Physicians' perspectives on processes for emergency mental health transfers from university health clinics to hospitals in Ontario, Canada: a qualitative analysis.

BACKGROUND: In Ontario, Canada, there is variability in how students experiencing a mental health crisis are transferred from university health clinics to hospital for emergent psychiatric assessment, particularly regarding police involvement and physical restraint use. We sought to understand existing processes for these transfers, and barriers to and facilitators of change.
METHODS: Between July 2018 and January 2019, we conducted semistructured qualitative interviews by telephone or in person with physicians working at Ontario university health clinics. We developed the interview guide by integrating an extensive literature review, and the expertise of stakeholders and people with lived experience. We analyzed the interview transcripts thematically. Analysis was informed by participant responses to a questionnaire exploring their perspectives about crisis transfer processes. We requested institutional policy and process documents to support analysis and generate a policy summary.
RESULTS: Eleven physicians (9 family physicians and 2 psychiatrists) from 9 university health clinics were interviewed. Ten of the 11 completed questionnaires. Policy and process documents were obtained from 5 clinics. There was variation in processes for emergency mental health transfers and in clinicians' experiences with and beliefs about these processes. Police were commonly involved in transfers from 7 of the 11 clinics, and in nearly all or all transfers from 5 of the 11 clinics. Handcuffs were always or almost always used during transfer at 2 clinics. Three major themes were identified: police involvement and restraint use can cause harm; clinical considerations are used to justify police involvement and restraint use; and pragmatic, nonclinical factors often inform transfer practices.
INTERPRETATION: The involvement of police and use of restraints in crisis mental health transfers to hospital were related to pragmatic, extramedical factors in some university health clinics in Ontario. Exploring existing variability and the factors that sustain potentially harmful practices can facilitate standard implementation of less invasive and traumatizing transfer processes.

Postsecondary students confront many stressors that may contribute to the emergence of mental illness.1 Longitudinal data suggest that rates of mental illnesses are increasing among Canadian postsecondary students, with more students reporting self-harm, suicidal ideation and suicide attempts over time.1 When a student presenting at a university-affiliated health clinic is assessed to be at imminent risk of harm, legislative tools may be used to mandate emergent psychiatric assessment. In Ontario, Canada, under the province’s Mental Health Act,2 a “Form 1 Application by Physician for Psychiatric Assessment” can be completed to initiate this process. When a Form 1 is issued, students must be transferred to a designated hospital facility for evaluation.

Although robust data are lacking, the annual number of student mental health transfers on some Ontario university campuses has increased substantially since 2014.3 However, the processes for such transfers are an underresearched topic, part of a larger gap in research surrounding the conveyance of people experiencing a mental health crisis to hospital from community-based points of care.4–8 There is substantial heterogeneity in transfer processes when a Form 1 is completed for students presenting at university health clinics in Ontario. In particular, the involvement of police and use of physical restraints for transfers vary between institutions8,9 and have been the subject of media scrutiny.3,10–12 These issues are important questions confronting health care professionals.9 Some have argued for police involvement and use of restraints on the basis that they provide a type of safety to patients themselves, to health care professionals and to those transporting patients to hospital.13 Opponents of routine police involvement and restraint use contend that such practices are traumatizing and stigmatizing, and perpetuate the criminalization of people with mental illness.7–9,14–17 Given historical and ongoing police violence against systemically marginalized groups, particularly racialized groups, the actual and potential harms of engaging police in the provision of mental health care must be acknowledged.18–20

Through qualitative analysis of interviews with physicians working at university health clinics, we sought to explore and understand existing processes for emergency mental health transfers in Ontario, and barriers to and facilitators of change. We were particularly interested in the role of police and use of physical restraints.

