Journal of the New Zealand Medical Association, 02-March-2007, Vol 120 No 1250
Improving care to stroke patients: adding an acute stroke unit helps
Carl Hanger, Valerie Fletcher, John Fink, Andrew Sidwell, Anne Roche
Patients with stroke have better outcomes if they are admitted to a stroke unit (SU);1–3 they have greater chance of being alive and independent and less likely to require long term institutional care at one year.1 Guidelines for the management of stroke advocate for the establishment of stroke units as well as organised stroke services throughout New Zealand.4
Christchurch Hospital is the acute hospital for Christchurch and surrounding North Canterbury and serves a catchment population of approximately 450,000 people.5 It is a university teaching hospital and has the regional neurosurgical unit. Approximately 800 people are admitted each year with a diagnosis of an acute stroke.
In Christchurch, a stroke rehabilitation unit (SRU) for older patients was opened in 2001 with documented benefits both for patients and the District Health Board (DHB).6,7
The Christchurch SRU was always envisaged as being one part of a larger integrated stroke service, with an acute stroke unit (ASU) and a community based specialist rehabilitation team planned from the outset.8
As the SRU is on a separate, geographically distant campus from the acute hospital, a combined acute and rehabilitation unit was deemed not feasible. Instead an ASU was established on the acute hospital site in October 2004 with an emphasis on making the transitions to other parts of stroke care as seamless as possible.
Prior to the establishment of the ASU, patients admitted with an acute stroke to Christchurch Hospital were treated on any one of six different medical wards or one neurology ward. These patients were under the care of general physicians (12 teams, with 2 teams admitting on any given day) or neurologists in approximately 80:20 ratio. The decision as to whether the admitting team was neurologist or general physician was based predominantly on an age cut-off (<65 years) but also comorbidities and presence of neurological complications. There were no formal protocols for the management of stroke used consistently across all areas, with the exception of thrombolysis.
Thrombolysis for selected stroke patients started in April 2002.9 These patients were admitted to the neurology high dependency area for post thrombolysis monitoring and care. A Clinical Nurse Specialist in Stroke (CNSS) position was established in 2002. Referrals to allied health professionals (AHPs) were made on individual basis by each clinical team.
Patients requiring further inpatient rehabilitation were transferred either to the SRU, or to the Brain Injury Rehabilitation Service (BIRS) for patients younger than 65 years. Both of these rehabilitation services are 5 kilometres away (in Cashmere and Burwood respectively) from the acute hospital (Central Christchurch city), but on different campuses. Outpatient rehabilitation services were also provided from these distant sites.
The ASU was established on 4 October 2004. This is a 15-bed unit embedded in one of the general medical wards (30 beds total). The ASU aims to take all stroke patients, irrespective of age, gender, and stroke severity.10 The patients are admitted under a general physician or neurologist as before, with the exception that only one (of the two) general medical teams on take each day admitted strokes—this reduced the number of general medical teams involved to six.
To develop and maintain consistency of care between the many different treating teams, and to across different geographic sites within the city, common protocols (e.g. hypertension, use of urinary catheters, early mobilisation), educational strategies, and documentation were developed. These were developed jointly by members of the multidisciplinary teams (MDT) in the ASU, SRU, and BIRS.
Thrombolysis for selected patients continues, but is now provided in the ASU. All stroke patients are admitted using a mutually agreed proforma which acts as a prompt, but is not a formal clinical pathway.11 The aim is to minimise duplication of information collection, whilst ensuring important information is not omitted.
Pre-stroke functioning is recorded only once in a joint “Life before Stroke” section and each professional group of the MDT supplements this with their own initial assessment proforma. Following these initial assessments, all of the MDT record ongoing progress in the same part of the notes. When a patient transfers to the SRU, this common documentation is continued as a contiguous, single set of notes.
Staff numbers were increased for AHP (physiotherapy by 1 full-time equivalent [FTE] position, occupational therapy by 1 FTE, social worker by 0.5 FTE, and speech language therapist by 0.6 FTE). Nursing or medical staff were not increased. The CNSS position predated the ASU, but it was always envisaged as an integral part of its establishment.
