Emergency Department Shift Change Is Associated With Pneumonia in Patients With Acute Ischemic Stroke
Background and Purpose—Emergency department (ED) nurses play a pivotal role in early acute ischemic stroke patient management. We hypothesized that patients exposed to ED nursing shift changes (SC) may develop pneumonia (PNA) more frequently and have worse early outcomes than do patients who have continuity of care until stroke unit admission.
Methods—Consecutive acute ischemic stroke patients presenting to our ED were studied using chart review and prospectively collected registry data. We evaluated the association of patient presence during an ED SC (ie, 07:00–08:00, 19:00–20:00) with length of stay in the ED, PNA rates, and early outcome measures (discharge disposition, modified Rankin Scale score, and death).
Results—Three hundred sixty-six consecutive acute ischemic stroke patients met the criteria. Of those, 54.9% were present during an SC. After adjusting for baseline National Institutes of Health Stroke Scale, admission glucose, and intravenous tissue-type plasminogen activator, patients present during SC were half as likely to be discharged home or to inpatient rehab (OR, 0.50; 95% CI, 0.26–0.96; P=0.04) and were 2.5 times more likely to develop PNA (OR, 2.54; 95% CI, 1.02–6.30; P=0.045). After additional adjustment for time in the ED, the difference in favorable discharge disposition was no longer significant, but SC was associated with 5 times the odds of PNA (OR, 5.35; 95% CI, 1.34–21.39; P=0.018) compared with patients with continuity of care.
Conclusions—In our center, acute ischemic stroke patients present during an ED nursing SC experienced higher rates of PNA and had decreased rates of favorable discharge disposition compared with patients with continuity of care. Strategies to prevent PNA and improve hand-off communication during SC may reduce this risk.
Emergency department (ED) staff are critically important in the delivery of acute stroke care before patients arrive in the stroke unit or intensive care unit (ICU). ED teams are trained to implement the National Institute of Neurological Disorders and Stroke guidelines to reduce delay in early acute evaluation and management, particularly when thrombolysis is indicated.1 Equally important is nursing staff that is familiar with the protocols used to administer safely tissue-type plasminogen activator (tPA) and the necessary interventions in the care of acute stroke patients. Complications after acute ischemic stroke (AIS) are common, lead to worse outcomes, and are largely preventable.2 Close monitoring of vital signs, blood glucose, level of consciousness, and neurological condition by ED nurses before stroke unit or ICU transfer is necessary for early recognition of deterioration and implementation of appropriate preventive measures.3,4
Time in the ED has been associated with worse outcomes and higher mortality among critically ill patients5 and patients with cerebrovascular disease.6 Depending on the center, patients may wait in the ED for several hours before transfer to the stroke unit or ICU. ED nurses, rather than neurology specialized nurses, provide acute care during this interval. The longer a patient spends in the ED, the more likely they are to experience a nursing shift change (SC). Management relevant to stroke patients, such as post-tPA monitoring, head-of-bed positioning, and safety of oral intake, must be transferred accurately and completely to the new nurses.7 Inadequate communication during a handoff at SC can lead to medical errors and worse patient outcomes.8 Recent research has suggested a number of strategies to correct this shortcoming and improve patient care.9 However, no studies of the impact of nursing SC have focused on patients with AIS.
Tulane's small but active stroke center frequently has delays in patient transfer to the neuro ICU or stroke unit. In accordance, we sought to determine whether presence in the ED during a nursing SC is associated with pneumonia (PNA) and poor outcome after AIS.