Methods

Study design and setting

We conducted a qualitative study exploring processes of emergency mental health transfers from university health clinics to hospital in Ontario. Our study included interviews with practising physicians working at student health clinics across the province from July 2018 to January 2019. Our thematic analysis of the interviews was informed by review of institutional policy and process documents, and a questionnaire completed by participants regarding their general beliefs about crisis transfer processes. The study was reported according to the Standards for Reporting Qualitative Research reporting guideline.21

There are 19 public university campuses in Ontario that are listed members of the Ontario University & College Health Association (OUCHA) (http://www.oucha.ca/list.php); 17 of these campuses operate university health clinics staffed by physicians. There are an additional 6 public universities that are not listed OUCHA members in the province.

Author reflexivity

A.C. worked as a primary care physician at a student health clinic for nearly a decade. She has been involved in the care of many students experiencing mental illness. S.N. is a disabled neurodivergent physician. He works in a student health clinic, where the bulk of his practice concerns the mental health of adolescents and young adults. G.N. is an undergraduate student and research assistant. She has lived experience of mental illness and mental health transfers between community sites and hospitals. J.Z. is an emergency department psychiatrist and researcher with qualitative research expertise. She often works with patients who have been sent to hospital from university health clinics after a Form 1 was issued.

A.C. and S.N. cofounded the informal “Best Practices for Mental Health Transfers Working Group” in the winter of 2018. This group advocated for the adoption of flexible processes for crisis mental health transfers at the University of Guelph, authored a commentary in Canadian Family Physician challenging routine use of restraints for such transfers,9 and performed a literature review and policy survey of transfer policies.8

Study sampling

Of the 19 university clinic facilities that are members of OUCHA, 16 were identified as initial recruitment sites. Two universities were excluded: 1 university did not have a university-operated health clinic, and 1 university-operated clinic did not employ physicians. Recruitment for participants from 1 OUCHA-member university clinic was not done per protocol in error (investigator omission). English-speaking primary care physicians or psychiatrists working in the OUCHA-member clinics were eligible to participate in the study. Nonphysicians were excluded, as they do not have the authority to invoke the Mental Health Act.2 The same eligibility criteria applied to physicians working at non-OUCHA university health clinics.

We established a preliminary target of 10–20 participants to optimize the representativeness of the sample and to achieve concept saturation, defined as the point at which none of the coders generated new codes during transcript analysis. 22,23 In thematic qualitative analysis, 10–20 participants are typically required to achieve concept saturation.22,23

We used the professional networks of A.C. and J.Z. to distribute recruitment materials electronically, via personal emails and the OUCHA Listserv. A.C. contacted clinics directly by telephone and email, and used standardized scripts to request the contact information of physicians identified as the clinic’s medical lead. A.C. also posted recruitment materials to the Canadian Physician Mom’s Group, a private Facebook group of more than 9000 physicians; this group was used for participant recruitment in studies by J.Z. (unpublished findings, 2022). Researchers in diverse fields have recruited from similar groups.24,25 Snowball sampling was used: A.C. asked participants whether they had contacts working in Ontario university health clinics who might be interested in participating. These sampling strategies allowed for the potential recruitment of physicians from non-OUCHA institutions.

Data sources and collection

We used 3 data sources: participant interviews, participant questionnaires, and institutional policy and process documents.

Participant interviews

To understand the variability in transfer processes, and physicians’ experiences with crisis mental health transfers of students, semistructured qualitative interviews were conducted with practising physicians working at student health clinics across Ontario. One author (A.C.) conducted the interviews using a study interview guide (Appendix 1, available at www.cmajopen.ca/content/10/2/E554/suppl/DC1) from July 2018 to January 2019. Written and verbal informed consent was obtained from all participants. No incentives were provided for study participation.

The interview guide was adapted from a template created by Dr. A. Ka Tat Tsang (Factor-Inwentash Faculty of Social Work, University of Toronto: personal communication, 2008). We developed the interview guide by integrating an extensive literature review, expertise of stakeholders and people with lived experience, and our expertise in qualitative research methodology. In conducting a literature search and scan of Ontario police department policy documents to write a review article describing policies and procedures for crisis mental health transfers from community clinics to emergency departments,8 A.C., S.N. and J.Z. established lines of questioning and sensitizing concepts that were used in the interview guide for this study. The lived experiences of members of the study team with mental health transfers also informed the development of the guide. Drawing on the knowledge base of our team, we formulated structured questions that were open and supportive to ensure that the interviewer and participant were able to engage in open exploration and generate thick descriptive data together.

