Observing the work of an urban safety-net psychiatric emergency room: managing the unmanageable

Abstract

Staff in the psychiatric emergency room (PER) have demanding jobs requiring a complex balance between the needs and safety of the individual and the community, systemic resources, and job responsibilities while providing timely, effective care. Little research exists concerning day-to-day work activities of PER staff, their interaction, and their perceptions of their work. This study explored the work of PER staff and the organisational context of the PER work setting. Observations of staff were conducted in the public spaces of a public urban PER using two observational techniques. The first was designed to measure the types of work activities staff engaged in and the time spent in these work activities (work task data). The second technique was the gathering of observational data by a peripheral-member-researcher (participant observation data). Analyses were conducted of both the work task and participant observation data. Results indicate that most PER staff time is spent in administrative and phone tasks, while less than a third is spent on direct clinical work. Four important issues for PER work were identified: a workload that is unmanageable, managing the unmanageable, bogus referrals and dumping and insurance problems. The PER remains the front-line of the medical and social service systems. Work done in these settings is of critical importance; however little attention is paid to the content and nature of the work. Our study demonstrates that staff of the PER face challenges on many levels as they struggle with the task of working with people presenting in psychiatric and social crisis.

Introduction

Psychiatric emergency rooms (PERs) are critical to both the mental health system and the social service networks of many communities and serve unique and significant functions within the mental healthcare system. The staff of PERs see millions of people struggling with severe mental illness and others in emotional crises. In 2004 in the United States there were 3.7 million visits to emergency departments for mental disorders and this number is on the rise, particularly in public urban PERs where the numbers of presenting patients has dramatically risen in the past two decades (Lipson and Koehler 1986, Rabinowitz 1995, Allen 1996, Claassen 2000). As federal, state and local funds have been cut, managed care has limited reimbursement, and access to other mental health and social services has been reduced, the net effect is to render the public urban PER as the last open door to services, acting as barometers for societal problems and pressures. Issues and challenges in the community are likely to be reflected in consumer presenting issues as well as staff experiences of working in the PER.

The formal provision of care in PERs includes the core activities of crisis intervention, assessment of risk and the appropriate disposition of people seeking services for a diverse array of urgent and non-urgent medical and social service problems. However, little is known about the nature of work in these settings. PER staff face the challenge of meeting the diverse and often acute needs of patients and their families, while coping with the enormous pressures on healthcare systems to assess risk and reduce costs. Perhaps most importantly, the decision to involuntarily hospitalise a patient, one of the few remaining areas of coercion in medicine, is most often made in a PER setting. The work of the PER is fast, intense, consequential, involving many actors, working with scant information of dubious accuracy. It is in these types of settings, with multiple actors, little knowledge and high stakes, that the influence of social characteristics and unrecognised biases are most likely to creep into medical decision making as people rely on decision heuristics.

In this paper we report on the findings of an observational study of a busy, public, urban American PER. Two types of observational data were gathered (work task and participant observation) to address two study questions: (1) what is the instrumental, task-oriented work of PER staff? and (2) what is the organisational context of the PER work setting and how do staff think about the work they do? These rich observational data allow us to examine the organizational context of work in this busy, public, urban PER, as well as staff perspectives of their work, as these are revealed in the everyday, problem-oriented, expressive commentary that is a consistent aspect of this work.

While there is a large body of research examining psychiatric emergency care, this research has focused primarily on two areas: the service utilisation patterns of patients presenting for care; and the clinical decision making of PER staff. We have been able to find only one study examining the instrumental, task-oriented work of the PER staff, their interactions with patients, or PER staff understandings of their work. Lipson and Koehler (1986) describe the staff sub-culture of a PER based on clinical experience. They describe how staff believe that they treat patients who the private sector would not even allow through their doors – that it is ‘them against the world’. Staff burnout is a recognised fact of life, with emergency psychiatry not seen as a long-term career. Few staff remain in the PER more than three or four years and staff who did remain were awarded less status among their colleagues since the expectation was that staff would leave.

