Clinical handover within the emergency care pathway and the potential risks of clinical handover failure (ECHO)

Lack of adequate standardisation

A frequently identified contributory factor is the absence of adequately structured handover processes. 26 Interviews conducted in an Australian hospital found that 95% of participants did not identify a formal procedure for shift-change handover. 38 A qualitative study comparing handover practices to pit stop practices in motor car racing concluded that handover had no clear procedures and was not supported by formal checklists. 42 A focus group-based study involving junior doctors found that shift handover was perceived as frequently being conducted in an ad hoc or chaotic fashion, and without obvious leadership. 43

Inadequate documentation and over-reliance on documentation

Another contributory factor discussed in the literature is missing and inaccurate documentation, or inadequate reliance on documentation. A study observing nursing handover of 12 simulated patients found that purely verbal handover resulted in information loss fairly quickly, whereas verbal handover supported by a typed handover sheet suffered only minimal information loss. 44 On the other hand, the use of such handover sheets may potentially make the handover more vulnerable by detracting from the focus on the most relevant items. 45 Over-reliance on medical records was reported in a study that investigated handover and communication between doctors and nurses. 46 This study found that often there was inadequate communication, and, as a result, there were disagreements on issues such as planned medication changes (42%), planned tests (26%) and necessary procedures (11%).

Non-verbal behaviour does not support building of shared understanding

Although the content of handover has been studied frequently, less is known about how non-verbal behaviour influences the quality of handover. A recent study in a number of US Department of Veterans Affairs (VA) medical centres investigated types of non-verbal behaviour in nursing and physician handover. The authors concluded that participants frequently adopted forms of non-verbal behaviour that may result in suboptional transfer of information. 47 Such forms of non-verbal behaviour included holding patient lists or other artefacts in such a way that they could not be seen by the other participant (‘poker hand’), not having a joint visual focus (‘parallel play’) and situations where the person giving the handover was standing while the other party was sitting, which resulted in hurried handovers with fewer questions (‘kerbside consultation’). The most productive form of non-verbal behaviour was reported to be the joint focus of attention, where both parties co-ordinate their verbal and visual attention jointly on an object.

Lack of organisational priority given to handover and absence of training

The literature suggests that a lack of organisational priority given to handover, and the absence of formal training in communication and handover both at universities as well as within health-care organisations are further barriers to the implementation of effective handover. 26 A recent interview study investigating transitions from primary care into hospital suggested that participants perceived handover as an administrative burden that took away time for their patient care duties. The study also found that handover and communication competencies were rarely taught, and that clinicians learned these skills ‘by being around and immersed in the clinical effort’. 48 A national survey of internal medicine training programmes in the USA found that 60% of these did not provide training in handover. 36 One study reports that junior doctors had not received any training in handover, and that, as a result, they had a narrow view of handover concerning only completion of outstanding tasks. 43 Arora et al. 49 present a competency-based approach to improving handover that entails the development of a standardised instructional approach to teach communication skills and the establishment of corresponding robust assessment systems.

Ambulance service A/hospital C

The graphical representation of the pathway for resuscitation patients is shown in Figures 2 and 3, and the representation for the pathway for patients with major injuries is shown in Figures 4 and 5. The complementary tabular descriptions are included in Appendix 1.

FIGURE 2.

Hospital C resuscitation pathway – part 1. DOB, date of birth; ETA, expected time to arrival; NIC, nurse in charge.

FIGURE 3.

Hospital C resuscitation pathway – part 2. ABX, antibiotics; BLDS, bloods (blood tests); CXR, chest X-ray; ECG, electrocardiogram; MEWS, Modified Early Warning Score.

FIGURE 4.

Hospital C major injuries pathway – part 1. ACC, ambulance control centre; BLDS, bloods (blood tests); DOB, date of birth; ETA, expected time to arrival; MEWS, Modified Early Warning Score.

FIGURE 5.

Hospital C major injuries pathway – part 2. ABX, antibiotics; BLDS, bloods (blood tests); CXR, chest X-ray; ECG, electrocardiogram; MEWS, Modified Early Warning Score.

In the resuscitation pathway, the first type of handover may occur when the ambulance service [either the AC or the ambulance control centre (ACC) having received this information from the AC] provides a pre-alert to the hospital ED, providing initial information about patient demographics, presenting complaint category, patient acuity and expected time of arrival. This handover provides an anticipatory and logistic function. The A&E resuscitation team uses this information to inform and to assemble the appropriate staff members and other teams (e.g. trauma team), to prepare the resuscitation bay and to ensure that necessary equipment is available. The next type of handover occurs when the ambulance arrives at the ED and the patient is brought into the resuscitation room. One AC member hands over to the senior clinician (in this organisation the term clinician is used refer to roles that make decisions on the treatment of patients, such as doctors of different grades, ACPs and ENPs) information including patient demographics, clinical history, mechanisms and details of injury, assessments done by the AC, any treatments undertaken, and the known medication history. The social history is not usually handed over at this stage. The handover should follow the ATMIST communication protocol. Although the handover is to the senior clinician, other team members listen in, and they may have questions themselves. This handover is considered the main handover between the ambulance service and the ED. It serves the dual functions of communication of immediately relevant clinical information and of transfer of responsibility for patient care. Sometimes, aspects of the handover are repeated when team members arrive after the handover. The AC member remains in the resuscitation room during the first part of the assessment in order to finish the paperwork and to be available for further questions. The patient report form (PRF) is an important tool for the handover, as it documents all the information that was available to the AC, even if some of this was not verbally handed over (e.g. social history). It provides an archival function, enabling future access to information. The second crew member registers the patient at reception, where another secondary handover occurs including patient demographics. This handover provides a logistic function.

The next handover occurs when a decision to admit has been taken, and the senior clinician makes a telephone referral to a clinician on the AMU. This telephone handover includes patient demographics, presenting complaint, observations and investigations done, any treatments that have been done and their effectiveness, as well as a request for referral, either explicit or implicit. This telephone handover is considered as another point at which there is a transfer of responsibility for patient care, although practically the patient remains under the care of the ED. In addition to the referral, there are two further handovers between A&E and AMU. The resuscitation nurse gives a nursing handover to the AMU co-ordinator on the telephone. This involves the communication of information including patient demographics, Modified Early Warning Score (MEWS), assessments and treatments done, specific requirements, and the time of arrival of the patient in the ED. This handover provides both an anticipatory and logistic function. The second handover occurs face to face, when an A&E nurse and the porter take the patient to AMU. This handover communicates the same information as the previous telephone handover but this time the handover is to the nurse who will be looking after the patient. This handover entails a full transfer of responsibility for patient care.

