Implementation of three innovative interventions in a psychiatric emergency department aimed at improving service use: a mixed-method study

Abstract

Background

Emergency department (ED) use is often viewed as an indicator of health system quality. ED use for mental health (MH) reasons is increasing and costly for health systems, patients, and their families. Patients with mental disorders (MD) including substance use disorders (SUD) and suicidal behaviors are high ED users. Improving ED services for these patients and their families, and developing alternatives to ED use are thus key issues. This study aimed to: (1) describe the implementation of three innovative interventions provided by a brief intervention team, crisis center team, and family-peer support team in a Quebec psychiatric ED, including the identification of implementation barriers, and (2) evaluate the impacts of these ED innovations on MH service use and response to needs.

Method

Using mixed methods with data triangulation, the implementation and impact of the three above-named ED interventions were studied. Quantitative data were collected from 101 participants (81 patients, 20 family members) using a user questionnaire and patient medical records. Qualitative data were gathered from focus groups (n = 3) with key intervention staff members (n = 14). The user questionnaire also included open-ended questions. Descriptive, comparative and content analyses were produced.

Results

Key implementation issues were identified in relation to system, organizational and patient profiles, similar to results identified in most studies in the ED implementation literature aimed at improving responsiveness to patients with MD. Results were encouraging, as the innovations had a significant impact for improved patient MH service use and adequacy of care. Services also seemed adapted to patient profiles. Family members were grateful for the help received in the ED.

Conclusions

Before implementing innovations, managers need to recognize the basic issues common to all new healthcare interventions: the need for staff training and strong involvement, particularly among physicians, development of collaborative tools especially in cases of potential cultural clash between staff and organizations, and continuous quality assessment. Future research needs to confirm the pertinence of these interventions, especially use of family-peer support teams in ED, as a highly innovative intervention. Broader ED strategies could also be deployed to improve MH services and decrease ED use for MH reasons.

Background

Emergency department (ED) use is often viewed as an indicator of health system quality, especially when patients make multiple yearly ED visits [1, 2]. ED overcrowding and misuse are recurring issues internationally and in Quebec (Canada), reflecting lack of access and continuity in ambulatory care [2, 3]. From 2005 to 2017, Quebec general ED use increased by 6% [4], with 2% of this increase attributed to mental health (MH) and substance use disorders (SUD) [5]. Higher ED use exacerbates wait times and decreases care quality and patient satisfaction [6]. ED use for MH reasons including SUD [7] and suicidal behaviors [8, 9] contributes substantially to ED overcrowding [10]. Individuals with MD also have higher prevalence rates for co-occurring physical conditions, further increasing their ED use [11]. High ED use is costly and strongly indicative of inadequate care, which in turn contributes to worse patient outcomes, staff turnover and general dissatisfaction [12, 13]. For family members, caring for a loved one with MD is associated with a 29–60% probability of enduring significant psychological distress [14], which may also lead to ED use.

In this context, the need to reduce ED use while improving services and responsiveness to patient needs is crucial [2]. Innovations may be introduced within ED by MH or SUD liaison agents, for example, or developed collaboratively with outside services, e.g. crisis stabilization centers. Three basic types of innovations implemented in general or psychiatric ED, but mainly the latter, were identified in the literature [6]. The first type, “brief intervention services” (e.g. ED liaison agents, flow strategies, case management) aimed to improve ED assessment time, treatment adequacy, and appropriate patient discharge and referral to ambulatory care [15]. The second, “crisis-related services”, (e.g. crisis teams or stabilization centers, home treatment teams) help MH patients in crisis to receive more rapid and adequate help at home or through alternative services, rather than ED. Both of these innovations were developed by hospitals or community organizations working in collaboration with ED [16,17,18]. The third type of ED innovation, less often used, concerns “peer-user support services” established to promote patient recovery [19, 20] and focused on the reduction of psychosocial burden. These services could also be deployed for families, yet no research on family-peer support services was identified in the ED literature.