Research examining other mental health service settings which work with patients in psychiatric crisis, such as acute psychiatric units (APUs), informs our understandings of the work of the PER; however these settings, and the work done within them, also differ in important ways from PERs. The results of numerous studies of APUs, conducted in multiple countries, identify similar struggles in these settings. In the landmark study of an American acute psychiatric unit (APU) Rhodes (1991) described the work done in the unit in great detail; including interactions with staff, patients, staff at other facilities, insurance companies, police, family members and others. She describes the work of the APU staff as the ‘people work’ defined by Goffman (1961), in which those who enter an institution become the ‘objects or products’ of the staff. A review of qualitative studies conducted in APUs (Lelliott and Quirk 2004) identifies three consistent findings. First, staff spend a great deal of time grappling with risk managements. Secondly, the nature of the response to risk or dangerous behaviour, is influenced by a complex set of factors including the types of behaviours patients present with, staff attitudes and the organisational context within which decisions are made. And finally, staff and patients are dissatisfied with the nature of care in the APUs. The research on staff in these studies focuses primarily on ward nurses who report dissatisfaction with the therapeutic alliances that occur in these settings.

These studies suggest that the organisational context of the APUs influences the treatment provided in these settings. Understanding the nature of work in the PER will help us better understand the experiences of PERs for both staff and patients. In addition, Glisson’s model for understanding the impact of service coordination on client outcomes (Glisson and James 1992) highlights how interventions aimed at improving outcomes by co-ordinating services are mediated by the attitudes and perceptions of staff, as well as the importance of the community context. They argue that organisational climate and staff attitudes are likely to be particularly salient influences on care in public human service organisations because of the poor worker attitudes found in these settings (Glisson and Durick 1988), as well as high rates of employee turnover and low job satisfaction (Solomon 1986). In their work in general emergency departments (EDs), differences in staff’s emphasis on providing emotional support to families using the ED were found to be a function of the organisational culture and climate of the ED in which they worked, and not based on individual level characteristics such as staff age, sex, experience or profession (Hemmelgarn et al. 2001: 104). In another study, nurses and other staff of a general ED who reported viewing their work conditions positively reported greater frequency of contact with patients with psychiatric problems and provided a more diverse set of treatments and interventions to these patients (Wright 2003). Thus, the organisational climate in the ED contributed to the attitudes towards psychiatric patients and the clinical care provided by the emergency department nurses and staff.

While there has been increasing attention paid to the role of social factors in mental health services research, it has most often focused on understanding the role of social factors in service utilisation or treatment outcomes. Research designed to better understand work and staff perspectives on work, is more limited and, while studies of APUs and general EDs have focused on the work in these settings, understanding the work of the PER, with its unique and critical function, remains an important and under-examined area. In this paper, we begin to address this gap, through our qualitative examination of work in a busy, public, urban PER.

Setting

All data collection took place at a PER in a busy, public, urban safety-net hospital in the Northeast United States. This PER receives more than 3,200 visits per year and serves a highly diverse patient population. The PER is a free-standing section of the hospital emergency room which evaluates and treats individuals in psychiatric crisis as well as providing general psychiatric consultation services to the general emergency department. On average three to five staff are on duty including psychiatrists, licensed master’s level clinicians, nurses, and psychiatric residents. There are no in-patient psychiatric beds in this safety-net hospital. Researchers collected observational data in non-clinical spaces, including the staff office space, hallways and the ER common area, in the PER for more than 50 hours over four months. The study focused on provider/patient interactions and, in order to most fully protect the confidentiality of patients in the PER, observational data were not collected in the four private treatment rooms.

Pilot study

We conducted a pilot study to determine the feasibility of using observational techniques to achieve the study aims. During the pilot study, two observers (AL and AW) conducted simultaneous observations. The observers took detailed notes during the observations and wrote up the notes with experience and observer impressions immediately after the observation period was completed. The two observers compared their notes and impressions. Initial reviews of observation data showed a great deal of information regarding staff interactions collected, but that observers were unable to track time spent on specific work tasks. The pilot study also demonstrated that because of ethical and clinical issues it was unfeasible for observers to follow PER staff around and track their activities. Thus, the study design was expanded and the study was conducted using two distinct observational techniques in order to address the study aims. These are described here as timed observational data of work tasks and participant observation data