The interviewer (A.C.) and J.Z. debriefed after each of the first 3 interviews and reviewed the transcripts to verify that the interview guide was allowing appropriate coverage of content areas. The interview guide was also revisited after 3 and 6 interviews by A.C. and J.Z. and amended to include new sensitizing concepts: previous experiences guiding future decisions; time and logistical pressures; and surprise that things could be otherwise.

The interview guide provided a clear set of instructions for the interviewer, permitting the collection of reliable, comparative qualitative data. Incorporating open-ended questions allowed further exploration of participants’ meanings, views and ideas. Interviews ended when the question list and topic discussion concluded.

Interviews were recorded digitally and transcribed verbatim. A.C. transcribed 2 interviews, a family physician with expertise in university student health care transcribed 1 interview, and the remaining 8 interviews were transcribed by a professional transcription service. Each transcript was checked against the audio recording to ensure fidelity. Person and place identifiers were removed before analysis.

Participant questionnaire

Participants completed an unvalidated questionnaire that elicited demographic, training and practice information, and surveyed their general beliefs about crisis transfer processes (Appendix 1). Participants rated 7 items by circling the response that best aligned with their point of view (“Disagree strongly,” “Disagree,” “Neutral,” “Agree” or “Agree strongly”). Participants could also provide descriptive information to assist the investigators in understanding their position. Questionnaire items were developed by A.C. and J.Z. using key concepts derived from the literature review and policy scan.8

A.C. distributed questionnaires to participants via email. Participants with odd enrolment numbers were asked to complete the questionnaire before the semistructured interview and those with even enrolment numbers were asked to complete the questionnaire after the semistructured interview.

Policy and process documents

Institutional policy and process documents were requested from participants during interviews. Policy and process documents were also requested from OUCHA member universities not otherwise represented, to allow for a more robust policy and process summary. A.C. sent scripted emails requesting these documents to the publicly available addresses of administrative personnel from university clinics not represented by the participants.

Data analysis

We analyzed the interview transcripts thematically.26 Theoretic perspectives, including the interpretivist paradigm, relativist ontology, subjectivist epistemology and social interactionist orientation,27,28 informed the analysis. Reality is complex and experienced uniquely (“interpretivist paradigm”). This reality is socially, intersubjectively and experientially created (“relativist ontology”).27 Each person’s understanding of the world is central to and influenced by their understanding of themselves and others (“subjectivist epistemology”).27 Investigators and participants are connected: as the inquiry proceeded, investigators and participants cocreated findings and knowledge through dialogue.27 Collective behaviours, beliefs and experiences were explored to understand how social interactions and behaviours contribute to the evolution of policies and processes (“social interactionist orientation”).28

Data collection and analysis occurred simultaneously in keeping with interpretive qualitative research practice.29,30 In the first step, transcripts were read and reread to obtain a broad understanding of physician perspectives about transfer policies and processes. The first 3 transcripts were read by all study team members, and ideas were noted through a memoing process. Memos were completed individually by A.C. after each interview and by S.N., J.Z. and G.N. after reading the transcripts. Memos about team discussions were also created.

Transcripts were then examined closely and initial codes inserted. A.C. and J.Z. open-coded each of the 3 transcripts independently and then met to review the coding and begin consolidating the coding tree. A.C. and J.Z. then coded the next 3 transcripts and met to discuss the codes and make connections between them. Each transcript was read, coded, reread and recoded as necessary by both A.C. and J.Z.; S.N. and G.N. offered feedback to the analytic process. After coding, the text subsumed under each code was reviewed and summarized, and an analytic memo was created to capture code content. A.C. and J.Z., with input from S.N. and G.N., grouped codes into potential themes and created a thematic map. Themes were refined and organized.

