Article Text
PDF
Abstract
Background The ESCAPE trial was a multicentre randomised controlled trial investigating the effectiveness of Chest Pain Unit (CPU) care. The process of CPU implementation and the activity of individual CPUs varied substantially between hospitals. The study reported here explored the organisational factors that influenced this variation.
Method A multiple case study approach was taken treating each site as a ‘case’. Six intervention sites were studied. Qualitative data were collected through interviews with key personnel at each site.
Results Activity of individual CPUs was not adequately explained by simple structural differences between hospitals, such as their size or location, or between CPUs, such as staffing and hours. Analysis suggested that the more active CPUs tended to have more of the following characteristics: being ‘primed’ by previous initiative or experience; appropriate leadership; a positive climate for innovation; established relationships between key staff/departments; role clarity amongst staff; an enthusiast for the development; and continuity of staffing. Role conflict, particularly between specialist nurses and others, was reported and had potential to interfere with development.
Conclusion Organisational factors were identified that could have impacted on the outcomes of the ESCAPE trial through, for example, delays in discharge, and missed recruitment opportunities. Complex interventions such as the ESCAPE trial are prone to the effects of local organisational issues, some of which could be predicted and planned for. Findings from single centre studies of complex interventions should be treated with caution before a decision is taken to implement in a new setting.
- Cardiac care
- care systems
- case study
- chest pain
- CHD
- emergency department management
- management
- organisational
- qualitative
- research, trauma
Statistics from Altmetric.com
- Cardiac care
- care systems
- case study
- chest pain
- CHD
- emergency department management
- management
- organisational
- qualitative
- research, trauma
Background
The Chest Pain Unit (CPU) provides protocol-driven care for patients with acute chest pain. The Effectiveness and Safety of Chest pain Assessment to Prevent Emergency admissions (ESCAPE) trial evaluated the effect of implementing CPU care at a variety of NHS hospitals.1 Whereas a previous single-centre trial had shown that CPU care could reduce hospital admissions, reattendances and readmissions,2 the ESCAPE trial showed mixed outcomes, with some evidence of increased admissions, reattendances and readmissions.1 All intervention sites developed the CPU as part of the trial, but the process of CPU implementation and patient throughput varied substantially between sites. This variation was not explained by simple structural differences between hospitals or CPUs, such as staffing and opening hours.3 The authors believe that variation in implementation of CPU care in the intervention sites was a critical factor in the mixed outcomes of the trial, hence exploring the reasons for this variation could shed light on the findings of the trial. The CPU is a complex intervention and its effectiveness is likely to depend on the organisational characteristics of the host institution; factors such as leadership, team working, attitudes to change, and structural and cultural characteristics of the departments involved. Therefore, this qualitative investigation was undertaken to gain a better insight into how these organisational factors may have influenced the activity and development of the CPU service and facilitated trial recruitment. Such insights are important because they may inform future service development or trials of organisational change in emergency medicine.
Methods
Seven hospitals set up CPU care in the ESCAPE trial. Six were involved in this study. Full details of the structures, processes and outcomes of CPUs, including the role of the CPU staff during the ESCAPE trial have previously been published.3 Briefly, each hospital established a CPU protocol involving serial electrocardiogram (ECG) recording and measurement of biochemical cardiac markers over 2–6 h, followed by exercise treadmill testing for selected patients. Those with positive tests were admitted to hospital and those with negative tests were discharged home. Patients were eligible for CPU care if they had chest pain that was potentially, but not definitely, due to acute coronary syndrome after initial clinical assessment, ECG and chest x-ray. Most of the units were based in or adjacent to the Emergency Department and run by Emergency Department staff. Operational hours and staffing varied.3 The number of patients receiving the full CPU protocol over the trial year varied between hospitals from 65 to 537, representing rates ranging from 1 to 7 per 1000 Emergency Department attendances.
For the organisational study reported here a qualitative methodology was used within a case study design treating each hospital as a ‘case’.4
Qualitative data were collected through semistructured interviews at the completion of the trial year at each site by a researcher independent of the ESCAPE trial (see Appendix 1 for the Interview Guide).
The core personnel involved in CPU care at the centres randomised to the intervention were Emergency Department medical staff, Cardiologists and the Cardiology and Emergency Department nurses staffing the CPU. The sample of 26 participants across all six target sites was comprised of Consultant Cardiologists (n=4), Emergency Consultants (n=6) and nurses (n=16). The trial medical lead and the lead CPU nurse for each site were represented in the sample. Other staff such as the wider Emergency Department staff were not included in this study.
