Adapted from a previous study on barriers and facilitators to care in ED (Fleury et al., 2019c), this framework included four dimensions contributing to high ED use: healthcare system features outside ED (e.g., access to services, adequacy of care); patient profiles (e.g., health and social issues, need for prompt care); professional practice (e.g., related to primary care with general practitioners or specialized care with psychiatrists); and ED organizational characteristics (e.g., access to care or adequacy of care in ED). The analysis accounted for the importance of each theme reported by patients (number, percentage). The rigor of the study was enhanced by the composition of the research team, which integrated ED clinicians with diverse expertise, by strategies that included adequate training and close monitoring of the research agents, and by the steering committee’s validation of instruments. The initial steps of identifying, grouping, and refining codes to fit the analysis grid involved a 90% inter-rater agreement procedure for 10% of the verbatim to minimize the impact of personal biases and assumptions. Saturation of the data was reached. A reflective journal was kept throughout the analytical process, and the team researchers, many of them with great ED use expertise, met to discuss individual considerations and identify possible blind spots.

Fig. 1
figure 1

Analytical framework: barriers and facilitators to high emergency department (ED) use among patients with mental disorders (MD), including substance-related disorders (SRD)

Full size image

Results

Description of Participants

Of the 450 ED users referred to the study, 50 could not be reached and 300 accepted to participate, for a 75% response rate. One patient withdrew from the study. Of the 299 patients in the final sample, 55% were women, 39% were between 30 and 49 years old; 57% had a post-secondary education; 82% were single, 46% unemployed; 47% had a household income of less than $20,000 (Canadian dollars), and 58% lived in rental housing (Table 1). In the 12 months before their interview, 44% had received a principal diagnosis of serious MD, 23% personality disorders, 24% common MD, 9% SRD without MD, while 54% reported suicidal behaviors, 45% co-occurring chronic physical illnesses, 38% MD-SRD, and 22% MD-SRD-chronic physical illnesses. Most patients (85%) reported they did not seek help from outpatient care for their MD before going to the ED. Finally, 58% of patients had a case manager.

Table 1 Characteristics of patients (n = 299)
Full size table

Barriers and Facilitators Explaining High ED Use

Table 2 shows barriers and facilitators explaining high ED use among patients with MD, including overall negative or positive experiences with either outpatient services, mental health clinicians, or ED services. Table 3 features quotes of patients as they describe the main barriers and facilitators to their high ED use. Regarding healthcare system features, most patients (56%) reported more negative experiences than positive ones with outpatient care. Roughly half the patients said the main reasons for their high ED use involved insufficient access to services (52%) and inadequacy of outpatient care (55%). Though identified by fewer patients (20%), continuity of care was considered protective against high ED use. Most patients reported long wait times for access to outpatient care, often exceeding 6 months, during which no alternative services were offered. Psychosocial follow-up care in community healthcare centers and access to family physicians or general practitioners in walk-in clinics, as well as to psychiatrists, was particularly hard to obtain. Patients had to deal with lengthy procedures for accessing outpatient care, with no or few services available during evenings, weekends or overnight. Help was also lacking from community-based organizations, including a lack of assistance for strengthening community integration or social networks. Patients reported poor communication and collaboration between primary and specialized care providers, including a failure to share medical records within the healthcare network. Patients also felt stigmatized, especially those with a criminal record and/or SRD, saying the health networks often did little to address the situation. They remembered being transferred too hastily from outpatient care to ED while in crisis. Conversely, patients identified barriers to high ED use that included frequent and flexible follow-up care, patient-centered biopsychosocial approaches (including best practices in medication management), and self-management tools (including crisis resolution techniques).