Questionnaire responses were summarized. Policy and process documents were reviewed as a means of clarifying practices and informing thematic analysis of the transcripts.

Data integration

The questionnaire responses and the policy and process documents were not analyzed formally. Questionnaire responses were discussed by A.C. and J.Z., the primary coders for the data. Questionnaire responses informed the development of the coding tree. A.C. and J.Z. reviewed the policy and process documents. Reflection on variable institutional contexts and respondent views related to transfers supported the development of a frame for analysis that would account for these differences.

Ethics approval

The study protocol was approved by the Hamilton Integrated Research Ethics Board.

Results

Eleven interviews, ranging in length from 27 to 50 minutes, were conducted with physicians from 9 university-affiliated health clinics (8 at OUCHA-member universities and 1 from a nonmember university). Ten interviews took place by telephone, and 1 participant requested an in-person interview. Table 1 summarizes the participants’ characteristics. More than half (7 [64%]) self-identified as female. Nine participants (82%) were family physicians, and 2 (18%) were psychiatrists. Over half (6 [54%]) had been working in student health for more than 10 years.

Table 1:

Demographic characteristics of participants*

Characteristic	No. (%) of participantsn = 11
Gender
Female	7 (64)
Male	4 (36)
Specialty
Family practice	9 (91)
Psychiatry	2 (9)
Years in practice
≤ 5	1 (9)
6–10	3 (27)
≥11	7 (64)
Years working in student health
≤ 5	3 (27)
6–10	2 (18)
≥ 11	6 (54)

Data were extracted from questionnaire responses for 10 participants and from the interview transcript for 1 participant, who did not complete a questionnaire.

Ten of the 11 participants completed questionnaires. Questionnaire responses are summarized in Table 2. Although the responses suggested a range of experiences, there was consensus that handcuffs are not required for transfer. The majority of respondents (8 [80%]) believed that involving police or security in mental health transfers is stigmatizing, and nearly all respondents (9 [90%]) believed that the routine use of restraints is stigmatizing. More than half (6 [60%]) felt that a clinician risk assessment informed decisions about whether restraints were used. There was a diversity of opinion about whether police are required for safe transfer, and perspectives varied widely regarding the skill and confidence displayed by police in conducting risk assessments to determine whether restraints are used.

Table 2:

Participants’ beliefs about mental health transfers (from questionnaire) (n = 10)

Item	Range*	Mode*†
Police or security are required for safe transfer	1–3	2 (n = 4), 3 (n = 4)
Handcuffs are required for safe transfer	0–1	1 (n = 6)
Using police or security officers for mental health transfers is stigmatizing	1–4	3 (n = 5)
Using restraints routinely for mental health transfers is stigmatizing	2–4	4 (n = 7)
Clinicians and individuals transporting students to hospital have a good working relationship	1–3	3 (n = 4)
Where police/security officers are involved in student transfers, a risk assessment by the clinician issuing a Form 1 is considered in determining whether restraints are used	0–4	3 (n = 6)
Where police/security officers are involved in student transfers, officers appear skilled and confident in assessing risk in order to determine whether restraints are used	0–4	2 (n = 5)

Response options ranged from 0 (“Disagree strongly”) to 4 (“Agree strongly”).

Most frequently occurring value.

Five institutional policy documents were collected, 4 from the participants and 1 from administrative personnel at an institution not represented by a participating physician. Information about an additional institution was obtained through email communication with administrative staff, as we were not successful in recruiting an interview participant and no written policy existed.

Based on data from the policy document review and interviews, the policy summary (Table 3) included 10 (59%) of 17 eligible OUCHA member university clinics and 1 (17%) of 6 nonmember university clinics. Clinics included were widely distributed around Ontario, with no notable geographic omissions. Institutions with large and small student populations were included. Only English-language institutions were represented. Police were commonly involved in transfers from 7 of the 11 clinics, and in nearly all or all transfers from 5 of the 11 clinics (Table 3). Alternatives included accompaniment by clinic staff or nonclinical support people on foot or by vehicle, and transfer by ambulance. Handcuffs were always or almost always used during transfers from a minority (2 [18%]) of clinics.

