Emergency Medical Service Hospital Prenotification Is Associated With Improved Evaluation and Treatment of Acute Ischemic Stroke

Emergency Medical Service Hospital Is Associated With Improved Evaluation and Treatment of Acute Ischemic Stroke Cheryl B. Lin, BS; Eric D. Peterson, MD, MPH; Eric E. Smith, MD, MPH; Jeffrey L. Saver, MD; Li Liang, PhD; Ying Xian, MD, PhD; DaiWai M. Olson, PhD, RN; Bimal R. Shah, MD; Adrian F. Hernandez, MD, MHS; Lee H. Schwamm, MD; Gregg C. Fonarow, MD Background The benefits of intravenous tissue-plasminogen activator (tpa) in acute ischemic stroke are time-dependent. Emergency medical services () hospital prenotification of an incoming patient with potential stroke may provide a means of reducing evaluation and treatment times and improving treatment rates; yet, available data are limited. Methods and Results We examined 371 988 patients with acute ischemic stroke transported by and enrolled in Get With The Guidelines Stroke from April 1, 2003, to March 31, 2011. occurred in 249 197 (67.0%) of -transported patients. Among eligible patients arriving by 2 hours, patients with prenotification were more likely to be treated with tpa within 3 hours (82.8% versus 79.2%, absolute difference +3.5%, P<0.0001, the National Institutes of Health Stroke Scale-documented cohort; 73.0% versus 64.0%, absolute difference +9.0%, P<0.0001, overall cohort). Patients with prenotification had shorter door-to imaging times (26 minutes versus 31 minutes, P<0.0001), shorter door-to needle times (78 minutes versus 80 minutes, P<0.0001), and shorter symptom onset-to needle times (141 minutes versus 145 minutes, P<0.0001). In multivariable and modified Poisson regression analyses accounting for the clustering of patients within hospitals, use of prenotification was independently associated with greater likelihood of door-to imaging times 25 minutes, door-to needle times for tpa 60 minutes, onset-to needle times 120 minutes, and tpa use within 3 hours. Conclusions hospital prenotification is associated with improved evaluation, timelier stroke treatment, and more eligible patients treated with tpa. These results support the need for initiatives targeted at increasing prenotification rates as a mechanism from improving quality of care and outcomes in stroke. (Circ Cardiovasc Qual Outcomes. 2012;5:514-522.) Key Words: stroke emergency medicine services thrombolytics hospitals registry Despite advances in the field, stroke remains the fourth leading cause of death in the United States and one of the leading causes for adult disability. 1 Treatment with intravenous tissue plasminogen activator (tpa) has been shown to significantly improve stroke outcomes; yet, the benefits of tpa are strongly time-dependent. 2 5 The American Heart Association (AHA)/American Stroke Association (ASA) have set guidelines for a 3- to 4.5-hour window from symptom onset to thrombolytic therapy and recommend rapid evaluation and treatment, with a targeted time of patient arrival to the hospital facility to intravenous tpa administration (door-to needle time) of 60 minutes. 2 Achieving timely evaluation and treatment of stroke is a challenge, and currently less than one third of patients with acute ischemic stroke treated with tpa receive treatment within 60 minutes of hospital arrival. 5 The use of emergency medical services () is a potentially important means of improving medical response and treatment for acute ischemic stroke. Prenotifying hospitals that a patient with potential stroke is being transported could facilitate earlier activation of the stroke evaluation team and play an important role in increasing the number of eligible patients receiving timely tpa therapy. National recommendations from the AHA/ASA and the National Association of Received January 17, 2012; accepted June 4, 2012. From the Duke National University of Singapore Graduate Medical School, Singapore, (C.B.L.); Duke Clinical Research Institute, Durham, NC (E.D.P., L.L., Y.X., D.M.O., B.R.S., A.F.H.),; Calgary Stroke Program, University of Calgary, Calgary, Alberta (E.E.S.); Department of Neurology, University of California, Los Angeles (J.L.S.); Stroke Service, Massachusetts General Hospital, Boston, MA (L.H.S.); Division of Cardiology, University of California, Los Angeles (G.C.F.). This mansucript was handled by Peter W. Groeneveld, MD as Guest Editor. The Editors had no role in the evaluation of the article or the decision about its acceptance. The online-only Data Supplement is available at http://circoutcomes.ahajournals.org/lookup/suppl/doi:10.1161/circoutcomes.111.965210/-/dc1. Correspondence to Gregg C. Fonarow, MD, Ahmanson-UCLA Cardiomyopathy Center, Ronald Reagan UCLA Medical Center, 10833 LeConte Ave, Room 47 123 CHS, Los Angeles, CA 90095-1679. E-mailgfonarow@mednet.ucla.edu 2012 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at http://circoutcomes.ahajournals.org 514 DOI: 10.1161/CIRCOUTCOMES.112.965210