Studies have investigated the implementation of MH innovations in ED [16, 21,22,23], yet the evidence on their results and efficacy is mixed [15, 24, 25]. Thus, further research is needed. Concerning implementation, studies on MH innovations have investigated system-level, organizational and individual (staff, patient, family) factors [23]. Factors hindering implementation at the system level due to lack of funding, access, continuity or required intensity of services were common to the three ED intervention types [21, 22, 25], whereas organizational-level barriers related more to problems with inter-professional collaboration and communication [24, 26]. A lack of ongoing quality assessment [21] and structured referral processes [15] were key implementation factors reported in brief intervention services. The risk of cultural clash between community organization or peer support staff providing services and health care professionals, particularly concerning recognition of professional expertise, was another important issue identified [27,28,29]. Training and supervision of peer support staff by health professionals were viewed as key indicators of success [30]. Regarding individual factors, interpersonal skills were identified as critical [24, 26]. Certain patients were difficult to engage, such as those with personality disorders or SUD [16, 23]. Concerning the impact of ED innovations, reduction in costs [25, 31, 32], ED wait times [33,34,35], ED use [6, 15, 35] and hospitalization rates [16, 18, 36] were most often cited, regardless of intervention type, while the adequacy of referrals generally improved [27, 37, 38] in brief intervention and peer-user support services. In terms of health outcomes, patient symptoms decreased for the three intervention types [25, 26, 36]. Although few studies have investigated patient satisfaction [23, 39, 40] or quality of life [19], results tend to be positive.

To our knowledge, no previous study has investigated these three types of innovative ED interventions in terms of their response to patient and family needs. Moreover, the combined experiences of managers, clinicians, patients and families have rarely been investigated using survey techniques and clinical records, nor has the adequacy of care been evaluated in this context. As mentioned, studies investigating the implementation of family-peer support teams in a psychiatric ED were not identified. Improving services for families should also be a priority, given their level of burden and major role in patient MH recovery [20]. Using mixed-methods with data triangulation, this study aimed to: (1) describe the implementation of three innovative interventions provided by a brief intervention team, crisis center team, and family-peer support team in a Quebec psychiatric ED, including the identification of implementation barriers; and (2) evaluate the impact of these interventions on MH services use and response to user needs.

Methods

Study setting

The study was conducted in the psychiatric ED of a MH university institute in Montreal. In 2019, 3995 patients made 5980 visits to this ED, which serves the local community (368,740 inhabitants) and patients from outside areas due to its strong reputation for providing specialized MH services. MH primary care in the territory is provided by local community health service centers, general practitioners and psychologists in private practice, and by community MH organizations (e.g. crisis centers, family and patient support groups).

Data collection

Participants were patients, family (i.e. users) or staff members providing the interventions. User members had to be at least 18 years old, and had visited the ED about 8 months prior to their inclusion in the research. Except for those served by the family-peer support team, participants had to give permission to transfer their medical records from the MH university institute to the research team. Intervention staff members referred patients or family members to researchers on a voluntary basis between January and December 2019, whereas recruitment from the family-peer support team was discontinued in May due to interruption of this service. At least 30 users for each of the three innovations were targeted for recruitment to the study, meeting minimal requirements for quantitative analyses. Recruitment was also planned for at least a one-year period to allow for the recruitment of an adequate number of users [41]. Interviews were held in a private room at the intervention offices or by telephone, and a small financial compensation was provided to patients. Staff participants were either the intervention managers or clinicians. All participants provided written informed consent. The research ethics board of the MH university institute approved the study protocol.

Data emanated from three sources: 1) a user questionnaire developed for this study, eliciting yes/no or 5-point Likert scale responses, and two open questions which users completed with interviewer assistance (supplementary file); 2) focus groups for staff involved in implementing the interventions (including a brief questionnaire for sociodemographic data); and 3) medical records for patients served by the brief intervention and crisis center teams. The user questionnaire covered: participant socio-demographic information; services received from the intervention including quality assessment; MH service use in the 12 months prior to ED visit leading to use of the brief intervention or crisis center teams, and in the 6 months after discharge from these interventions. The user questionnaire also included some clinical data not covered in medical records (i.e. perceived MH and physical health) or data that may be under-reported in medical records such as SUD, where two questions were introduced (Did you have a drinking or drug problem in the past 12 months?). For SUD data, this question and information reported in the medical records were merged. The two open questions for users concerned reasons for ED visit leading to the innovative interventions used, and an assessment of their key components. The user questionnaire required 30 min to complete. Focus groups with staff respondents used an interview guide developed for this study, which described or evaluated the following: the innovations (e.g. mandate, functioning), possible system or organizational barriers, and patient profile issues. Staff sociodemographic data were also obtained (e.g. age, sex, seniority). The focus group lasted 90 min. The user questionnaire and interview guide for the staff focus groups were developed by the research team based on the literature on innovations implemented in ED and were revised by key decision makers with strong ED expertise involved in the study.

