Prehospital and Emergency Department Delays After Acute Stroke
The Genentech Stroke Presentation Survey
Background and Purpose—Patient delays in seeking treatment for stroke and delays within the Emergency Department (ED) are major factors in the lack of use of thrombolytic therapy for stroke. The Genentech Stroke Presentation Survey was a multicentered prospective registry of patients with acute stroke. The study was designed to characterize prehospital delays and delays within the ED.
Methods—Patients with stroke symptoms presenting to 48 EDs participating in a clinical trial of acute stroke therapy were enrolled prospectively. A 1-page data form was completed from patient interviews and medical records.
Results—A total of 1207 subjects were entered into the study. Ninety-four percent of the 721 subjects with complete data had a diagnosis of stroke or transient ischemic attack, 13% were black, 50% were female, and 67% were aged >65 years. The median time from symptom onset to ED arrival was 2.6 (interquartile range 1.2 to 6.3) hours. The median time from ED arrival until CT scan completion was 1.1 (0.7 to 1.8) hours, and the total delay time (symptom onset until CT scan completion) had a median of 4.0 (2.3 to 8.3) hours. Patients who arrived by emergency medical services had significantly shorter prehospital delay times and times to CT scan. Age, race, sex, and educational level did not appear to affect prehospital delay times.
Conclusions—Despite its limitations, this large geographically diverse study strongly suggests that the use of emergency medical services is an important modifiable determinant of delay time for the treatment of acute stroke.
- emergency medical services
- emergency service, hospital
- stroke onset
- stroke, acute
- tomography, x-ray computed
Despite the approval of thrombolytic therapy for the acute treatment of stroke, only a small percentage of patients receive this therapy.1 There are number of factors that prevent the use of this treatment more frequently. However, the most important is that patients do not present early enough to be treated within 3 hours of symptom onset.2 Patient delay may be even more important with the recent understanding that the efficacy of thrombolytic therapy with recombinant tissue plasminogen activator (rtPA) appears to be directly related to the time from symptom onset to treatment, even within the 3-hour window.3 A number of studies have examined the factors that lead to delays in the treatment of stroke patients, including delays before and after arrival at the Emergency Department (ED).4 5 6 7 8 9 However, the majority of these studies have been small (<200 subjects) and confined to one geographic or medical location. The Genentech Stroke Presentation Survey (GSPS) was a large multicentered study designed to investigate patient delays in seeking care after stroke as well as delays in diagnostic studies within the ED. This large study, with its diverse patient population, will provide important information for future strategies to reduce delays in stroke treatment.
Subjects and Methods
The GSPS was designed as a prospective registry of patients presenting with acute stroke to the ED. The 48 EDs who participated in the study were all concurrently enrolling patients in a clinical trial of rtPA treatment for acute stroke during the 3- to 5-hour window.10 From March 1997 to December 1997, all patients presenting to the study EDs with symptoms of acute stroke were enrolled in the GSPS. Trained staff from each study center completed a 1-page survey on information related to symptom onset, ED arrival, and care. Information was gathered from the patient or surrogate when possible and from the ED record itself. Each center was provided with training and written standardized instructions on how to collect the data. The forms were edited on the local level and then processed centrally at Genentech, Inc.
Symptom Onset Time
Symptom onset time was defined as the time the patient was last known to be without symptoms or at baseline. If the patient awoke with symptoms, symptom onset time was defined as when the patient went to sleep or was last known awake without symptoms.
Prehospital delay was defined as the time from symptom onset until the earliest documented time in the ED (usually triage or registration).
Delay to CT Scan
Delay to CT scan was defined as the time from ED arrival (see above) until the completion of the CT scan. This was sometimes approximated by the return of the patient to his/her bed or bay. This time did not necessarily include interpretation of the CT scan.
Time to Neurologist Consult
Time to neurologist consult was defined as the time from ED arrival until arrival of the neurology consultant in the ED.
Total delay was defined as the time from symptom onset to completion of the head CT scan.
Patients were eligible to participate in the present study if they were aged ≥18 years and presented to the ED with signs or symptoms of acute stroke. Patients were excluded if their symptom onset was >24 hours. Patients were also excluded if they were subsequently enrolled in the clinical trial of rtPA for the 3- to 5-hour window.10 (This had minimal effect, because the majority of patients presenting in the 3- to 5-hour window were ineligible for the clinical trial and hence eligible for the present study.) Patients enrolled in other clinical trials or who received rtPA outside of clinical trial were eligible for enrollment. Investigators were limited to a total enrollment of 40 subjects at each site to maximize geographic diversity.
Each study center obtained local institutional review board approval for the study. Some centers required written informed consent for the survey, but most allowed undocumented verbal consent.
