Abstract
Objectives:
To report on the perceived barriers surrounding the use of telepsychiatry for emergency assessments and our approach to overcoming those barriers to achieve successful implementation of a program to increase access to emergency psychiatric assessment in a Canadian urban setting.
Methods:
We conducted a survey of emergency care staff to inform the implementation of an emergency telepsychiatry program in the urban setting of Winnipeg, Manitoba, where hospitals have variable on-site emergency psychiatric coverage. We analyzed survey responses for perceived barriers we would need to address in implementation. We employed implementation strategies for each barrier and scaled the program to three sites over the first year. Data from the first year were collected including number of telepsychiatry assessments, reasons for referral, wait time, and percentage of patient transfers avoided.
Results:
Survey respondents (N = 111) had little prior exposure to telepsychiatry, but the majority were open to its use for emergency psychiatric assessments in the region. We identified three categories of perceived barriers: clinical, logistical/technical, and readiness barriers. Implementation planning addressed each barrier, and a hub-and-spoke program was launched. After the first year, the program had one hub serving three spokes, and 243 emergency telepsychiatry assessments had been completed. After 12 months, we were avoiding 65% of patient transfers.
Conclusions:
By conducting a user survey to identify perceived barriers, and addressing these during implementation, we successfully scaled our emergency telepsychiatry program across our region. Our report of this experience may benefit others attempting to implement a similar program.
Introduction
The use of telemedicine for emergency care is expanding with growing evidence supporting its acceptability, clinical use, and cost effectiveness. 1,2 Telemedicine for psychiatric emergency care is a viable option; however, use in these settings remains uncommon. 3 Moreover, while the focus of telemedicine tends to be for care provision to rural and remote communities, some urban settings have adopted emergency telepsychiatry as a solution when psychiatric services may be centralized and/or limited. 4 For example, South Carolina established a statewide emergency telepsychiatry service in response to the fact that only 32% of the emergency departments had access to an on-site psychiatry team. 5 The centralized, hub-and-spoke program completed over 9,000 telepsychiatry assessments in the first 4 years of operation. An analysis that matched individuals presenting to hospitals participating or not participating in the program demonstrated significantly lower rates of admission and hospital costs and higher rates of postemergency follow-up for individuals seen at the telepsychiatry sites. 6 Limited discussion of telepsychiatry for emergency care in Canada exists in the literature, although some data support cost savings for emergency suicide risk assessment to remote communities. 7
Similar to many health service innovations, 8 the widespread uptake and sustainability of telemedicine remain a challenge. 9,10 Since the use of telemedicine in emergency psychiatry settings is more novel, 3 there is little discussion of the implementation challenges in these settings which may differ from routine outpatient mental health care. Some emergency telepsychiatry programs, such as the South Carolina service, 6 and a rural hub-based model in New South Wales, Australia, 11 have provided detailed program descriptions in the literature to assist other jurisdictions with implementation. Shore et al. 12 described emergency management guidelines for telepsychiatry based on experiences across three emergency telepsychiatry programs in the United States and Australia, serving a range of populations including rural communities and Veterans. Although these guidelines and program descriptions are helpful, any new service must consider not only the specific innovation and supporting evidence but also the influence of human-based resources and the impact of local contextual factors. 8 A more robust reporting of service planning and implementation is rarely found in the published literature and offers valuable information for others intending to adopt similar innovations. 10 In general, there is little information available to inform the implementation of telepsychiatry in Canadian health systems. 13
This article describes the preparation for and implementation of a regional emergency telepsychiatry program in an urban setting in Manitoba, Canada. We conducted a user-based survey to identify perceived barriers to implementation and describe the approaches used to address those barriers. We also report preliminary data on the performance of the program over the first year of operation.