Data from the third source, patient medical records, were collected 12 months before the designated ED visit and 6 months after discharge from the intervention, and included: patient diagnoses, illness acuity at ED visit before intervention referral, specialized outpatient services used, ED visits and hospitalization. Diagnoses were based on the International Classification of Diseases Ninth Revision (ICD-9) for outpatient services and on the Tenth Revision (ICD-10-CA) for ED visits and hospitalization. MD included: common MD (depressive disorders, anxiety disorders, adjustment disorders), serious MD (schizophrenia spectrum and other psychotic disorders, bipolar disorders), personality disorders, and SUD. Co-occurring MD and SUD, and MD and chronic physical illnesses (e.g. cancer, diabetes) were also considered (including data on SUD reported in the user questionnaire). Illness acuity was measured with the five-level Canadian Triage Acuity Scale [42], used to determine treatment priority in ED and ranging from level 1 (urgent situation with vital risk) to level 5 (least urgent).

Analysis

This study was based on mixed methods using a sequential explanatory design [43], where qualitative findings (i.e. description of the innovation and implementation issues) are used to explain and complement the quantitative data, especially the implementation outcomes or impacts, taking into consideration patient profiles (user questionnaire). Regarding quantitative data, descriptive statistics were first produced, including percentages for categorical variables and mean values for continuous variables. Comparative analyses were conducted to test before-after differences in the interventions using chi-square tests or Fisher’s exact tests for categorical variables, and t-tests for continuous variables. The qualitative phase used a content analysis approach, where themes identified in the literature, or based on the interview guide or open questions framed the analysis, including the possibility to code emergent themes. The qualitative analysis included a four-step process [44] as follows: 1) audio-recording of interviews and verbatim transcription; 2) preliminary readings; 3) selection and definition of classification units for open-questions (patients or family members) or dimensions of the focus group interview guide (staff); and 4) separation of content into units of meaning.

Two research team members independently read the interviews with users and staff. The coding and qualitative analyses (users, staff) were produced by one team member, and validated by a second one for 30% of the content. Interrater reliability was validated at roughly 90%. To reduce the information, quantitative data (for each innovation and type of user) were summarized in tables, while the qualitative data for staff or users were summarized in a short report taking into account convergences and divergences among the innovations. This process was validated by the entire research team. Triangulation of the quantitative and qualitative data was then conducted [45], integrating user profiles (questionnaires and medical records), the three innovations including functioning and implementation issues (focus groups), services received by users, as well as other user outcomes or impacts of the innovations (questionnaire and medical records).

Results

Sample description (based on user or staff member questionnaires)

Of the 119 patients or family members referred by the intervention staff, 101 participated for an overall response rate of 85%: 44/48 (92%) from the brief intervention team, 37/43 (86%) from the crisis center team, and 20/28 (71%) from the family-peer support team. Average patient or family member age was 41 years old; 62% were female; 69% single and 75% had a college education (Table 1). Half worked, but only 35% full time. For 44%, family income was from Can$20,000 to $69,999. Most (81%) had close friends, three on average. Eighty-five percent of brief intervention and crisis center team participants reported a diagnosed MD: 52% common MD, 25% serious MD, 25% personality disorders, and 17% SUD; 12% reported co-occurring MD-SUD, and 2% MD-chronic physical illnesses, while 48% had experienced suicidal behaviors. Over one-third (38%) of innovation patients or family members rated their MH, and nearly a quarter (24%) their physical health, as poor or fair. Overall, crisis center intervention patients had worse living conditions than those served by the brief intervention team (e.g. fewer in autonomous housing; less education and employment; more single; more with serious MD, personality disorders, suicidal behaviours, or co-occurring MD-SUD). Response rates for intervention staff were 93%: 5/6 professionals from the brief intervention team, 5/5 from the crisis center team, and 4/4 from the family-peer support team participated in a focus group. Most staff were women (93%), 41 years old on average with 14 years of experience, 11 years of which in their current position.

Table 1 Patient and family member sociodemographic and clinical characteristics

Description of the three innovations (based on focus groups with staff members and statistics reported from the user questionnaire)

Implemented in 2016, the brief intervention team included a part-time psychiatrist (18 h/week), a full-time nurse and an administrative agent, who managed 160 patients yearly during working hours for an average of six visits per patient over 5.5 months (Table 2). The ED referred patients assessed as priority 2–3 in the triage, patients from territories not served by the MH university institute, or others without MH follow-up. All patients received brief psychosocial interventions including crisis resolution. Medication was managed for 95% of them. After discharge, 61% were referred to outpatient services: 34% to public primary care, 32% to community organizations, and 25% to specialized services.