For patients with major injuries there is not normally a pre-alert. The AC typically provides a brief status update to their control centre, communicating basic details about presenting complaint category, patient acuity and expected time of arrival to allow monitoring and oversight of ambulance activities. This information is also available to the hospital ambulance liaison officer (HALO), who is stationed in the ED. The first handover to the ED occurs when the AC arrive at the ED with the patient. This handover is a verbal handover to an assessment nurse at the rapid assessment triage (RAT) point, and it potentially includes all of the clinical and social information that the AC has gathered, for example patient demographics, clinical and social history, symptoms, and any assessments and treatments done. The PRF supports the handover. The assessment nurse triages the patient and instructs the AC to take the patient to a dedicated cubicle. This handover is perceived to include a transfer of responsibility for patient care, and this is documented by signing the PRF. In parallel to this handover, the second crew member registers the patient and does a basic handover to the receptionist, providing a logistic function. The assessment nurse then goes to do an in-depth assessment of the patient. There may be optional, additional handovers depending on the acuity of the patient and the workload and availability of the cubicle nurse. The assessment nurse may prioritise the patient with the physician or may instruct the cubicle nurse to undertake certain activities. The latter serves the purpose of delegation of aspects of patient care. A further optional handover may occur if the cubicle nurse provides an update on the patient to the clinician who comes assess the patient. The next handover occurs when a decision to admit is taken, and the clinician refers the patient to AMU through a telephone handover to the AMU nurse co-ordinator. On the AMU side, this handover is supported by a SBAR checklist. This is again perceived as a transfer of responsibility for patient care, providing both an anticipatory and logistic function. The final handover occurs when an A&E nurse takes the patient to AMU and hands over to the nurse looking after the patient. This handover entails a full transfer of responsibility for patient care.

Ambulance service B/hospital D

The resuscitation pathway is identical to hospital C as far as the basic steps are concerned. There are small variations in terms of tools used. For example, there is a dedicated ATMIST reporting form next to the telephone where pre-alerts are received. However, in most aspects the resuscitation pathway is standardised across organisations.

The pathway for patients with major injuries is shown in Figures 6 and 7. The model adopted at this site differs slightly from the model adopted at hospital C. This site does not normally have a dedicated HALO. A HALO is present in situations of high demand, only. The information displayed on the ambulance computer screen prior to the arrival of the ambulance is not commonly used. The handover from the AC is to a dedicated senior nurse co-ordinator and may include all information that the AC have gathered. The nurse co-ordinator records on a pro forma basic triage details, such as patient demographics and triage category. This sticker is given to the AC to take with the patient. This handover is perceived to include the transfer of responsibility for patient care, and also provides a logistic function enabling assessment of the impact on the resources in the department. In order to improve ambulance turnaround times, the AC is supposed to leave the patient in the assigned cubicle and return to their vehicle. In practice, there may be an informal secondary handover from the AC to the nurse who is going to look after the patient. This handover provides the function of highlighting aspects of the patient’s care that are of particular importance. Another difference is the referral from A&E to the EAU, which is a handover from the A&E clinician to a doctor on EAU. This is perceived to involve the transfer of responsibility for patient care. The clinician will then inform the A&E nursing staff of the decision to admit. This serves the purpose of delegation, prompting the senior A&E nurse to start to arrange the transfer of the patient to EAU. The senior nurse will liaise with the senior nurse on EAU. This handover provides an anticipatory and logistic function.

FIGURE 6.

Hospital D major injuries pathway – part 1. GP, general practitioner; NIC, nurse in charge.

FIGURE 7.

Hospital D major injuries pathway – part 2. ABX, antibiotics; BLDS, bloods (blood tests); CXR, chest X-ray; ECG, electrocardiogram; NIC, nurse in charge.

Hospital E

The pathway for resuscitation patients follows the same general model as those at hospitals C and D.

There are, again, some variations for the pathway for patients with major injuries. The pathway is represented in Figures 8 and 9. There is a handover from the AC to the senior nurse co-ordinator at the nurses’ station. This handover follows the same pattern as at the other sites, and it includes the transfer of responsibility for patient care and also provides a logistic function. At this site, there is a dedicated assessment room, and the AC transfer the patient to this room after the handover. There is an assessment team consisting of a qualified nurse and a HCA, who assess and triage the patient. There is not normally a handover to the assessment team. Following the assessment and depending on the acuity of the patient, there may be an optional prioritisation of the patient with the nurse co-ordinator, who, in turn, will prioritise the patient with the doctor. Upon the decision to admit, the referral takes place as at hospital D as a telephone handover between the A&E clinician and the clinician on CDU. This referral includes the transfer of responsibility for patient care. As before, there are subsequent handovers between the senior nurse on A&E and the senior nurse on CDU, as well as the face-to-face handover between the A&E nurse and the receiving CDU nurse.

FIGURE 8.

Hospital E major injuries pathway – part 1. DOB, date of birth; ETA, expected time to arrival; NIC, nurse in charge.

FIGURE 9.

Hospital E major injuries pathway – part 2. ECG, electrocardiogram; FY2, foundation year 2 doctor; NIC, nurse in charge.

Discussion

A summary of the findings of the pathway description is given in Table 10.

The process walks, observations and process mapping sessions provide a detailed representation and understanding of how handover is linked to clinical practice, and the different goals and functions it can serve. There are numerous situations where patient-related information is handed over, and this takes place between people with different backgrounds and potentially different goals. For example, the handover between paramedic and senior A&E nurse involves potentially conflicting goals. The goal of the paramedic is to communicate relevant clinical and social history. The goal of the receiving A&E nurse, on the other hand, is to assess the criticality of the patient and determine the impact on the resources available in the department.

Handover may involve only the communication of patient-related information or it may involve also the transfer of responsibility for patient care. Pinpointing where this transfer of responsibility occurs is sometimes difficult. For example, when ambulances are in a queue and waiting to hand over the patient to the senior nurse, they are on the premises of the ED but have not formally handed over responsibility. Likewise, when a patient is referred on the telephone, the patient may remain on the premises of A&E for some time, and will be cared for by A&E staff even after the handover has taken place and the patient is shown on the computer system as under care of another specialty.

The resuscitation pathway is similar across sites and exhibits a higher level of standardisation than the pathway for patients with major injuries. Communication is often supported by protocols (e.g. ATMIST) and checklists for recording information. This may be due to the acuity of the patient, and the availability of national supporting guidelines and protocols.

