Article Text
PDF
Abstract
Background Emergency Care Practitioners (ECPs) are operational in the UK in a variety of emergency and urgent care settings. However, there is little evidence of the effectiveness of ECPs within these different settings. The aim of this study was to evaluate the impact of ECPs on patient pathways and care in different emergency care settings.
Methods A pragmatic quasi-experimental multi-site community intervention trial comprising five matched pairs of intervention (ECP) and control services (usual care providers): ambulance, care home, minor injury unit, urgent care centre and GP out-of-hours. The main outcome being assessed was patient disposal pathway following the care episode.
Results 5525 patient episodes (n=2363 intervention and n=3162 control) were included in the study. A significantly greater percentage of patients were discharged by ECPs working in mobile settings such as the ambulance service (percentage diff. 36.7%, 95% CI 30.8% to 42.7%) and care home service (36.8%, 26.7% to 46.8%). In static services such as out-of-hours (−17.9%, −30.8% to −42.7%) and urgent care centres (−11.5%, −18.0% to −5.1%), a significantly greater percentage of patients were discharged by usual care providers.
Conclusions ECPs have a differential impact compared with usual care providers dependent on the operational service settings. Maximal impact occurs when they operate in mobile settings when care is taken to the patient. In these settings ECPs have a broader range of skills than the usual care providers (eg, paramedic), and are targeted to specific clinical groups who can benefit from alternative pathways of care (such as older people who have fallen).
Trial Registration No ISRCTN22085282 (Controlled trials.com).
- Emergency care practitioners (ECPs)
- extended roles
- effectiveness
- prehospital
- emergency care systems
- admission avoidance
- nursing
- paramedics
Statistics from Altmetric.com
- Emergency care practitioners (ECPs)
- extended roles
- effectiveness
- prehospital
- emergency care systems
- admission avoidance
- nursing
- paramedics
Introduction
Many activities traditionally undertaken by doctors are increasingly being carried out by allied healthcare staff with extended training.1–3 One example of an extended role within this setting is the Emergency Care Practitioner (ECP), created in order to contribute a more appropriate response to patient needs in emergency and urgent care settings and meet the workforce challenges facing the health service.4 5
ECPs are from paramedic and nursing backgrounds and complete a nationally developed core competency programme. The programme includes training in taking a patient history, making a physical examination and ordering relevant investigations and tests, such as x-rays. Under certain conditions ECPs can also administer and supply medication. In cases where further investigation or treatment is required, ECPs have the relevant skills and pathways open to them to refer patients to other health and social care professionals, where appropriate.6
On completion of the core competencies ECPs are able to operate as autonomous practitioners in a range of prehospital, primary and acute care settings (such as emergency ambulance, urgent care centres and primary care out-of-hours). The main focus of the ECP role is to improve the patient experience and pathway of care in these settings, particularly by discharging patients at scene or by referring to the most appropriate care practitioner, reducing unnecessary attendances at emergency departments or avoidable admissions.6 7 However, the relative impact of ECPs operating in different settings on patient pathways and care is unknown.
To facilitate a robust evaluation, a pragmatic quasi-experimental multi-site community intervention trial was designed as the best alternative to a randomised controlled trial. The trial consisted of five different studies integrated within a mixed method framework.8 This paper reports the methods and results of the core study that examined the impact of ECP working on the primary focus of the role to improve patient pathways and care.
Methods
The study received ethical approval through Scottish MREC (06/MRE00/20).
Participating services
The trial compared five intervention services employing ECPs to deliver emergency and unscheduled care with five matched control services employing usual care providers to deliver care. All NHS trusts employing ECPs in England and Scotland were invited to participate. Of those trusts responding, potential ‘intervention’ trust sites were selected on the basis of their heterogeneity of service delivery of ECP care. ‘Control’ trust sites that did not employ ECPs, but were in close geographical proximity (ie, within the same or in a neighbouring county) and which offered the same service configurations as the intervention trusts, were then selected. Following a process to map services between intervention and control trust sites, five intervention and five control services (matched pairs) were selected for inclusion in the trial (table 1). Exact matches in terms of service configurations were made for four of the five pairs. In the remaining pair (pair 3) there was no urgent care centre in close geographical proximity and the best match was a nurse-led emergency department.
