What does it take to provide clinical interventions with temporal consistency? A qualitative study of London hyperacute stroke units

What is known about this topic?

• Hyperacute stroke units display no variation by day or time of admission in 3-day mortality or disability at hospital discharge, and no variation by day or time of admission in delivery of ‘front-door’ interventions, such as stroke nursing assessment, brain scanning and thrombolysis measures.
• Hyperacute stroke units show significant variations by day and time of admission in other interventions (including timely consultations and assessments with stroke specialists and therapies, and length of hospital stay).

What this analysis adds

• Consistency is achieved by adapting and extending roles, creating continuities between different times of day, building relationships and trust, and prioritisation of ‘front-door’ interventions. However, these adaptations have unintended consequences.
• Variation persists due to inconsistencies in medical, managerial and allied health professional staffing by time of day, delivery of therapist assessments and barriers in repatriation processes.

Introduction

Seven-day provision of consistent, high-quality, urgent and emergency care settings is an international research and policy priority.9,11,29,61,78,79 This has motivated 15 years of attention on the ‘weekend effect’ and other temporal variations,22 scrutinising mortality and other outcomes for those admitted at the weekend compared with those admitted during the week.6,7,9,31,80–83 Varying features that may explain differential outcomes include staffing levels31 and patient mix,25 leading to widespread interest in studying weekend organisational features.22,84

In Chapter 3 we showed that HASU patients’ mortality and disability outcomes are consistent according to the time of their admission. However, there were some variations in the delivery of care. In this chapter, we set out to explore how acute care interventions are delivered with temporal consistency, and why some care interventions are resistant to temporal consistency, by examining the organisation of services at different times of day and week.

Methods

This section is based on a paper by Black and Ramsay et al.85 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

Our methods are reported in accordance with COnsolidated criteria for REporting Qualitative studies (COREQ) reporting guidelines.49

Recruitment and sample

Our recruitment procedures are detailed in Chapter 2, Recruitment and consent. The particular data sample relating to this chapter is described in Tables 2 and 3.

Non-participant observations

We conducted non-participant observations at least four times at each HASU site, two visits on weekdays, one in the evening during the week and one at the weekend, with additional visits to confirm or add to our findings. We collected data on various aspects of HASU activity likely to influence care provision, guided by clinical interventions in our quantitative analysis and our initial observations (Table 9).

Purposive sampling of observations is described in Methods.

We initially structured observations so that availability of staff and key processes (e.g. handover and ward rounds) were covered. Subsequently, we targeted our observations towards our emerging analysis, such as following patients’ journeys from the ED to the HASU ward, and shadowing the nurse in charge and consultants to understand their roles.

We obtained global written consent for observations from service leads and from a selection of staff at team meetings. Thereafter, verbal consent was given by staff members for researchers to make observations of staff meetings and of care provision.

Interviews

Staff were sampled purposively to ensure coverage of a range of professional roles within all eight HASUs, with perspectives on the clinical interventions65 (e.g. interpreting brain scans, thrombolysis, therapy assessments), including medical, nursing, therapy, and administrative or managerial staff (see Chapter 2, Sampling).

Researchers approached staff during the observation work and through existing contacts with study information, stating that participation would be confidential and anonymously presented in publications. Interviews were conducted with fully informed, written consent and in private settings, according to a semistructured topic guide (see Report Supplementary Material 1), and with the aim of understanding temporal variations from the staff perspective, including typical daily activities and attitudes towards working in and out of hours. Interviews lasted between 20 minutes and 1 hour. No field notes were taken, but each interview was discussed with the other team members. Interviews were audio-recorded and transcribed verbatim.

Analysis

Quality and trustworthiness of the analysis were maintained through a select number of joint observations and interviews, to sensitise ourselves to pertinent contextual issues, and regular reflexive discussions between the researchers. Our five-step analysis process is outlined in Chapter 2, Qualitative analysis. For this chapter, we mapped the results against the variations in practice across the 7-day week found in our analysis of the performance data for care provision in London HASUs set out in Chapter 3. In particular, our analysis focused on the distinction between ‘front-door’ clinical interventions (which were provided consistently) and other interventions (which were not).