Data collection I: timed observational data of work tasks

Two observers sat in designated public spots in the PER from where they were able to observe all staff clinical and administrative activities. It is important to note that while they could not hear or observe staff-patient interactions, their vantage points allowed them to see when staff entered or exited a treatment room so that clinical time could be measured. Initial reviews of the pilot data indicated that staff activities could be broken down into four categories: phone work, clinical activities, administrative activities, and other activities (i.e. personal conversations, food breaks). Clinical activities were noted when staff were working face to face with patients, either in protected clinical spaces such as treatment rooms, or in public spaces such as the ER waiting room. These categories were further broken down into activity subcategories (Table 1). The final data collection instrument was created when no categories were added or changed after two observation sessions. The amount of time for each observation was also recorded and a separate page was used for each staff member, noting identifying information (code name, job role), date and time of the observation, and the observer’s rating of how busy the PER was (how many patients were listed on the PER census board, general 1–5 rating of the level of staff activity). In addition, a data collection protocol was created, with detailed category definitions and specific instructions for timing staff activities. In order to ensure data was collected across all possible shifts, both on weekdays and at weekends, a sampling matrix was designed to capture work task time observations for 24 hours including four hours for each of three shifts of the weekdays and weekends. Data were collected during April and May of 2006.

Table 1.

Staff activities

	Total	Master’s Level Clinicians	Psychiatrists	Nurses	Residents
Percentages of total activities
Clinical	16.3	11.9	9.2	29.1	15.0
Administrative	58.4	47.0	70.1	48.8	66.9
Phone	10.9	9.5	18.9	8.9	6.4
Other	14.4	31.6	1.0	13.4	11.7
Percentages of clinical activities
In PER	46.2	51.0	72.2	15.7	46.0
In pediatrics	8.3	32.2	0	1.0	0
In acute	24.2	5.8	22.2	27.1	41.7
In trauma	8.2	0.5	0	28.6	3.6
In waiting room	3.0	1.0	0	10.2	0.7
Doing labs	1.1	0	0	4.3	0
In other areas	9.0	9.6	5.6	13.1	7.9
Percentages of administrative activities
Notes	43.3	42.1	41.0	33.2	57.0
Consulting with PER staff	11.1	7.2	2.9	27.9	6.3
Consulting with other ER staff	45.6	50.7	56.1	38.9	36.7

For each observation period the observer had a data collection sheet and a stopwatch to time activities. At the beginning of each session, the observer wrote a code name for the 3–4 PER staff to be observed on the top of a data collection sheet. By using two observers, PER staff were observed in both administrative and clinical activities.

Timed observational data were analysed by grouping staff members into categories based on the role in the PER including: psychiatrists, master’s level clinicians, nurses, and psychiatric residents. Timed observational data were first analysed to determine the amount of time accounted for by the recording of work tasks during the observation periods. For example, during observation of staff at a specific PER over a two hour period, the observers may not have seen them for the entire two hours because PER staff travel to other parts of the hospital, were blocked from view, or for other reasons. In order to be confident about descriptions of work task time, a high percentage of observation time must be explained. Overall, the data collected allowed us to account for 92 per cent of PER staff time during observation periods. This time varied slightly across staff role (100% of the work time of psychiatrists, 91% of nurses, 90% of medical residents and 89% of master’s level clinicians was explained), across shift (100% of the morning shift, 92% of the afternoon shift, and 87% of the evening shift) and across observer’s ratings of an activity scale (93% for somewhat slow, 88% for average, and 84% for somewhat busy). There was less variation across day of the week (91% for weekdays, 90% for weekends) or the number of patients on the nursing station census board (95% for no patients, 89% for one or two patients).