Pairs of services included in the study
Participants
Informed consent was obtained from all study participants prior to recruitment. Within each pair of services all patients presenting with emergency or urgent complaints that were eligible to be seen by ECPs and presented to either the intervention or the control services between May 2006 and August 2007 were included in the trial. The patients ECPs were eligible to see were determined by the setting in which they operated and local protocols developed by individual services. Additional information on eligible patients in each study pair is contained in the main study report.8
Data collection
Details of patient management at the time of the acute health episode and disposal pathway immediately after the episode were collected from either electronic or paper clinical records. These data were extracted by the research team from clinical records anonymised and forwarded to them by staff in each site. The details included: mode of presentation (eg, ambulance, self present, primary care referral); incident location (eg, home, public place); presenting complaint; investigations performed (eg, electrocardiograph, radiology, urine and laboratory tests); treatments given (eg, analgesia, antibiotics, wound management, application of dressings, written advice); disposition (admitted, discharged, referred to primary care or other outpatient services), discharge diagnosis and episode time.
Data management
Databases were designed by the research team for storing all data relating to the acute health episode using Access 2000 software. On completion of coding and cleaning of the database the data were imported into the SPSS Version 14.0 statistical software package and then into STATA Version 9.0 for further analyses.
Sample size
The study was designed to evaluate the impact of ECP schemes for patients, professionals and services. There is no single primary outcome by which the ‘success’ of the new role could be determined; rather this was a multi-dimensional assessment using several quantitative patient indicators. Nevertheless, if major changes existed in patient care and disposal pathway the study design was sufficiently sensitive to detect them. Consequently, the trial had a recruitment target of n=600 intervention and n=600 control patients in each of the participating pairs of trust sites. Within each site, this gave 90% power at α=0.01 to detect effect sizes of 0.3SD and adjusting for case-mix differences in potential confounders such as age and sex. A type 1 error rate of 0.01 was selected to allow testing approximately five outcomes at the 5% significance level.
Data analysis
Primary outcomes
Patient disposal pathway
A number of referral options were available to intervention and control services, such as primary care, community care, emergency department and direct hospital admission. These referral options were categorised into three groups (discharged, urgent referral and non-urgent referral) and compared:
The percentage of patients:
discharged following consultation (those who were discharged without either an appointment or advice to see another health professional).
urgently referred to hospital (these were patients who were referred immediately to either the emergency department (ED) or direct admission to a hospital ward)
non-urgently referred to primary or community care (these were patients who were referred to either their general practitioner, a community professional or a hospital outpatient department at a specified time).
Secondary outcomes:
undergoing any investigation at their initial episode
receiving any treatment at their initial episode
episode time. Defined as the time from the patient first contacting the included study service to the end of their contact with that service (recorded in minutes).
All outcome measures were analysed in the same manner. First, the treatment effect within each pair of services was calculated. For the analysis of episode times, a log-transform was applied to the data and the results presented as time ratios (which can be interpreted as the ratio of the median episode time in the ECP compared to control). For all other outcomes the difference in percentages was calculated. The overall effect, averaged across all five pairs, was derived using a random-effects meta-analysis.9 The difference in percentages (or time ratios for episode time) with their corresponding 95% CIs are presented for each outcome. The summary (pooled) treatment effect (with 95% CI), the I2 statistic and the statistical test for heterogeneity are also shown. The test of heterogeneity indicates whether the difference between intervention and control are consistent across the five pairs of services, and high I2 values (near 100%) indicate a high percentage of the overall variance is due to differences among the pairs.
As this was a non-randomised design, further analyses were performed to assess whether any differences in outcomes between intervention and control services could be due to characteristics of the services other than the intervention. The data were re-analysed using analysis of covariance (episode time) or logistic regression (all other outcomes), adjusting for age, gender and presenting complaint (categorised as trauma vs non-trauma). Age was modelled by including a quadratic term (ie, the square of the age) as this was determined to be the best fit following a fractional polynomial regression. For each outcome and each service the unadjusted OR (time ratio for episode time) was calculated and was compared to the OR (time ratio) derived from the relevant regression model to assess whether the findings still held after adjustment for these factors.
Results
A total of 5525 patients were identified across the five pairs of services as being eligible for inclusion in the study (n=2363, 42.8% intervention; n=3162, 57.2% control). Figure 1 describes the trial profile of eligible patients. Routine clinical data were available for all the patients identified, but in some cases these were incomplete. The characteristics and distribution of the samples achieved across the five pairs of services are shown in table 2.
The trial profile of eligible patients.
Baseline patient data by intervention and control within pairs of services and overall
Trial outcomes
Patient outcomes are presented in table 3. This shows the distribution by intervention and control service pair and overall. Significant heterogeneity (p<0.001) was observed for all outcomes and therefore the results focus on differences observed within each pair.