On the other hand, the pathway for patients with major injuries exhibits greater variation across sites, although not being completely dissimilar. Different models are adopted at the different sites for receiving patients from the ambulance service, for example handover to an assessment nurse who also looks after the patient or handover to the nurse co-ordinator, who is responsible primarily for managing the department and patient flows. The latter model is adopted to support the logistic function, i.e. in order to provide a better overview of demand and capacity, and to speed up ambulance turnaround times. There are also different models for referring patients from ED to acute medicine. An A&E clinician always gives the referral but it can be given to either a doctor or a nurse in acute medicine.

This systematic description of handover along the emergency care pathway embeds handover within the overarching patient journey, and illustrates the different types of handover that are present as well as the goals and motivations that may be involved. Such a representation is useful to investigate further both the risks that arise from handover failures (see Systematic risk analysis, below) and the way the verbal communication actually takes place (see Content and language form of handover, below).

Ambulance service A/hospital C

At this site, 10 handover failure modes and associated risks were perceived to be of particular significance (risk scores of ≥ 12). These are discussed in more detail in Appendix 3. A brief list of the top five handover failure modes is shown in Box 3 for illustration.

Participants related a number of these failure modes to a lack of capacity within the hospital, in terms of either staff or beds (but there may be other reasons too). For example, delays in handover of patients from AC to A&E staff may occur because of queues upon arrival; the fact that medical registrars on AMU do not accept patient referrals straightaway may be due to a lack of available beds on AMU and the resulting more stringent ‘gatekeeping’ behaviour; and the inadequate verbal handover of the nurse accompanying the patient during transfer to AMU owing to unfamiliarity with the patient is usually down to the fact that the nurse who had been looking after the patient is busy elsewhere and unable to accompany the patient. Overcrowding in the A&E department also means that nurses have to make more difficult decisions about whether or not to prioritise patients with the clinician, and a lack of experience may contribute to situations where patients are not prioritised adequately.

Further commonly identified contributory factors are described in the vignette below (Box 4) describing the risk arising from the failure to prioritise the patient with the clinician.

Ambulance service B/hospital D

At this site, nine failure modes and associated risks were perceived to be of particular significance (risk scores of ≥ 12). These are discussed in more detail in Appendix 3. A brief list of the top five handover failure modes is shown in Box 5.

The lack of capacity and delays due to queues are perceived to contribute to risks identified at this site as at the previous site. In addition, handover for resuscitation patients was perceived to carry significant risks owing to the criticality of the patient condition, where even short delays could have severe adverse consequences. On the one hand, this makes the pre-alert an important step providing an anticipatory function as it allows the A&E department to prepare for the patient. Failure to give a pre-alert either because the ambulance is very close by or because the criticality of the patient may not have been recognised may lead to situations where there can be delays in getting the right people to the resuscitation room or freeing up space. On the other hand, owing to the perceived need to act quickly, A&E staff may get hands-on as soon as the patient arrives, which makes the handover from the paramedic more vulnerable as important information may not be heard or consciously processed.

Further commonly identified contributory factors are described in the vignette below (Box 6) describing the risk arising from delayed ambulance handover.

Hospital E

At this site, nine failure modes and associated risks were perceived to be of particular significance (risk scores of ≥ 12). These are discussed in more detail in Appendix 3. A brief list of the top five handover failure modes is shown in Box 7 for illustration.

As with the other two sites, the highest-ranking risks often relate to capacity and resource issues, such as the unavailability of resuscitation beds, the unfamiliarity of the nurse accompanying the patient with the patient’s story and the resulting inadequate verbal handover, possible gatekeeping behaviour owing to lack of capacity on CDU, and the missing of important information due to getting hands-on too quickly in situations of perceived urgency.

Further commonly identified contributory factors are described in the vignette below (Box 8) describing the risk arising from failed referral of a patient from A&E to acute medicine (medical registrar does not accept referral).

Discussion

A summary of the main lessons from the risk analysis is provided in Table 11.

The output of the risk analysis at each site shares many similarities with the other sites. The handover failure modes that have been identified apply across all three sites. There were no failure modes with significant associated risk that were particular to just one of the sites. The causes for the failure modes were also similar across the sites. This may be one of the most important lessons of the risk analysis. Many of the identified failure modes are linked causally to capacity and resource issues. As described, inadequate patient flow may lead to overcrowding in the ED, giving rise to several potential handover failures: delays in AC handover, more difficult prioritisation decisions, and inadequate patient transfer handover due to unfamiliarity with the patient. In addition, in order to manage patient flows better, handover from the AC may be taken by a senior nurse with an overview of capacity of the whole ED. The senior nurse has different information needs, which may result in information such as social history not being communicated or not being consciously heard. Inadequate patient flow into the hospital further contributes to overcrowding. This may be caused by resource constraints on the wards themselves. Appendix 3 provides a more in-depth description of the risks identified, the possible contributory factors, and mitigations that have been proposed during the FMEA sessions.

Differences existed in the assessment of risk associated with failure modes, i.e. in the estimation of the likelihood of occurrence and the severity of the consequences. This represents a limitation of FMEA when used for the risk analysis of sociotechnical systems. As described earlier, FMEA is traditionally used as one of several methods to conduct the risk analysis of technical systems. For such systems failure modes and corresponding failure rates are often sufficiently well documented to ensure consistency. For example, failure rates of computer processors or bit flips of memory chips are readily available. On the other hand, it is significantly more difficult to provide consistent estimation on the likelihood of a person not listening attentively during handover. The estimation of the severity of the consequences is difficult both for technical as well as for sociotechnical systems. However, in health care this is exacerbated by the need for consideration of the patient condition. Depending on the patient condition, one could actually look at radically different ‘systems’, in as much as the patient condition represents a major factor in determining the consequences of a failure. Participants repeatedly stressed that if a patient was sufficiently vulnerable, even the most insignificant delays or errors could be sufficient to trigger a catastrophic turn of events. The conversations that participants had when determining the severity of the consequences provided insights that their understanding and perception of the severity of consequences were fundamentally similar, even if they assigned a different risk score across sites.

Frequency of handover communication content and language forms used

The three types of interorganisational and interdepartmental handover selected for analysis of the content and language forms used were (1) resuscitation handover from paramedic to senior ED doctor; (2) major injuries handover from paramedic to senior ED nurse; and (3) the telephone referral of patients from ED clinician to acute medicine doctor or senior nurse. At hospital E, only referrals were studied, as the ambulance service did not provide permission to involve their staff in the study.