Lin et al Emergency Medical Service 515 Emergency Medical Services Physicians endorse stroke prenotification. 6,7 Prior smaller studies and those performed outside of the United States have suggested an association between stroke prenotification and more rapid assessment, faster access to imaging, and shorter door-to needle times. 8 12 ; however, there is little contemporary data that establishes the association of prenotification with im - proved timeliness of evaluation and treatment of stroke in a broad cohort of patients with stroke from throughout the United States. Using data from the AHA/ASA s Get With The Guidelines- Stroke and GWTG-Stroke registry, our study goals were to evaluate the association of prenotification with acute ischemic stroke evaluation and treatment, including door-to- -imaging times, door-to needle times, onset-to door times, and rates of tpa treatment in eligible patients. WHAT IS KNOWN Treatment with intravenous tissue plasminogen activator has been shown to significantly improve outcomes in acute ischemic stroke; yet, the benefits are highly time-dependent. Emergency Medical Service hospital prenotification of an incoming patient with potential stroke may provide a means of reducing evaluation and treatment times and improving treatment rates; however, available data are limited. WHAT THE STUDY ADDS Drawing from data collected from all regions of the country, Emergency Medical Service hospital prenotification was demonstrated to be associated with significantly more eligible patients with acute ischemic stroke being treated with tissue plasminogen activator therapy. Emergency Medical Service prenotification was also associated with shorter door-to-imaging times, shorter door-to needle times, and shorter symptom onset-to needle times in patients with acute ischemic stroke. These results support the need for initiatives targeted at increasing Emergency Medical Service prenotification rates as a mechanism from improving quality of care and outcomes in stroke. Methods Data collected from hospitals participating in AHA/ASA s GWTG Stroke were used for this study. GWTG Stroke is an initiative undertaken by the AHA/ASA to improve the quality of care delivered to patients with stroke and transient ischemic attacks through changes implemented at the hospital level. Details of the design and conduct of the GWTG Stroke program have previously been described. 5,13 GWTG uses a Web-based patient management tool (Outcome Sciences, Inc) to collect clinical data on consecutively admitted patients, provide decision support, and enable real-time online reporting features. After a yearlong pilot phase, the GWTG Stroke program was made available to all hospitals across the United States in April 2003. 5,13 Each participating hospital received either human research approval to enroll cases without individual patient consent under the common rule or a waiver of authorization and exemption from subsequent review by their institutional review board. Outcome Sciences, Inc serves as the data collection and coordination center for GWTG. The Duke Clinical Research Institute serves as the data analysis center and has an agreement to analyze the aggregate, de-identified data for research purposes. Hospital site personnel were trained to collect data on consecutive patients admitted with the principal clinical diagnosis of acute stroke or transient ischemic attack (TIA) by prospective clinical identification, retrospective identification through the use of discharge codes, or a combination. Methods used for the prospective clinical identification of cases involved regular review of a combination of data sources, including emergency department admission logs, ward census logs, intensive care unit logs, and neurology service consultations. Methods used for the retrospective clinical identification of cases included regular surveillance of discharge codes, specifically International Classification of Disease, Ninth Revision, codes 433.xx, 434.xx, and 436 for ischemic stroke. The eligibility of all acute stroke admissions was confirmed before chart abstraction. 5,13 Patient data (such as demographics, medical history, onset time of stroke symptoms [defined as last known well time], mode of arrival, arrival time, in-hospital diagnostic studies, treatments and procedures, time of initial imaging study, tpa treatment initiation time, tpa complications, in-hospital mortality, and discharge treatments, counseling and destination) were abstracted by trained hospital personnel. The National Institutes of Health Stroke Scale (NIHSS) was used to index stroke severity, when documented. Data collectors at hospital facilities indicated mode of patient arrival as from home/scene, hospital transfer, private transport/walk in, or not documented. If applicable, they further recorded if there was prenotification to hospital for the patient in question, based on explicit documentation anywhere in the records that advanced notification by had occurred and that this notification included that the patient had a suspected stroke, through the use of the word stroke or any documentation of signs and symptoms consistent with stroke. All patient data were de-identified before submission. Data on hospital-level characteristics (eg, bed size, geographical region, teaching status) was accessed from the American Hospital Association. 