Trial outcomes (paired services)
Primary outcomes
Patient disposal pathways
Overall, the majority of patients were discharged (57.3%, n=3165) with a further 24.4% being referred urgently to hospital (n=1347) and 14.3% having a non-urgent referral (n=789).
Discharged
For each service pair, the probability of being discharged between the intervention and control services is shown in figure 2. Significantly more patients were discharged in the intervention services (ECP) in the ambulance and care home service pairs. In the GP out-of-hours and urgent care centre pairs, significantly more patients were discharged by non-ECP staff in the control services. In the MIU pair, the percentages discharged were similar for the intervention and the control services.
Rates of discharge, urgent and non-urgent referral by trial setting.
Referrals
The proportion of patients urgently referred to hospital or non-urgently referred to primary or community care is also shown in figure 2. Significantly more patients were urgently referred by non-ECP staff in both the ambulance service and care home pairs. In the minor injury unit pair, however, significantly more patients were urgently referred by ECP staff.
In four of the five pairs (ambulance service, GP out-of-hours, urgent care and care home) ECPs made significantly more non-urgent referrals. In the minor injury unit pair significantly more non-urgent referrals were made by control staff.
Secondary outcomes
Investigation
The clinical information received on patient care was used to extract information about which and how many investigations patients received at their initial care episode. Overall, only 13.5% of patients seen in the study (n=745) received an investigation. Of these, n=349 (6.3%) received an x-ray.
The percentage of patients undergoing at least one investigation was significantly higher in the intervention service (ECP) rather than control service (non-ECP) in the ambulance and GP out-of-hours pairs. The percentage of patients undergoing investigation was significantly higher in the control services for the minor injury unit and urgent care centre pairs. In the care home pair there was no statistically significant difference in the percentage of patients undergoing an investigation (table 3).
Treatment
The clinical information received on patient care was used to extract information about treatments patients received at their initial care episode. Overall, 43.6% of patients (n=2408) received some treatment for their complaint, with a further 15.9% (n=876) being given advice alone. The most common treatment was the administration of a prescription.
Those receiving any treatment (excluding advice alone) were distinguished from those receiving no treatment. The percentage of patients receiving a treatment was significantly higher in the intervention service (ECP) in the ambulance and care home pairs. The percentage of patients receiving a treatment was significantly higher in the control services for the urgent care pair. A higher (but non-statistically significant) percentage of control patients received treatment in the minor injury unit pair. In the out-of-hours pair similar numbers of patients received a treatment in intervention and control services (table 3).
Patient time
Total episode time for the initial care episode was recorded from routine data submitted by each site.
The difference in episode time was significant for all pairs of services. In the ambulance and care home pair, the time spent with patients was significantly longer for the intervention services (ECP). In the GP out-of-hours, urgent care centre and minor injury pairs the time was longer for the control services (non-ECP).
Validation
Controlling for possible confounding factors of age, sex and presenting complaint (categorised as trauma vs non-trauma) in a logistic regression model (analysis of covariance for episode time) gave very similar results to the unadjusted results for all study outcomes.
Discussion
Principal findings
This study demonstrates that ECPs are having a differential impact when compared with the usual care provider and this impact varies between the different paired service settings. In the ambulance and care home paired services, patients were significantly more likely to be discharged by an ECP than by usual care providers. Further analysis of patients who were referred showed that significantly more patients were urgently transported either to the ED or to a hospital ward by non-ECP staff in both the ambulance service and care home pairs.
By contrast, in the out-of-hours service and urgent care service pairs, an ECP delivering care is significantly less likely to discharge a patient than the usual care provider and is significantly more likely to refer a patient urgently to either the ED or direct hospital admission.
ECPs in the ambulance and care home pairs were more likely to undertake an investigation and provide a treatment to patients than the usual care provider. Perhaps unsurprisingly, this meant that in both these paired services episode times were significantly longer for the ECP services than the usual care provider services.
Some opposing trends were apparent in the minor injury unit and the urgent care service pairs.
Implications
There are various examples of extended roles for healthcare practitioners currently in the UK NHS. However, the impact of new roles (and the most effective ways of implementing them) on patients are poorly understood. The development and implementation of ECP working represented an ideal example to examine the impact of new roles in healthcare. The ECP role has evolved around the UK within different service settings and within those systems, ECPs may work across multiple boundaries depending on how they have been commissioned locally.10 This flexibility in how ECPs are utilised nationally presented an opportunity to evaluate the role and provide some evidence for the impact they are having on patient pathways and care. The five pairs of services in this study were selected on the basis of their heterogeneity in terms of models of service delivery of ECP care. The fact that the results from the analyses of the patient data are also heterogeneous demonstrates that there are important lessons to be learnt about how to maximise the impact of ECPs.