Data collection II: participant observations

One observer (AW) spent over 50 hours in the PER across each of the three shifts per day, each day of the week, and each week of the month in order to account for variation in PER utilisation patterns (Arfken et al. 2002, Sobel et al. 1998). This observer is best described as having been a ‘peripheral-member-researcher’ as described in Adler and Adler’s (1994) typology of researcher membership roles. In this role the researcher is described as observing and interacting ‘closely enough with the members to establish an insider’s identity without participating in those activities constituting the core of group membership’ (1994: 380). In fact, the researcher was known to many of the staff from earlier unrelated projects. The researcher assisted the staff in small, non-professional ways such as helping with computer problems or getting staff coffee. The researcher also engaged in conversation with staff. The researcher asked staff clarification questions. Often staff would also volunteer their explanations and thoughts to the observer. These systematic observations of staff interactions focused on conversations and exchanges that took place as part of ordinary work activity. The observations were descriptive and focused on what was seen and heard, as well as what was said to the observer. What people said and did was recorded by the observer as well as physical descriptions of people and the setting. In addition, information known to influence what occurs in the PER (Arfken et al. 2002, Diehel et al. 1981) including staffing patterns, the weather, number of patients seen in the PER during the shift, and any occurrence of a high profile event in the community involving a person with mental illness in the past two weeks was noted.

For these observations, the observer was positioned in the nurses’ office, the hallways and waiting areas, but again, not present in any private clinical space. While the observer could see when staff were in treatment rooms with patients, the only directly observable staff-patient interactions took place in public areas of the PER. During weekly meetings of the research team notes were reviewed to determine whether new information was continuing to be collected through the use of the structured observations. Once no new information was being collected, we determined that a saturation point of observational data had been reached (Glaser and Straus 1967). The study was approved by the hospital’s Institutional Review Board, and public notices explaining that observations were being conducted were posted around the emergency room.

Observational notes were immediately typed and formatted into transcripts for the software HyperRESEARCH (Researchware v2.8, Hesse-Biber et al. 1994). In the coding process codes representing a category or theme found in the data were attached to corresponding segments of text. Each transcript was evaluated by the set of thematic codes developed from the transcripts by the PI, CoPI and Project Co-ordinator. The coding process involved three steps of data coding. In step one, the three researchers read through the transcripts and underlined important words or phrases and developed a set of thematic codes based on the issues and themes of relevance. In step two codes were put directly into the text in the form of an abbreviation of a category name that represents the category or theme (code) found in the data. To improve inter-rater reliability, the three researchers each coded four transcripts, reviewed the codes, and when agreement was consistently high, the project co-ordinator coded the transcripts. In the third step, to explore whether or not linkages existed between and/or among particular themes we used HyperRESEARCH software, a text retrieval program geared toward in-depth exploration of data.

Results

What is the instrumental, task-oriented work of the PER staff? What do PER staff do?

The timed observational data indicate different distributions of activities across each of the major clinical and administrative staff categories (clinical, administrative, phone, and other) and subcategories (Table 1). Most staff time was consumed by administrative activities while a limited amount of time was spent in clinical activities. Master’s level clinicians spent 47 per cent of their time doing administrative work and 12 per cent of their time in clinical activities while psychiatrists spent 70 per cent of their time doing administrative work and nine per cent in clinical work. Variations are also seen across staff role with sub-categories of activity patterns demonstrating multiple staff activities (Table 1). Differences in the types of administrative work are also presented in Table 1. For example, residents (the physicians in training on the staff) spent 57 per cent of their administrative time on notes, while nurses spent about a third of their time making notes.

Finally, these timed data allowed us to examine a common theme in shift work in greater detail. Our preliminary analyses of the data from the larger study of PER staff perspectives suggested that a common theme among staff is the belief that staff of ‘other shifts’ do less work or leave behind work for the next shift at shift change (‘dumped work’). These timed observations allow us to examine patterns of work for evidence of ‘dumped work’. A similar proportion of time across morning, afternoon and evening shifts was spent on administrative tasks (61%, 50% and 53% respectively). These analyses demonstrate that the type of work and time spent on task seems to vary very little across shifts despite perceptions among the staff (see Table 2).

Table 2.

Activities across shifts

	Total	Mornings	Afternoons	Evenings
Clinical	16.3	17.0	21.4	17.0
Administrative	58.4	61.1	50.0	52.7
Phone	10.9	5.6	10.0	9.4
Other	14.4	16.4	18.9	21.1

What is the organisational context of the PER work setting and how do staff think about the work they do?

The systematic observations of staff interactions were examined to understand the context of work in the PER and staff perceptions of their work. The findings identified four themes related to the PER work setting: the workload is unmanageable, managing the unmanageable, bogus referrals and dumping, and the role of health insurance.