14 There were 936 702 patients with acute ischemic stroke from 1635 hospitals enrolled between April 1, 2003, and April 2, 2011 (Figure). Patients who had in-hospital strokes (n=15985) and those transferred in from other acute care facilities (n=94 061) were excluded for the purpose of this study. Of the remaining patients, 404 031 did not arrive by from home/scene. Another 50 637 (12.0%) patients who were transported by had missing or nonapplicable prenotification data and were therefore excluded, leaving 371988 -transported patients that constituted the study population. Statistical Analysis Sociodemographic factors, clinical variables, hospital characteristics, and quality of care measures were compared between patients transported by with prenotification versus those without. Categorical and continuous variables were reported in percentages and median (25 th and 75 th percentiles) as appropriate. The acute ischemic stroke process of care, including door-to imaging time 25 minutes (overall and confined to patients presenting within 3 hours of symptom onset), door-to needle time for intravenous tpa 60 minutes, and onset-to door time for intravenous tpa 120 minutes, use of intravenous tpa in eligible patients without documented contraindications arriving within 2 hours and treated within 3 hours, and door-to imaging, door-to needle, and onset-to needle times as continuous variables were compared between patients transported by with prenotification versus those without. The Pearson Chi 2 and Wilcoxon rank-sum test were used to compare the binary process of care measures and continuous time measures, respectively. The time intervals and proportion of eligible patients evaluated or treated within recommended time windows were also estimated, using modified Poisson regression with the generalized estimating equations (GEE) approach to account for clustering of patients within hospitals. For the time variables, log transformation

516 Circ Cardiovasc Qual Outcomes July 2012 Figure 1. Population flow diagram. The study population s derivation and key exclusions are shown, with the final study population further divided by emergency medical service () use and prenotification data. was performed first, and then regression analysis was conducted for the log-transformed time variable, with the geometric mean reported. The analyses were also examined with multivariable modified Poisson regression models and adjusted for the potential confounder variables that may have impacted the association of prenotification with stroke evaluation and treatment, including age, sex, race/ethnicity (white, black, Hispanic, Asian, and other), medical history (including atrial fibrillation/flutter, coronary artery disease/ prior myocardial infarction, carotid stenosis, diabetes mellitus, hyperlipidemia, heart failure, hypertension, smoking status, prior stroke/tia, peripheral vascular disease, and prosthetic heart valve), insurance status (Medicare, Medicaid, no insurance, or other, including health maintenance organization/veterans Affairs/private), arrival during regular working hours (defined as Monday to Friday between 7AM and 6PM), hospital bed size, geographic region (Midwest, Northeast, West, South), teaching status, average number of ischemic stroke/tia patients annually, and average number of patients treated with tpa annually. These relationships were also examined with multivariable logistic regression models using the GEE approach to adjust for within-hospital clustering. Because the degree of stroke severity is a potential factor impacting both prenotification and stroke care but was not available in all patients, sensitivity analyses were performed on the subset of patients with complete NIHSS data available (N=199 154), including NIHSS as an additional covariate in the multivariable modified Poisson regression models using GEE. Similar analyses were conducted to investigate the relationship between prenotification and complications of intravenous tpa therapy (symptomatic intracranial hemorrhage within 36 hours and major bleeding). GEE was similarly employed in these models to account for within-hospital clustering, and the same set of previously mentioned potential confounders were used for risk adjustment. All tests were 2-tailed, with P<0.05 considered as the level of statistical significance. All statistical analyses were performed using SAS version 9.2 software (SAS Institute). The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written. Results This study analyzed 371 988 patients with acute ischemic stroke from 1585 participating sites arriving by transport from 2003 to 2011. Of patients transported by during the 8-year study period, hospital prenotification occurred in 249 197 (67.0%). When the analysis is confined to patients arriving by within 4.5 hours from symptom onset, prenotification occurred in 97 511 of 135 308 patients (72.1%). Among patients arriving with symptom duration 2 hours, the rate of prenotification was 72 407 of 99 145 patients (73.0%). Patient and hospital characteristics for those with prenotification compared with those without are provided in Table 1. Among -transported patients, was more likely to prenotify hospitals for patients who were younger, white, male, and had a history of atrial fibrillation. Patients with a history of previous stroke/tia, diabetes mellitus, peripheral vascular disease, hypertension, dyslipidemia, or heart failure were less likely to have prenotification. Time from symptom onset to hospital arrival was shorter in patients with prenotification. prenotification also tended to occur for patients with higher NIHSS scores and those arriving during facility off-hours. Hospital characteristics also differed among those patients with and without prenotification (Table 1). Differences were observed in timing of evaluation, timing of treatment, and the portion of eligible patients treated with tpa between -arriving patients with and without prenotification. Patients with prenotification underwent more rapid evaluation and, when treated, more timely tpa administration (Table 2). Among patients arriving within 3 hours of symptom onset, the median door-to imaging times were less (26 minutes versus 31 minutes, P<0.0001) and the portion of patients with door-to imaging times within 25 minutes was higher with prenotification compared with patients without prenotification (48.8% versus 40.5%, P<0.0001). With prenotification, the onset-to door