Table 2 Patient and family member use of interventions and perceived quality of mental health (MH) services

The crisis center team, created in 1987 and reactivated in recent years, included one clinician working in the ED on a half-day/week basis, who transferred ambulatory patients to the crisis center after ED evaluation and discharge. A community organization offering short-term accommodation and community crisis follow-up, with 12 full-time professionals on staff, managed the team. About 150 ED patients were referred to the crisis center team yearly. They included patients prioritized as 4 or 5 in the ED triage, those in crisis from bereavement, separation, job loss, etc., not considered dangerous and treatment compliant. Most (78%) received three-day crisis accommodation on average, and 66% received a mean of three psychosocial interventions over an average 8-week period.

Finally, the family-peer support team started in 2017 with inclusion of a family-peer staff member in the ED in conjunction with the community organization, which included 9 full-time staff offering MH information, individual family or group support, respite services and outreach to families of patients with MD who presented directly at the ED. Flyers available in the ED provided further information about the service. The family-peer support team reached about 35–40 families monthly until May 2019 when the service terminated for lack of resources. Families of ED patients were offered psychosocial support, either single sessions (usually one only) or telephone support. Most (80%) were referred to the family-peer support group for additional follow-up. For all three interventions, service quality was rated as adequate: from 2.99 for the family-peer support team to 2.72 for the brief intervention team, based on a 5-point Likert scale.

Factors hindering implementation of the innovative interventions (focus groups with staff members)

According to the qualitative results, patients who used the ED and one of the three interventions did so because MH services had not responded adequately to their needs for access to or continuity of services. They identified the following MH system failures: insufficient knowledge of MD among family doctors, lack of specialized MH services, unacceptably long wait times for psychiatrists or primary care psychosocial services in local community health service centers, lack of outpatient MH services or insufficient intensity in follow-up (Table 3).

Table 3 Box quotations

According to staff, system-level barriers included underfunding for MH, especially ambulatory care, and community follow-up, which forced teams, particularly the brief intervention team, to support patients longer than expected. Community organizations such as the crisis center and family-peer support teams were disproportionately underfunded, which limited potential ED partnerships, especially for the crisis center, and contributed to the termination of the family-peer support innovation.

Organizational challenges affected inter-organizational collaboration, frequent staff turnover, engagement of psychiatrists, operations and space, and culture clash. The MH university institute covered territories that operated differently, while the brief intervention team had to refer patients to providers in these diverse localities, which involved important challenges. Staff turnover particularly affected ED managers and physicians. Physicians were numerous, as most work in the ED part time, yet this hindered their involvement with the new interventions. Physicians also seemed especially uninterested, with very few attending information or management sessions on the innovations. Operational and space challenges mainly affected the crisis center and family-peer support teams. Crisis center staff had to arrange for patient transfers from the ED, often involving long waits for patient discharge. Few confidential and quiet spaces were available in the ED to the crisis center and family-peer support teams for patient or family member follow-up and psychosocial support. As they represented community organizations, ED professionals tended not to recognize sufficiently the expertise of either teams, which reduced the number of patient referrals to the crisis center. As for the family-peer support team, the integration of family perspectives and coordination with ED staff were not viewed as optimal.

Concerning patient profiles, those with antisocial personality disorders and problems with the law were evaluated as difficult to treat, as were family members who became aggressive or overly critical of peer treatment. Some ED patients also forbade clinicians from contacting their families about their conditions or from integrating family into their treatment. Others had little contact with their family.

Innovation impact

Based on the qualitative investigation included in the user questionnaire, patients and family members mainly underlined that staff were compassionate and sensitive to them, that they listened carefully and genuinely helped. Patients also reported receiving rapid treatment; the treatment steps were explained, subsequent appointments set, and information on MH services made available. In the brief intervention team especially, medication effects were adequately explained, and comprehensive treatments offered, including psychosocial interventions. Patients especially appreciated their short respite periods in the comfortable and friendly environment of the crisis center, as well as the ability to reach clinicians on a 24/7 basis. As with the family-peer support team, crisis services were much appreciated for their easy access from the ED and the possibility to hold consultations during the wait period at the ED (Table 3).