Staff across all categories described internal and external forces that result in the work of the PER as overwhelming and unmanageable but they also described different methods used to ‘manage the unmanageable’.

The workload is unmanageable

Staff discussions among themselves reflect the shared belief that the workload is unmanageable, and highlight a number of factors that contribute to the unmanageability including the number of people using the PER services, the extent of their medical and social need, and staffing patterns. As one staff member described the patient load in the PER to the observer:

‘We’ve got 25 people here, there are more people than there are beds, and we can’t take any new patients without affecting the care of the patients who are already here. Right now there are 14 people waiting in acute and even more en route.’

It was common for staff members to struggle with getting medical consultations (consults) for patients seeking PER care. As was observed, the phone rings and a mental health clinician answers it. After listening, he angrily states ‘there are still two consults that are left over from yesterday!’ and after hanging up, walks out of the door.

And finally, staff members discussed their frustration with having to grapple with a patient flow that they cannot control:

The resident notifies the attending physician that there is another patient who will be coming in shortly. A psychiatrist responds ‘we shouldn’t even be doing this – we’re the only hospital open right now, we’re on pre-divert, and we have no beds open on the acute side. I heard about this person – supposedly there’s a ‘fantastic’ (sarcastically referring to poorly done) evaluation coming with her’. A nurse asks ‘so the next time, what should I do when they call with someone coming over?’ The psychiatrist responds –‘Just have them wait until I can talk to them, but yeah, we’ll eventually have to take them anyways’.

The workload is also seen as unmanageable because of contextual factors not within the control of the PER staff including time of the day, day of the week, week of the month, weather, local events and school vacations. Staff members referred to the effect of different days, as in ‘The time of month really matters’, alluding to the first of the month when checks come out or ‘things also get worse when it gets really hot. You see a lot more violence and trauma coming in when it’s hot.’ One staff member described to the observer that Fridays were the day of the week when both staff and patients are more anxious:

‘This is because you can’t just wait to see what happens tomorrow –we do that a lot during the week in outpatient, just telling people to come back tomorrow. On Fridays we have to find something before 5pm, when it’s too late.’

Managing the unmanageable

There are few formal ways that the PER staff can manage the flow of patients, and staff improvise distinctive, and often informal strategies for managing the impossible workload. One strategy is to use informal diversion. Some emergency rooms are able to go to ‘diversion’ status but this busy, urban PER rarely uses that status and thus staff have developed informal ways of diverting patients, particularly those they consider inappropriately using services. For example, this happened when over-burdened staff were unable to tell people they could not come into the PER but they could provide information that might discourage people from seeking care in their PER. One seasoned staff member was observed explaining to a more junior colleague:

‘Just tell whoever is referring that it will take us a really long time to see them, especially if it is an elderly case. Get the information, but try to see if there is anywhere else that can take them….If she comes in, there is no space. She’ll be in the hall, it will be hours before we can see her.’

Staff also used psychological mechanisms to reduce the stress associated with working in the PER. As one staff member said to another describing their use of cognitive reframing in coping with their negative feelings towards patients:

A mental health clinician remarks to a nurse, ‘People think I’m uncompassionate about my job here, but I just modulate my compassion.’ The nurse replies, ‘the more exposure you get the more you see that some people aren’t using resources very well, and are taking away from others.’ The clinican responds, ‘Nice reframe’.

Bogus referrals and dumping

The PER staff believe that not only do their patients view the PER as the last open door to services, but other providers and services do so as well, inappropriately referring or dropping off patients at the PER. Staff describe patients as being ‘dumped’ when a referral is seen as clearly not appropriate and the patient is perceived as having been difficult for another provider or agency to deal with. One staff member described a referral from the school system to another staff member by stating:

‘There is a wicked bogus referral in pedi [pediatrics]. The school sent him over because he was being oppositional and didn’t want to go to dance class. Kids get upset, they send them here.’

Other institutions, such as nursing homes, are also described as using the PER in ways that are not in the patient’s best interest. As a mental health clinician stated to a nurse:

‘She was born in 1909. I mean what the fuck? Why are they sending her to us? Someone should be taking care of her. I tried, I just said to them that we are having a really hard time placing elderly folks.’