Table 3:

Description of transfer process and policies at the clinics when Form 1 is issued, and representative quotes (n = 11)

Type of process; institution no.	Transfer process	Police or campus police involved	Handcuffs used	Representative quote
Fixed
Emergency response is activated, with police transporting student to hospital in most cases
1*	Campus police (special constables of municipal police) are contacted to transport student. In rare cases, 9-1-1¶ is called, and municipal police convey student.	Always	Formerly always; now discretionary use	We will contact our campus police service; they will come, and again, there will be … a joint assessment of the situation, and typically students are not needing to be handcuffed anymore. (Participant E1)They [police] don’t use any restraint procedures unless the situation indicates that, and it is done in as low-key and as kind of student-friendly and gentle a way as possible. (Participant E2)
2*	Campus police (special constables of municipal police) are called to transport student.	Always	Almost always	They [police] mostly … 9 times out of 10, will apply handcuffs to a patient, which can be a very traumatic experience. (Participant D)
3*†	Campus police (special constables of municipal police) are called to transport student.	Always	Always	Every time that I’ve called, they [police] have handcuffed the patient. And zero times did I think it was necessary. … I remember having a conversation … with the police officers to maybe consider not handcuffing, because the patient was totally willing to go, but … they said “no” in each circumstance. (Participant K)
4*†	9-1-1 is called, and police or paramedics, or both, convey student.	Almost always	Never	The nurse arranges [for] the police to come. (Participant B) That horrified me. … Handcuffs? … I hadn’t even thought of handcuffs. (Participant B)
5*†	9-1-1 is called, and a mobile crisis team (police and mental health worker) or paramedics, or both, convey student.	Majority of cases	Rarely	The time there was [handcuffs] … I think it was out of necessity. … The person was verbally resistant before … police arrived, but then, when police arrived, they were a little more physically resistant, so it was out of necessity they used restraints. … But otherwise it’s never been discussed because I think it was just clear it wasn’t needed. (Participant I)
6*	9-1-1 is called, and police or paramedics convey student.	Almost always	Rarely	We started specifically requesting for police instead of paramedics … and then our experience has been if that’s available, they do send a mental health officer, or an officer with some mental health training. And we have usually had pretty good success. (Participant A)I can’t ever remember handcuffs being used. (Participant A)
Emergency response is activated, with ambulance conveying student to hospital in most cases
7*	9-1-1 is called, and student is most often transported by paramedics. In rare instances, where safety concerns are identified, police become involved in transfers.	Rarely	Rarely	[We] would call an ambulance, and usually they will come to the university and then they will take [the student] from there. If we have any concerns about [the student] wanting to leave, or [if we] feel unsafe, we call security, which is on campus. That has happened quite a few times where we have just had security waiting until the ambulance comes and takes the person to the hospital. (Participant H)Definitely no, nothing really that we have seen in terms of restraints or anything like that. (Participant H)
8†‡	Campus police (special constables of municipal police) are called. Campus police call 9-1-1 and wait in clinic until paramedics arrive. Paramedics convey student to hospital.	Not specified in protocol	Not specified in protocol	–
Flexible: students are often accompanied by clinic staff, with discretionary involvement of police or paramedics
9*	Student is accompanied to hospital by clinic staff in the majority of cases. Occasionally, student is accompanied to hospital by friends or family. In rare cases, on the basis of safety concerns, 9-1-1 is called, and police or paramedics, or both, convey student.	Rarely	Rarely	Usually what will happen if someone is really, really distressed, whether they’re certified or not, [clinic staff] will escort them over to the emergency department. (Participant C1)The options can be the patient going with one of our nursing staff, [taking] them over; the patient being escorted by the police; and the patient being escorted by a family member or friend. Those would be really the 3. Or when I say one of our nurses, also some other nonnursing staff, like a clinic manager will sometimes take students to the hospital on a Form 1. (Participant C2)
10*†	Student is accompanied to hospital by clinic staff in the majority of cases. Occasionally, student is accompanied to hospital by friends or family. In some cases, on the basis of safety concerns, 9-1-1 is called, and police or paramedics, or both, convey student.	About 50% of the time	Rarely	For those patients who are seeking help and recognize that they need help and who accept our assessment that they should be [on a] Form 1, because they are a risk to themself or to others, we offer them … transportation that we arrange, and an accompaniment with one of our staff people. We’ll actually send a nurse with a patient to the [emergency department], and hand over the patient at the [emergency department] to a nurse and triage at the [emergency department]. (Participant J)
Form 1 not used
11‡§	NA	NA	NA	–