Each discipline is expected to see every patient within one working (Monday–Friday) day, thus negating the need for written referrals. Whilst acute stroke treatments are an important component of the ASU work, the ethos is also to provide consistent, skilled nursing and allied health professional care and to begin the rehabilitation and education processes early12—in line with the 5 key components of stroke units (SU).13
This emphasis on SU care has greater potential community benefit than a pure medical model.14 An interdisciplinary approach for each patient is fostered with regular team meetings. To facilitate nursing involvement in rehabilitation, therapy is performed on the ward where possible.
All staff were given the opportunity to update their stroke skills. In particular, nurses were trained to screen for dysphagia and are involved in a programme of continuing professional development coordinated by the CNSS. Close linkages with the SRU and BIRS were emphasised, and the stroke specific documentation is common to all three units.
Following the introduction of the ASU, we wished to assess whether it gave additional benefit over and above the gains accrued by the SRU.6,7 Benefit may be assessed in two main ways. One is to look at overall patient outcomes such as function, domicile, and survival, whereas the other is to look at consistency of key clinical care processes.15
As consistency of care (particularly in the acute phase) was thought to be poor, we have chosen to focus on the latter approach in this study. Using an internationally recognised stroke audit tool16, this paper addresses the question—“Does the addition of an ASU give improvements in stroke care over and above the benefits already accrued by a SRU?”
The Royal College of Physicians of London stroke audit tool (RCPLSA) was used to assess the process stroke care.16 This was designed for retrospective case note review and has been shown to have high inter-rater reliability.17 It has two main sections: (1) casemix indicators and (2) 48 clinical process audit questions covering 12 broad areas of care.
As the last section covered five aspects of care after hospital discharge, which are not recorded in the hospital clinical record, we did not collect data on these five areas. The response to each standard is recorded as either Y (meets standard), N (does not meet standard), or not appropriate. Criteria for the latter category are tightly defined. Overall compliance (%) for each standard is defined as Y*100/(Y+N), thus excluding those not appropriate.
Prior to ASU implementation, function was only assessed in the two rehabilitation hospitals (using the Functional Independence Measure [FIM]). Since 4 October 2004, functional abilities are also routinely measured at CH using the FIM. As the two questions on function in the casemix section of the audit require a Barthel Index (BI), and not the FIM, these two questions were also omitted.
Discharge coding data (ICD-10 codes I61-I64, I67.2, and I67.5-9) were used to identify both the “before” and “after” cohorts. All patients admitted to Christchurch Hospital (CH) between 1 December 2003 and 29 February 2004, formed the “before” group. The “after” group consisted of a selection of patients (two-thirds selection, by omitting every third patient) admitted to CH between 1 February 2005 and 31 March 2005. Both groups included some patients transferred to Princess Margaret Hospital (PMH). As the audit aimed to assess stroke care throughout all three hospitals, admission to ASU was not a prerequisite in the “after” cohort.
Categorical and continuous casemix variables were compared using Chi-squared (χ2) analysis and Student’s t-test respectively.
The study was approved by the Upper South A Regional Ethics Committee (URA/05/02/004).
During the 12-month period from 4 October 2004—3 March 2005, 648 patients with an acute stroke were admitted to the ASU, with a mean (median) length of stay (LOS) of 8 (6) days. 305/648 (47.0%) were transferred to PMH or BIRS for further inpatient rehabilitation. During the same period, 735 patients were discharged from CH with a diagnosis of acute stroke. Thus approximately 88% of all strokes were admitted to ASU.
Casemix variables for the two cohorts that were audited using RCPLSA are shown in Table 1. Patients in the post-ASU cohort were significantly more likely to spend the majority of their hospital stay in a SU and less likely to have multiple ward transfers.
This group had a higher proportion of women, and had more disabling strokes as indicated by a lower level of consciousness on admission, and a worse continence status at 1 week. There were no differences in discharge destinations between the cohorts.
The results from the audit of process of care variables are shown in Table 2. Significant improvements in process of care were shown for the post-ASU cohort (compared to pre-ASU) in 27 of the 43 areas recorded.
Table 1. Casemix of the two audited groups
This audit has shown that important PoC for stroke patients improved following the introduction of an ASU. These benefits were additional to those already gained by having a SRU.6,7 These results also indicate that changes made at the “front end” of stroke care can improve a wide range of processes throughout the whole inpatient stay. The gains were not limited to just one professional group’s work but were across all disciplines.
