STROKE is a major cause of disability and

Transcription

1 218 STROKE DELAYS Morris et al. STROKE CARE DELAYS EDUCATION AND PRACTICE Time Delays in Accessing Stroke Care in the Emergency Department DEXTER L. MORRIS, PHD, MD, WAYNE D. ROSAMOND, PHD, ALBERT R. HINN, MD, REBECCA A. GORTON, MD Abstract. Objective: To delineate components of delay within the hospital ED for patients presenting with symptoms of stroke. Methods: A prospective registry of patients presenting to the ED with signs or symptoms of stroke was established at a university hospital from July 1995 to March The ED arrival time, time to being seen by an emergency physician (EP), time to CT scan, and time to neurology consultation were obtained by medical record review. Results: The median delay (interquartile range) from ED arrival to being seen by an EP for the 170 eligible subjects was 0.42 ( ) hours. The median delay to CT scan was 1.88 hours ( ) and the median delay to neurology consultation was 2.42 hours ( ). Age, race, sex, and hospital discharge diagnosis had little influence on delay. Subjects arriving by emergency medical services (EMS) had a significantly shorter time to being seen by an EP (0.33 vs 0.50 hours) when compared with those who arrived by other means. Time to CT scan was shorter by 0.5 hours for patients arriving by EMS as well. These differences persisted when stratified by out-of-hospital delay times. Conclusions: These data suggest that arriving by EMS is associated with shorter times to being seen by an EP and receiving a CT scan. The influence of EMS on delays associated with rapid medical care of stroke patients reaches beyond the out-of-hospital transport phase. Key words: delay in care; stroke; emergency medical services; cerebral infarction; rapid care; CT scans. ACADEMIC EMERGENCY MEDICINE 1999; 6: STROKE is a major cause of disability and death in the United States, with more than 500,000 new strokes and 150,000 deaths occurring annually. 1 There is now evidence that thrombolytic therapy for acute stroke can improve patient outcomes in some groups, 2 and other neuroprotective agents are on the horizon. 3 The efficacy of these new therapies is markedly time-dependent, placing renewed importance on reducing delays in stroke care. Out-of-hospital components of delay have been reported, but less attention has been paid to delays in stroke care occurring in the ED itself. 4 6 From the Departments of Emergency Medicine (DLM, RAG) and Neurology (ARH), School of Medicine, and the Department of Epidemiology (WDR), School of Public Health, University of North Carolina, Chapel Hill, NC. Received September 17, 1998; revision received November 12, 1998; accepted November 15, Presented in part at the SAEM annual meeting, Chicago, IL, May Supported by a Medical Alumni Endowment Fund grant from the University of North Carolina at Chapel Hill. Address for correspondence and reprints: Dexter L. Morris, PhD, MD, Department of Emergency Medicine, CB 7594, Ground Floor Neurosciences, University of North Carolina, Chapel Hill, NC Fax: ; dmorris@med.unc.edu To delineate components of delay within the ED, we conducted a study of patients with strokelike symptoms presenting to a university hospital ED. The objective of the Delay in Accessing Stroke Healthcare (DASH) study 7 was to quantify components of both out-of-hospital and ED delays in the care of acute stroke. The goal of these investigations was to refine the knowledge base from which effective intervention strategies might be developed. This report specifically examines delays within the ED. METHODS Study Design. The methods of the DASH study have been previously reported. 7 A prospective registry of patients presenting to the ED with signs or symptoms suggestive of stroke was established between July 1995 and March The institutional review board on human subjects in research at the University of North Carolina (UNC) at Chapel Hill approved the study protocol. Subjects or next of kin gave informed consent to be interviewed. Study Setting and Population. The study was conducted in the ED of UNC Hospitals, Chapel