Table 2 Barriers and facilitators to high emergency department (ED) use for patients with mental disorders (MD) including substance-related disorders (SRD)
Full size table
Table 3 Examples of quotations from study participants regarding barriers and facilitators to high emergency department (ED) use for patients with mental disorders (MD) including substance-related disorders (SRD)
Full size table

Patient profiles that justified high ED use were mainly associated with acute health and social issues. About one quarter of patients reported using ED for suicidal ideation or self-injury, social problems (e.g., financial, housing or food difficulties) or crisis situations (e.g., death or separation of a loved one), intoxication or psychotic episodes, and medication issues (e.g., side effects). In particular, patients without a usual physician, either family physician or psychiatrist, had to use ED for medication renewals. Nearly one third of patients identified the lack of information concerning their conditions, of services offered by community-based organizations, and of options to facilitate crisis management and recovery as justifications for their high ED use. Roughly one fifth of participants had used ED involuntarily, as they were brought there either by the police, through court order, or by a relative. A few others (5%) mentioned using ED “out of habit”, or because their medical records were at that hospital. But what mainly drove patients to ED was their need for prompt access to care and rapid and adequate support, neither of which were available in outpatient care.

Overall, few patients found professional practice to be a reason for their high ED use. Less than one fifth of them expressed concerns regarding the attitude of clinicians, the quality or adequacy of care, or issues around accessibility. Over half the patients (54%) reported more negative experiences than positive ones with clinicians. Specifically, they criticized general practitioners and crisis centers for their lack of expertise on MD, especially complex and severe MD. Patients felt that care professionals tended to refer them to ED too swiftly when they were in crisis or needed access to a psychiatrist. Patients also thought general practitioners lacked interest in treating MD, or in listening to their needs and concerns. They also offered too few psychotherapy sessions, which were mostly available through psychologists working in the private sector yet scarcely accessible to patients without insurance. Even patients with insurance faced yearly thresholds in monetary terms and in the number of sessions covered, both of which often fell short of their needs. Psychiatrists were criticized for not providing sufficient follow-up, for not making themselves available when patients were in crisis, and for not sufficiently coordinating care with other psychosocial resources. A minority of participants reported having a weak therapeutic alliance with their psychiatrist: listening skills, giving patients information about their conditions, treatment options (including medication), and involvement in recovery planning were all considered less than optimal. Hospitalized patients felt they did not receive enough individual or group therapy, and faced the possibility of premature discharge without sufficient follow-up care, contributing to the revolving door syndrome. By contrast, addiction services, intensive case management, and assertive community treatment were perceived as protective against ED use by the few patients who were provided these services. Patients felt these services gave them strong alliances with clinicians, who listened to them without judgment, were readily available when needed, and fully responded to their needs, including by resolving crises. Addiction services were especially found to be more available after hospital discharge.

Regarding ED organizational features, patients expressed more overall satisfaction (72%) than dissatisfaction with the care received in these settings. However, most criticized ED’s discharge process for leaving them without adequate referrals or alternatives for follow-up care. They felt ED staff were not knowledgeable enough about available resources in the community. Conversely, most patients viewed ED clinicians as highly competent, responsive to their needs, and felt they had received adequate treatment in ED even though the wait was too long to see a physician. They perceived ED to be understaffed, and clinicians frequently overworked or unavailable. They mostly disliked requiring a referral from the general ED of a general hospital before being transferred to the psychiatric ED. The lack of beds in ED sometimes required patients to wait several days on stretchers before being transferred to inpatient units or psychiatric care. Nearly half found that ED staff lacked empathy – especially patients with serious MD, who felt staff stigmatized them and didn’t listen to them. Some reported being restrained too soon and kept under close supervision. More than a third mentioned they didn’t receive sufficient explanation in the ED regarding their MD conditions, prescribed medications, and treatment options; they also lacked emotional support. About one quarter reported high physician turnover in ED and insufficient collaboration among them, resulting in differing diagnoses and treatments for the same patient from one physician to another. Some patients considered the ED environment not friendly, comfortable, or safe enough, notably due to a lack of surveillance (especially at night), lack of adequate rooms or quiet spaces, or insufficient separation between patients with serious MD or in crisis and patients with milder conditions.