As reported in patient medical records or the user questionnaire, overall MH service use diminished significantly among patients after discharge from the brief intervention and crisis center teams: 57% of participants used MH services post-discharge compared with 72% prior (Table 4). Regarding specific services, 69% of patients consulted general practitioners before using the interventions, compared with 41% afterward; six outpatient hospital services were provided before and four after the use of the interventions; 42% of patients used the ED before vs 25% after discharge, and 35% were hospitalized before vs 19% after discharge. Crisis center team patients were followed-up in MH services more often than patients with the brief intervention team. In fact, for the 6-month period after discharge from the crisis center team, no significant decrease was observed in overall service use for this specific group (76% before vs 70% after, compared with 68% vs 45% for the brief intervention team). However, a significant reduction in outpatient hospital service use (7 before vs 5 after) and hospitalization (52% vs 30%) were recorded. Regarding adequacy of care, patients, particularly those with the brief intervention team, claimed that MH services moderately addressed their needs, with a significant positive change after discharged from the interventions as compared with before (2.65 vs 2.22 overall, and 2.70 vs 2.05 for the brief intervention team). Regarding reasons why needs were not met, 89% of patients identified system problems (e.g. accessibility, quality of services), 59% individual reasons (e.g. no time, stigmatization), and 18% preferred to manage on their own.

Table 4 Mental health (MH) service use

Discussion

This study aimed to describe the implementation and impact of three innovative interventions designed to improve service use in a Quebec psychiatric ED and better respond to the needs of patients and family members. The three interventions examined in this study differed in certain respects from interventions described in the literature. As a new type of innovation, the brief intervention team was similar to the brief case management strategy, identified in the literature, in providing both medication and psychosocial interventions [46]. This innovation also resembled the flow strategy [15], and liaison psychiatry services implemented mainly in general ED [33], given the importance of appropriate referrals to community MH service providers for this clientele [23]. However, these interventions served patients without current MH follow-up and those residing outside the hospital territory, and not only patients with co-occurring MD-SUD or specific physical illnesses as found in the literature for the other related interventions. Rather, the team described in this study served a clientele without family doctors, helping them to avoid inappropriate ED use. The crisis center and family-peer support teams, as community services that collaborated closely with ED staff, met the recommendations of the American Psychiatric Association for potentially reducing ED and hospitalization [27] by offering viable alternatives [16, 47]. The major difference between the crisis center team intervention in this study and home treatments teams described in the literature was that treatment was provided in a small community organization, which offered a homelike, friendly environment [17]. Moreover, clients of the crisis center team were experiencing psychosocial crisis more than MH crisis, as envisaged in the original model [18]. While peer users in ED were the subject of a clinical demonstration project [48], the inclusion of a family-peer in the ED has not been previously reported as already mentioned. The family-peer represents an interesting strategy for reaching families and offering them coping tools to better support loved ones with MH issues and their own wellbeing, a strong advantage that was reported by all group study members interviewed. Family-peer staff played the same role as peer-user staff at the ED or in similar health teams, counseling peers through their own MH experiences [20].

Regarding factors that hindered implementation identified in this study through focus groups, most were previously reported in the ED literature and related mainly to heath system issues: lack of resources for adequately meeting patient needs, long delays in accessing psychosocial or psychiatric services, and insufficient intensity of MH follow-up [31]. ED use is often reported in the literature as resulting from lack of MH expertise among family practitioners [49,50,51] and was observed in this study by both users and staff members. Community services like the study crisis center and family-peer support teams were also disproportionately identified as underfunded [31], contributing to high ED use and incentivizing the brief intervention team to support patients longer than planned. In the MH university institute where the study took place, roughly half of ED patients were prioritized as 4 or 5 at triage [42]; that is, they were less urgent cases more appropriate for ambulatory care than the ED.

Regarding organizational issues, staff members underlined communication problems between staff and organizations, often related to their distinct cultures, values and practices as observed in other studies [25, 27, 28]. While recovery is now central to MH treatment, the lack of appreciation for peer expertise among health professionals reported in our study is frequently mentioned in the literature [19, 20], as is the problem of staff turnover [52,53,54] and insufficient involvement in innovations [23], all important barriers to implementation also identified in this study. Lack of private space in the ED for patient or family member consultations, reported in the qualitative findings, was previously identified in a study of brief intervention services [23]. Concerning individual barriers, difficulties in treating patients identified by staff participants were like those identified in previous studies as affected by stigma, e.g. patients with co-occurring MD-SUD or personality disorders. MH professionals in the ED, like those in primary care, are often considered reluctant to treat patients with more serious or complex MD [55]. Finally, the desire of families to be engaged in the care of their loved ones was strongly hindered by patient confidentiality issues according to our findings, as in previous research [56].