Lin et al Emergency Medical Service 517 Table 1. Baseline Patient and Hospital Characteristics by Status Age, median (25 th to 75 th percentile), y n=249 197 No n=122 791 76 (64 84) 76 (64 84) Male, % 46.6 45.7 Race/ethnicity, % White, non-hispanic 74.6 70.2 Black 13.8 18.1 Hispanic 5.6 5.4 Asian 2.3 2.2 Other 3.7 4.1 Insurance status, % HMO/Private 26.1 26.7 Medicare 55.2 54.4 Medicaid/military/VA 5.7 6.7 Self/none 3.6 3.8 NIHSS, median (25 th to 75 th percentile) NIHSS, % 7 (3 15) 6 (3 13) >25 2.9 2.1 21 25 4.0 2.8 16 20 6.1 4.4 11 15 7.8 5.9 6 10 12.2 10.0 0 5 23.4 22.5 Not documented 43.5 52.4 Arrival off-hours, %* 52.5 52.1 Time from symptom onset to arrival, median, minutes Medical history, % 113 (55 340) 150 (60 445) Atrial fibrillation/flutter 23.4 22.3 Prosthetic heart valve 1.6 1.6 Previous stroke/tia 34.3 35.1 Coronary artery disease/prior MI 30.5 30.0 Carotid stenosis 4.3 4.4 Diabetes mellitus 31.3 33.3 Peripheral vascular disease 5.1 5.7 Hypertension 80.7 81.8 Smoker 18.1 18.0 Dyslipidemia 39.6 40.4 Heart failure 6.0 6.6 Medications prior to admission, % Antihypertensive 70.5 71.1 Cholesterol reducer 37.9 38.2 Diabetic medication 23.2 24.6 (Continued) Table 1. Continued Hospital characteristics Annual volume of ischemic stroke/tia admissions, % n=249 197 n=249 197 301+ 45.7 42.8 101 300 45.6 46.0 0 100 8.7 11.1 Annual volume of tpa administration, % 11+ 23.2 19.4 7 10 35.2 32.8 0 6 41.6 47.9 Hospital size, median (25 th to 75 th percentile), beds Hospital type, % 362 (260 546) 365 (250 524) Academic 53.4 60.9 Hospital region, % West 20.9 12.6 South 35.3 34.7 Midwest 19.3 15.3 Northeast 24.5 37.5 indicates emergency medical services; HMO, health maintenance organization; VA, Veterans Affairs; NIHSS, National Institutes of Health Stroke Scale; TIA, transient ischemic attack; MI, myocardial infarction; tpa, tissue plasminogen activator. *Arrival at the hospital that did not occur during Monday to Friday, 7AM to 6PM. NIHSS values were recorded in 199154 patients. Sex was missing in 0.08%, race/ethnicity in 0.06%, medical history in 6.5%, teaching status in 4.4%, and bed size in 6.5%. times were lower (113 minutes versus 150 minutes, P<0.0001), and more eligible patients were treated with tpa within 120 minutes of symptom onset (31.9% versus 29.5%, P<0.0001) compared with those patients without prenotification. There was a greater portion of eligible patients arriving within 2 hours who were treated with tpa within 3 hours compared with patients without prenotification (73.0% versus 64.0%, absolute difference +9.0%, P<0.0001). Adjustment for potential confounding variables and accounting for the correlation of data within hospitals demonstrates that prenotification was independently associated with earlier imaging and more rapid tpa administration (Table 3). prenotification was associated with improved timeliness in imaging among patients arriving within 3 hours (door-to imaging time 25 minutes) and improved timeliness of tpa administration by both door-to needle times (door-to needle time 60 minutes) and onset-to needle times (onsetto needle time 120 minutes). prenotification was also independently associated with more eligible patients with acute ischemic stroke being treated with tpa, adjusted rates 71.8% versus 62.2% (absolute difference +9.6%, P<0.0001), with and without prenotification (Table 3). After multivariate adjustment, prenotification patients arriving

518 Circ Cardiovasc Qual Outcomes July 2012 Table 2. In-Hospital Evaluation and Treatment of Patients With Acute Ischemic Stroke With Compared With Patients Without Door-to imaging time, n, median (25 th to 75 th percentile), min (in patients arriving 3 h) No 76 459 28 220 Absolute Difference (95% CI) P Value 26 (16 45) 31 (18 56) 235 min <0.0001 Door-to imaging time 25 min, (in patients arriving 3 h), % 48.8% 40.5% +8.3% (7.6 8.9) <0.0001 Door-to imaging time, n, median (25 th to 75 th percentile), min 230 430 112 580 42 (22 83) 55 (28 103) 13 min <0.0001 Door-to imaging time 25 min, % 30.9% 22.4% +8.5% (8.2 8.8) <0.0001 Door-to needle time, median (25 th to 75 th percentile), min 78 (60 100) 80 (60 103) 2 min <0.0001 Door-to needle time 60 min, % 27.0% 25.9% +1.1% (0.0 2.1) 0.0583 Onset-to needle time, median (25 th to 75 th percentile), min 141 (115 169) 145 (116 170) 4 min <0.0001 Onset-to needle time 120 min, % 31.9% 29.5% +2.4% (1.2 3.5) <0.0001 tpa Rx rate (arrive by 2 h, treat by 3 h), n/n, % 22 305/30 541 7193/11 244 73.0% 64.0% +9.0% (8.0 10.1) <0.0001 indicates emergency medical services; RR, relative risk; CI, confidence interval; tpa, tissue plasminogen activator; and Rx, treatment. Table displays crude rates and times. within 2 hours of symptom onset had greater odds of receiving treatment with intravenous tpa within 3 hours from symptom onset. Similar results were obtained using the multivariable GEE models (online-only Data Supplement Table I). When the analyses were repeated with adjustment for NIHSS, using data from the 199 154 patients with NIHSS documented, the results were similar, though differences in tpa treatment were more modest (Table 4). Among eligible patients arriving by 2 hours, patients with prenotification were more likely to be treated with tpa within 