By using shared protocols and documentation, together with close linkages between the units, this ASU complements the existing SRU and goes some way towards the goal of an overall coordinated stroke service.4,18
Previous studies from both New Zealand and overseas have shown significant deficits in stroke care.19–22, 37 with inadequate assessment or treatment in acute phase, and poor attention to secondary prevention strategies on discharge. More recent studies have shown improvements in care processes,23 clinician attitudes,24 and (in the last 5 years) development of SUs in some District Health Boards (DHBs).6,23,25,26
This paper shows that with the development of an overall inpatient stroke service (with most patients under the care of a stroke specialist team), these PoC can be further improved.
Whilst the initial assessments have greatly improved (with exception of visual attention), further improvements are still required. There was a documented plan to manage hypertension in the longer term in only 75%, despite recent studies and guidelines advocating more aggressive blood pressure management 4,27. Others have noted similar deficits in hypertension plans, although overaggressive management in some older frail patients may cause postural hypotension later.23
Whilst screening for swallowing difficulties improved between the cohorts, only ¾ of the after sample were screened within 24 hours of admission. Some of those who were not screened were not admitted to ASU. Thus we need to have alternative options to assess swallowing when patients are cared for in non-SU settings. Some processes for important, yet non-life-threatening consequences of stroke (continence, cognition, mood, and carer needs) have improved, but there remains considerable room for further improvement. Attention in these areas may improve quality of life for stroke patients.
The gains above have been achieved without extending length of stay (LOS) in hospital. The mean LOS in this study is similar, or lower, than comparable studies.15,22,23,28–30 This is in keeping with international SU literature, where better quality outcomes are achieved with similar LOS.1 The return home rate is also comparable or better.23,25,31
Our reported 30-day mortality (24–25%) appears high compared to New Zealand studies, but these have reported in-hospital mortality only.15,22,23,25 Our 30-day figures are comparable to those reported in the Auckland Stroke Study.32
There is not uniform agreement whether an improvement in documented PoC translates into better patient outcomes. McNaughton et al 22 found that there is a relationship between PoC and outcomes, but that this relationship is weak and complex. We agree with their comments on complexity. However their conclusions on the strength of relationship were derived from audits using the older RCPL stroke audit tool.20This was heavily biased towards the medical management of stroke.
The newer version (used in this study)16 is much broader and looks at many multidisciplinary processes over 12 key clinical domains. Each process was chosen because it was sensitive to variations in quality of care; was thought to be relevant to clinical outcomes; and (where possible) were based on the UK guidelines for stroke.33
Kwan11 found that improvements in quality of documentation and PoC were associated with fewer complications. In two recent studies, greater adherence to quality PoCs were associated with improved patient outcomes (reduced complications or mortality, and trend to improved independence at home).34,35 Similarly better performance on process measures was strongly associated with better survival amongst community dwelling older adults.36
Some secondary prevention PoC (such as antiplatelet therapy and blood pressure management) have long-term beneficial outcomes,27 but these are hidden when measuring short term functional or domicile outcomes. Thus we believe that the improved PoC demonstrated here will have a very positive longer term impact on the quality and consistency of care given to our stroke patients.
The development of an ASU at Christchurch Hospital occurred within a larger context, with a plan to develop an overall stroke service.4 A pilot community-based stroke rehabilitation team commenced in March 2006, and is the third (and equally important) component of such a service.
ASUs should not occur in isolation and need close links with rehabilitation, community, vascular surgical, and general medical services. This will ensure the patient gets the appropriate care at the right time and transitions through the various stages of their illness with minimal disruption.1,18
What has made the difference? Was it just better documentation? This may be a partial answer as each discipline has initial assessment proforma, which may have prompted better recording. Consistency of care given by staff (with an interest and expertise in stroke) in a SU is another critical component.
During the planning and implementation of the ASU, considerable emphasis was placed on developing all of the key elements of a successful SU.13,37 These elements include acute assessment procedures; early management policies (not just medical treatment); rehabilitation (including coordination of multidisciplinary team (MDT) care, regular carer involvement, and MDT meetings); and building staff expertise with regular education and training.