2 ACADEMIC EMERGENCY MEDICINE March 1999, Volume 6, Number Hill, NC. This tertiary care facility has an adult ED volume of 35,000 visits per year. Trained ED staff nurses enrolled 201 patients presenting with any of the following symptoms: sudden unilateral weakness or numbness of face, arm, or leg; sudden dimness or loss of vision; loss of speech; sudden severe headache; or dizziness. There were no age or other entry criteria other than willingness to give informed consent. These broad eligibility criteria, which were based on symptoms at presentation and not ED or hospital discharge diagnoses, resulted in the enrollment of many patients who ultimately were not diagnosed as having stroke. This, however, allowed for the characterization of ED delays among a full spectrum of patients presenting with stroke-like symptoms. Measurement. A total of 25 ED nurses participated in patient enrollment and interviews. All were trained by the project coordinator to identify cases and complete a structured interview lasting approximately 20 minutes. After the patients were discharged from the hospital, one trained medical record abstractor obtained information about the timing of diagnostic tests and therapeutic procedures as well as the final hospital discharge diagnosis. Cases were classified as stroke if they had International Classification of Diseases Clinical Modification (ICD-9-CM) discharge diagnosis codes Patients with a code of 435 were classified as having a transient ischemic attack (TIA). Data Analysis. Twenty-nine patients were excluded who reported delays from onset of symptoms to arrival in the ED of 36 hours or more and two were excluded who did not have ED arrival times available. During the nine-month enrollment period, approximately 50% of all the stroke patients admitted to the hospital (including those not seen in the ED and therefore not eligible for the registry) were enrolled in the study. Time to seeing an emergency physician (EP) was defined as the time from arrival in the ED (first time recorded in ED, usually triage or registration) until first being seen by an EP (earliest time of first physician encounter recorded by physician or nurse). Time to CT scan return was defined as the time from ED arrival to the patient s return from CT scan (from nursing notes). In this ED, the CT scanner was located 100 feet from the ED, so patient return times were likely very close to CT scan completion times. Time to seeing a neurologist was defined as the time from ED arrival until first contact with a neurologist (from nursing notes or consult sheet). Differences in times were tested statistically using a nonparametric analysis of variance. RESULTS Of a total of 170 patients included in the analysis, 78 (46%) were discharged with a diagnosis of stroke, 16 (9%) with a diagnosis of TIA, and 76 (45%) with diagnoses other than stroke or TIA. The majority of the strokes (76%) were ischemic. Those subjects discharged with a diagnosis other than stroke or TIA had diagnoses that included migraine, brain neoplasm, drug intoxication, other neurologic disorders, and ill-defined conditions. The mean age of the subjects was 67 years, 55% were female, and 41% were black (Table 1). Emergency medical services (EMS) was used for transport for 49% of the subjects, 81% had a CT scan while in the ED, and 79% were seen by a neurologist in the ED. The patients with a discharge diagnosis of stroke or TIA were more likely to arrive by ambulance (58% vs 49%), have a CT scan done in the ED (94% vs 65%), and be seen by a neurologist in the ED (96% vs 58%) than were those patients with discharge diagnoses other than stroke or TIA (Table 1). The times to seeing an EP, CT scan return, and seeing a neurologist are shown in hours by diagnostic groups in Table 2. For all the subjects, the median time to being seen by an EP was 0.4 hours, with an interquartile range of 0.2 to 0.8 hours. There was little difference in these times by specific diagnosis. The median time to CT scan return for all the subjects was 1.9 hours, with an interquartile range of 1.3 to 2.7. Although the subjects with a discharge diagnosis of stroke appeared to have a shorter median time to CT scan return (1.7 hours) than either those with TIA (2.2 hours) or those with other diagnoses (1.9 hours), the differ- TABLE 1. Characteristics of the Patients with Stroke, Transient Ischemic Attack (TIA), or Other Conditions Discharge Diagnosis Stroke (n = 78) TIA (n = 16) Other (n = 76) Total (n = 170) Age mean ( SD) 67.5 ( 14.4) yr 70.8 ( 9.9) yr 66.4 ( 15.3) yr 67.3 ( 14.5) yr Gender female 55.1% 50.0% 55.3% 54.7% Race black 42.3% 37.5% 39.5% 40.6% Arrival by ambulance 57.7% 37.5% 43.4% 49.4% CT scan done in ED 93.6% 93.8% 64.5% 80.6% Neurology consultation done in ED 96.2% 93.8% 57.9% 78.8%