Overall, patients and family members in this study appreciated the three innovative ED interventions, as in related studies [16, 23, 39]. We found that empathy and active listening skills among staff, including peers, had direct benefit for patients or family members, as previously reported [39]. Studies have also emphasized that good patient-carer therapeutic relationships improve care continuity [57, 58] and patient functioning [46]. These aspects were also confirmed by staff members who identified these qualities as facilitating factors for these types of innovation. In this study, patients received timely treatment and follow-up by innovation staff following ED discharge, contributing to patient satisfaction. Research shows that the shorter the wait time the more patients were satisfied with services [59, 60]. Biopsychosocial treatments including medication management were previously found to increase patient satisfaction [46], as also evidenced in the present brief intervention team. Surveys have found that patients in crisis prefer to be treated in a friendly, homelike environment like that offered in the study crisis center team rather than in ED or hospitals [30, 36].

The reduction in patient service use after discharge from these innovative services coincided with previous findings [6, 15, 18]. ED use diminished particularly among patients with the brief intervention team, similar to results in liaison psychiatry services [33], flow strategy [15] and brief case management studies [61]. Consultations with general practitioners also significantly diminished and perceived adequacy of care increased. To our knowledge, these results have not been previously reported, suggesting that the frequency and type of help received responded adequately to patient needs and improved their MH conditions. Studies on innovations similar to those of the brief intervention team have reported decreased symptoms [26], suicidal behaviors [22, 34] and psychosocial problems [46]. Since most patients (61%) were given MH referrals following the brief intervention, it may be hypothesized that these referrals were adequate and that MH network functioning was effective [62, 63]. However, an alternative explanation for this decrease in service use may be that patients never accessed services, that discontinuity in care occurred or that patient non-adherence to services increased after discharge from the intervention and referral – however the study found that patients were generally satisfied with services. As with use of the crisis center team in this study, a significant reduction in use of outpatient hospital services and in hospitalization after discharge have been reported [18, 35]. These results corresponded to the profiles of patients with more serious MH using in the study more the crisis center, who usually are also treated in specialized services and experience frequent hospitalization [18, 47]. Research on adequacy of help among patients confirms the difficulty of treating patients with more serious and complex MH profiles, as in the present group, and in meeting their needs [16]. Thus, they often require specialized, ongoing, and intensive outpatient care.

Limitations

This study had certain limitations. First, as the sample was selected from a psychiatric ED in a Quebec MH university institute, and on a voluntary basis, the findings may not be generalizable to all ED, and particularly not general ED or those located in semi-urban or rural areas. Second, the innovations targeted few ED patients compared with the overall yearly patient volume in the hospital ED. More effort is needed to promote and improve ED efficiency and response to patients with MH needs and their family members. Also, the small sample size for users of each ED intervention prevented the generation of more complex statistics, especially for the family-peer user group. Third, it was also difficult to assess the family-peer support team, as no research on this type of intervention was identified in the literature. Finally, the six-month assessment period for investigating outcomes may be insufficient to ensure that results are sustained.

Conclusion

This study was original in focusing on brief interventions in a Quebec psychiatric ED for a clientele without family doctors, and on crisis prevention and the use of family-peer support teams for improving MH services and for reducing ED use and hospitalization among MH patients. Implementation issues identified in this study were similar to those reported elsewhere in the context of ED or in the general implementation literature. Managers need to better identify key issues before embarking on implementation initiatives: improving staff training and actively involving them in new interventions, particularly physicians; developing collaborative tools geared especially to preventing cultural clashes between staff and organization, and encouraging continuous quality assessment. Reduced patient service use after discharge reflected the positive impact of these innovative interventions, leading to greater adequacy of care for patients as well as improved use of outpatient MH services, which seemed adapted to patient profiles. Family members were grateful for the help they received in the ED, even though this type of intervention would need to be more broadly investigated. Future research needs to confirm the pertinence of these interventions and promote broader ED strategies to improve access to MH services and decrease ED use for MH reasons.