3 hours (82.8% versus 79.2%, absolute difference +3.5%, P<0.0001). prenotification was independently associated with more eligible patients with acute ischemic stroke being treated with tpa, adjusted rates 82.3% versus 78.1% (absolute difference +4.2%, P<0.0001), with and without prenotification. Adjustment for potential confounding variables and accounting for the correlation of data within hospitals demonstrates that prenotification was independently associated with earlier imaging, more rapid tpa administration, and more eligible patients treated with tpa, as shown in Table 5. Similar results were obtained using the multivariable GEE models (online-only Data Supplement Table II). No significant association was found between prenotification and complications from intravenous tpa, such as symptomatic intracranial hemorrhage 36 hours, or lifethreatening systemic hemorrhage 36 hours, or any serious Table 3. Unadjusted and Adjusted Rates, Times, and Relative Risks for Stroke Evaluation and Treatment in Patients With Compared With Patients Without Measure Door-to imaging time, median, (in patients arriving min 3 h), min Door-to imaging time 25 min (in patients arriving 3 h), % Unadjusted Rates and Times No P Value Adjusted Rates and Times No Difference (95% CI) P Value Adjusted Relative Risk RR 95% CI P Value 24.0 31.2 <0.0001 23.5 30.4 6.9 min (5.6 8.1) <0.0001......... 48.5% 38.2% <0.0001 49.7% 39.6% +10.2% (8.7 11.6) <0.0001 1.26 1.21 1.30 <0.0001 Door-to imaging time, median, min 39.5 54.3 <0.0001 39.8 54.3 14.4 min (12.6 16.3) <0.0001......... Door-to imaging time 25 min 31.9% 21.5% <0.0001 32.5% 22.3% +10.2% (9.0 11.3) <0.0001 1.45 1.39 1.52 <0.0001 Door-to needle time, median, min 78.2 83.0 <0.0001 76.2 80.7 4.5 min (2.8 6.2) <0.0001......... Door-to needle time 60 min 24.5% 21.1% <0.0001 25.9% 22.7% +3.3% (1.8 4.8) <0.0001 1.14 1.07 1.22 <0.0001 Onset-to needle time, median, min 140.0 145.0 0.0002 140.0 143.8 3.8 min (1.6 6.0) <0.0001......... Onset-to needle time 120 min 31.5% 28.0% <0.0001 31.5% 28.3% +3.2% (1.8 4.6) <0.0001 1.11 1.06 1.17 <0.0001 tpa Rx rate (arrive by 2 h, treat by 3 h) 65.9% 56.2% <0.0001 71.8% 62.2% +9.6% (7.8 11.5) <0.0001 1.16 1.12 1.19 <0.0001 indicates emergency medical services; CI, confidence interval; min, minutes; h, hours; Rx, treatment. Modified Poisson regression with robust variance estimators to account for clustering of patients within hospitals. Variables included in multivariable models were age, sex, race/ethnicity, prior medical history of atrial fibrillation/flutter, coronary artery disease/prior myocardial infarction, carotid stenosis, diabetes mellitus, hyperlipidemia, heart failure, hypertension, smoking status, prior stroke/transient ischemic attack, peripheral vascular disease, and prosthetic heart valve, insurance status, arrival during on/off hours, hospital characteristics of bed size, geographic region, teaching status, average number of patients with ischemic stroke/transient ischemic attack annually, and average number of patients treated with tissue plasminogen activator annually. The adjusted times represent the geometric means.

Lin et al Emergency Medical Service 519 Table 4. In-Hospital Evaluation and Treatment of Acute Ischemic Stroke Patients With Compared With Patients Without Among Patients With NIHSS Documented (N=199 154) Door-to imaging time, n, median (25 th to 75 th percentile), min (in patients arriving 3 h) Door-to imaging time 25 min, (in patients arriving 3 h), % Door-to imaging time, n, median (25 th to 75 th percentile), min No 54 983 17 838 Absolute Difference (95% CI) P Value 24 (15 38) 26 (16 43) 2 min <0.0001 54.8% 48.5% +6.3% (5.4 7.1) <0.0001 132 374 54 462 32 (18 63) 41 (22 80) 9 min <0.0001 Door-to imaging time 25 min, % 69.3% 60.5% +8.8% (8.3 9.2) <0.0001 Door-to needle time, n, median (25 th -75 th percentile), min 77 (59 99) 79 (60 101) 2 min <0.0001 Door-to needle time 60 min, % 27.7% 26.9% +0.8% (0.4 2.0) 0.1787 Onset-to needle time, median (25 th to 75 th percentile), min 140 (114 168) 145 (116 170) 5 min <0.0001 Onset-to needle time 120 min, % 32.4% 29.8% +2.6% (1.4 3.9) <0.0001 tpa Rx rate (arrive by 2 h, treat by 3 h), n/n, % 19 476/23 531 6017/7594 82.8% 79.2% +3.5% (2.5 4.6) <0.0001 indicates emergency medical services; NIHSS, National Institutes of Health Stroke Scale; tpa, tissue plasminogen activator; and Rx, treatment. Table displays crude rates and times. tpa complication on unadjusted or adjusted analyses (Table 6). Similar results were found in the sensitivity analyses adjusting for NIHSS (data not shown). Discussion Our analysis supports the role of prenotification as a potentially important but underused means to improving rapid triage, evaluation, and treatment of patients with acute ischemic stroke. Although previous studies have shown that use may speed therapy in selected settings, 8 12 our study expands these findings to the stroke patients from all regions of the country and diverse hospital settings. Among patients transported via, we found prenotification to be independently associated with more patients undergoing timely evaluation, shorter onset-to needle and door-to needle times, and more eligible acute ischemic stroke patients being treated with tpa. Despite the proven benefits of timely administration of tpa for acute ischemic stroke and national targets for timely evaluation and treatment, many hospitals still find difficulty in meeting these performance goals. 