3 220 STROKE DELAYS Morris et al. STROKE CARE DELAYS ences were not statistically significant. The median time to neurology consultation for all the subjects was 2.4 hours, with little difference among diagnostic groups. Race, sex, and age did not affect time to seeing an EP or time to CT scan return in statistical analyses (data not shown). The use of EMS, on the other hand, was associated with shorter delays to being seen by an EP and to receiving a CT scan. Those subjects using EMS had a median time to seeing an EP of 0.33 hours (interquartile range ), compared with 0.50 hours (interquartile range ) for those who did not. Time to CT scan return was 0.54 hours shorter among those using EMS (see Figure 1). Both differences were statistically significant. The elapsed time from the onset of symptoms and arrival to the hospital ED (out-of-hospital delay) may influence times within the ED itself. To examine this, times to seeing an EP and CT scan were categorized by out-of-hospital delay less than or equal to three hours and more than three hours. The three-hour cutoff point was chosen to represent the current upper limit for tissue plasminogen activator (t-pa) use. As expected, those subjects arriving in three hours or less had shorter times to seeing an EP (median 0.37 hours, interquartile range ) compared with those arriving later (median 0.47 hours, interquartile range ). The median time to CT scan return was also shorter. Delay times characterized by outof-hospital delay and EMS use are shown in Figure 1. Times to seeing an EP and CT scan return were shorter for those arriving within three hours, particularly if they arrived by EMS. For example, the time to seeing an EP for those arriving by EMS within three hours of symptoms was 0.32 hours (interquartile range ) compared with 0.43 hours for those arriving by EMS more than three hours after the onset of symptoms. Most notable in these numbers is that arrival by EMS shortens time to seeing an EP and time to CT scan return independently of out-of-hospital delay times. Identical relationships were seen when the analysis was done using patients with a discharge diagnosis of stroke only. As expected, a longer time to seeing an EP was associated with a longer time to CT scan return. This is shown in Table 3 where time to CT scan return is presented by quartile of time to seeing an EP. The difference between the lowest and highest quartiles was one hour. TABLE 2. Delay Intervals for the Patients with Stroke, Transient Ischemic Attack (TIA), or Other Conditions Discharge Diagnosis Stroke (n = 78) TIA (n = 16) Other (n = 76) Total (n = 170) Time from ED arrival to seeing an emergency physician (hr) Median (25th, 75th percentiles) (0.33, 1.17) (0.33, 1.17) (0.12, 0.75) (0.20, 0.75) Mean ( SD) 0.56 ( 0.55) 0.76 ( 0.49) 0.58 ( 0.69) 0.59 ( 0.61) No. missing Time from ED arrival to CT scan return (hr) Median (25th, 75th percentiles) (0.37, 2.57) (1.37, 2.69) (1.58, 3.25) (1.25, 2.67) Mean ( SD) 1.88 ( 1.39) 2.02 ( 0.78) 2.49 ( 1.32) 2.10 ( 1.33) No. missing Time from ED arrival to seeing a neurologist (hr) Median (25th, 75th percentiles) (1.50, 3.42) (1.50, 4.50) (1.45, 3.75) (1.50, 3.48) Mean ( SD) 2.63 ( 1.67) 3.09 ( 1.74) 2.86 ( 1.91) 2.76 ( 1.75) No. missing TABLE 3. Time to CT Scan Return by Quartile of Time to Seeing an Emergency Physician First Quartile (<0.20 hr) Second Quartile ( hr) Third Quartile ( hr) Fourth Quartile ( 0.75 hr) Time from ED arrival to CT scan return median (range) (hr) 1.63 ( ) 1.54 ( ) 1.96 ( ) 2.67 ( ) N (missing) 26 (12) 32 (8) 26 (9) 24 (10) p < 0.001, Kruskal-Wallis one-way nonparametric analysis of variance.