The interviews were transcribed, coded for site and category of participant, and entered into QSR NVIVO V.7 qualitative data analysis software. Template analysis5 was used to organise the data into themes and sub-themes. A coding template was developed using an input-process-output model; a descriptive framework for understanding and predicting group or team effectiveness and a useful structure for understanding the relationships between complex factors affecting team performance.6 7
Relationships between and within individual cases and cases grouped by level of CPU activity, and between and within categories of respondents were explored. Sites are referred to as high, medium or low (patient) throughput sites. This label is simply made on the basis of the number of CPU patients per site during the trial year (low: 65–91; medium: 188–201; high: 484–537)3 and does not imply any judgement of quality or effectiveness of the service.
Results
Examples of the key themes are presented here under the input, process and output headings (table 1).
Main themes of the organisational analysis
Inputs - organisational readiness
It was apparent that the sites were at different stages of ‘readiness’ but all expressed a view that the management of acute chest pain, and low-risk chest pain particularly, needed to be addressed and saw this as an ideal opportunity to meet an as yet unmet need.
‘…we were aware that most of our chest pains were being admitted even though many of them didn’t need to be admitted and we needed to come up with a system… to screen and rule in or out in A&E depending on their risk, but we didn't have any money' (Emergency Consultant - medium throughput site).
This quote is typical and suggests the notion of an ‘opportunity in waiting’ reflected across most of the sites. However, the data suggest differences in the extent to which the ground was primed for this innovation prior to the trial. Factors such as existing relationships, the reported attitudes to and history of change, and the nature of the existing service were salient.
Most respondents talked of the challenges that change can bring, and there was a common view that much depended on the personal characteristics of those involved.
‘I think it’s polarised between the ones who see change as being potentially useful and those who don't want to change because they're comfortable where they are…and it's one of the difficult things to overcome really' (Emergency Consultant - low throughput site).
Where a positive experience of change was reported this tended to be associated with the quality of the existing relationships, particularly between Cardiology and the Emergency Department, illustrated by two contrasting examples.
‘The relationship with Cardiology is very good; we’re practically handcuffed to them… it works very well, there's no stress on that relationship, that's all very positive' (Emergency Consultant - high throughput site).
‘I wouldn’t say it's very close… it's not a close relationship, not at all. Cardiology tend to keep themselves to one side and as does A&E…there isn't a lot of overlap between the two' (Chest Pain Nurse - low throughput site).
Inputs - team characteristics
The idea that this intervention required the involvement of a team that crossed departmental and professional boundaries was shared widely; respondents talked of a ‘team effort’. However, there were differences in how the ‘team’ was viewed and the extent to which the key players saw themselves as engaging as a team. Core members of the team involved included Cardiology and Emergency Medical Lead and the Chest Pain Nurses. In the high throughput sites it was reported that there was clear and active participation from each of the three core groups, in others one or other of these three groups was less involved:
‘I think I facilitated the trial coming here in the first place… but then once the decision was taken to go ahead with it then I had remarkably little to do with it except providing the A&E side of the data’ (Emergency Consultant – medium throughput site).
Views were expressed about the characteristics of team members. The Chest Pain Nurses were from a Cardiology setting in all but one site and would be expected to have an appropriate range of skills and experience for the role. However, there was a view that this skill and experience needs to be accompanied by additional qualities to be effective in the Emergency Department.
‘… you can’t just take a coronary care nurse and put them into this kind of role though. Whilst the background is important you need to add skills to that experience that they won't necessarily have because they are working in quite a different environment in an A&E department.' (Cardiology Consultant - high throughput site)
A factor influencing the level of activity of the CPU may have been the enthusiasm and commitment of team members.
‘(The chest pain nurse) is the engine, and if you took them out of it…there are people to do that role, and that works okay, but I think (they do) a lot more than just seeing the patients, and I think they’d be very hard to replace…as long as you had someone to replace them who had a similar drive and commitment, that would be fine, but I wouldn't underestimate their importance; it works as well as it does because of them' (Emergency Consultant - high throughput site).
Although there was little evidence of the Chest Pain team becoming integrated into the wider Emergency Department team in which most of them work, in the high throughput sites the data suggest a blurring of the boundaries to some extent and a very close working relationship with the Emergency Department staff while still retaining a very clear and separate identity within the service.
‘We still see ourselves as cardiology, but working within A&E… we get support from a lot of the A&E staff…a lot are very supportive of the role’ (Chest Pain Nurse – high throughput site).
By contrast in one of the sites the Chest Pain Nurses felt isolated as a team and very separate from the Emergency Department within which they work.