Table 12 shows the number of audio-recordings that were collected for each type of handover at the three sites, and the number of audio-recordings included in the analysis. Table 13 shows the mean duration for the different types of handover at each site (low and high values shown in parentheses). Some audio-recordings could not be included owing to excessive background noise, poor recording quality or technical failures (technical failures refer to, for example, telephone handovers during which only utterances of one party could be heard due to problems with the in-ear microphone used to record telephone conversations). All suitable audio-recordings were included in the analysis. Suitable means that the conversation could be meaningfully reconstructed during the transcription process, with the exception of individual inaudible words or very short utterances.

Tables 14 and 15 provide overviews of the frequency with which particular content was communicated during each type of handover, and the frequency of the language forms used.

Qualitative interpretation of frequency counts

Review of the frequency counts provided a number of qualitative insights.

Standardisation of handover communication

As described above, ambulance service handover may be standardised more easily than referrals. The ATMIST communication protocol had already been implemented by the participating ambulance services. Referrals may be more difficult to standardise but they may still benefit from a structured approach. This issue will be described in more detail with examples in this short section.

The AMU at hospital C had adopted a checklist-supported communication protocol for referrals from the ED. As a result, referrals at this site usually followed a particular pattern, often guided by the receiving AMU nurse:

(*PID, patient identifier.)

An experienced clinician in the ED can pre-empt what he or she is going to be asked, and therefore the handover may start with the clinician verbally filling out the AMU form for the AMU sister.

(*PID, patient identifier.)

At hospital E no standardised approach had been implemented. Below the full transcript of a referral from ED to CDU at hospital E is shown. The initial part of the transcript illustrates that the communication partners do not share an understanding of how the conversation should unfold. As a result a number of clarifying questions are used to repair the conversation flow. For the major part, the conversation is guided through questions from the registrar on CDU receiving the referral. There are several instances where the ED doctor is unsure what information is required and needs to revert to asking for clarification. Throughout the conversation it appears that the communication partners are navigating from utterance to utterance without being able to develop a shared understanding of what is going to be discussed next. The CDU registrar deduces and summarises the reason for admission only about two-thirds of the way through the conversation.

(*CRP, C-reactive protein; INR, international normalised ratio.)

The transcript also illustrates that the duration of the handover and the number of questions that are asked, are not necessarily markers of quality. Senior clinical decision-makers will often give a more detailed referral to a specialty team in less time. In addition, interruptions from the specialty can sometimes throw the inexperienced referrer off their tracks, and prolong and confuse the handover process. It is less likely that there are extended discussions about whether or not to accept the referral of a patient in a top-down hierarchical handover (ED consultant to medical registrar) than with an ED foundation year 2 junior doctor to a medical registrar.

Documentation

The major tensions inherent in documentation described by interview participants are summarised in Table 22.

The documentation of information is an important part of clinical handover and it contributes to the delivery of high-quality care. There is a drive at organisational level across organisations to document as much as possible (‘if it’s not documented, it didn’t happen’). This is both for legal as well as for quality assurance purposes. In addition, there is a perception at the management level that the number of verbal handovers can be reduced if everything is documented comprehensively (‘no point in them [AC] replicating handover several times’). This would save time and allow, for example, the ACs to get back out on to the road more quickly. On the other hand, there is the recognition, especially among practitioners directly involved in patient care, that one cannot simply rely on documentation. Documentation can be missing, temporarily unavailable, inaccurate and incomplete. Documentation may not highlight or prioritise the important bits of information, and subtleties or concerns may not be documented. Most importantly, documentation is unidirectional and, therefore, does not provide the opportunity to discuss and clarify areas of particular concern or interest. This is important to staff who take on responsibility for patient care. For professional and personal accountability, gathering accurate and up-to-date information in a dialogue is perceived as highly valuable (‘get things clear in my head’).

One participant from the ED describes this tension, where they also indicate the importance of experience in determining the amount and ultimately the quality of documentation. In this view, documenting everything is a naive and suboptimal approach. Experience provides people with the skills to determine and articulate what is important.

You’ve got an organisational system, which ideally should allow all these ports of transfer of information. So they should allow the actual chat between paramedic and the nurse or the paramedic and the doctor and should also allow the written conveying of that information. The formal part, the recorded part must be there for legal purposes. [. . .] In both my management and my clinical role, I would insist that both are always done. To be able to effectively do that is experiential. So the person with little experience is going to be required or going to have a necessity to record an awful lot of normal information. Because he or she, or the computer. Because computers are doing a lot of this sort of thing now which is a most dire system. You know, it really is. But it’s a computer. It’s a computer saying no! And it’s a very simplistic thing. Then you get an untrained person, they’re going to have a simplistic approach. And the simplistic medical or nursing practitioner, you take every single bit of information and you record it. The consultant’s approach, the senior person right at the end, is going to record the important positives and the important negatives. He or she can justify his clinical actions, justify the treatment and will hopefully represent exactly what the patient said.

The way documentation is used is also influenced by departmental capacity and workload. Writing comprehensive notes takes time, and people may write less when they feel that they are under time pressure. This may lead to omissions. Equally, the person taking on responsibility for patient care may not read the notes or only consult them superficially owing to the perceived pressure of seeing the patient quickly.

A second participant from the ED describes this trade-off, and also emphasises again the importance of experience over a simplistic standardised approach that may be inappropriate depending on the circumstances.

You can read stuff very quickly. I think you can probably read information quicker than you can hear it if someone’s speaking to you. So the quickest way would be to read what somebody’s written. But that then takes longer for the person who actually has to write it down because it takes longer to write than to say. So it’s really looking to all different time pressures because everybody is under a time pressure. You can’t have a standard formula for all the information because for a 21-year-old, otherwise fit and healthy person, you probably don’t want to know the social information you might need to know for someone who is house-bound with elderly relatives at home. So you can’t standardise and say this is what must be given. In many ways, it’s experience knowing what’s important and which is the information that you really do need to get.

Verbal communication

The major tensions inherent in verbal communication described by interview participants are summarised in Table 23.