Patients are also dumped to the PER from other parts of the hospital as well as from other medical facilities reflecting the PER as the last open door to services. Two staff members discussed a dump from a nearby hospital:

‘A patient was section 12d [involuntarily transported] to Hospital X, and they declined her, so she is on her way over here.’ A nurse responds, ‘no way – they can’t do that!’ The psychiatrist states, ‘she’s on her way over right now.’

Another describes a patient dumped from a different part of the Emergency department without adequate information to provide appropriate care:

‘Do you know what’s up with patient XXXX? They just threw him in here from pedi, and gave his information to a resident, who left.’

Insurance

In most urban safety-net hospitals there are many problems with patient insurance coverage. However, this PER, and others like it, provides needed psychiatric emergency care regardless of insurance status, and thus staff often experience the frustration of trying to sort out the complicated insurance status of their patients. Patients often either have no insurance, or have problems with the insurance they do have. As one staff member described to another:

‘What’s the main issue?’ A mental health clinician responds, ‘ The main issue beyond the CHINS [Child In Need of Supervision or Services], substance use and suicidal ideation is that she has no insurance.’

One of the greatest challenges for the minority of patients who do have insurance is that they have often ‘used up days’ available for mental health inpatient care on their insurance. As described below, this was true for both those with private insurance as well as state and/or federally funded insurers, such as Medicaid or Medicare. A nurse describes a patient in the PER, ‘he’s either high or manic. He’s looking for an inpatient admission, but his Medicare days are used up, so it’s going to be a challenge’.

Susan Stefans (2006) has recently written that,

Discussion

Emergency departments and their staff share some of the same attributes of people with psychiatric disabilities they see – they go from crisis to crisis, with insufficient structural support and very little public understanding of the pressures they face. They inhabit a culture that few outsiders can comprehend (2006: xi).

In this paper, we take a step towards addressing this concern as we examine the work done in busy, public, urban PERs, staff perspectives of work, and the organisational context of these important settings. These perspectives on work reveal themselves in the commentary on the ordinary, fraught, crisis management that is enmeshed with the work itself.

We collected specific data on work task time and conducted observations of staff interactions in the busy, chaotic settings of PERs. The observational protocol developed to collect data on different types of staff activities as well as amount of time spent on them generated data which explained more than 90 per cent of the observation period. These data show that most time is spent in administrative and phone tasks while approximately a third or less of staff time is spent on direct clinical work.

Differences in staff roles were evident from the findings on the different types of clinical work performed in the PER. Nursing staff were typically called upon to handle not only psychiatric issues in the PER but also in the larger ER where they consult on trauma cases, assist with rape victims, and work with the families of patients. Their time is spent in multiple areas of the ER but due to the administrative structure of the ER, is focused primarily on adult cases. In contrast, PER Master’s level clinicians float between the adult PER as well as the pediatric side of the ER, conducting psychiatric assessments of both adults and children.

Importantly, we have demonstrated that among the PER staff an average of 16 per cent of staff time is spent in clinical interactions and that those with the highest level of clinical training (psychiatric attendings and residents) spend the least amount of time with patients. Conversely, 70 per cent of staff time is spent in administrative or phone work, with psychiatrists spending 89 per cent of their time on these types of work. This is similar to descriptions of nursing work in acute psychiatric wards. In these settings, nurses are described as spending the greater part of their time in the ‘virtual wards’– that is nurses are shifted to engage in organisational or administrative work which minimises therapeutic work with patients in acute ward settings (Deacon 2003). While in acute psychiatric wards, Deacon argues that patients benefit from and appreciate nurses’ involvement in paperwork and other non-clinical activities, it is not clear that similar dynamics occur in PERs. Ware et al. (2000) in their study of clinicians in a community mental health centre (CMHC) and their experience of managed care, have noted that proliferation of paperwork is a characteristic of a managed care system in ‘which only the clinical provider can guarantee the information that justifies the need for continued care’ (2000: 15). They highlight the multiple administrative structures in the clinic they study. While both acute ward and community health centres differ in many ways from PERs, this common struggle with the demands of increased administrative and organisational burden with the desires of staff to provide care they see as therapeutic has been a consistent finding across service settings and international boundaries. Unlike acute wards, in this PER there was no direct access to psychiatric inpatient beds. This intensified the importance of staff work to ‘find beds’ for patients. And while neither the PER nor community health centres have direct access to psychiatric inpatient beds, the acuity and needs of patients presenting at the PER differ dramatically from patients receiving outpatient care at the CMHC. In the PER, on top of the administrative burden created by the various payers, staff often have duplicative documentation requirements, must make phone calls to other institutions to secure psychiatric placement and contact collateral sources of information. In addition to paperwork for document need, it is often only the psychiatrists who can either gather needed information or provide the information needed by managed care organisations and other service providers to guarantee the appropriate placement or continued care of the patients in the PER. Thus, psychiatrists appear to be spending significant time on the phone to make the time spent by PER staff during the patient visit meaningful, as they defend their disposition or treatment plans to parties outside the PER.