5 Although delays from the time of stroke symptom onset to patient arrival (prehospital delay) remains a barrier to timely tpa administration, a substantial portion of patients arrive within a time window in which intravenous tpa could be administered. 13 Identifying and implementing effective strategies to facilitate the rapid evaluation and treatment of patients with acute ischemic Table 5. Unadjusted and Adjusted Rates, Times, and Relative Risks for Stroke Evaluation and Treatment in Patients With Compared With Those Without Among Patients With NIHSS Documented (N=199 154) Measure Door-to imaging time, median, (in patients arriving min 3 h), min Door-to imaging time 25 min (in patients arriving 3 h), % Unadjusted Rates and Times No P Value Adjusted Rates and Times No Difference (95% CI) P Value Adjusted Relative Risks RR 95% CI P Value 19.8 24.5 <0.0001 19.9 24.4 4.5 min (3.4 5.5) <0.0001......... 56.3% 47.3% <0.0001 56.7% 48.4% +8.3% (6.8 9.8) <0.0001 1.17 1.14 1.21 <0.0001 Door-to imaging time, median, min 28.3 38.8 <0.0001 29.3 38.7 9.4 min (7.8 11.1) <0.0001......... Door-to imaging time 25 min 43.2% 31.6% <0.0001 44.1% 33.7% +10.4% (8.9 11.9) <0.0001 1.31 1.26 1.36 <0.0001 Door-to needle time, median, min 76.8 81.2 <0.0001 75.1 79.1 4.0 min (2.1 5.7) <0.0001......... Door-to needle time 60 min 25.4% 21.9% 0.0002 26.7% 23.4% +3.3% (1.6 5.0) 0.0003 1.14 1.06 1.22 0.0003 Onset-to needle time, median, min 140.0 144.0 0.0006 139.4 142.7 3.3 min (1.0 5.6) 0.0041......... Onset-to needle time 120 min 32.2% 28.3% <0.0001 31.9% 28.5% +3.4% (1.9 4.9) <0.0001 1.12 1.06 1.18 <0.0001 tpa Rx rate (arrive by 2 h, treat by 3 h) 80.0% 75.2% <0.0001 82.3% 78.1% +4.2% (2.8 5.7) <0.0001 1.05 1.04 1.07 <0.0001 indicates emergency medical services; CI, confidence interval; RR, relative risk; tpa, tissue plasminogen activator; Rx, treatment. Modified Poisson regression with robust variance estimators to account for clustering of patients within hospitals. Variables included in multivariable models were National Institutes of Health Stroke Scale and the variables listed in Table 3.

520 Circ Cardiovasc Qual Outcomes July 2012 Table 6. Complications of tpa Use in Patients With Compared With Patients Without : Unadjusted Rates and Adjusted Relative Risk tpa Complication Rates (Unadjusted) No P Value Adjusted Relative Risk Relative Risk 95% CI P Value Symptomatic intracranial hemorrhage 36 h 5.8% 6.0% 0.4020 0.97 0.87 1.08 0.6095 Life-threatening or serious systemic hemorrhage 36 h 1.4% 1.3% 0.2705 1.11 0.88 1.39 0.3770 Other serious complications 36 h 2.1% 1.9% 0.1686 1.15 0.94 1.51 0.1741 Any serious complications 11.6% 11.5% 0.7793 1.02 0.95 1.10 0.5512 indicates emergency medical services; CI, confidence interval. Adjusted for variables listed in Table 3. stroke is critical. 6 Prearrival notification is one such strategy, and prior studies have suggested that this approach may facilitate more timely stroke evaluate and care. 8 12,15 These smaller studies generally found prenotification was associated with faster imaging and imaging interpretation times and increased patient eligibility for tpa. Abdullah et al identified a study population of 118 patients arriving at the Massachusetts General Hospital emergency department by within 6 hours of symptom onset, with an prenotification rate of 37%. 15 This analysis found significantly shorter door-to imaging times and increased occurrence of thrombolysis in the prenotification group compared with no prenotification. Multivariate linear regression modeling found that advance notification by was associated with a 23% shorter door-to imaging times. 15 In one of the largest studies to date, Patel et al studied stroke patients in North Carolina (n=13 894) and found that prenotification increased the number of patients undergoing imaging 25 minutes and imaging interpretation 45 minutes compared with those patients arriving by private transport. 12 Our findings establish the benefits of prenotification in a much larger and diverse group of patients and demonstrate not just an association with improved imaging times but improved timeliness of tpa administration and more eligible patients treated with tpa compared with those patients arriving by without prenotification. This GWTG Stroke study also demonstrates that prenotification is generally underused in contemporary US practice. Our results suggest that improved rates of use and prenotification for patients with ischemic stroke may improve the speed of patient evaluation, allow for more eligible patients to be treated with tpa, and facilitate more timely tpa administration. 6 As the benefits of tpa are highly time-dependent and, for every minute a large vessel stroke goes untreated, as many as 1.9 million neurons and 14 billion synapses are potentially lost, even small differences in door-to needle and onset-to needle times may produce clinically relevant differences in stroke functional outcomes. 4,16 These collective data support the need for targeted improvements in rates of prenotification and awareness of the benefits of prenotification, as part of local, regional, and national initiatives to improve stroke care and outcomes. Increasing the portion of patients evaluated and treated with tpa in a timely fashion requires a systems approach that is operative at 3 major levels. 