Materials and Methods
Setting
Winnipeg is the capital city of Manitoba, a centrally located province in Canada. The city spans 464 km2 and has a population of approximately 750,000, home to roughly half of all residents of Manitoba. Winnipeg has six hospitals with 24/7 services. As a result of a series of significant changes in the system, psychiatric support has been very limited at three of these hospitals. After-hours psychiatric coverage had been largely removed in 2013 following the opening of the Crisis Response Centre (CRC), a stand-alone 24-hr facility providing voluntary emergency mental health services to the entire city in a stepped care model. Additionally, the appointment of dedicated daytime emergency psychiatrists was not realistic because personnel were limited, and volumes did not support it. Partial daytime on-site support persisted at two of the three hospitals, provided by the sites’ inpatient psychiatrists but was often restricted to conducting involuntary assessments under the Mental Health Act. A protocol was initiated to transfer voluntary individuals requiring psychiatric assessment from the other hospitals to the CRC and involuntary patients to an emergency department with on-site psychiatric support. If inpatient beds were not available at the time of assessment or later assessment was deemed necessary, patients returned to their originating site to wait, sometimes transferred a second or third time for a disposition to be determined. These transfers created vulnerabilities in the system, at times leading to cases that attracted negative media and government attention. Starting in 2017, additional consolidation of health-care services at those three hospitals was planned, including the phased conversion of two emergency departments to urgent care centers, and the closing of two psychiatric inpatient units to merge resources at fewer facilities. Coincident with these changes, further reduction in daytime emergency psychiatric staffing occurred or was anticipated. This prompted the need for a strategy to increase access to psychiatry and reduce patient transfers.
In the 6 months leading up to the launch of the urgent telepsychiatry program at the end of 2017, the three hospitals with limited psychiatric support saw a combined average of 427 mental health and addiction presentations per month, with an average of 33 (7.7%) transferred to the CRC for psychiatric consultation. During the same period, the CRC saw an average of 651 presentations per month, with an average of 74 (11.6%) being referred to the on-site psychiatry team for assessment. The CRC psychiatry team was averaging three to four assessments per day, well within the capacity of a two- to three-member team (psychiatrist plus physician assistant and/or psychiatry resident) from 08:00 to 00:00 hr.
Technology
Manitoba has a secure telehealth network, MBTelehealth, established in 2001. 14 MBTelehealth uses a secure link to provide videoconferencing for health-care services, continuing education, and meetings to approved sites across Manitoba. The network is compatible with local privacy standards for personal health information storage and exchange and can be accessed through fixed or mobile equipment that is wired in or remotely through desktop and personal devices over a secure web-based platform linked to individual user accounts. Within approved programs, urgent, unscheduled videoconference calls can be initiated between sites by dialing the network call number assigned to the referring site. Technical support is available 24/7 for urgent care programs.
Emergency Telepsychiatry Planning and Implementation
We started with a user survey to assess perceived barriers for an emergency telepsychiatry service in the Winnipeg region. This informed our implementation approach that included rapid iterative testing of the model and collection of feedback to guide modifications before widespread rollout.
User survey
A web-based survey targeting frontline staff working at the CRC and all hospitals within the Winnipeg Region was developed and hosted on SurveyMonkey™. In June 2016, survey invitations (N = 290) were emailed to practicing emergency physicians, psychiatrists and psychiatry residents (years 2 to 5), psychiatry emergency nurses, physician assistants supporting the psychiatry team, and administrators (CRC only). The survey remained open for 1 month with a reminder email sent 10 days prior to the survey closing. To improve response rate, the survey invitation offered a chance to win one of the three $25 gift cards as an incentive. Participants were informed at the beginning of the survey that participation was voluntary and anonymous, and no personal identifying information was collected.
The survey collected data on demographics and previous telehealth exposure (overall and for psychiatric care in rural and urban settings specifically), willingness for an emergency telepsychiatry service, and perceived barriers to implementation of a local emergency telepsychiatry program. Demographics included age, gender, job title, and the number of years worked in their role. Respondents were asked to indicate by yes or no if they were open to a local emergency telepsychiatry program. Respondents were then asked to identify the top three barriers they perceived to implementation of a local emergency telepsychiatry program using open-text responses. The web survey had one question per page, and participants could review and change their answers throughout the survey. Prior to the survey launch, the survey was reviewed by a small group of stakeholders and a communications expert resulting in some minor wording changes.
Survey responses were exported from SurveyMonkey for analysis. Demographics and telehealth exposure were descriptively analyzed using Microsoft Excel. Qualitative content analysis 15 was conducted for open-text responses on perceived barriers. All responses were independently coded by two members of the team and reviewed by a third member with expertise in telehealth and implementation. Initial coding was inductive, and codes were clustered into higher order categories and subcategories. Discrepancies were discussed between team members until consensus was reached. Barriers were quantified at the subcategory level, and for each discipline, the proportion of barriers in each subcategory was calculated.