Results

This section is based on a paper by Black and Ramsay et al.85 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

London HASUs deliver ‘front-door’ clinical interventions, including brain scans, thrombolysis and dysphagia assessment, without significant temporal variation.65 LOS showed significant variation without a clear temporal trend, assessment by a stroke physician within 12 or 24 hours varied by time of day and assessments by therapists varied by the day of the week. Care quality was worse for patients admitted on a Friday.

In the following sections, we present (1) factors that may explain why ‘front-door’ interventions showed temporal consistency, as well as some unintended consequences of the strategies London HASUs employed; and (2) factors that may explain why other interventions and outcomes showed significant temporal variations (Table 10).

Factors influencing temporally consistent care

We found that there was no variation between the HASUs in terms of adherence to the London stroke service standards. The London stroke service standards required that HASUs have 24/7 availability of staff who are trained to assess eligibility for and deliver thrombolysis; this was normally completed by a team of medical and nursing staff who were on standby to immediately attend patients at the ED when alerted to new potential thrombolysis patients by the ED team (Figure 11).

FIGURE 11 Thrombolysis pathway. This figure is adapted from Black and Ramsay et al.85 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure. (Figure not reproduced here; see the original article.)

Thrombolysis was administered by the stroke team when appropriate, after a battery of clinical and imaging tests had been conducted. The time taken to complete this process is one of the core performance metrics by which each treating unit is measured. Each HASU attempted to make this pathway as efficient as possible, by having a designated ‘thrombolysis team’ irrespective of clinical activity elsewhere in the stroke service and through local arrangements with radiology, to guarantee that 100% of stroke patients potentially eligible for thrombolysis were scanned within the next CT scan slot. However, as detailed in Table 11, the staff supporting ‘front-door’ activity reduced both in number and in seniority in the evenings and at the weekends. As a result, the HASU teams made significant adaptations out of hours, including extending roles and responsibilities, and introducing processes to ensure continuity between different times of day.

Adapting and extending roles

At all times of day, stroke specialist registrars and senior house officers (junior doctors) made decisions about thrombolysis with consultant telephone support (for additional data see Report Supplementary Material 2, Theme 1a). However, observations and interviews confirmed that, in some HASUs, at night, the consultant was called for positive confirmation only if the registrar thought a patient eligible for thrombolysis. This was seen as a positive educational opportunity for junior doctors, but it placed greater responsibility on senior nurses to decide whether or not to admit the patient to the HASU in cases when the stroke diagnosis was unclear.

Creating continuities between different times of day

Rapid movement of patients through the hyperacute stroke pathway required quick resolution of problems, and transfer of detailed information between team members who work at different times of the day and week (see Report Supplementary Material 2, Theme 1b). Key mechanisms took the form of handover meetings, MDT meetings and ward rounds:

Bay nurse leads on his patients, going through the discharge sheet that everyone has and focuses on any particular issues that have arisen, or that the next nurse needs to be aware of [. . .]

Other ways of creating continuity included shifting or staggering staff rotas (to bridge gaps between shifts) and extending therapists’ hours into the early evening.

Building relationships and trust

Delivering ‘front-door’ interventions consistently depended on rapid decision-making, an important facilitator of which was the development of strong and trusting relationships with allied disciplines, in particular with the ED and neuroradiology (see Report Supplementary Material 2, Theme 1c).

Hyperacute stroke unit staff worked alongside ED staff to assess the patient and there were often overlaps when a patient’s diagnosis was unclear or when multiple patients presented at once. HASU staff felt that ED clinicians also valued this close relationship:

We are popular with the ED team [. . .] we are one of the few teams where you’ve got a consultant down there sweating away with them.

Hyperacute stroke unit staff also had an important relationship with radiology, with prioritised access to CT scans as outlined in the London stroke service standards. The HASU team often conducted initial interpretation of CT scans, which reduced decision-making delays. HASU staff felt that this self-sufficiency strengthened their relationship with ED and radiology staff.