Discussion

This study explored perceived barriers and facilitators to high ED use among patients with MD. These patients were quite vulnerable, which partly justifies their repeated ED use, and their profiles resembled those of high ED users found in previous research (Armoon et al., 2022; Buhumaid et al., 2015; Slankamenac et al., 2020). Roughly half had serious MD, suicidal behaviors, and chronic physical illnesses, while more than a third had co-occurring MD-SRD. Compared to the general Quebec population, about four times more patients with MD lived alone (82% vs. 17%) (Statistics Canada, 2017), nearly twice as many were unemployed (36% vs. 64%) (Statistics Canada, 2017), a third less owned their own homes (20% vs. 61%) (Statistics Canada, 2017), and their annual median income was less than one third that of the Canadian population (less than $20,000 vs. $60,000 in Canadian dollars) (Statistics Canada, 2017). Very few (15%) had consulted an outpatient service before deciding to use ED, accounting for their more negative experiences with outpatient care and clinicians in the treatment of MD. However, their experience with overall care provided in ED was quite positive, further justifying their high ED use. Patients focused more on reporting barriers than facilitators to care, as they experienced high levels of unmet needs. The barriers mentioned in relation to outpatient care were more numerous, with much improvement needed in ED care as well. Patients commented less on the dimensions involving patient profiles and professional practice.

Providing appropriate care to high ED users brings major challenges to the mental healthcare system, as the health issues this population faces are often too complex to manage in primary care. Patients, and especially those with serious MD, personality disorders or SRD, highlighted the general practitioners’ lack of expertise, comfort, or interest in providing them with follow-up care – a factor that contributes to high ED use, according to the literature (Castillejos Anguiano et al., 2019; Fleury et al., 2021; Wakeman et al., 2016). In Quebec, access to family physicians remains a key issue. In 2022, about 23% of the general population don’t have one (Institut de la statistique du Québec, 2023). However, general practitioners are the main providers used when having a MD (Norton et al., 2018; Stephenson, 2023). More than patients with MD only, those with co-occurring MD and chronic physical illnesses were found to have less chances of receiving mental health services (Jolles et al., 2015) or adequate MD treatment (Menear et al., 2014) by general practitioners in primary care settings.

Study participants reported using ED because they perceived it to be the most accessible and appropriate service for managing health conditions requiring immediate care. While ED are used by close to half of patients with incident MD as their first point of contact for treating their MD (Fleury et al., 2019a; Kurdyak et al., 2021), ED are also used as a last resort by the most vulnerable populations when access to outpatient care is quite restricted (Nesper et al., 2016). ED clinicians, however, may perceive repeated ED use as a failure in treatment, potentially leading to a lack of sensitivity or empathy on their part as well as stigma against high ED users, as reported in previous studies (Digel Vandyk et al., 2018; Wise-Harris et al., 2017). Few of the study patients commented on the efficacy of professional practice overall (i.e., technical aspects of care) (Campbell et al., 2000), even in psychiatric care. This may be explained by the fact that few of them received continuous outpatient care, so more criticisms were aimed at the clinicians’ attitude (lack of empathy, poor listening skills, failure to involve patients in decisions concerning their treatment, etc.). A better therapeutic alliance between patients and outpatient clinicians may reduce unjustified ED use. Facilitators identified by patients highlighted the key role of the clinician/patient therapeutic alliance, of frequent and flexible follow-up care, of developing patient-centered approaches (Langberg et al., 2019) that respond more fully to patient needs, and of crisis and self-management tools. Patient profiles, especially patient in crises involving suicidal behaviors or intoxication, accounted for considerable ED use, which could possibly have been avoided if they had the support of a case manager in such situations. Almost half of the high ED users in the study (42%) did not have access to case management. Having a case manager could definitely be recommended to reduce ED use.