Note: NA = not applicable.

Information was drawn from interview transcripts.

Information was obtained from policy and process documents.

No physician respondent.

Clinic director reported Form 1 use to be rare.

Emergency telephone contact number in Ontario.

Major themes

Three major themes were identified in the interviews: police and restraints cause harm to students experiencing a mental health crisis; police involvement and restraint use are justified on the basis of patient considerations; and transfer processes are often informed by extramedical factors. Representative quotes are presented in Table 4.

Table 4:

Quotes supporting themes

Theme; subtheme	Representative quote*
Police and restraints cause harm to students experiencing a mental health crisis
Police involvement is problematic generally	A lot of people have had some very negative interactions with the mental health care system and the justice system, that’s for sure. (Participant C1)People already have enough trouble being in hospital, but to have to be taken in handcuffs … out of the building and loaded up in a police cruiser and taken half a block, it seems brutal and traumatic for the patient, and sends all the wrong messages about a caring, supportive environment. (Participant C1)
Police involvement may interfere with future treatment	They [patients] are very suspicious and hard to engage. (Participant C1)The idea that you’re breaking trust with a vulnerable person can have huge impacts on care down the road. (Participant D)We have run into people who have either come back for another reason and are clearly unwell, or who have come back with significant reluctance, saying … “I am only here for X, Y, Z; I am not going to tell you all this because of what happened last time.” … Having been placed on the Form 1 and transferred sometimes will, I think, prevent people from coming back. (Participant D)
Police involvement and restraint use are justified on the basis of patient considerations
	[We are] balancing the safety of the student with what’s going to be most comfortable for them and finding the right balance there. And I know it is potentially not a great experience to be escorted by police, but definitely when it’s really necessary for their safety, then it really does make sense. (Participant C2)If we have any indication that someone may be violent toward other people or is actively psychotic … we engage the police at all times. (Participant J)Paramedics aren’t really trained to go after a patient and chase them down. Not that that is something that happens frequently, but it’s still something … you have to be concerned about, in potential worst-case scenarios, and how that could turn out. (Participant B)
Transfer processes are often informed by extramedical factors
Rationale for police involvement	I think there may have been some concern based on … union responsibilities and roles for the staff that were involved, that it was outside of their roles. (Participant E1)I think the main risk is if the student decides to flee the situation and our staff wouldn’t really be able to make them go to [the emergency department]. And then … if something went wrong and the student ended up hurting themselves, how would that affect the staff that was unable to really do that job properly? (Participant C2)In the past, we used to send a counsellor or a nurse with them [the student] in a taxi, and we found that to be too time-consuming because they might end up in the emergency department for 5 hours waiting to be seen. So, they [clinic administration] changed that policy to us calling 9-1-1. (Participant H)We’re quite busy in our clinic … so … one part of it is, do we have staff that can leave, and usually they [staff] will wait with the students until the students get seen. (Participant C2)My feeling is that police probably are the right group to do the transfer. … The reason I think police [is] just because I think it’s faster, and sometimes that’s important because it’s not a pleasant experience often for patients to be sent to hospital on a Form 1. (Participant I)
Extramedical rationale for restraint use	It was really just concerns from the police standpoint of their liability, and that was the main issue. (Participant J)They [police] mostly … 9 times out of 10, will apply handcuffs to a patient, which can be a very traumatic experience. And so discussions that we have had with the constables about whether or not that should be done are typically met with, “You know what, we have to cater to the highest potential risk.” (Participant D)The campus police say that they are following the guidelines of [municipality name] Police Service, which say, “Use restraints every time.” And my impression — this has not been said to me, but my impression is — that they are always supposed to use restraints, but there are a few officers who go against … the commanding officer’s request. They make a decision in the moment, and … I am not sure that that would be supported by their organization. (Participant D)