This is a retrospective, observational study evaluating the association of ED nursing SC on pneumonia and outcome in patients following AIS. ED patients for this study were identified using data from a single hospital stroke registry that includes all patients diagnosed with AIS in the ED or during admission. The study examined consecutive patients in the registry for whom the discharge diagnosis was AIS from July 2008 to June 2010 and who met the inclusion criteria of admission through the ED and available time data for analysis. Patients transferred from outside hospitals and with in-hospital strokes were excluded. Prospectively collected registry data and the original medical records were then reviewed and data abstracted for analysis. The intravenous (IV) tPA treatment rate for AIS in our stroke center is 27.9%, and stroke diversion is not permitted. Tulane Medical Center has 352 licensed beds, including an 8-bed stroke unit/neurological ICU, staffed by university physicians. The ED has a maximum capacity of 28 beds (16 adult and 12 pediatric) and is staffed by physicians who are board certified by the American Board of Emergency Medicine. There is no emergency medicine residency training program. The ED physicians are not involved in the care of stroke patients once they are admitted to the stroke service. ED nurses, stroke service resident physicians and/or nurse practitioners, and a stroke service attending physician are the only providers involved in the care of admitted patients with stroke in the ED. The ED nurses change shift twice daily between 07:00 to 08:00 and 19:00 to 20:00. The nursing SC is a verbal report, given near the patient's bedside, but does not include formal review of stroke-specific orders, such as head-of-bed positioning, status of dysphagia screening, and/or speech therapy recommendations for diet.
In our center, all patients with presumptive stroke are given nothing by mouth after arrival to the ED. Standard orders for patients who are not treated with IV tPA include an antiplatelet agent, but only after cleared for oral medications by dysphagia screening performed by the patient's nurse. A dysphagia screen only clears a patient for medications with small amounts of water. If the patient does not pass the dysphagia screen, oral medications are held unless an oro- or nasogastric tube is placed. Only after formal evaluation by a speech therapist is an order for a diet given. Competency training in this protocol is required for all new nurse hires and annually for all hospital nurses.
Demographics, vascular risk factors, baseline National Institutes of Health Stroke Scale (NIHSS) score, labs, treatment with IV tPA, and date/time of ED arrival, were collected prospectively by a trained nurse coordinator as part of our registry. Other variables of interest to this study, including the date/time that patients were transported from the ED to be admitted to stroke unit/neurological ICU or medical/surgical unit, occurrence of complications, and modified Rankin Scale (mRS) score on discharge were collected retrospectively from the electronic medical record and were added to the existing stroke database. Patients were coded as present during a SC if they were in the ED during the 07:00 to 08:00 (morning SC) or 19:00 to 20:00 (night SC) hours. To determine whether there was a bias related to the “weekend effect,” SCs that occurred from Friday night through Monday morning were coded as weekend SC.10 Neuroworsening was defined as an increase in the NIHSS score of 2 or more points within a 24-hour period. PNA was defined as a new infiltrate on chest radiography with appropriate clinical signs and symptoms. Additional chart review was conducted to determine whether the PNA was present on admission, developed in the first 72 hours of admission in lower lung zones in nonintubated patients (aspiration PNA), or developed after 48 hours of intubation and mechanical ventilation (ventilator-associated PNA unless intubated because of respiratory failure after aspiration). Favorable discharge disposition was defined as discharge to home or to an inpatient rehabilitation center. Good functional outcome was defined as discharge mRS 0 to 2. Based on a previous study in stroke patients,6 we dichotomized length of stay (LOS) in the ED as short (<5 hours) and long (≥5 hours).
The data collection was performed by stroke researchers trained by the principal investigator and/or data manager with formal definitions for all data elements collected. All providers responsible for the care of stroke patients in the ED are certified to perform NIHSS evaluation by the American Heart Association/American Stroke Association online training program. Only physicians who are certified in the mRS examination scored the examination based on independent examination and review of therapy notes. Physicians scoring the mRS were blinded to LOS in the ED and exposure to SC. Though interrater and intrarater agreement analyses were not performed, data inconsistencies were adjudicated by the principal investigator, and data validation was performed and reconciled during quality assessment.
The primary outcome measure was the proportion of patients achieving a favorable discharge disposition (home or inpatient rehabilitation). Secondary outcome measures included rates of neuroworsening, PNA, death, and proportion of patients with good functional outcome (mRS 0–2 on discharge). These outcomes were compared in patients with and without an SC while in the ED. Among patients with an SC, outcomes were compared in patients with morning versus night SC and weekday versus weekend SC. We also compared outcomes based on total time in the ED. This study was approved by the Tulane University Institutional Review Board.