The nature of the work of PER staff is seemingly much like Sisyphus’, constantly facing a flow of patients with multiple and complex needs, reaching dispositions, only to start again. Our findings show that staff struggle to manage the unmanageable workload and develop both informal rules and psychological strategies to assist with this. The strategies described by PER staff to manage the amount of work, as well as the flow and pace of their work, include the use of informal diversion tactics and the use of ‘waiting’ to deter people from presenting at the PER. In addition, the staff develop psychological mechanisms to cope with the strain of the PER work setting. This is similar to Rhodes’s (1991) findings in her study of an Acute Psychiatric Unit (APU). Rhodes examined how staff viewed their work and found that the ‘situated knowledge’ shared by staff influenced many aspects of the work they did and the services they provided. Most staff felt that they were ‘mandated to do impossible things’ (1991: 117). Frequent users or ‘migrating birds’ of the APU services were constant reminders to the staff of the inability or refusal of others (families, other institutions, community care facilities, or homeless shelters) to take care of their patients. This continually reinforced to the staff that the exceptional needs of their patients exceeded the resources available to help them. The mandate of the APU was to ‘keep people moving’ (Rhodes 1991). As she states, staff shared beliefs that ‘producing useful individuals might be impossible, but producing empty beds was not’ (1991: 172).

Staff identify many system issues which impact on their work and ultimately their interactions with patients. Similarly, there are a growing number of studies which have highlighted the importance of system and contextual factors on the care provided in acute wards across multiple countries and service systems. Factors such as level of staffing, experience of staff, and the types of demands on staff all appear to play a role in risk management in acute wards (Lelliot and Quirk 2004). In addition, in our work, the system issues of bogus referrals, dumping patients and insurance status in many ways contribute to the context of the PER, the sense that staff are the last line of defence, the last people still fighting to provide services under the most difficult of circumstances. PER staff represent a diverse group of workers both in training and income. Hodson (1991) most often describes ‘dirty work’ as related to low-wage workers; these staff, through relations with each other and the PER patients, attach meaning to their work and strive for dignity and autonomy in jobs that outsiders perceive as ‘dirty work’. Observations of discussions of the relationship between the medical staff of the emergency department and the PER staff themselves, suggest that the PER staff perceive the medical staff to see their work as ‘dirty work’, working with patients they do not want to deal with. Brown (1989) found in his examination of the ‘dirty work’ of the psychiatric staff of a community mental health centre (CMHC), that ‘dirty work’ was designated based on tasks or services provided to external, non-psychiatric agencies, including evaluations of pre-release prisoners, serving the homeless mentally ill who were not part of the CMHC catchment area, and determining psychiatric disability status. In the PER, work that the staff do not believe is capitalising on their strengths, skills and training remains ‘dirty work’; for example the bogus referrals which are seen as an inappropriate use of the PER services.

Ware and colleagues (2000) have described the use of innovative and informal practices to manage the unmanageable in terms of ‘resistance’. In their work in the mental health centre, they found many clinicians who reacted with hostility to the ‘affront’ of managed care and the limitations on their professional autonomy created by these forms of reimbursement. Certainly, the staff of the PER often attempt to resist the complex forces that pressure them from all directions as they work to best meet the needs of the patients who present to them. Unlike staff in a clinic setting, however, their options for resistance are more limited by the factors we have described.