Service implementation
Based on survey findings, available guidelines, 12 and descriptions of services elsewhere, 6,11,16 we designed a hub-and-spoke emergency telepsychiatry service and created an implementation plan informed by the survey findings on perceived barriers to success. We then used a variety of implementation strategies to initiate, refine, sustain, and scale the service. We tracked referral volumes, wait time, transfers, and overall workload at the hub for the first year following program launch in the last week of December 2017.
Results
User Survey
Of 290 surveyed individuals, 111 responded providing an overall response rate of 38%. Response rates were highest among psychiatry residents (26/42, 62%) and psychiatrists (33/65, 51%). Conversely, emergency physicians and psychiatry emergency nurses had lower response rates (33/138, 24% and 14/38, 37%, respectively) but still represented all hospitals in the region. Responses were obtained from 2/2 physician assistants and 3/5 CRC administrators. Demographics and previous telehealth exposure among survey respondents are summarized in Table 1. Age range varied by discipline as did respondent gender, reflecting the actual age and gender distribution of the staff groups in Winnipeg. Similarly, most psychiatrists and emergency physicians were in their role for more than 10 years, while most psychiatry residents and psychiatry emergency nurses were newer to the roles. A minority of respondents had used telehealth previously for psychiatric care, yet the vast majority were open to the idea of a local emergency telepsychiatry program.
Demographics and Telehealth Exposure by Respondent Discipline.
Note. N = 111. Presented as number (%).
a Defined as a remote community that does not have year-round road access.
Qualitative content analysis of responses regarding perceived barriers to a local emergency telepsychiatry program produced three categories of barriers, each with subcategories (Table 2). The three categories of perceived barriers were “clinical,” “logistical and technical,” and “readiness.” Clinical barriers included issues regarding the ability to determine patient suitability for remote assessment, the ability to remotely perform the clinical assessment including accuracy and rapport, and the ability to gather and document required information. This category also included medicolegal concerns related to Mental Health Act assessments and provider responsibility. Logistical and technical barriers included barriers related to coordination of assessments, patient flow, telehealth costs, physician billing, and technological problems. Readiness barriers were preparedness factors involving people, both from a staffing and user engagement perspective, and technology including availability of both equipment and space. Barriers were distributed differently across respondent discipline (Figure 1). Psychiatry residents had proportionally more logistical concerns and concerns regarding information sharing. Psychiatry emergency nurses, psychiatrists, and psychiatry residents had a similar proportion of concerns about accurate assessments, staffing, and people readiness. Physician assistants were concerned only about accurate assessments and logistical challenges. Administrators wanted to ensure that technology and people readiness were addressed.
Qualitative Analysis of Perceived Barriers Reported by Respondents.
Note. N = 111.
Barriers by subcategory, displayed as the proportion of all barriers reported among each respondent discipline.
Service Implementation
Iterative approach and scale-up
In response to the barriers identified in the survey, we applied various implementation strategies (Table 3), drawing on established implementation approaches and principles. 17,18 In late December 2017, our hub-and-spoke service opened to referrals from a single low-volume site that had no on-site psychiatry support and strong leadership with a lens for quality and innovation. With anticipated low referral volumes and in the absence of staffing or funding to create a dedicated emergency telepsychiatry team, we opened the service 08:00 to 17:00 hr Monday to Friday to align with consistent psychiatric coverage by a core group of experienced emergency psychiatrists. These psychiatrists were on salary and provided team-based daytime psychiatric care at the CRC (the hub), including seeing the patients transferred from other sites. After hours (until midnight) and weekend coverage at the CRC is offered by over 40 psychiatrists and psychiatric residents on a rotating schedule, we opted not to engage this group until the procedures were tested and refined. In March 2018, we expanded to a larger volume referral site when psychiatric coverage to that emergency department was reduced, and in October 2018, we added the third site. All transfers and telehealth assessments were tracked in real time. For all sites, we held an onboarding meeting (in person or by videoconferencing) with key stakeholders and a follow-up meeting 1 to 2 months later where an audit of all assessments completed at that site was reviewed. Here, successes were celebrated, and improvement strategies were discussed. Following this meeting, each new site was invited to attend regular check-ins held every 4 to 6 weeks with representation from all spokes and the hub.