‘We’re stuck on the outskirts in a sense waiting for them to beckon us in…when the need arises' (Chest Pain Nurse - low throughput site).
This isolation was not necessarily accompanied by tension, rather there was simply very fixed lines of demarcation; this is your work this is my work.
Inputs - role boundaries
There was a perception among respondents of significant tensions arising from role boundaries between the CPU staff and Emergency Department staff, junior ward medical staff, and ECG staff. Crossing boundaries was experienced as undergoing something of a ‘right of passage’ by some staff and several of the nurses talked in terms of having to prove themselves before being accepted.
‘When we first went down to A&E…it was quite intimidating and we did come across quite a lot of friction, but when we got established and proved our worth then that was fine, they miss us when we're not there’ (Chest Pain Nurse - medium throughput site).
Prior expectations of role based on profession and speciality were reported as creating tensions that were keenly felt:
‘… there was resistance…at a local level with some of the A&E (medical) staff and the nursing staff who regarded these people as interlopers and weren't giving them the clinical support’ (Emergency Consultant - medium throughput site).
‘One barrier is being accepted in A&E…a major barrier…it's a power thing, they don't want someone else going…don't come in telling us what to do’ (Chest Pain Nurse - medium throughput site).
‘Clinical investigations felt that it wasn't our job…we shouldn't be doing ETTs…’ (Chest Pain Nurse - medium throughput site).
Working through these difficulties delayed the establishment of the CPU in some cases, which would have impacted on recruitment. Such tensions have been reported before in new or specialist roles and may have resulted from a perception of threat or from lack of clarity of the role, particularly where there is ambiguity over lines of responsibility; for example
‘…one of the areas where the protocol fell down was that you never knew quite who was referring to the nurses…often you'd get this conflict about who's responsibility the patient is…the patient will think why am I being examined by two different people and it does create a problem’ (Emergency Consultant - medium throughput site).
Process - leadership
A simple but often cited continuum8 can be used to characterise leadership in terms of the degree of participation in decision making by the workforce, from ‘manager centred’ at one end to ‘subordinate centred’ at the other. There were examples from both ends of this spectrum here. Interestingly, the two sites with the highest levels of activity cited examples that seem to place them at opposite ends of this continuum.
The style of leadership may have some impact on the degree of autonomy of the CPU staff. Although all respondents talked in terms of high levels of Chest Pain Nurse autonomy, there were differences in decision making latitude that in some cases seemed to reflect the prevailing model of leadership. An example of this is differences in the restrictions on decision making for discharge.
‘You needed a senior doctor, you needed a consultant to discharge the patient, and they were quite adamant about that even with the trial protocol they wanted a senior doctor to discharge’ (Chest Pain Nurse - high throughput site).
By contrast
‘Once the patients are referred to them they really take ownership of them, and they are extremely effective in terms of running them through the protocol, correlating the results, doing the discharge themselves…’ (Emergency Consultant - high throughput site).
As avoiding admission and reducing bed occupancy were important outcomes for CPU care this may have been influential.
There were also examples of leaders crossing departmental boundaries to establish relationships and to facilitate a team approach at some sites.
‘I think being involved in the discussions and the education that goes on down there is very important so that the A&E staff buy into your vision…if you want to develop something in A&E you have to get down to A&E in order to win hearts and minds’ (Cardiology Consultant - medium throughput site).
Such ‘boundary spanning’ leadership activity may have contributed to establishing the intervention.
Process - operational delivery
Key areas of difficulty for some were recruitment, getting timely biochemical tests, and discharge.
‘The most difficult issue…was with the junior doctors getting around to discharging patients quickly and I think we had big issues in getting people out at the point where they could have gone, I think if the junior doctors had complied then we could have cut stay times massively, at weekends and out of hours our discharge rate was not what it could have been’ (Emergency Consultant - high throughput site).
Recruitment to the trial occurred almost exclusively during the time that the Chest Pain Nurses were on duty and was, in the words of one respondent, very ‘presence driven’:
‘when we weren’t here nobody took it on…we're not here all the time, and when we're on annual leave no-one covers, patients that could have gone to the Chest Pain Unit were trotted across to the MAU for overnight stay’ (Chest Pain Nurse – low throughput site).
‘…we’ve probably missed a lot of patients, people are just not aware enough of ESCAPE and think about it enough I think. And I don't think that's missing people in sending them home, I think that's missing people as in referring them on to the inpatient teams' (Chest Pain Nurse – low throughput site).