There are several limitations to this study. The timed observational data were collected during a two-month period in the spring, a period which may have affected patient flow and impacted on the staff’s work. Observations at other times of the year may show different work effects. In addition, the observations were conducted in public areas of the PER only in order to address the many challenges to conducting research in busy, public, urban PERs; including, the need not to disrupt patient care, concerns around confidentiality, and respect for staff workloads. We cannot draw conclusions about staff interactions with patients or other staff in clinical settings as we did not conduct observations of clinical interactions. As noted, the PER setting was a busy, urban setting and limited our ability to understand non-urban settings. Finally, these data do not allow us to examine the impact of staff work tasks on their perceptions of their work and work environment.

These data are some of the first observational data gathered in this challenging and important setting. The procedures used of using multiple observers, systematic and repeated observations over varying conditions and the structured form for notes, helped to increase the validity of the observational data collected (Adler and Adler 1994). An additional strength of this study is the trustworthiness of the findings. When the results of this paper were presented to and discussed with PER staff they reported that the findings were an accurate reflection of some of the challenges inherent to working in the current community mental health system. Filstead (1970, as cited in Patton 2002: 53) explains that it is essential to validity and reliability to envisage the context as the subject does, rather than impose the researcher’s perspective, while simultaneously maintaining rigorous instrumentation and procedures.

The research team developed a co-operative working arrangement with the clinical staff at the PER prior to the start of the study. Study staff frequently attended PER staff meetings to share preliminary findings with the PER team and to answer questions about the study. This co-operative arrangement served several purposes including facilitating the conduct of the research, to function as a form of member-check, and for the staff to consider how to use these data towards having program needs met. As findings were presented, staff were able to clarify questions for the research team and to confirm meanings of conversations and events observed. Staff of the PER were able to use these data on the types of work and the timed observational data to push for changes in the staffing patterns of the PER.

Conclusion

This observational study of staff activities in one PER is important for several reasons. Despite the critical need to better understand what occurs in psychiatric emergency settings, conducting research in this setting is highly challenging. Retrospective studies of chart review data do provide one solution to this challenge; however, chart review analyses fail to provide the rich data needed to better understand staff activities in, and the work structure of, the PER. The work of the staff of a busy, urban PER creates unique challenges for staff. Staff have a complex, multilevel (i.e. patient/workload/mental health system) view of their environment. These observations of the amount of time spent on PER activities as well as PER staff descriptions of their work highlight the manner in which PER manage their complex work tasks and interact with patients, other PER staff and the larger community of service providers. The work of the PER is service-work, and as Hopper (2006) notes ‘The paradox of service-work is that the delicate matter of reaching out to those left behind, made redundant or declared deranged, must somehow be managed by people paid to do it’ (2006: 221) Those seeking care at this urban PER are often those left behind, and the staff of the PER do somehow manage to serve their clients.

The President’s New Freedom Commission on Mental Health has identified that there is a gap in our knowledge about acute psychiatric care in the United States. Given the critical importance PERs play in communities and the mental healthcare system, this knowledge gap must be addressed so policy makers, mental health professionals, and consumers of mental health services and their families, can develop programs and services which best meet the needs of our communities. These data are a critical step towards a richer understanding of the instrumental, task-oriented work of the PER.

Finally, this study contributes to the growing body of work which focuses attention on the importance of workplace context, including organisational culture, organisational climate and work environment, both for service workers and those receiving services. An international review of work in the acute ward highlights the shift in focus from individual patient characteristics to system and organisational factors in better understanding risk (Lelliott and Quirk 2004). Similarly, research on psychiatric emergency care should expand our focus beyond the individual clinical characteristics of those who seek care in these settings, to better understand the work done in PERs and the social context of such work. While this study does not examine the impact of these work setting characteristics on individual patient outcomes, it is clear that the unique nature of the PER work setting impacts on patient care in important ways. More work is needed to better understand the impact of the issues identified; the nature of the work, staff-patient relationships, and system issues on staff and patients; our findings have however identified these as critical issues.