4 ACADEMIC EMERGENCY MEDICINE March 1999, Volume 6, Number Figure 1. Median hours (interquartile range) from ED arrival to being seen by an emergency physician or returning from receiving a CT scan by use of emergency medical services (EMS) and out-of-hospital delay, defined as time from onset of symptoms to arrival in the ED. *Time from ED arrival to being seen by a physician. Time from ED arrival to returning from receiving a CT scan. DISCUSSION Until recently, patients presenting to the ED with symptoms of stroke were often not considered urgent unless they had specific airway or cardiac problems. This was based on the fact that there was no specific treatment for acute stroke. With the advent of a time-dependent treatment for stroke and the anticipation of additional treatments in the future, this attitude is changing. A major force for this change has been the guidelines developed by the National Institute of Neurological Disorders and Stroke, published in The guidelines state that patients with stroke symptoms should be evaluated by a physician within 10 minutes of arrival at the ED and a CT scan should be initiated within 25 minutes and interpreted within 45 minutes of arrival. In the present study, the median time to being seen by a physician for all the subjects was 25 minutes, with only 25% of the patients being seen within 10 minutes. The median time to CT scan was not determined directly in this study, but the median time for the patient to return from CT was 113 minutes. During the course of the study an acute stroke treatment protocol was in place, which included t-pa, with ongoing training of EPs and nursing staff, but there was no special stroke team or hospital-wide acute stroke program. The delay times seen in this study are comparable to those reported elsewhere, although the methodologies were different. In a retrospective review of 112 patients presenting within six hours of symptom onset in Houston, Bratina et al. found a mean delay of 28 minutes to being seen by a physician (median delay was not reported) and a mean delay of 100 minutes until a CT scan was obtained. 5 Delays for 86 patients arriving by EMS only were studied in Cincinnati, an area with an active stroke program. The average time from arrival to seeing an EP was about 15 minutes and the average time to CT scan was about 57 minutes. 9 Factors that are related to delay times are important because they may represent opportunities to reduce delays or identify groups that are particularly in need of intervention. The demographic variables examined in this study, race, age, and sex, were not related to ED delays. This is reassuring but needs to be examined in other medical settings because these factors have been related to utilization of thrombolytic therapy for acute myocardial infarction. 10,11 Ultimate diagnosis did not appear to influence ED delay times, particularly time to being seen by a physician. This result was anticipated because it is often difficult for the EP to tell whether a patient has had a stroke or not from a chief complaint or a triage note. The use of EMS by patients with stroke symptoms has been shown to markedly decrease delay from onset of symptoms to arrival in the ED. 7,12,13 The data reported here suggest that use of EMS

5 222 STROKE DELAYS Morris et al. STROKE CARE DELAYS also reduces time from arrival in the ED to being seen by a physician. The median time to being seen by an EP was 20 minutes for those who used EMS vs 30 minutes for those who did not. Time to CT scan return was also shorter (90 minutes vs 130 minutes, respectively). A likely contributing factor to the reduced in-hospital delay is the fact that patients arriving by EMS are most often placed directly in a bed and are available to be seen by an EP immediately. This is in contrast to a patient arriving by the front door who first must be triaged before he or she can physically be brought to a bed. Registration delays or lack of recognition of the urgency of the problem at triage would only add to the delay. Another explanation of the shorter delay times for patients arriving by EMS may be the level of urgency (real or perceived) that the EP places on seeing the patient. It is possible that the mere fact of arriving by EMS raises the physician s or ED staff s level of concern about the patient. Alternatively, patients arriving by EMS may be more severely ill and thus naturally would be seen by an EP sooner. This study does not directly measure severity of stroke. We did, however, collect Glasgow Coma Scale scores and found them not to be related to the time from hospital arrival to being seen by an EP. Other data suggest that severity does have a role in rapid response within the ED. Kothari et al. found that patients arriving by Basic Cardiac Life Support (BCLS) transport had longer times to seeing EPs than did those arriving by Advanced Cardiac Life Support (ACLS) transport, indicating that the severity of the patient s illness cannot be dismissed as a factor influencing delay. 