The comments from both respondents suggest that the trial struggled to become embedded in the service and, to some extent, lacked visibility.
Process - continuity of staffing
During the trial there were a number of personnel changes that may have had an impact on CPU activity. Two sites particularly experienced the loss of key staff during the trial and reported general difficulties with staffing in the Emergency Department. These were also the two sites that saw the fewest patients. In addition, more widespread staffing difficulties were reported:
‘At that time, when this trial was started we didn't have all our permanent staff with us, so we were basically running on the locums’ (Emergency Consultant - low throughput site).
Output - role expansion and service development
Most felt that there was what one respondent referred to as ‘added value’ from being an intervention site for the trial.
‘I think it's succeeded more than we could ever have expected, it hasn't just succeeded in doing what the trial set out to do…it's actually allowed us to have a proper integrated chest pain service’ (Emergency Consultant - medium throughput site).
At a number of sites, the CPU staff had become increasingly involved in the assessment and management of both intermediate and high-risk patients and respondents gave numerous examples of how they felt they were impacting on the care of all those presenting with chest pain or associated cardiac problems.
‘…with the patients identified at high risk of intervention, they have been very useful in making sure we know about those…whereas in the past nothing much happened, a patient would sit on MAU receiving suboptimal care by today's standards and would then land on the cardiology ward 4 days later…so the chest pain nurses have been instrumental in making sure those patients come to our attention within 24 hours’ (Cardiology Consultant - medium throughput site).
Other areas where respondents felt that participation in the trial had led to wider benefits included the transfer of knowledge to Emergency Department staff, facilitating the care of acute heart failure patients, and establishing nurse-led arrhythmia clinics
The data also suggest that some felt the trial facilitated boundary spanning and that this may have a positive impact on service development.
‘It's one more area that draws us together as departments because the departments get used to working together in these cross boundary services, it's not just one area hanging onto another but we actually have to interact closely to operate successfully so that's a benefit.’ (Emergency Consultant - high throughput site).
Outputs - impact on team members
For some there was a tension between wanting to deliver an effective service and the personal cost that this may require. A number of the respondents felt that they had made sacrifices to ensure the service developed.
‘(We) both have taken a pay cut to do this job… so we have lost heaps of money to do it and what made the difference for us is the pair of us know how well this works’ (Chest Pain Nurse - medium throughput site).
The extent to which this perceived sacrifice was offset by leadership support may have been influential.
‘The key thing for me is to create pride and ownership in the people doing the job, because there is no other reward you can influence, I have no control for example over pay, what you can have some influence over is empowering them so that they get satisfaction from seeing that they make a difference’ (Cardiology Consultant - medium throughput site).
Discussion
The factors that determine the success of a complex service are difficult to measure in the context of a randomised controlled trial.9 This study has identified a number of characteristics that may explain the variation in activity seen in hospitals participating in the ESCAPE trial and are summarised here as characteristics most likely to support the activity of the CPU during the trial. These are largely recognisable from the wider literature as those characteristics supporting change and innovation.10–12
The concept of ‘readiness’ is one that appears across the change literature and refers to the extent to which an individual or system is willing and ready for change and can be considered from the point of view of general or a specific change.10 Here, it appears that the higher activity sites may have been primed for this innovation by previous projects or developments. Organisations that have a positive history of innovation are also more likely to be successful at change in future projects.10 The literature gives possible explanations for the perceived resistance to change reported by some respondents, which include poor dissemination of information, threats to power, seeing change as imposed, competing tasks, and lack of relevance to the individual.11 12 Limitations in the study, in particular not interviewing a wider range of staff from the relevant departments, limits conclusions about why particular sites gave positive change histories, but of the factors that emerge from these data the literature would suggest that the reported effective leadership and staff involvement in some sites are important factors.10 13
Service development is more likely when leadership is present and is appropriate to the setting.13 In the two highest activity sites it is likely that the key leadership behaviour was at opposite ends of the leadership spectrum, yet both appeared to be effective. It is likely that at the time of the trial the prevailing styles of leadership in the two sites had been present for some time, were expected by the respective teams and were effective within their own setting.
An aspect of leadership behaviour that may have been important is ‘boundary spanning’. Boundary spanning is behaviour that crosses traditional organisational boundaries14 15 and is associated with successful problem solving.
Positive tension-free relationships between key staff and departments were more apparent in the more active sites, although not exclusively. Where tensions had arisen during the trial they were generally associated with role boundaries between staff. Role boundaries, particularly those experienced by nurses in specialist roles have been widely reported and there are many similarities between this literature and the experiences of some of the staff here. In the Exploring New Roles in Practice (ENRiP) study,16 17 Read et al explored new roles in nursing and allied health professions and found that conflicts were more likely to arise where there was a lack of clarity over roles and responsibilities.