A differential impact of the ECP was found on patient care and disposal pathways, which appears to be both dependent on the service setting they are operational in and also on the usual care provider that they were replacing. ECPs appear to be having a maximum beneficial impact on the patient journey when they operate in mobile settings (care travelling to the patient) and when they replace usual care providers with a narrower range of skills than themselves (eg, paramedic). This impact is both in discharging significantly more patients and urgently referring significantly fewer patients to the hospital. Where usual care providers have a higher or comparable level of skill (such as a GP, nurse practitioner) the impact of the ECP is lessened. Where the ECP model of working does not contain a mobile element (eg, urgent care centre) their ability to impact on the patient journey is compromised.
Other studies of extended skill roles within the emergency care settings found that patient management was also affected by the role. Previously, ECPs were found to have carried out fewer investigations, provided more treatments and were more likely to discharge patients home, than the usual providers.11 12 Non-conveyance rates among ECPs were also higher than for non-ECP ambulance patients.12 ECPs and paramedic practitioners (a precursor to the ECP role) specifically treating older people were also found to discharge more patients at scene thereby having a significant impact on ED attendances and hospital admission rates.13 14 In the ED setting, nurse practitioners have been shown to be safe and deliver care to patients with minor injuries comparable with junior doctors.15 However, this role is limited in scope and therefore more expensive.16 17 There is an expectation that role substitution will have an impact by reducing costs while maintaining quality of patient care.18 A cost-effectiveness analysis carried out as part of the wider study found that only ECP services that had a mobile element, reduced costs compared with usual care providers as a consequence of a reduction in ED and hospital admissions in patients seen by ECPs.8 Previously, paramedic practitioners assessing and treating older people produced a non-significant cost saving when compared with standard ambulance response, principally because of a significant reduction in ED and hospital admissions.19
Limitations
Models of service delivery were peculiar to each of the five pairs of services participating in this pragmatic clinical trial. Therefore, findings for these services individually may not be generalisable to other similar service settings. However, nationally, the participating services were not considered atypical in how ECP services had developed in some sites and not in others. The ECP and non-ECP services included in this study were recruited purposively to enable comparisons of ECP working with non-ECP working in matched health service and spatial settings to be undertaken in a systematic way. Patient follow-up was limited to initial disposal after the acute trial episode, and therefore the degree to which patients were discharged and then reattended services with the same clinical problem is unknown.
It was not possible to design a randomised study in the context of ECP working because the services were already fully operational. Instead, we had to take a pragmatic approach and sought to examine the impact of ECP working from a ‘whole systems’ perspective available from combining several different research methodologies within the same trial design.7 In respect of the findings of this particular study, although some imbalances were observed between the patient characteristics in the ECP and non-ECP groups in the different settings, the differences observed after adjusting for these factors were very similar to the unadjusted differences.
Summary
The paper has reported the methods and findings of the core study in a multi-perspective trial of ECP working. The study has shown that ECPs can have clear benefits to patients within certain emergency and urgent healthcare settings. However, ECP working is not a panacea and a beneficial impact of the role on patients may not transfer equally across all settings or health professionals. ECP training and skills are intended to equip practitioners to see a selected but limited range of clinical complaints and patient groups. The extended role should be considered a way of supplementing existing roles by releasing doctor and other resource time, (such as unnecessary ambulance journeys, or ED attendances) for more appropriate targeting of these services to patient need.
Future research
Further research could focus on interventions to assist the ECPs in maximising their impact in specific client groups such as older people requiring emergency care, a key area where demand is rising and services are poorly configured to match needs. Additional analysis of these data will be undertaken to investigate the relative impact of ECPs for specific client groups. Further research is also needed to establish the cost-effectiveness of ECPs in comparison to usual care providers.
Acknowledgments
The authors wish to thank the NHS staff who assisted with this study.
References
Footnotes
Funding The study was funded by the United Kingdom National Institute for Health Research (NIHR) programme on Service Delivery and Organisation (SDO). The study funders approved the final protocol and design for the study. However, the views expressed here are those of the authors alone. Funder reference SDO 2005/98.
Competing interests None.
Ethics approval The study received ethical approval through Scottish MREC (06/MRE00/20).
Provenance and peer review Not commissioned; externally peer reviewed.