Healthcare system features and aspects of ED organization that patients reported to explain high ED use were mainly related to insufficient accessibility, adequacy and continuity of care, including lack of collaboration between ED and other providers. Numerous patients were forced to use ED to overcome the lack of outpatient services, thus compounding their high ED use. While previous Quebec mental health reforms were specifically aimed at improving access to services, long waitlists remained for both psychiatric care and psychosocial services, this in spite of the fact mental health teams were deployed in Quebec’s community healthcare centers, and that case management teams were reinforced (Fleury et al., 2016; Ministère de la Santé et des Services sociaux (MSSS), 2021). And although Quebec’s public health insurance does not cover services provided by private psychologists (Bartram, 2019), many initiatives were deployed to encourage greater access to psychotherapy (Vasiliadis et al., 2015) and to reinforce the deployment of stepped-care initiatives (Gouvernement du Québec, 2022; Roberge et al., 2015). Meanwhile, the opening hours and services of community-based organizations like crisis and suicide prevention centers or self-help groups, which could substantially help high ED users, have often been reduced due to underfunding (Réseau communutaire en santé mentale (COSME), 2019); Vérificateur général du Québec, 2023). Concerning adequacy and continuity of care, literature has repeatedly emphasized the importance of close patient monitoring and of patient-centered or recovery-oriented approaches (Dell et al., 2021) as ways of preventing high ED use, especially for patients with serious MD or co-occurring MD-SRD-chronic physical illnesses (Barker et al., 2020; Bischof et al., 2021). Collaborative care (Kappelin et al., 2022) may also be more vigorously promoted in the follow-up of high ED users, this in order to better respond to their multiple needs and to improve integrated patient care and teamwork. Increasing the patients’ social networks through rehabilitation and daytime activities might also be suggested to better integrate them into their communities, social determinants being key to the recovery of patients with MD (Podogrodzka-Niell & Tyszkowska, 2014; Reid et al., 2019).

Organizational aspects of ED could also be improved, like promoting a better attitude of staff towards patients (e.g., less stigmatization) and deploying strategies to increase follow-up care after ED discharge. Among the practices best known for helping reduce repeated ED use are assertive community treatment (Penzenstadler et al., 2019), intensive case management (Tsai et al., 2022), permanent supportive housing for homeless individuals with MD (Aubry et al., 2020), short-stay crisis units (Anderson et al., 2022), intensive home treatment (Lamb et al., 2020), and community crisis services (Siskind et al., 2013). Other initiatives involving care plans (Abello et al., 2012), peer support (Mulvale et al., 2019), and brief interventions (Stanley et al., 2015) have also shown promising results in terms of reducing ED use, and could be further tested in different ED settings (Gabet et al., 2023). Staff competence and knowledge regarding the services most appropriate for patient care, and care coordination are also key for effective teamwork in ED, as is continuing cross-training for staff (Perreault et al., 2009). These initiatives may also be more widely implemented.

Limitations

Some study limitations are worth mentioning. First, while high ED users were often affected by stigma, it is possible that some participants may have preferred not to disclose pertinent information regarding their experiences in ED or other healthcare services. Second, the responses of some high ED users may have reflected memory bias. Third, as all ED in this study included a psychiatric care unit, were located in large urban territories and operated under a public healthcare system, the results cannot be generalized to other types of hospitals, territories or contexts. Fourth, men and youth patients were underrepresented in the sample, and the study was based exclusively on the patients’ perception of their high ED use. Fifth, the perspectives of ED clinicians and relatives of high ED users were not explored, although their perceptions would be very pertinent and may have differed from those of the high ED users themselves. Finally, the various barriers and facilitators to high ED use that were identified, as well as their respective impact, cannot be considered equivalent.

Conclusion

This study was innovative in that it identified barriers and facilitators to high ED use based on the perspectives of a large sample of patients with MD. Overall, patients identified more barriers than facilitators in accessing outpatient care or for adequacy of care. For the most part, ED were operating in silo, that is, insufficiently connected to outpatient care, whether psychiatric or primary care. This context was found to be inappropriate in ways that favored the revolving door syndrome among high ED users. Improving ED interventions in collaboration with outpatient care may thus be prioritized to reduce high ED use, though ED should not replace outpatient care for these most vulnerable patients. Better funding of outpatient mental health services, more training and collaborative care among general practitioners in the treatment of MD might also be suggested, along with more robust anti-stigma campaigns and incentives to increase MD treatment, especially in primary care. Enhanced referral protocols, follow-up care after ED use and transfer processes to outpatient care, care plans and case management programs should be highly recommended for high ED users, with a view towards breaking the cycles that lead to high ED use.