A sense of shared ownership and partnership between Emergency Medicine and Cardiology was clearly seen in the two most active sites and one of the medium-activity sites. In the two low-activity sites the trial was more exclusively hosted and managed by the Emergency Department with varied degrees of support from Cardiology, but there was less of a sense of partnership. The concepts of partnership and of shared vision are seen as important to successful team working7 and innovation.18
A feature of both the high-activity and medium-activity sites was the continued presence of one person who was very active in maintaining and promoting the trial. Such persons are referred to as ‘champions’19 in the innovation literature and their presence is thought to be extremely influential.20 However, where this drive and enthusiasm is present there may be a danger of over-reliance on an individual who may not be able to sustain high levels of commitment without cost.
Continuity of staff is reported as being influential in successful innovation, particularly continuity of leadership.21 The two low-activity sites had continuity problems in that key members of staff either left or were away for an extended period during the trial. In addition these two sites reported recruitment problems for doctors at one site and nursing staff at the other. Continuity was not a problem at the high-activity sites and they may have benefited from this.7
Limitations of study
First, the data are cross-sectional rather than longitudinal. This allows only a snapshot into the experience of taking part in the trial, and respondents may reflect differently on the experience at the end of the trial from the beginning or the middle. The second is that data were collected only from those closely involved with the trial. It would have been interesting to hear the voice of those who interacted with the trial but who were not part of it. The restriction to the sample arises from resource limitations that necessitated focusing the data collection on those who would have the greatest insight into the factors that facilitated or limited the CPU development. Finally, the interviewer had in the past worked in a similar field and this may have been known to some respondents; how this may have influenced responses is unknown. Despite these limitations, the authors feel that it is important to report these findings.
Conclusion
A range of characteristics may support the establishment of complex interventions such as the CPU and these reflect the wider innovation literature. A number of organisational factors were identified that could have impacted on the outcomes of the ESCAPE trial through, for example, delays in discharge and laboratory tests, and missed recruitment opportunities. Some reported significant barriers arising from role boundaries; the extent to which these tensions impacted upon the trial is difficult to identify although it seems to have been partly mitigated where the relationship between the senior A&E and cardiology staff was well established and where there was supportive leadership. Respondents at most sites felt that being an intervention site for the trial had produced positive outcomes for service development, which is an interesting observation and suggests that the trial may have acted as a trigger for improvement. Some of the issues reported here are predictable and could be considered prior to the establishment of the service. For example, crossing professional and organisational boundaries may lead to tensions between staff and may prove difficult unless there is active participation and support from all relevant groups. What is clear from the results of the ESCAPE trial is that findings from single-centre studies of complex interventions should be treated with caution before a decision is taken to implement in a new setting. Complex interventions are prone to the effects of local organisational issues, as shown in this study, and these findings may be particularly relevant to those instigating or supporting such innovations and for staff establishing a role in a new setting.
Acknowledgments
We would like to thank the staff at the participating hospitals for their help and contribution.
Appendix
Interview Guide
Reflecting on the current service
What are your roles and responsibilities with regard to acute chest pain services?
How has the current service changed over the past year?
What are the strengths and weaknesses of the current service?
What future developments are planned?
What are the key challenges for the development of chest pain services in this setting?
Exploring roles and boundaries
Who has the greatest influence on the development of chest pain services?
Have there been any changes to roles in the development of the chest pain service; if so in what way?
How much autonomy do the Chest Pain Nurses have?
Managing change
What was the most recent significant change to the service?
What or who triggered this change and how was it introduced?
What factors facilitated this change?
What factors hindered this change?
Leadership and decision making
How would you describe the dominant management and leadership style in the service?
How would disagreements or conflicts normally be managed between those involved in delivering the chest pain service?
What is the relationship between the chest pain services and other parts of the hospital; for example, the medical wards, laboratories, outpatients?
Footnotes
Funding The escape trial was funded by NHS Service Delivery and Organisation R&D Programme (reference SDO/41/2003).
Competing interests None.
Ethics approval This study was conducted with the approval of the Thames Valley REC (REC reference 05/MRE12/11).
Provenance and peer review Not commissioned; externally peer reviewed.
Disclaimer The views and opinions expressed in this article are those of the authors and do not necessarily reflect those of the NIHR SDO programme or the Department of Health.
Linked Articles
- Primary survey