Prioritisation of ‘front-door’ interventions by staff

Hyperacute stroke unit staff’s enthusiasm was an important facilitator of sustained performance in delivering ‘front-door’ interventions (see Report Supplementary Material 2, Theme 1d). Interviews and observations (e.g. of the urgency with which staff responded to calls to attend the ED) suggested that HASU staff relished the early stages of acute stroke care and the potential to see rapid positive outcomes:

When the stroke happens we have to work fast: run to ED, do everything within four hours to ensure that the patient can be thrombolysed, and I’ve seen the patient like almost dead [. . .] we were able to save the patient’s life because immediately we were able to assess, go for CT, thrombolyse the patient [. . .] to me it’s great work.

Unintended consequences of adaptations

Although extending roles out of hours was seen as an opportunity for staff development, some interviewees felt that the threshold for admission to the HASU was weakened (see Report Supplementary Material 2, Theme 1e). Junior doctors were seen as more risk averse than consultants, thus admitting more patients unnecessarily out of hours and in turn placing greater strain on the service at a time when it is particularly difficult to move patients from the HASU:

Some doctors, they will send them in to you, put the pressure on you to take that patient [. . .] in the morning when the consultant sees the patient, this patient has not had a stroke [. . .] once the patient gets in here, it’s difficult to send the patient back to the wards, that becomes a big problem, they’re here one week, they’re still waiting for medics to take the patient over.

Factors influencing temporally inconsistent care

Variations in outcomes and delivery of clinical interventions in London HASUs are summarised in Table 12. We explain these findings in terms of reductions in both the number and the seniority of medical, managerial and allied health professionals out of hours, and reductions in repatriation options out of hours.

Variations in medical, managerial and allied health professionals by time of day

Variations in delivery of interventions by time of day but not by day of week (see Table 12) are likely to have been influenced by variations in HASU activity and staffing variations that derived from the London stroke service standards (see Report Supplementary Material 2, Theme 2a).

Although some staffing levels were lower at night, the number of patients arriving also reduced, giving patients a better chance of being admitted to the HASU within 4 hours:

We know less people have strokes overnight [. . .] so we know it’s like going to be quieter, but from a staffing perspective, that’s why we’ve done the 24-hour thing, so there is always those amount of staff on.

Patients were assessed during the consultant-led ward round, which the standards required to take place daily and which commonly occurred between 09.00 and 12.00. Undergoing consultant assessment within 12 and 24 hours therefore depended on patients reaching the ward during this period. For example, if a patient arrived at 03.00, their first consultant assessment would be likely to take place during that morning’s ward round (≈6-hour wait). If they arrived at 15.00, it would be likely to occur the following morning (≈18-hour wait):

After the ward round has finished and we’ve tidied up a bit, yes, you’re less likely to come back and see a case, unless it was very urgent or some unusual type thing.

Variations in delivering therapist assessments

Therapists generally worked ‘in-hours’ shifts, so patients arriving at hospital in the morning were unlikely to be assessed until the following morning (see Table 11) after a board round, when the team make daily decisions about individual patients’ care. This meant that patients arriving at hospital between 08.00 and 09.00 were the least likely to be assessed within 24 hours.

The London stroke service standards specified therapy staffing levels that fully covered only 5 days per week and, therefore, HASUs faced a decision on how best to use these limited resources: some chose Monday to Friday because of traditional working patterns on these days (see Report Supplementary Material 2, Theme 2b). The resultant gap in therapist coverage at weekends explained why patients admitted on a Friday were less likely to undergo therapy interventions within 72 hours of arrival (whereas patients admitted at the weekend were more likely to be assessed on the following Monday or Tuesday). Therapists described feeling rushed on a Friday, as they struggled to get through their workload before the weekend:

Yeah, I think it can be very stressful on a Friday, just if patients are going home when there’ve been a few discharges at the same time it can get quite complicated trying to co-ordinate a lot of family members, patients, staff to fill in documentation, social work, making sure a care package has gone in and completing lots of referral forms.