9 The time from the onset of symptoms may also play a role in the ED staff s perception of severity. Regardless of their beliefs about the usefulness of t-pa for ischemic stroke, staff may regard someone whose symptoms have been present for only an hour as being different from someone whose symptoms have been present for more than a day. Indeed, the data reported suggest that the time to being seen by an EP is shorter for those presenting within three hours of the onset of symptoms (22 minutes vs 28 minutes for those presenting after three hours). However, arrival by EMS still resulted in shorter times to seeing an EP, regardless of the length of out-of-hospital delay. Time to CT scan return followed similar trends. LIMITATIONS AND FUTURE QUESTIONS A major limitation of this study is the retrospective collection of EP and CT scan times. About 15% of the times were missing, and in the absence of direct monitoring, it was not possible to validate the times that were available. Prospective, independent collection of physician times would be necessary in any future study to address this problem. It may be possible for future studies to collect actual CT scan time by using the time that is recorded electronically on the imaging equipment. A second limitation is that this study was done in one hospital. Practices and patient populations vary considerably, and it is necessary to confirm these findings in other settings. Finally, time to neurology consultation was difficult to interpret and should be treated tentatively. There was no information on whether the delays were related to the time it took to initiate a contact with a neurologist or delay in arrival after notification. CONCLUSIONS The results from this study suggest that there is significant room for improvement in reducing times to being seen by an EP and obtaining a CT scan for patients presenting with symptoms of stroke. While there are many approaches to this goal, this study suggests that patient characteristics were not the major determinants of delay. However, patients arriving by EMS had significantly shorter times to seeing an EP and CT scan return than did patients who arrived by other means. If these results are confirmed in future prospective studies, programs aimed at increasing the use of EMS would seem appropriate. Use of EMS has also been associated with an almost 50% shorter out-of-hospital delay time. 7 This, in conjunction with the shorter delay times in the ED that may be influenced by EMS use, could result in significant reductions in total elapsed time from symptom onset to definitive treatment. References 1. American Heart Association. Heart and Stroke Facts: 1997 statistical supplement. Dallas: AHA, National Institute of Neurological Disorders and Stroke rt- PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995; 333: Barinaga M. Finding new drugs to treat stroke. Science. 1996; 272: Ferro J, Melo T, Oliveira V, Crespo M, Canhao P, Pinto A. An analysis of the admission delay of acute strokes. Cerebrovasc Dis. 1994; 4: Bratina P, Greenberg L, Pasteur W, Grotta J. Current emergency department management of stroke in Houston, Texas. Stroke. 1995; 26: Malik M, Gomez C, Tulyapronchote R, Malkoff M, Bandamudi R, Banet G. Delay between emergency room arrival and stroke consultation. J Stroke Cerebrovasc Dis. 1993; 3: Rosamond W, Gorton R, Hinn A, Hohenhaus S, Morris D. Rapid response to stroke symptoms: the Delay in Accessing Stroke Healthcare (DASH) study. Acad Emerg Med. 1998; 5: National Institute of Neurological Disorders and Stroke. Rapid identification and treatment of acute stroke. In: Marler J, Jones P, Emr M (eds). Proceedings of a National Symposium on Rapid Identification and Treatment of Acute Stroke. Bethesda, MD: National Institutes of Health; NIH Publication

6 ACADEMIC EMERGENCY MEDICINE March 1999, Volume 6, Number No , Kothari R, Barsan W, Brott T, Broderick J, Ashbrock S. Frequency and accuracy of prehospital diagnosis of acute stroke. Stroke. 1995; 26: Weitzman S, Cooper L, Chambless L, et al. Gender, racial, and geographic differences in the performance of cardiac diagnostic and therapeutic procedures for hospitalized acute myocardial infarction in four states. Am J Cardiol. 1997; 79: Maynard C, Althouse R, Cerqueira M, Olsufka M, Kennedy J. Underutilization of thrombolytic therapy in eligible women with acute myocardial infarction. Am J Cardiol. 1991; 68: Azzimondi G, Bassein L, Fiorani L, et al. Variables associated with hospital arrival time after stroke. Stroke. 1997; 28: Alberts M, Bertels C, Dawson D. An analysis of time of stroke presentation after stroke. JAMA. 1990; 263:65 8. REFLECTIONS Pediatric Intubation Dr. Wesley Curry intubating a pediatric patient in the Pamona Valley Medical Center Emergency Department. Photograph by ELSBURGH CLARKE, MD, Rehoboth Beach, Delaware.