Handover is usually conducted verbally, either face-to-face or over the telephone, even although there is now a trend to rely on documentation in order to reduce the number of handovers and the time required. Participants described many advantages of verbal communication that enables them to provide better care to patients. Verbal communication goes beyond documentation and provides added value by conveying subtleties and information that may not be immediately relevant to the patient’s acute condition, but which may be of importance later on (‘small things that may seem unimportant, but aren’t’), by allowing highlighting and prioritisation of the important pieces of information, and by providing information first hand (‘straight from the horse’s mouth’). As it is a dialogue, it also allows for questioning, feedback and education. Verbal communication also fosters collaboration between colleagues, and the resulting personal familiarity with colleagues is perceived as an important facilitator in providing high-quality care across departmental boundaries. On the other hand, participants also pointed out many problems they faced with verbal communication. It relies on memory and it usually conveys only what the sender perceives as the important issues. The communication can be unstructured, rushed, and jumping from topic to topic without providing a clear picture of what is required. Verbal communication often takes place in less than adequate locations, where there are frequent interruptions, high levels of noise and activity, and little privacy.

One participant from the ED describes the tension that is inherent in verbal communication when referring a patient by contrasting this with the alternative form of electronic referrals, pointing out the potential risks of not having the possibility of a discussion.

Then there’s the actual transmission of the information of course, which obviously there’s the amount of communication that you have, the method of it and the records of it, so our systems at the moment for instance, we’re trying to move with some of our referrals to an electronic system where we would be able to transmit information electronically, because what happens at the moment is a lot of the conversation that’s happening is just repetition of information that’s already available. Now, there are two sides to it. As the manager of this area I can say, yes this will smooth the process and make it much quicker. Looking at the system I would say, yes but that also means that the person, if there isn’t any element of discussion or challenge at the point of referral, potentially these inaccuracies won’t be picked up, so there’s kind of – that’s something that can improve systems, but also a risk can be accentuated by that as well.

Verbal communication is also strongly affected by departmental capacity and workload. On the one hand, verbal communication allows people to prioritise their work and manage flow and their own workload more efficiently. On the other hand, however, the lack of capacity and flow may mean that the communication is delayed (e.g. due to queues or unavailability of one party) or may even be skipped altogether. The lack of capacity can also lead to situations where the verbal handover is given by somebody who is not familiar with the patient, and who is thus unable to provide that added value. This may also lead to frustrations on the part of the receiving party (‘we can all read’). The verbal communication can also be perceived as taking away time that could be spent on actual patient care.

A second participant (from ED) describes how a trade-off could be made between relying on documentation only and having the verbal communication that allows providing added information. The trade-off, in this instance, relies on the subjective notion of ‘being worried’, i.e. on the clinical judgement by and experience of the person giving the handover about what is of importance.

If I’m seeing the patient first, I’m quite happy just to look at that initial documentation. When I want a verbal handover is when they’re worried about a patient because if they’re worried, I want to know and I want to know earlier. I don’t just want to happen to find it. So I like to know that if somebody is concerned, they will alert me otherwise I think it’s too much by chance that you find out that your patient is sick whereas if that’s already been recognised, then you go in then with a heightened sense of urgency and then heightened awareness.

Transfer of responsibility

The major tensions inherent in the transfer of responsibility described by interview participants are summarised in Table 24.

Clinical handover entails both the communication of information and the transfer of responsibility for patient care. The explicit transfer of responsibility ensures that there are no gaps in the provision of care. Participants receiving handover described the transfer of responsibility in terms of personal accountability (‘it’s my patient’), and their desire and professional responsibility to gather all relevant information (see verbal communication). The same holds for the party giving the handover, who perceive it as their professional duty to communicate all the information they have gathered. When the transfer of responsibility involves verbal communication, this provides also the opportunity to have a discussion, review treatment plans, and to ensure that the transfer of responsibility is appropriate and in the patient’s best interest (‘ensuring the patient goes to the right place, first time’).

Participants from ED described the referral of patients from ED to other departments as ‘a difficult conversation’ and ‘not a handover, but a bargaining tool’, which may result in the transfer of responsibility for patient care being refused. Participants from ED and from acute medicine provided differing explanation for this experience. From an ED perspective the refusal to accept patients by specialists or the ‘selling of patients’ that they have to do may be caused by ‘boxing’ or ‘gatekeeping’ behaviour of specialist wards and their concern for their own work. Selling of patients in this context means that the person giving the referral perceives that they have to make the patient appear attractive to the specialist or highlight those aspects of the patient’s condition that make the patient particularly relevant to the specialty. They report that ‘A&E is seen to create work for others’ and that there is an absence of a pulling mechanism from wards. Such pulling mechanisms would entail that specialists highlight appropriate patients within the ED and will proactively organise the forward planning of their care. They also described situations where junior doctors receiving the referral were unfamiliar with organisational policies for admitting patients that provided clear guidelines. On the other hand, participants from acute medicine described the conversation around patient referral as an opportunity to discuss and to ensure that the patient goes to the right place. They also described situations where handover was poorly structured and confusing, which may cast doubts on the appropriateness of the referral. They also reported on a tendency of junior doctors in A&E to ‘medicalise’ reasons for admission even if the admission is primarily for social reasons or if there is a large degree of uncertainty.

One participant from ED describes the difficulties they experience when referring patients that may not fit a particular specialty unambiguously. This can lead to discussions and refusal to accept the patient, which ultimately leads to delays and situations of crowding. The solution they offer is being assertive and being backed by a formal trust policy on referrals and admissions into hospital.

The classic thing is medicine has become so boxed and every specialty in the hospital has made their box as small as possible and they put up as big a wall as they can around it, so we’ve got our upper GI [gastrointestinal] surgeons, we’ve got lower GI surgeons, we’ve got little finger toenail surgeons almost. It’s just ridiculous. We, as the attending clinicians, have to make a decision about who is most appropriate. And if we’re wrong, which we will be, they then send on to the next team that they think. But it’s classically patients who fall in between. So that GI bleed. Is it lower GI or upper GI? So should a surgeon take that or is it Medicine? So you can end up with patients waiting in the ED, and that’s what’s classically has happened across EDs, across the country. For hours and hours and hours, no one makes a decision. So we have to make a decision which way they should go. It’s still a problem but we try and force the issue by doing these techniques. The patient is coming in, I say they are coming in under you, that’s been agreed by the Chief Operating Officer of the Trust [name], and you need to come and see them now. Occasionally they fight back. Occasionally I have very difficult conversations with consultant colleagues. And they say ‘Oh, this is all about 4 hours’ and I say, ‘Yes, it is about 4 hours.’ But that’s really about quality of care of patients.