Participants coded with the same letter represent the same clinic.

Police and restraints cause harm to students experiencing a mental health crisis

Each participant identified harms of police involvement. Participants expressed a common understanding that mode of transfer should be dictated by clinical factors, with restraints used rarely.

Two subthemes were identified: police involvement is problematic generally, and police involvement may interfere with future treatment.

Police involvement is problematic generally

Involvement of police and the use of handcuffs were described as traumatic and criminalizing. Participants reported that negative consequences of police involvement were amplified when restraints were used.

Police involvement may interfere with future treatment

Participants shared concerns that police involvement and use of restraints could deter patients from accessing future care and make clinical interactions more challenging. Engaging police could also fracture the therapeutic relationship. Participants reported having experienced these sorts of feared outcomes.

Police involvement and restraint use are justified on the basis of patient considerations

Despite the attendant harms of police involvement and use of restraints, participants believed that there are circumstances in which transfer by police and, in rare cases, use of restraints are necessary. Notions of patient safety informed decision-making about mode of transfer. Clinical factors were also weighed. Assessments about risk of violence, probability of clinical deterioration and potential for patient elopement influenced decision-making. In the presence of such risks, police were identified as the most suitable personnel to conduct transfers by virtue of their training.

Transfer processes are often informed by extramedical factors

This theme included 2 subthemes: rationale for police involvement and extramedical rationale for restraint use.

Rationale for police involvement

At 6 of the 11 health clinics included in the policy summary, formally codified or well-understood processes dictated that campus or municipal police be contacted for all transfers under the Mental Health Act. At 2 clinics, activation of the province-wide emergency telephone number (9-1-1) when a Form 1 was enacted led to engagement of either police or emergency medical services (ambulance). At 2 clinics, flexible processes permitted staff or support people selected by the student to accompany the student to hospital. At the final clinic, the director indicated that no process existed because crisis mental health transfers were rare.

Two participants had experienced shifts from flexible to more constrained transfer practices. Participants identified barriers to the adoption of flexible practices. Barriers included notions of job roles and staff liability concerns.

Participants cited workflow and wait time considerations as reasons for discontinuing staff involvement in student transfers. In settings in which physicians were able to exercise discretion about mode of transfer, these same pragmatic considerations were identified as reasons for choosing police over transfer by a mobile crisis unit or ambulance, or staff accompaniment. One participant posited that expeditious care may be of paramount importance, even if it entailed police involvement.

Extramedical rationale for restraint use

The routine use of physical restraints during crisis mental health transfers was uncommon, occurring at 2 of the 11 clinics. Where routine handcuffing was practised, participants understood this to be a consequence of police policies aimed at avoiding liability and ensuring student safety. Strict adherence to police policies was rationalized on the basis that restraints ensure a secure transfer. Participants hypothesized that compliance with policies reduced professional risk to police officers.