6,7 First, increasing patient stroke symptom recognition and rapid activation of is important to address. Community education initiatives should be undertaken to inform the public of the outcome benefits associated with use. 6 Second, adequate training of staff in the proper use of stroke screening instruments for timely identification of ischemic stroke together with hospital prenotification is vital. As the first form of medical contact, trained personnel have the opportunity to stabilize the patient, prenotify the nearest hospital equipped to administer tpa in the appropriate circumstances, and provide rapid transport to this facility. 6,7,11 Previous studies have suggested that such training should focus on defining the minimum thresholds and type of data required for activation of hospital-based stroke teams. 6 Furthermore, standardized tools necessary for acute stroke detection in the ambulatory setting should be developed and implemented across the board. 6,15 Finally, implementation of systems of care within the receiving hospital is crucial. Approaches to provide timely evaluation and treatment include organized stroke teams, written protocols for acute triage and patient flow, single call systems to activate all stroke team members, computed tomography (CT) or magnetic resonance imaging (MRI) scanner clearance as soon as center is made aware of an incoming patient, locating the CT scanner in the Emergency Department, storage and rapid access to thrombolytic drug in the Emergency Department, collaboration in developing treatment pathways among physician, nurses, pharmacists, and technologists from Emergency Medicine, Neurology, and Radiology, and continuous data collection to drive iterative system improvement. 5 prenotification facilitates earlier activation of stroke team personnel and the preparation of imaging modalities before patient arrival. 6,11 Appropriate response within hospital facilities is necessary to derive the benefits of prenotification. Limitations There are several limitations to this study that need to be acknowledged. First, although GWTG Stroke covers >1000 US hospitals, participation is voluntary and could, in fact, select for centers with greater interest in quality improvement and, thus, better process performance than non- GWTG Stroke centers. 5,13 Second, as with all registries, the data presented here are dependent on the accuracy and completeness of abstraction from the medical records. In some patients, prenotification may have taken place but not

Lin et al Emergency Medical Service 521 been documented in the medical record. In other patients, specific stroke patient prenotification may have been documented but not actually occurred. The assessment of prenotification was confined to present or absent and additional information that may have been communicated to hospitals such as prehospital stroke screening results and type of symptoms were not captured. We also do not have any data on the type of first responder, the level of training of the rendering prehospital providers, or characteristics of the agencies. Additionally, participating hospitals are instructed to include all consecutive admissions for ischemic stroke; however, since these processes are not audited, the potential exists for selection bias. Although our analysis has identified a number of patient and hospital characteristics that may impact stroke care, there are likely other important influencing factors not captured. Residual measured and unmeasured confounding variables may have influenced some or all of the findings. NIHSS was not documented in 33.4% of patients; however, findings were similar in models with and without NIHSS, although the absolute difference in tpa treatment in eligible patients was smaller in models adjusting for NIHSS. As a result of the large sample size, some small differences in absolute terms and of uncertain clinical significance are still highly statistically significant. Finally, measures of stroke functional outcomes such as modified Rankin score at discharge or 90 days are not available in this study. Conclusions Among patients with acute ischemic stroke, transported by to GWTG Stroke participating hospitals, prenotification is independently associated with more timely imaging and administration of tpa, and a greater portion of eligible patients with acute ischemic stroke receiving tpa. Despite these benefits, prenotification occurred in only two thirds of patients with acute ischemic stroke arriving by. These findings support the need for initiatives targeted at improving prenotification rates nationally as a mechanism for improving acute ischemic stroke quality of care and outcomes. Sources of Funding The GWTG Stroke program is provided by the American Heart Association/American Stroke Association. The GWTG Stroke program is currently supported, in part, by a charitable contribution from Bristol-Myers Squibb/Sanofi Pharmaceutical Partnership and the American Heart Association Pharmaceutical Roundtable. GWTG Stroke has been funded in the past through support from Boeringher- Ingelheim and Merck. Disclosures Dr Smith serves as a member of the GWTG Science Subcommittee and receives research support from the National Institutes of Health (NIH) (NINDS R01 NS062028), the Canadian Stroke Network, the Canadian Institutes of Health Research, and the Heart and Stroke Foundation of Canada, and has served on an advisory board for Genentech. Dr Saver serves as a member of the GTWG Science Subcommittee; as a scientific consultant regarding trial design and conduct to CoAxia, Concentric Medical, Talecris/Grifols, Ferrer, PhotoThera, BrainsGate, SYGNIS, and ev3; received lecture honoraria from Ferrer; and is an employee