Implementation Strategies to Address Perceived Barriers Identified in User Survey.
The original model
Requests for emergency telepsychiatry assessment were initiated by the emergency physician at the spoke site to the psychiatrist at the CRC hub; if both agreed that the patient was suitable to be seen remotely, available information was shared, and the assessment time was determined. Guidelines on suitability and process were posted at all sites. All patients required accompaniment by a health-care provider throughout the assessment, with collaborative disposition management. Recommendations were either implemented directly by the hub team with support from the referring site (e.g., hospital admission) or communicated verbally to the most responsible provider (e.g., medication changes requiring prescription). The psychiatric assessment report was forwarded to primary-care providers and other specialists.
Preliminary program performance
In the first year of operation, 243 assessments were completed, averaging 31 per month over the last 3 months. Of the 243 assessments, 223 were done on the same day as the request, with a median time from consult request to assessment of 2.1 hr (interquartile range = 1.1 to 3.1; target = 3 hr). The primary reasons for referral provided by the emergency physician at the time of request in descending order of frequency were (1) risk assessment, predominantly suicidal ideation with or without an attempt (n = 157, 64.6%); (2) psychosis (n = 48, 19.8%); (3) mood and/or anxiety disorder (n = 28, 11.5%); (4) substance use (n = 5, 2.1%); and (5) confusion (n = 3, 1.2%). Workload at the hub had increased by 42% between the 6 months preprogram (mean 3 to 4 assessments/day) and the second 6 months of program operation (mean 5 to 6 assessments/day). In part, this reflected increases in total presentations at the hub (20% increase) as well as some increase in workload from the spokes which saw declines in on-site support and an 8% increase in total mental health and addiction presentations. The percentage of transfers avoided increased from 0% preprogram to 65% in December 2018 (Figure 2). In some cases, patients still preferred voluntary self-transfer to CRC, or staffing limitations and patient factors made a telehealth assessment difficult.
Hub workload plotted by month, including on-site consults, transfers, and telehealth for the 6 months prior to the emergency telepsychiatry program and during the first year of operation. The bar graph represents counts plotted on the primary axis and the line graph is percentage of avoided transfers plotted on the secondary axis. The arrows designate the timing of the onboarding of the 3 sites to the emergency telepsychiatry program. Transfers and telehealth assessments are displayed for all three sites combined.
Current state
At the time of writing this article, our emergency telepsychiatry coverage is provided by the hub-based team 7 days a week, 08:00 hr until approximately 21:00 hr. This team continues to concurrently provide psychiatric coverage to the hub site. There is some partial on-site early day psychiatric coverage (08:00 to 12:30 hr) persisting at two of the three spoke sites; one of them remains an emergency department, the others have converted to urgent care centers. Accompaniment during the telehealth assessment is now flexible, allowing patients to be accompanied by a range of health-care providers or to be seen alone if deemed safe to do so.
Discussion
In this article, we have described our approach to the implementation of an emergency telepsychiatry program in a large urban region. Given the growing comfort with telemedicine, as well as increasingly robust evidence supporting its use, it has been proposed that we may be approaching a tipping point for its adoption whereby it is no longer seen as innovative but rather a pragmatic solution to real problems with access and care delivery. 9 As such, we found that the majority of respondents to a user survey said that they were open to an emergency telepsychiatry program in the context of limited on-site psychiatric support. In a study of a mental health emergency care program that provides phone and video assessment to rural emergency departments in Australia, emergency department staff felt much more confident and supported managing mental health presentations with specialists available remotely. 19 Although less common in urban settings, emergency telepsychiatry programs have demonstrated improved care access and outcomes in United States and Australian contexts. 4,6,12
The primary objective of our survey was to elucidate the perceived barriers among the users who would be essential to the success of any local emergency telepsychiatry program. Although there is literature on barriers to technology adoption in health care 8,20 and specifically to telemedicine adoption 21,22 to draw on, the act of surveying our users was itself an implementation strategy. We presented these findings to our stakeholders and developed implementation strategies to respond to the barriers identified by the frontline staff. These barriers were not altogether surprising or unexpected; other studies have found that staff raise concerns about the ability to conduct assessments accurately and with rapport, technological problems, and perceived potential negative impacts on work flow. 21,22 In general, user willingness to do things in a new way is often identified as a significant barrier to change in health-care settings. 8 Barriers around patient suitability, staffing, and impact on patient flow are more unique in the emergency setting owing to the high pace and increased acuity. One barrier often discussed that we did not find in our survey results was a concern about confidentiality. 21 This may be due to familiarity with our provincial telehealth network which is securely monitored and widely used. Similarly, we did not uncover the same extent of legal and licensure issues that have been reported, 21 which may be a reflection of our local system practices and publicly funded health-care system in Canada compared to other jurisdictions. Of interest, other studies have reported that attitudes toward telemedicine often differ between those who have and have not used it. 22 We had several psychiatrists who were self-proclaimed “skeptics” prior to our program, who later became “converted” after they had conducted a small number of emergency telepsychiatry assessments.