Patients with stroke were previously managed on many different wards, with varying degrees of expertise and interest. Our inclusive admission policy has increased the numbers of patients who were cared for in a SU environment (RCPLSA data 69%, but approximately 88% from discharge coding data).
During the implementation phase there was some pressure to develop a clinical care pathway for patients with acute stroke. This was resisted, as evidence for a strict pathway is equivocal.11,30,38,39 The team also resisted the concept of a mobile team of stroke experts—as this has been shown to be an inferior model to a geographically distinct SU.3
Instead, our team focused on a geographically distinct SU concept together with development of a stroke proforma, shared documentation (and protocols) across disciplines and units, and ensuring that all the key elements of a SU were developed.12,13,35 Such an approach allows the staff to use their expertise and experience to make judgements, without the tight confines of a care pathway.
Using a historical control group in a before and after study has limitations. Many different processes may have changed, rather than just the intervention being tested. In this case, the intervention is not just the setting up of an ASU, but also the change in documentation and ethos of all stroke care. The magnitude of the changes found, suggests that the ASU and associated changes in PoC have made a significant impact over and above other background changes.
Whilst the audits were done at a similar time of year for both cohorts, the pre-ASU audit did include the Christmas holiday break, which could potentially have reduced therapist availability for this cohort. However any negative effect from this is likely to be mitigated by the larger number included in this cohort (N=119). The assumption that what is documented reflects the PoC delivered has already been discussed earlier.
It is disappointing that we have not been able to show improved patient outcomes in this study. Like the profitability of any new business, we expect these improved PoCs to translate into better outcomes over time. However, the finding of similar clinical outcomes despite the post-ASU cohort having more severe strokes (and hence more challenging to rehabilitate and successfully discharge) is encouraging.
The reason for this change in casemix is not clear. Whilst the sampling was non random, sample sizes were considerably larger than those used for each trust in the UK National Sentinel Audit.33 A trend to increased ambulatory care for patients with TIA or minor stroke may be another factor with primary care access to DHB funded CT scanning for stroke now available in Canterbury. It is feasible that subtle drowsiness is better recognised in the ASU and hence lower levels of consciousness are now being recorded. This might account for an apparent worsening in casemix, but a similar explanation would not account for the worsened continence status.
Setting up the ASU was challenging. One problem was the perceived heaviness or dependency of acute stroke patients and hence the ASU was not seen as a desirable place for nurses to work.25 A counter to this was giving nurses greater roles and a chance to specialise, as well as the strong sense of teamwork that developed. Other clinical staff held the view of “Why do we need a SU? We all look after stroke patients and do it well”.
The pre-ASU audit shows that PoCs were not done well and needed to be improved. Furthermore the literature is clear that SU care is superior to general ward care,1,2,4 and patients prefer SUs.40
Changes in service delivery in a large organisation are slow and difficult to achieve. This ASU is no exception to that, taking at least 5 years from conception to fruition.8 Fiscal restrictions were real, but the set-up costs are small compared to the much larger costs of poor stroke outcomes such as institutional care.41.
In summary, adding an ASU to complement an existing SRU gave major improvements in PoC across many different facets of stroke care. We believe this is one step closer to the ideals of an overall coordinated stroke service, as recommended in stroke guidelines, and better care for patients with stroke.
Conflict of interest statement: All authors were involved in the set up of, and/or continue to work in, the Acute Stroke Unit.
Author information: H Carl Hanger1; Valerie Fletcher1,2; John Fink3; Andrew Sidwell1,2; Anne Roche1,2
Acknowledgements: The authors thank all the dedicated staff who have been involved in the establishment and running of the ASU. In particular, we acknowledge the huge efforts of Alison Gallant (Clinical Charge Nurse), Christine Pithie, and Mary Griffith (Clinical Nurse Specialists). Elder Care Canterbury have also played a key role in assisting the reorganisation of stroke care within the Canterbury District Health Board. The authors also acknowledge the useful comments of the anonymous reviewers.
Correspondence: Dr H Carl Hanger, Older Persons Health, The Princess Margaret Hospital, PO Box 800, Christchurch. Fax: (03) 3377823; email: firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals