Delirium screening in a stroke unit by nurses using 4AT: Results from a quality improvement project

Abstract

Aim

To assess the feasibility of delirium screening with the screening tool 4AT conducted by stroke unit nurses.

Design

Observational.

Methods

Patients with confirmed acute stroke admitted to the stroke unit at Bærum Hospital, Norway, from March to October 2020, were consecutively recruited. Nurses performed delirium screening using the rapid screening tool 4AT within 24 h of admission, at discharge and when delirium was suspected, and filled out a questionnaire assessing their experiences with the delirium screening. A geriatrician validated the delirium diagnosis.

Results

In all, 62 patients were included, mean age 73.3 years. 4AT was performed according to protocol in 49 (79.0%) and 39 (62.9%) patients at admission and discharge respectively. Lack of time (40%) was reported as the most common reason for not performing delirium screening. The nurses reported that the felt competent to carry out the 4AT screening, and did not experience it as significant extra workload. Five patients (8%) were diagnosed with delirium. Delirium screening performed by stroke unit nurses seemed feasible and the nurses experienced that 4AT was a useful tool for this purpose.

AIM

Delirium is associated with poor prognosis. Early identification and prevention of delirium may improve outcomes in patients suffering from stroke. Stroke guidelines emphasize that stroke care teams should be aware of delirium. Yet, delirium screening is not part of routine practice in many stroke units. We aimed (1) to assess the feasibility of delirium screening conducted by nurses and (2) to estimate the delirium prevalence among patients admitted to a stroke unit.

BACKGROUND

Previous studies have shown that stroke is associated with an increased risk of delirium (Shi et al., 2012). Early identification and prevention of delirium may improve outcomes in patients suffering from stroke, and guidelines emphasize that stroke care teams should be aware of delirium (The Norwegian Directorate of Health, 2017). Multicomponent non-pharmacological interventions may prevent delirium and reduce the risk of severe delirium in hospitalized patients. Thus, patients suffering stroke should be carefully reviewed for risk and precipitating factors for delirium and be provided supportive interventions with the aim to reduce delirium promoting factors (e.g. dehydration, pain, sleep disturbance). Yet, delirium screening is not part of routine practice in many stroke units. In addition, few studies have addressed delirium in stroke units, and the prevalence of delirium in previous studies varied from 10% to 50% (McManus et al., 2007; Pasinska et al., 2018; Sheng et al., 2006; Zipser et al., 2021).

Delirium is associated with prolonged hospitalization, increased risk of institutionalization, higher mortality and increased risk of subsequent cognitive impairment (McManus et al., 2007; Zipser et al., 2021). Despite these serious consequences that may complicate patient trajectories, delirium often remains undetected. Hence, systematic screening of stroke populations may improve the care and outcomes (Shi et al., 2012; Song et al., 2018). The 4 ‘A’-test (4AT) is a well-proven and widely used delirium screening test. It is developed for a busy clinical setting and is designed to be performed by a range of healthcare practitioners, and previous studies have suggested that 4AT is a reasonable choice for delirium screening in acute stroke (Lees et al., 2013).

This quality study aimed to explore the feasibility of delirium screening with the 4AT performed by stroke unit nurses. To the best of our knowledge, this has not been studied previously. Secondly, we aimed to assess the prevalence of delirium in a Norwegian stroke unit by stroke severity.

MATERIALS AND METHODS Participants

This was a quality study performed at the Stroke Unit of Bærum hospital, a non-university hospital in the capital region of Norway. We consecutively recruited patients from 5 March to 29 October 2020, with the study set on hold from 16 March to 26 April and from 19 July to 14 August due to the COVID-19 pandemic and summer holidays respectively. We invited all patients aged ≥18 years with stroke according to the American Stroke Association definition, defined as a neurological deficit attributed to an acute focal injury of the central nervous system by a presumed vascular cause, including ischemic, intracranial and subarachnoid haemorrhage with or without sufficient neuroimaging to participate regardless of premorbid cognitive status (Sacco et al., 2013). Patients or next-of-kin (if the patient was unable to take a position) who refused to give informed consent or had short life expectancy/receiving palliative care were excluded from the study.

Measurements

The 4AT is a rapid delirium screening tool. The test assesses alertness, cognition, attention and acute change or fluctuation of the variables above (Bellelli et al., 2014).

All patients were screened for delirium by a nurse using the 4AT within 24 h of admission to the stroke unit, and at discharge. In addition, the nurses were instructed to perform delirium screening when they recognized symptoms of delirium. The test requires little special training to perform according to the manual. However, before the project started, the nurses underwent brief training consisting of a session on a day seminar for the nurses. In addition, a nurse experienced in 4AT joined/watched the first time 4AT was performed. If a 4AT score of 4 or greater was obtained, a geriatrician was summoned and diagnosed delirium according to The Diagnostic and Statistical Manual of Mental Disorders (DSM-5) criteria. In addition, every patient was evaluated daily by a geriatrician. After completing the study, the stroke nurses (n = 15) answered a questionnaire about how they had experienced the implementation and use of 4AT in daily practice. National Institutes of Health Stroke Scale (NIHSS) is a tool used to objectively quantify focal neurological symptoms caused by stroke. In this study, we used the NIHSS performed at hospital admission. We defined NIHSS ≤ 5 as minor stroke (Yakhkind et al., 2016).

The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) screening tool is a short questionnaire filled out by the next of kin, designed to assess cognitive impairment in the elderly and was used to assess pre-stroke cognitive function in this study (Jorm & Jacomb, 1989).

Ethics

Written informed consent was obtained from all patients or next kin if the patient was not able to sign. The trial was approved by the Vestre Vikens hospital trust data protection officer and conducted according to guidelines of the Declaration of Helsinki.

Statistics

Continuous variables are presented as mean ± standard deviation and categorical variables as numbers and percentages (%). We used Student's t-test for mean of continuous variables and Pearson's Chi-square independence test for categorical variables. Logistic regression analysis was used to explore the association between different independent variables on the incidence of delirium. Results are presented as unadjusted and adjusted odds ratios (ORs). The final model was adjusted according to the predefined variables age and sex. The two-sided significance level was set at p < 0.05. All statistical analyses were performed using IMB SPSS Statistics version 25.0.

RESULTS

In total, 62 out of 110 eligible patients (56%) with a confirmed stroke diagnosis were included in the study, of which 57 (92%) had ischaemic stroke. Mean age was 73.3 (SD 13.7), and 24 (39%) were women. Mean NIHSS was 3.5 (SD 3.6), and 46 (74%) had a minor stroke. Figure 1 provides an overview of patients included in the study.

FIGURE 1. Flowchart of study population.

Nurses performed delirium screening using 4AT within 24 h of admission in 49 (79.0%), at discharge in 39 (62.9%) and at least once during the hospital stay in 56 (90.3%) patients, suggesting that delirium screening with 4AT by stroke unit nurses using was feasible in this study. Every nurse (n = 14) that completed the 4AT training answered that they felt competent to carry out the 4AT screening, did not experience it as significant extra workload and felt that 4AT was a useful clinical tool. When the 4AT was not assessed according to protocol, lack of time (40%) was the most common reason given. Assessments of acute change or fluctuating course (sub-item 4) were experienced as the most difficult to score.

Only two patients had a 4AT score of ≥4 within 24 h of admission, indicating delirium, and the diagnosis was confirmed by a geriatrician in both cases. Another three patients developed delirium during the hospital stay, resulting in a total prevalence of delirium of 8.1%. No other cases of delirium were detected during the hospital stay. While delirium occurred in only one out of 46 patients (2%) with minor stroke, delirium seemed to be more common among patients with major stroke (four out of 16, 25%). In addition, six patients (9.7%) had a 4AT-score of 2–3 during their hospital stay, indicating some form of cognitive impairment. The patients with minor stroke who suffered a delirium had an IQCODE score indicating cognitive impairment (>3.3) before the stroke. Clinical characteristics according to presence or not of delirium are shown in Table 1.

TABLE 1 Study characteristics.

	Total	No delirium w/in 24 h/at discharge	Delirium w/in 24 h/at discharge	p
Total included with confirmed stroke	62	57	5
Women (%)	24 (39)	23 (40)	1 (20)	0.66
Mean age (SD)	73.3 (13.7)	72.8 (14.2)	77.8 (5.9)	0.44
NIHSS at admission, mean (SD)	3.5 (3.6)	3.0 (3.0)	9.0 (4.8)	<0.001
Minor stroke (NIHSS ≤ 5) (%)	46 (74)	45 (79)	1 (20)	0.16
Type of stroke (%)				0.02
Ischaemic	57 (92)	53 (93)	4 (80)
Intracerebral haemorrhage	5 (8)	4 (7)	1 (20)
Treated with alteplase (%)	16 (26)	15 (26)	1 (20)	0.75
Treated with mechanical thrombectomy (%)	5 (8)	4 (7)	1 (20)	0.87
Premorbid cognitive impairment (%)	7 (11)	5 (9)	2 (40)	0.17
Length of stay in days (%)
Mean (SD)	6.4 (3.3)	6.3 (2.9)	8.6 (6.6)	0.14
0–3 days	7 (11)
4–7 days	38 (61)
≥8 days	17 (27)
Complication (%)
Falls	2 (3)	2 (4)	0
Infections	6 (10)	5 (9)	1 (20)
4AT score at day 1 (%)
0	47 (76)	46	1
1–3	11 (18)	10	2
≥4a	4 (6)	0	2
IQCODEb	3.2 (0.6)	3.1 (0.5)	3.7 (0.8)	0.03
Discharged to (%)				0.046
Home	31 (50)	30 (53)	1 (20)
Rehabilitation	26 (42)	23 (40)	3 (60)
Nursing home	3 (5)	2 (4)	1 (20)
Transferred to another hospital	2 (3)	2 (4)	0

Note: w/in 24 h/at discharge delirium screening and diagnosis was performed within the first 24 h after admission and/or at discharge.

Abbreviations: NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation.

^aScore ≥4 indicates delirium.

^bMissing in one patient.

In logistic univariate regression analyses, both NIHSS as a measure of stroke severity and pre-stroke IQCODE score separately predicted delirium. The results did not change after adjusting for sex and age (Table 2). When dichotomizing NIHSS (NIHSS ≤ 5 vs NIHSS > 5), moderate to severe stroke was associated with delirium (OR 15.0, 95%CI 1.5–146.9 and p < 0.05). Adjusting for age, sex, and IQCODE score affected the result only slightly (OR 11.7, 95% CI 1.0–131.5 and p < 0.05).

TABLE 2 Predictors of delirium.

	Unadjusted	95% CI			Adjusteda	95% CI
	OR	Lower	Upper	p	OR	Lower	Upper	p
NIHSS	1.54	1.12	2.12	0.01	1.73	1.11	2.71	0.02
IQCODE	9.94	1.44	68.56	0.02	8.50	1.12	64.37	0.04

Abbreviations: CI, confidence interval; IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio.

^aAdjusted for sex and age.

DISCUSSION

This quality study suggests that nurse conducted delirium screening with 4AT is feasible in stroke units. The validity and feasibility of the 4AT in stroke patients should be evaluated further in larger, prospective studies, but our experience suggests that the screening tool was easily implemented and carried out during daily care by nurses, and useful to detect delirium in patients with stroke.

4AT was largely performed according to the protocol, and lack of time or oversight were the most common reason of protocol violation. We experienced that the implementation of 4AT in the stroke unit increased awareness of delirium, and that the routinely screening was not perceived as significantly increased workload by the nurse staff. If delirium was detected, patients were monitored, followed and treated according to NICE guidelines on management of delirium (National Institute for Health and Care Excellence, 2010), with focus on adequate sleep, unambiguous environment, communication, education and reorientation.

The prevalence of post-stroke delirium was 8%, which is in the lower range compared to previous studies (Pasinska et al., 2018; Shi et al., 2012; Zipser et al., 2021). The variation in delirium prevalence could, at least partly be explained by different delirium screening approach and differences between the studied patients cohorts with regard to stroke severity. In our study, delirium rarely occurred in patients with minor stroke, but one out of four patients with major stroke suffered a delirium.

Three out of five cases of delirium occurred after 24 h of admission. Fleischmann and co-workers found that only one out of four delirium cases developed within 24 h after symptom onset, and approximately 50% of the delirium cases would be missed if the screening was performed only once (Fleischmann et al., 2021). This emphasizes the importance of repeated assessments, and suggest that delirium screening conducted only once is insufficient. Consequently, we now routinely perform screening daily for the first 72 h, and if suspicion of change in cognition during the hospital stay.

Hence, we cannot rule out that we have overlooked delirium cases, as screening was not performed daily and by nurses only, and as 4AT was used as the sole screening tool. Our conservative approach with few measurements was chosen to avoid a heavy workload on the nurses. However, the study showed that 4AT was not perceived as resource-intensive, and as a result, daily screening should be attempted in clinical practice. Further, the sub-item number 4 (Acute change or fluctuating course) in the 4AT was perceived as the most difficult to score, especially in the first meeting with the patient when background information was lacking. This finding underpins the importance of repeated screening, and may have contributed to the low prevalence of delirium in our study cohort. The 4AT have favourable properties for delirium screening in an acute stroke setting as it in comparison to other delirium screening tests contains cognitive test items, and the tool has shown good sensitivity and specificity in previous studies (Fleischmann et al., 2021). The fact that the patients received daily visits from geriatricians also reduces the risk of overlooked cases.

Further, we excluded patients receiving palliative care, of whom many had suffered a moderate or severe stroke, as stroke severity is associated with delirium. Finally, only 56% of the eligible patients consented to participate in this study. We experienced that the inclusion of patients with cognitive impairment was challenging and believe that this may have contributed to an underestimation of the delirium prevalence in our study.

In line with previous studies (Shaw et al., 2019; Zipser et al., 2021), we found that predisposing factors for post-stroke delirium were pre-stroke cognitive impairment and stroke severity.

Several study limitations needs to be considered, in addition to those addressed above. This quality study was neither designed to assess the validity of the inter-rater reliability of the 4AT, nor powered to explore predictors of delirium, and the small number of participants limited statistical testing. We cannot rule out a selection bias as women, and moderate and severe stroke patients were underrepresented, possibly reducing the generalizability of the results. Finally, important strengths of the study are that we have only included patients with a confirmed stroke diagnosis, that 4AT was carried out during daily care by nurses in a real-life setting, and that a geriatrician confirmed the delirium diagnosis in real time. Finally, we included old patients with cognitive impairment and delirium; groups that are often excluded from clinical studies.

CONCLUSION

This quality study suggests that delirium screening with 4AT performed by nurses in a stroke unit is feasible and may increase the awareness of delirium among the stroke unit staff. If possible, daily screening with 4AT is preferable.

AUTHOR CONTRIBUTIONS

All authors contributed to the study conception and design. TJ, GH, EG and NJ performed data collection, and HIH performed the analysis. HIH together with MM wrote the first draft of the article and all authors commented on previous versions of the article. All authors read and approved the final article.

ACKNOWLEDGEMENTS

We thank all dedicated nurses at the Stroke Unit, Bærum Hospital, for all their efforts.

FUNDING INFORMATION

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

CONFLICT OF INTEREST STATEMENT

The authors have no relevant financial or non-financial interests to disclose.

DATA AVAILABILITY STATEMENT

The datasets used during the current study are available from the corresponding author on reasonable request.

ETHICAL APPROVAL

The trial was approved by the Vestre Vikens hospital trust data protection officer (20/01673-1), and conducted according to guidelines of the Declaration of Helsinki.