Participants described as problematic those situations where the transfer of responsibility was implicit and unclear. This can lead to situations of patients being ‘stuck or lost in the system’. Participants described several such scenarios. When ACs are queuing owing to a lack of capacity in the ED, there may be an unclear allocation of responsibility for the patient’s care. On the one hand, the patient is physically still on the ambulance stretcher. On the other hand, the patient is physically already in the ED. Handover may not have been given yet, but paramedics are not trained to provide extended care in the ED. This situation gets even more complicated when a rapid triage has been conducted, but no explicit transfer of responsibility has taken place. Similarly, the elimination of handover between paramedics and the cubicle nurse in favour of a single handover to the nurse co-ordinator or triage nurse, may lead to situations of patients being left in cubicles without the nurse realising and patients being left unattended (‘patient saying “is anybody going to come and see me?” ’).

The referral of patients from ED to other departments may also lead to situations of unclear allocation of responsibility. Participants described situations where patients had been referred by telephone, but remained in the ED for an extended period of time until a specialist was seeing them. Even more confusing situations arise when patients have been seen by the specialist (‘eyeballing’) but are waiting for a bed to become available on the ward or where they have been admitted on to a ward but remain under the care of the ED consultant.

A participant from an ambulance service describes the problem of allocation of responsibility when there are delays at the ED and ACs are queuing. For them, the responsibility for care should transfer as soon as the ambulance arrives at the ED. In their view, the solution to achieving this lies in clear guidelines, education and greater collaboration between organisations.

In fact, I’ve just come from a meeting with one of our hospitals in the region where we’ve been exploring the issues of ambulance delays, and one of the issues that I’m certainly now very comfortable with is that I think all acute hospitals are on the same page as us as an ambulance service and they understand that responsibility does commence as soon as patients arrive and that it’s the responsibility of all of us to work jointly to ensure that that handover process takes place as rapidly as possible to free the crews up so that they can get back out to the community and undertake their primary role. [. . .] Clearly potential tension lies at an operational level. My observations are that, at a strategic and tactical level within the organisation, there’s no issue. People understand the issues and responsibility. The reality, I think, where things break down, is there is failure in the operational plans or implementation of effective handover. That’s the level where things can break down. I think one of the things that contribute to that is a lack of understanding, particularly amongst emergency nursing staff, about where the duty of care begins and just simply ignoring a patient that’s arrived for half an hour, if they’re in the ambulance entrance or still on the vehicle, doesn’t offset their requirements in terms of responsibility. So I think, when people understand that at an operational level then there’s much better engagement. Clearly one of the things that I’ve been anxious about is to try and ensure as far as we’re able to make sure there isn’t tension at operational level and so this is where education and working together jointly becomes, I think, very important.

Goals

The major tensions inherent in the different goals that handover may serve are summarised in Table 25.

Different actors have different motivations and these can create tensions in the handover process. Handover plays an important role in managing patient flows. Participants described the importance of identifying more seriously ill patients quickly, and of determining the impact on departmental resources and capacity. For example, the senior nurse receiving handover from paramedics prefers to have a quick handover in order to understand the impact on the department and to decide where the patient needs to go (‘I want to hear three words’). Handover to a single dedicated nurse co-ordinator ensures that the co-ordinator can construct and maintain this departmental overview, and thereby provide better management of patient flow. On the other hand, from a clinical perspective handover also helps staff in assessing their own workload and in prioritising their activities accordingly. In addition, both parties involved in handover are motivated by personal and professional accountability. In the case of paramedics this means, for example, that more information than the nurse co-ordinator wants to hear will be communicated or that the paramedics will seek out the nurse looking after the patient to convey this additional information (‘seeing the patient through’), even if from a managerial perspective this may be perceived as needlessly replicating handover. Similarly, the nurse looking after the patient would like to hear the full story to enable them to provide high-quality care, rather than having to rely on documentation only (‘it’s a responsibility you take on’).

A participant (from ED) reflected on the informal practice of paramedics giving a second handover, although this was discouraged at organisational level.

And I think that’s why you still do get the verbal handover to a certain point because I think the Paramedics want to know that somebody realises that their patient has arrived. So I think they want to know that they have completed that care. And I think it is just professionalism isn’t it, that you want to know.

Staff from different departments and organisations recognise that they have to work together and trust one another in order to provide best possible care. The party receiving handover has to work with the information that is provided, and the sender should describe subtleties, concerns and uncertainties (‘give an honest rundown’). Ideally, staff try to anticipate each other’s information needs and act accordingly, for example by doing all observations before transferring a patient. Participants described the need to understand the pressures and capacity limitations faced by other departments and organisations. There is a tension between having to trust one’s colleagues (‘take what the GP [general practitioner] says as gospel’) and avoiding duplication (‘patients don’t want to be asked the same things over and over’), but also knowing that the stories they can get from, for example, the paramedics may be quite different from the patient’s story (‘decisions cannot be based on information from ambulance service alone’). In addition, the patient’s condition may have changed over time. So, on the one hand, people have to trust each other, but on the other hand they are not relying on others uncritically.

A participant from ED describes this tension, and they provide a reflection of how they address the required trade-off in practice by verbally checking with the patient and adding to the documentation selectively, rather than simply copying what was documented before.

You can check the information verbally but, if it’s correct, you don’t need to write it out. I have literally seen people who have got my notes and copied them. That isn’t them being personally accountable. That’s them being good at copying. They are not actually necessarily checking what I’m doing, they are just copying. That is totally and utterly redundant and should not happen. [. . .] I will often go through and say, say an example of diabetes, ‘I believe you are diabetic. How long have you been diabetic?’ And add in the extra information I am getting. Or sometimes I’d say ‘It says that you’ve had a stroke.’ And they’d say ‘No, I haven’t’, in which case, then again, I will say that they haven’t. So it’s more that the time which you are really wasting is when you are documenting in writing information that is already in writing. It doesn’t mean that you can’t check it verbally. And, if you agree with what’s already been said verbally, you can just write ‘as above’. So then actually you can say that you agree with that. It doesn’t mean that you are no longer personally accountable. And then you can add in the extra information that you’ve added, your subtleties or whatever. And I think that everybody knows that a patient tells slightly different stories every time, so therefore you can actually put in your different subtleties. The pain that they’re describing. How they described it to the first person may be slightly different to the way they describe it so I would then document how they’ve described it to me. And that will all help to build up a picture. So that would be as far as people who say it shouldn’t happen because they’re not now personally accountable. You can still make yourself personally accountable by checking it verbally and documenting any discrepancies.

Trust among colleagues is put under pressure by time performance targets. Participants described situations where people may use purposeful misinformation and particular keywords in order to force others to prioritise and accept patients. As a result, there may be a degree of mistrust between individuals and departments, for example wards asking questions that appear inappropriate from an ED perspective or refusing to accept a patient. This was discussed above in relation to the transfer of responsibility during referral. The influence of targets will be discussed further below as part of the second theme, the management of patient flows across organisational boundaries.