Statistical analysis was performed using Predictive Analytics SoftWare (PASW) Statistics 18.0 (SPSS Inc.). Categorical data were presented as frequencies and compared using Pearson χ2 or Fisher exact test where appropriate. Continuous data were presented as mean±SD or median (minimum-maximum) and compared using independent-samples Student t test or Mann-Whitney U test, respectively. The association between SC and each prespecified complication or outcome was estimated using logistic regression with and without adjustment for stroke severity (NIHSS), admission glucose, IV tPA, PNA, and time in ED. Probability values of <0.05 (2-sided) were considered statistically significant.
We assessed our registry of 441 records of patients with AIS. Twenty-one patient records were excluded from analysis for in-hospital stroke, 20 records for bypassing the ED and 34 records with incomplete time data, leaving 366 records of ED patients for analysis. More than half of the patients (54.9%; 201/366) were present in our ED during an SC (Table 1). The median time in ED for the SC group was significantly higher than for no shift change (NSC) group (518 versus 238 minutes; P<0.0001). Median baseline NIHSS, admission glucose, and IV tPA treatment rates were lower in the SC group compared with the NSC group. As baseline NIHSS, admission glucose, and IV tPA treatment have been shown to affect patient outcome and were unbalanced in SC and NSC groups, logistic regression was performed adjusting for these variables.
The SC group had twice the rate of PNA (9.5% versus 4.8%; P=0.07) of the NSC group. After adjusting for NIHSS, glucose, and use of IV tPA, patients present during an ED SC were 2.5 times more likely to develop PNA (OR, 2.54; 95% CI, 1.02–6.30; P=0.045). After adjusting further for time in the ED, ED SC was associated with 5 times the odds of developing PNA (OR, 5.345; 95% CI, 1.336–21.389; P=0.018). Aspiration PNA developed in the first 72 hours in only 5 SC patients and 2 NSC patients. After adjusting for NIHSS, glucose, and IV tPA, SC patients were less likely to be discharged to home or to inpatient rehabilitation (OR, 0.50; 95% CI, 0.26–0.96; P=0.04) when compared with NSC patients. However, after adjusting further for PNA (OR, 0.57; 95% CI, 0.29–1.12; P=0.10) or ED LOS (OR, 0.866; 95% CI, 0.269–2.791; P=0.866), the difference in favorable discharge disposition was no longer significant.
Patients with morning SC had significantly higher baseline NIHSS scores, longer inpatient LOS, higher rates of neuroworsening and PNA, and lower rates of mRS 0 to 2 at discharge (Table 2). Patients who experienced a weekend SC spent less time in the ED when compared with the weekday SC group (median 428 minutes versus 538 minutes; P=0.04). Patients with weekend SC had significantly higher baseline NIHSS scores. There were no differences in IV tPA treatment rates based on time of day or day of the week.
Median LOS in the ED for the entire sample was 351 minutes (45–3140). There were weak correlations with relation to ED LOS and patient outcome measures: discharge disposition (r=−0.144; P=0.01) and mRS score at discharge (r=−0.131; P=0.01). After adjusting for baseline NIHSS, glucose, and IV tPA treatment, time spent in the ED was not an independent predictor of favorable discharge disposition (OR, 1.00; 95% CI, 0.999–1.001; P=0.79) or discharge mRS 0 to 2 (OR, 1.00; 95% CI, 1.000–1.001; P=0.37). Patients who had long ED LOS (≥ 5 hours; 59.6%; 218/366) had lower median baseline NIHSS (5 versus 7; P=0.007) and rates of IV tPA (18.8% versus 41.2%; P<0.0001) than did those with shorter ED LOS (Table 3). There were no differences in outcome measures when comparing these 2 groups. Of patients present in the ED for <5 hours, patients with SC were nearly 4 times more likely to develop PNA when compared with patients without an SC (OR, 3.533; 95% CI, 1.062–11.754; P=0.07). As shown in Table 3, there were no other significant differences in baseline demographics, stroke characteristics, or outcome measures among SC and NSC patients present in the ED for <5 hours.