Interpretation

In this study of physicians working at university health clinics in Ontario, conducted mainly in 2018, participants were generally not aware of the existence of formal institutional policies governing emergency mental health transfers. They reported that, in most settings, informal processes dictated how transfers occur. Police involvement and use of handcuffs, although understood by the participants to be harmful, were standard in some clinics. At other clinics, students were routinely transported to hospital by ambulance, and, at yet others, clinic staff generally conveyed students to hospital, with police involvement rare. Some participants believed that police involvement and restraint use may be conditionally necessary to prevent or manage elopement, violence or clinical deterioration. Participants noted that pragmatic concerns related to workflow and human resources capacity underpin the continued reliance on police. Notions of risk — inextricably linked to stigmatization of mental illness — and of liability are layered on inflexible processes to make involvement of police and use of restraints normative in some settings.

Police involvement and use of restraints in the provision of crisis mental health care relate to limited system resources; gaps in comfort and training; notions of risk and safety; and concerns about liability.8 Police involvement and restraint use can be traumatic, with the death of people in crisis occurring in the worst-case scenarios.5–8,14–18 People may experience fear and stigmatization when police are involved and handcuffs applied, which can engender mistrust and deter future help-seeking. 5–8 The potential for adverse outcomes is magnified when police are engaged in crisis mental health care, particularly for patients with intersecting systemically marginalized identities.18–20 Alternatives to police exist in the form of mobile crisis teams or paramedic services in some provincial or territorial, national and international jurisdictions.6,8

There is a dearth of information about processes for transferring patients requiring emergent psychiatric assessment from community-based points of care to hospital.8 Advocates in Canada and elsewhere have advanced visions of crisis services and supports that do not involve police at all.3,10,11,16,18–20 Our exploration of physicians’ experiences with transfer processes at Ontario university clinics highlights existing variability, affirms the known harms of police involvement and use of restraints, and clarifies factors influencing police involvement and handcuff use.

Limitations

We interviewed physicians from 47% of eligible OUCHA-member clinics and obtained information about transfer practices from administrative staff at 2 additional clinics. One OUCHA-member clinic was not contacted directly per protocol in error; however, because our sampling process produced multiple pathways to recruitment, a representative from that institution had other opportunities for inclusion. We were able to recruit a participant from only 1 of 6 non–OUCHA-member clinics. Although the included institutions were geographically distributed around Ontario, and institutions with both large and small student populations were included, only English-language universities were represented.

Inclusion of more participants, including clinicians at French-language universities, clinicians assessing transferred patients in hospital and the personnel involved in transfers, might have permitted a richer analysis of more varied experiences. It is possible that concerns about existing processes motivated some physicians to participate, and that some positive perspectives about mental health transfer processes were thus not included. Physician perspectives may diverge in important ways from the perspectives of other clinical team members.

We did not examine the transfer experiences of students, who can best speak to the impact of various practices. This study did not explore decision-making about the completion of Form 1. Clinician opinions about the role of police in transfers may be related to differences in how the Form 1 is used in practice. We plan to conduct further research on this important topic. In designing our study, we did not explicitly consider how intersecting contextual factors, such as race and racism, may influence people’s experiences of police involvement and use of handcuffs. We acknowledge that this represents a substantial omission and that it should be addressed in future work.

Conclusion

Processes for transferring students experiencing a mental health crisis to hospital varied between university health clinics in Ontario. Decisions to engage police and use handcuffs were sometimes constrained by formally articulated policies or otherwise well-understood policies. Clinical factors, workflow considerations, and rhetoric of safety and liability also influenced decision-making. Some participants shared experiences with flexible processes whereby clinical staff accompany the student to hospital in most cases. Such processes permit more dignified and less stigmatizing transfers. In rare instances, in the context of safety or elopement concerns, police are contacted; in such cases, discretion is exercised about the need for restraint use.

To avoid piecemeal policy implementation, and to develop and implement coherent and consistent best practices for mental health transfers province-wide, collaboration is necessary among municipal and provincial governments, university administration, university bodies, nongovernmental organizations, hospitals, hospital associations, medical bodies, police departments and organizations, and emergency response units and organizations. People with lived experience must be engaged meaningfully. To gain a thorough understanding of risk and liability, the harms of police involvement to people experiencing a mental health crisis must be considered.