Capacity and patient flow

Handover contributes to the efficient management of capacity and patient flows. In situations where there is a lack of capacity or inadequate patient flow, there may also be a negative effect on handover. The relationship between patient flow and handover is summarised in Table 27.

Participants related handover to capacity and patient flow issues by describing the purpose of handover as supporting the understanding of demand and capacity at a departmental level, as supporting the controlling of the flow of patients, and as enabling them to prioritise their own activities and patients (at an individual level). For example, a stated reason for having a dedicated senior nurse receiving all ambulance service handovers for patients with major injuries was that it provided an overview of ‘what’s coming in’ and allowed the creation of ‘awareness of the impacts on the rest of the department over the rest of the shift’. As described above, this goal of handover may be at odds with the goals of paramedics giving the handover, who may feel that ‘there’s so much more to tell’ than simply how sick the patient was. Even before ambulances arrive the provision of advance notification through pre-alerts or through the ambulance information system contributes to the anticipation of demand and the preparation for it in the ED in order to maintain patient flow.

The lack of capacity and inadequate patient flows may create multiple problems for handover. For example, participants described that problems occur when people are not answering their bleeps because they are in theatre (surgeons) or when the appropriate person is temporarily not around. In these situations handover cannot take place and the patient needs to remain in the ED or may be admitted without handover. In either case, this may lead to situations of unclear allocation of responsibility for patient care or delays in moving the patient onward. A similar problem can occur when handover has taken place, but the patient cannot be moved onward because there are no free beds. Another major problem described by participants is the situation where a nurse, who is not familiar with the patient, may need to transfer the patient and then give a handover that is perceived as simply reading off the notes and does not provide any added value. This occurs when the nurse who had been responsible for the patient’s care is busy with other duties and asks another nurse to take the patient in order to free up the bed and keep patients moving.

The lack of capacity also leads to ambulance queues. This is described as a potentially serious problem. On the one hand, ambulance queues may result in multiple and increasingly filtered handovers as one crew hands over their patient to another crew. This leads to situations where one crew may be handing over several patients with whom they are unfamiliar and where they cannot provide any additional verbal information. On the other hand, as crews are queuing, there may be situations of unclear allocation of responsibility, as already described above. Participants from the ambulance services described as the most serious problem resulting from queues the loss of capacity to respond to emergencies in the community as the ambulances are stuck at the ED.

A participant from an ambulance service describes the trade-off that needs to be made between potentially less than ideal handover and freeing up crews to get back out into the community based on a risk assessment of the patient’s condition.

So we’ll deploy an ambulance liaison officer, particularly if we’ve got multiple patients or multiple vehicles stacking and that is to support the management of those patients. To provide oversight, if necessary, to take over the care of those patients if necessary because one of the things we practice as an ambulance service when we’ve got stacking is to risk-assess those patients and see if we can free up those crews and get the crews to double up and monitor two patients, fully accepting that that’s less than ideal. We we’re in this situation we’re operating in a decompensated circumstance so, in other words, we need to modify. Essentially, we’re effectively managing patients as we would in a major incident where we are sort of streamlining processes but clearly we’re having to, out of necessity, to address unmet clinical need in the community. So of course, the other side of things that I’m very cognisant of is that the risks of having to accelerate clinical handover or to streamline the handover process and release crews back into the community. The risks of that, in my view, are far less than having a patient with chest pain or a potentially life-threatening issue with no clinician available to support or resuscitate them. So that’s the balance. But if the system is well resourced and well managed the key to this in terms of avoiding these systems, is having upstream management measures in place to prevent the queue occurring in the first place.

The same participant continues by describing that queues resulting from increased levels of demand are predictable, and that they should be accounted for by adequate resources in order to avoid having to make such trade-offs in the first place.

So, one of the things that I consider to be crucial in terms of risk mitigation is that, bearing in mind that these surges in demand are predictable, they occur at predictable times of day. Then the critical thing from my point of view is to make sure there’s an adequate nursing resource to be able to take responsibility for these patients as soon as they arrive. And I would suggest in the ratio of no more than five patients to one – probably three to one would be the right level. So modelling where there’s been a single Escalation Nurse may not be adequate if there are 10 or 15 patients waiting to be off-loaded and handed over and again, this is areas of practice that we’ve tried to share with colleagues.

Further problems resulting from a lack of capacity include poor documentation as people ‘write less when busy’, which contributes to the tension of people being asked to rely increasingly on documentation in order to avoid duplication of handover. The lack of capacity may lead to abuse of the handover process, for example certain keywords may be used during handover to force admission of a patient. This could set off the admitting team on a wrong path and lead to delays and complications in patient care. It contributes to tensions between the need to trust one’s colleagues and the knowledge that the system may be abused to ‘get the patient through the front door’. Familiarity with colleagues makes this tension manageable. However, this is threatened by high turnover of staff and the increasing reliance on agency staff, another capacity issue.

Time-related performance targets

Time-related performance target are an important quality improvement tool. However, the pressures created by the targets may introduce risks to patients and contribute to tensions between departments and organisations. The relationship between targets and handover is described in Table 28.

Emergency departments, and more recently ambulance services, have to meet time-related performance targets to ensure that patients get seen in the ED within a reasonable amount of time and that ambulances can get back out on to the road in order to be able to respond to emergencies. Meeting targets acts as powerful motivation. Participants with management responsibility described targets as an important quality improvement tool that ‘provides significant corporate focus’.

Frontline staff also felt quite strongly about time-related performance targets, but cautioned against too many pressures. In combination with increasing patient numbers and more junior staff responsible for their care this poses additional significant challenges. Some participants even expressed significant negative attitudes towards the resulting system (‘I hate it!’). For example, the target introduced for ambulance services was described as leading to a process ‘designed entirely around the target’, which may threaten the quality of care and which people feel is inappropriate. This can lead to situations that create patient safety risks, such as when a patient is simply left in the ED and the nurse is unaware.

A participant from ED describes the tensions in perception between clinical and management staff, and between the ambulance service and the ED. They describe a situation from their clinical experience when the focus on targets led to a hazardous situation for the patient.