An exploratory analysis of the association between IV tPA use and baseline characteristics, exposure to SC, time in ED, and patient outcome demonstrated significant differences in patients who were and were not treated with IV tPA (Table 4). Patients who were treated with IV tPA had more-severe strokes (per NIHSS), experienced fewer SCs in the ED, spent less time in the ED, and had worse outcome than did patients who were not treated with IV tPA.
Our study found that an ED nursing SC was associated with PNA in patients with AIS even after adjusting for time in the ED and baseline imbalances in factors known to contribute to outcome. Aspiration, a common cause of PNA in stroke patients, especially in those experiencing dysphagia and poor control of oral secretions, has been consistently associated with worse outcome.11 Given that stroke severity is associated with risk for developing PNA poststroke12 and patients who were present during a SC had significantly lower baseline NIHSS scores, they should have had lower risk of PNA. The frequency of new aspiration PNA within the first 3 days of admission (8.5% in patients with ED SC and 4.2% in NSC patients) was uncommon, limiting attribution of the higher rate of PNA to events during the ED stay. A power calculation estimates that almost 900 patients in each group would be needed to detect a difference in groups. Regardless, these results suggest aspiration may be an early event and may increase morbidity even in lower-severity stroke patients. It is also possible that PNA may be an effect of preadmission aspiration related to “wake-up” stroke. Because patients with morning SC had more-severe strokes, more neuroworsening, and PNA, and worse functional outcomes than did patients with night SC, the sickest stroke patients present in early morning hours and are in the ED during the morning SC. The almost 2-hour difference in transfer times out of the ED may reflect limitations on transfer secondary to stroke team rounding times that facilitate opening stroke unit/neurological ICU beds. Additional study will be necessary to clarify these issues.
A recent study of critically ill patients with acute ischemic and hemorrhagic stroke demonstrated that patients who spent more than 5 hours in the ED had a 4-fold increase in the odds of poor outcome.6 In our study, univariate analysis did not find significant associations between ED LOS and outcomes. This may be caused by our restricted sample, as we excluded patients with intracerebral hemorrhage, who, on average, have higher morbidity and mortality. However, ED LOS must be related to discharge disposition in our sample, because SC was no longer an independent predictor of disposition once adjusting for ED LOS. Furthermore, in our sample, patients who were in the ED more than 5 hours had significantly less-severe strokes, reducing the likelihood of poor outcome.
Patients who were present during a SC had significantly lower rates of IV tPA treatment. It is more likely that post-tPA patients received prioritization for a stroke unit/neurological ICU bed, resulting in a shorter time in the ED and lower likelihood to experience an SC, than that the SC resulted in lower rates of IV tPA treatment. In our center, patients with the highest risk for worsening and highest demand for nursing attention (more-severe stroke and IV tPA treatment) reached the stroke unit more quickly than did other patients. This likely represents appropriate bed utilization.
Our study has several important limitations. The retrospective design renders determination of causality impossible. Relevant information on SC procedures and hospital bed allocation processes may not be ascertainable. Other factors, not evaluated in the present study, may contribute to the higher frequency of PNA in patients who experience a nursing SC in the ED. The retrospective assessment of the mRS score for some patients may limit the validity of this outcome measure, but mRS scores were determined blinded to the exposure of the SC. In addition, our single center experience makes generalizability somewhat limited. To the best of our knowledge, this is the largest reported study of time in the ED and SC in ischemic stroke patients. However, the small sample size may be limiting our ability to detect differences, should they exist, especially in subset analyses.
The higher rate of PNA seen in patients present during a SC raises the question—could structured methods of alerting oncoming nurses to the status of dysphagia screening and head-of-bed-position orders in these patients reduce stroke-associated PNA? Although little can improve bed availability, efforts to minimize ED LOS and SC turnover may improve stroke outcome. Additional study examining the impact of SC on outcome in patients with AIS is still needed.
- Received December 31, 2010.
- Accepted June 16, 2011.
- © 2011 American Heart Association, Inc.
- Johnston KC,
- Li JY,
- Lyden PD,
- Hanson SK,
- Feasby TE,
- Adams RJ,
- et al
- Hilker R,
- Poetter C,
- Findeisen N,
- Sobesky J,
- Jacobs A,
- Neveling M,
- et al