of the University of California, which holds a patent on retriever devices for stroke. Dr Liang is a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Dr Olson is a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Dr Hernandez is a member of the Duke Clinical Research Institute, which serves as the AHA GWTG data coordinating center. Dr Hernandez is a recipient of an AHA Pharmaceutical Roundtable grant (0675060N); has received a research grant from Johnson & Johnson; and has received honorarium from AstraZeneca and Amgen. Dr Peterson has received research grants from Eli Lilly and Company, Johnson & Johnson, and Bristol-Myers Squibb, Sanofiaventis, and the Merck/Schering-Plough partnership. Dr Peterson serves as principle investigator of the Data Analytic Center for GTWG. Dr Schwamm serves as chair of the GWTG Steering Committee; serves as a consultant to the Research Triangle Institute and to the Massachusetts Department of Public Health; and serves on the Steering Committee for Lundbeck s DIAS4 clinical trial. Dr Fonarow receives research support from the National Institutes of Health (significant) and is an employee of the University of California, which holds a patent on retriever devices for stroke (significant). The other authors report no conflicts. Role of the Sponsors: The industry sponsors had no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript. References 1. Towfighi A, Saver JL. Stroke declines from third to fourth leading cause of death in the United States: historical perspective and challenges ahead. Stroke. 2011;42:2351 2355. 2. Adams HP Jr, del Zoppo G, Alberts MJ, Bhatt DL, Brass L, Furlan A, Grubb RL, Higashida RT, Jauch EC, Kidwell C, Lyden PD, Morgenstern LB, Qureshi AI, Rosenwasser RH, Scott PA, Wijdicks EF. Guidelines for the early management of adults with ischemic stroke. Circulation. 2007;115:e478 e534. 3. The National Institute of Neurological Disorders and Stroke rt-pa Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333:1581 1587. 4. Lees KR, Bluhmki E, von Kummer R, Brott TG, Toni D, Grotta JC, Albers GW, Kaste M, Marler JR, Hamilton SA, Tilley BC, Davis SM, Donnan GA, Hacke W; ECASS, ATLANTIS, NINDS and EPITHET rt-pa Study Group, Allen K, Mau J, Meier D, del Zoppo G, De Silva DA, Butcher KS, Parsons MW, Barber PA, Levi C, Bladin C, Byrnes G. Time to treatment with intravenous alteplase and outcome in stroke: an updated pooled analysis of ECASS, ATLANTIS, NINDS, and EPITHET trials. Lancet. 2010;375:1695 1703. 5. Fonarow GC, Smith EE, Saver JL, Reeves MJ, Bhatt DL, Grau-Sepulveda MV, Olson DM, Hernandez AF, Peterson ED, Schwamm LH. Timeliness of tissue-type plasminogen activator therapy in acute ischemic stroke: patient characteristics, hospital factors, and outcomes associated with door-to needle times within 60 minutes. Circulation. 2011;123:750 758. 6. Acker JE 3rd, Pancioli AM, Crocco TJ, Eckstein MK, Jauch EC, Larrabee H, Meltzer NM, Mergendahl WC, Munn JW, Prentiss SM, Sand C, Saver JL, Eigel B, Gilpin BR, Schoeberl M, Solis P, Bailey JR, Horton KB, Stranne SK; American Heart Association; American Stroke Association Expert Panel on Emergency Medical Services Systems, Stroke Council. Implementation strategies for emergency medical services within stroke systems of care: a policy statement from the American Heart Association/American Stroke Association Expert Panel on Emergency Medical Services Systems and the Stroke Council. Stroke. 2007;38:3097 3115. 7. Crocco TJ, Grotta JC, Jauch EC, Kasner SE, Kothari RU, Larmon BR, Saver JL, Sayre MR, Davis SM. management of acute stroke prehospital triage (resource document to NAP position statement). Prehosp Emerg Care. 2007;11:313 317. 8. Bae HJ, Kim DH, Yoo NT, Choi JH, Huh JT, Cha JK, Kim SK, Choi JS, Kim JW. Prehospital notification from the emergency medical service reduces the transfer and intra-hospital processing times for acute stroke patients. J Clin Neurol. 2010;6:138 142. 9. Kim SK, Lee SY, Bae HJ, Lee YS, Kim SY, Kang MJ, Cha JK. Prehospital notification reduced the door-to needle time for iv t-pa in acute ischaemic stroke. Eur J Neurol. 2009;16:1331 1335.

522 Circ Cardiovasc Qual Outcomes July 2012 10. Mosley I, Nicol M, Donnan G, Patrick I, Kerr F, Dewey H. The impact of ambulance practice on acute stroke care. Stroke. 2007;38:2765 2770. 11. Rajajee V, Saver J. Prehospital care of the acute stroke patient. Tech Vasc Interv Radiol. 2005;8:74 80. 12. Patel MD, Rose KM, O Brien EC, Rosamond WD. Prehospital notification by emergency medical services reduces delays in stroke evaluation: findings from the North Carolina stroke care collaborative. Stroke. 2011;42:2263 2268. 13. Fonarow GC, Reeves MJ, Smith EE, Saver JL, Zhao X, Olson DW, Hernandez AF, Peterson ED, Schwamm LH; GWTG-Stroke Steering Committee and Investigators. Characteristics, performance measures, and in-hospital outcomes of the first one million stroke and transient ischemic attack admissions in get with the guidelines-stroke. Circ Cardiovasc Qual Outcomes. 2010;3:291 302. 14. American Hospital Association. American Hospital Association Hospital Statistics 2009. Chicago, IL: American Hospital Association; 2009. 15. Abdullah AR, Smith EE, Biddinger PD, Kalenderian D, Schwamm LH. Advance hospital notification by in acute stroke is associated with shorter door-to computed tomography time and increased likelihood of administration of tissue-plasminogen activator. Prehosp Emerg Care. 2008;12:426 431. 16. Saver JL. Time is brain quantified. Stroke. 2006;37:263 266.