Given a limited amount of literature on emergency telepsychiatry and little experience in our local context, we implemented slowly and rapidly sought to identify needed refinements to the service. This was facilitated by our implementation leads also being at the frontline providing the service. For example, based on an early incident with an agitated patient we modified the protocol for assessing suitability and determining security presence. This was critical because an initial concern among our staff related to patient acuity in the emergency setting and suitability for a telepsychiatry assessment. We did not want to restrict referral criteria which could limit the reach of the service, so we instead made changes to when security involvement would be requested and to what level, the role of the support staff at the referring end, and protocols for conducting assessments and obtaining collateral information. Through regular site check-ins and responsiveness to feedback, we persevered after an incident that may have otherwise led to early staff disengagement and possibly program failure. We have not since experienced any major incidents, although we remain vigilant to this potential. The most significant unintended consequence of the program was the amount of psychiatry emergency nurse (or other emergency department staff) time required to accompany the assessment. Although this helps with patient experience and safety, we have since added options to optimize resources including when patients may be unaccompanied or when assessments can be more focused and collaborative. This is a shift in role definitions for both emergency staff and psychiatrists who do not often work this collaboratively in our system.
Limitations to the generalizability of this report include the unique situation in Winnipeg at the time of this program being established; that is rapid change and uncertain psychiatric coverage across the city. As a result, the top-down levers for change were stronger than they may be in other settings or at different times. In our survey on perceived barriers, we had a modest response rate with respondents being representative of the groups surveyed, but some opinions may have been missed. Patient and family input was not obtained, although other literature reports that patient experience of telepsychiatry is very positive. 1 And while we do not know exactly which of our implementation strategies worked, the approach was not resource intensive and was carefully designed to address the concerns of necessary users.
We plan to conduct a comprehensive evaluation of our emergency telepsychiatry program to determine its impact in our publicly funded health-care system. With additional changes happening in our local system, we expect our model to be sustained with ongoing need and leadership buy-in. Improvement initiatives will focus on reducing nursing staff burden and challenges such as multiple electronic medical records for documentation and outdated communication technology (facsimile, pagers). Experience in other jurisdictions may help us in this regard; for example, the development of a web portal to connect sites and track referrals in the North Carolina Statewide Telepsychiatry Program. 23 Beyond sustaining our current program, we see opportunities for local scale to our mobile crisis team or ambulance services, 24 as well as to the rest of the province where access is an even bigger problem due to geographical barriers.
Footnotes
Authors’ Note
Jennifer Hensel and Reid Graham have contributed equally. Data available from authors upon request.
Acknowledgments
The authors acknowledge the implementation team at MBTelehealth for their support with this initiative. Jacqueline Wong provided support for survey development and administration. Site leads and health-care providers at participating hospitals and the Crisis Response Centre were pivotal to the success of the program.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Jitender Sareen receives royalties from UPTODATE for work unrelated to this project.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Survey participation incentives were provided by the Department of Psychiatry at the University of Manitoba. Jennifer Hensel was partially funded by the University of Manitoba Department of Psychiatry Frank J. Clancy Memorial Research Fund. Jitender Sareen is partially funded by the Canadian Institutes of Health Research Foundation grant (#333252).