That is a problem. We know it’s a problem. They [ambulance service] just say ‘It’s your problem, why are you keeping our paramedics too long’. It’s not us, it’s them pressing their button and getting out there. So there are issues around timing, and there are organisational pressures, which I don’t think you should put too much pressure on, because actually they ensure quality. But the ambulance service is great. They do a superb service but they have managers who are pressing their guys to be back out on the road, back out on the road, back out on the road. Too much pressure inevitably will cause quality of care to drop. I’ve seen a patient two or three months ago who was just left on the spinal ward in a cubicle. Our nurse didn’t pick it up. We had put in nursing governance issues. The patient was left on the spinal ward but none of us knew about it except the handover people. That’s because the paramedics had been told to get out and leave the boards here. In the past, we would always have a policy where they were always logged off by the paramedics. So organisational pressures to achieve organisational targets, like paramedics attending to patients within 8 minutes or whatever, are pressures, which [inaudible] can impact on the quality of handover and have done. And will continue to do. We guard against that. The paramedics guard against that. But you may have some people sitting in offices upstairs here or sitting in offices in [ambulance service] who don’t appreciate our efforts.

Frequently participants described how they used their professional judgement in order to balance the tension of having to meet the target and delivering high-quality care to the patient they are responsible for. This can lead to the adoption of an informal working practice and a discrepancy between work-as-imagined (i.e. prescribed by policy) and work-as-is.

A participant from an ambulance service described how they perceive the time frame set by the target as the point by which they have to be back out on the road. They will use this time available according to their own judgement in order to ensure good care.

We’ve got 15 minutes from when we arrive to when we should handover. [. . .] And then 15 minutes from when we’ve handed over to when we’ve finished our paperwork and we’ve come clear. [. . .] So we’ve only got a 30-minute window here. [. . .] So I look at it as, OK as long as I press that button I will wait around and talk to the nurse, as long as I’ve cleared in 30 minutes, how I spend my time here is up to me.

Similar attitudes were expressed towards the ED breach target. This may lead to what are perceived inappropriate referrals in order to meet the target, and it may threaten trust among colleagues from different departments. Participants also described that it may create negative perceptions towards A&E and their staff (‘wards don’t welcome A&E nurses’), thereby threatening interdepartmental collaboration. Participants also cautioned that meeting the target can lead to poor patient experience and that their responsibility was to ‘provide quality experience for the patient, not just moving them quickly’.

Collaboration across departmental and organisational boundaries

The problems around handover and patient safety are frequently linked to inadequate patient flows and lack of capacity (see above). These cannot be addressed where they present themselves, alone. An approach supported by the whole system is required. Greater collaboration and communication across departments and organisations through engagement at all levels and through joint working may offer possible ways for achieving this. The relationship between collaboration across boundaries and handover is summarised in Table 29.

The actors involved in providing emergency care to patients belong to different departments and organisations. This means that they have their own respective concerns and pressures, as well as their respective local cultures. This can at times lead to very strong views about the perceived lack of concern or unwillingness to collaborate by the other parties. For example, ED staff described their frustrations with specialists from other disciplines whom they perceive to be only ‘interested in their own work’. The tensions that may arise from this have already been described in detail above when looking at verbal communication and the lack of patient flows. Similarly, participants described their dissatisfaction with general practitioner (GP) attitudes and the provision of out-of-hours service. This may lead to predictable peaks in demands in the ED, for example on a Monday when patients who had been feeling sick over the weekend are sent in large numbers to the ED. These in turn cause ambulance delays and flow problems. Participants described that the solution to such problems will probably not be found in increasing the numbers of staff or improved local handover practices. The problem would need to be addressed at a system level, by engaging with GPs, where the problem is created.

Closely related to this is the set-up of the health system and the drivers who are in place. One participant from the ED suggests that the drivers would need to be reconsidered and realigned in order to create appropriate incentives for engagement and collaboration at the system level.

The key thing that needs to change though in the NHS is the drivers. The drivers are all in the wrong place. The money has to follow the patient. It doesn’t at the moment. The money is given to the self-interest groups who use it in their own way, to the GPs at the moment and to the hospital doctors to an extent. It is going back to basic simple things. You know, we’re not going to improve the quality of care until we sort out what’s driving things. I have problems with handovers because things that my consultant colleagues are judged on are nothing to do with the quality of patients that arrive in the department. The only person that’s really judged on sorting out emergency care is the emergency department. So no one else is bothered. They do not give a toss about it. So they don’t do anything about it. GPs couldn’t care less if I have 200 patients arriving at half past whatever. You know our numbers have gone up 10% in the last 6 months. They are the largest number of patients in our department. No one gives a damn about it. It doesn’t bother them what’s happening here. [. . .] You’ve actually got to change the drivers in the system. So while the drivers are set up as they were, these issues, these pressure points, are going to get worse. And they’re going to get worse at the final deliveries of care.

Participants also described their own attempts at finding solutions to these problems. A participant from the ambulance service describes engagement with people at all levels as a key ingredient in fostering greater collaboration, and in developing whole system solutions.

In fact I’ve just come from a meeting with one of our hospitals in the region where we’ve been exploring the issues of ambulance delays and one of the issues that I’m certainly now very comfortable with is that I think all acute hospitals are on the same page as us as an ambulance service. [. . .] The challenge is managing surges in demand and trying to secure the cultural awareness that this is an issue and that we all do have a responsibility to ensure that our crews get back out into the community quickly. [. . .] This is why I have felt it’s been critically important to engage at the highest level within the organisation so we’ve had engagement at Chief Executive and Medical Director level. [. . .] The reason that that is critical is that, in order to maintain flow in these circumstances, you actually need the whole system supporting so it requires good operational management but also actually requires clinical buy-in from the in-patient team.

Joint working was proposed as an additional way of strengthening collaboration and creating a shared awareness of risks to patient safety. This could include joint, part-time appointments between the ambulance service and the ED, or ambulance service staff being integrated within the ED. This would create staff with experience of working both in pre-hospital as well as hospital-based care.

Participants described the need to tackle the problem where it arises, before it actually becomes a problem (‘if there’s a queue, the system has failed’). This requires communication and collaboration between all stakeholders involved in order to develop a system that is able to maintain flow. For example, one way of reducing ambulance queues at the ED and the resulting problems for handover and quality of care is to reduce the number of necessary conveyance to hospital. Participants from the ambulance service described protocols they had established with GPs for patients with a degree of uncertainty about their past medical history. The protocol ensures there can be a discussion with the patient’s GP, who may choose to take on responsibility for that patient or advise that they be escalated, admitted or they may choose to arrange an alternative care pathway. Similar arrangements have been proposed by participants from the ED in order to ensure senior clinical input prior to referral of a patient by their GP. One participant nicely summarised what was needed was ‘a systems approach – how do we provide emergency care services for people who live in this area’.