Other HASUs spread their limited therapy resources into the weekend, which was reported to have a beneficial outcome on discharge figures; however, participants reported a change in priorities. Assessment of new patients dominated, and therapeutic work or talking to families was diminished. Furthermore, by spreading therapist resources into the weekend, HASUs reduced therapist capacity during the week. Therapists in almost all HASUs suggested that existing attempts to cover weekend working resulted in reduced prioritisation of therapeutic activity, suggesting that no current staffing model observed permitted consistently sufficient therapy provision.

Factors influencing length of stay

Patients admitted at the weekend in London had a longer hospital stay. This related to a number of factors (see Report Supplementary Material 2, Theme 2c). As patients admitted at weekends were less likely to be seen by therapists, patients could potentially fail to have their rehabilitation and nutritional needs assessed for 3 days in a row:

If you’re nil by mouth on the Friday when you come in, say at half past four [. . .] you could technically be nil by mouth until Monday.

Reduced therapeutic capacity could also delay discharge and, therefore, extend LOS:

[. . .] you think, ‘Well, this person can’t go because they need a physio, and we could have discharged them on the Saturday but they have to wait till Sunday or even Monday, which [. . .] can cause a problem sometimes if we need beds [. . .] and that person’s then spent another day potentially in hospital that they potentially don’t need to.

Hyperacute stroke unit staff suggested that social services, care homes, stroke units (acute rehabilitation units) and community rehabilitation units were significantly less likely to accept new cases at weekends. This restricted the timing of discharges from the HASU, often leading to longer stays. Input from social services and early supported discharge (a service designed to accelerate the discharge home of patients in hospital86) was important in ensuring that patients returned home or to care homes quickly once sufficiently recovered. However, social services had extremely limited weekend operation, which prevented liaison during the weekends to prepare packages of care or transfer patients.

Discussion

This section is based on a paper by Black and Ramsay et al.85 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

To our knowledge, this is the first qualitative study about the organisation of stroke care with respect to temporal variation. This study reports qualitative data that help explain the findings presented in Chapter 3. Consistent provision of clinical interventions was underpinned by (1) junior nursing and medical staff extending their in-hours responsibilities to cover key decision-making roles, such as that of the thrombolysis nurse; (2) intervening to bridge potential gaps caused by shift-working (staggering rotas, holding meetings to share information); and (3) HASU leadership building trust and respect across staff both within HASUs and with key specialties elsewhere in the hospital (such as the ED and neuroradiology). Key issues leading to temporal variation in care provision included reductions in medical, managerial and allied health professions, and significantly reduced options for repatriation to other acute services and community services, at night and at the weekend. Variations resulted in greater pressure on the ward from low thresholds for admission at night, dilution of staff capacity and bottlenecks in repatriation pathways. Some of these effects were mitigated by strategies to create ward space and expedite discharges on a Friday, but these strategies had a number of unintended negative consequences in terms of patient outcomes.

The London stroke service standards were an important influence on the delivery of clinical interventions, whether consistent or inconsistent. When standards required 24/7 availability of staff (e.g. nurses), aspects of care associated with these staff groups tended to be delivered consistently, regardless of time of day or day of week. When standards required that a key activity was conducted on a daily basis (such as the consultant-led ward round), the likelihood of patients undergoing the associated intervention varied significantly according to when they were admitted. Finally, when standards specified staffing levels to cover only 5 days (as with therapies), it was not possible to provide interventions consistently over 7 days, regardless of local adaptations employed.

The strengths of our study are founded on detailed data collection in each London HASU at different times of day, both during the week and at weekends, providing a rich picture of the realities of organising and providing a high-performing acute care system. Observing and comparing how eight sites organised themselves in different ways to meet the same standards affords generalisability to our results. There were several limitations to the study. First, we did not study any hyperacute stroke services operating within a different service model (whether centralised or non-centralised). The lack of a comparator limited our confidence that our findings explain 24/7 care per se (as compared with a centralised model of care). Second, we did not interview all potentially relevant professions within the studied organisations (e.g. pharmacy, emergency medical practitioners). Therefore, our perspective on important working relationships beyond the HASU was based on HASU staff perceptions, although they were supported by our own observations, particularly of ED co-ordination. Finally, these services develop constantly and some aspects of provision, such as staffing levels both within and beyond the HASUs, are likely to have changed.