Patients with missing data for symptom onset time or ED arrival time were excluded from the main analysis. Median times were reported primarily because the skewed distribution of the data and its truncation made the means less descriptive. A logistic regression model was used to help evaluate contributions of various factors associated with delay.
A total of 1207 subjects were entered into the present study. However, 483 of these had missing data for either the symptom onset time or ED arrival time. The demographic characteristics and diagnosis for the subjects included and excluded from the analysis are shown in Table 1⇓. The 2 groups were very similar, but the excluded subjects had 10% more individuals who were aged >65 years.
The demographics of the subjects included in the analysis are shown in Table 1⇑. Thirteen percent of the subjects were black, 50% were female, 31% did not have a high school diploma, and 37% were enrolled at academic hospitals. Ischemic stroke was the most common diagnosis (73%), with 13% of the patients being diagnosed with a transient ischemic attack (TIA) and 8% with intracerebral hemorrhage. Six percent of the patients had diagnoses other than stroke or TIA. Twenty-two of the patients in the study were treated with rtPA.
The median prehospital delay time in the present study was 2.6 hours, with an interquartile range of 1.2 to 6.3 hours (Table 2⇓). The mean prehospital delay time was 5.4 hours. The distribution of these delay times is shown in Figure 1⇓. Twenty-four percent of the patients arrived within the first hour, 21% arrived within the second hour, and a total of 56% arrived within 3 hours of symptom onset.
The median time from ED arrival to CT scan was 1.1 (interquartile range 0.7 to 1.8) hours, with a mean delay time of 1.9 hours. The median total delay time was 4.0 (interquartile range 2.3 to 8.3) hours, with a mean total delay time of 6.8 hours. The time to neurology consult had a median delay of 3.1 (interquartile range 1.2 to 8.2) hours. Forty-seven percent of the patients had a total delay time of ≤3 hours.
To investigate the factors associated with delay, prehospital delay, CT scan delay and total delay times are broken down by several patient characteristics in Table 3⇓. Of note, compared with black subjects, white subjects appeared to have longer prehospital delays but shorter delays to CT scan. Patients arriving by emergency medical services (EMS) had almost half the prehospital delay and three fourths of the delay to CT scan than did patients arriving at the ED by other means.
Patients alone at symptom onset had a median prehospital delay of 3.6 hours compared with 2.1 hours for those who were not alone. As expected from the definition of symptom onset time, those who awoke with symptoms had longer prehospital delays. Those patients with an elevated diastolic blood pressure on ED arrival and those patients at an academic institution also appeared to have longer prehospital delays. No obvious geographic trends were noted (north/south), but this was not examined in detail.
Because some of the factors studied were interrelated, a logistic model based on categorization of prehospital delay <2 hours and CT scan delay <1 hour was used to further define factors important in delay. Only 4 factors (race, sex, arrival by EMS, and awakening with symptoms) contributed substantially to the model. These are shown in Table 4⇓. Compared with those arriving by other means, those arriving by EMS were almost twice as likely to arrive at the ED in <2 hours. They were also 1.6 times as likely to get their CT scans in <1 hour. Those patients awakening with symptoms were a third as likely to arrive at the ED in <2 hours as opposed to those who did not awake with symptoms. None of the other factors had statistically significant relationships with delay time in this multivariate model.
The relationship between EMS use and delay is further described in Figure 2⇓, in which the percentages of patients arriving at the ED in <2 hours and having their CT scans in <1 hour are displayed by whether they arrived by EMS. Fifty-one percent of the patients arriving by EMS had prehospital delays <2 hours; 33% of those not arriving by EMS had prehospital delays <2 hours. Those arriving by EMS also had significantly shorter times to CT scan.
In the present study, the median prehospital delay was 2.6 hours, with approximately half the patients arriving at the ED in <3 hours. This figure is somewhat lower than that reported in several other studies. For example, a study performed in Houston, Tex, consisted of 241 patients with a median prehospital delay time of 4.5 hours,9 and one in Cincinnati Ohio included 119 patients with a delay time of 5.7 hours.5 Part of the reason for the shorter prehospital delay times in the present study is that the study was truncated at 24 hours, thus cutting off the small but significant number of patients who may have come in >24 hours after symptom onset. In a small study of delay in which enrollment was truncated at 36 hours,7 the median delay was 2.8 for stroke patients, which is closer to that observed in the present study.
Although several factors have been associated with shorter prehospital delay times,5 8 11 only arrival by EMS and awakening with symptoms were significantly related to prehospital delay in the present study. The median prehospital delay time for those patients arriving by ambulance was 1.9 hours shorter than those who arrived by other means. Clearly, this time differential represents much more than time saved in the actual transport. Rather, use of EMS likely prevents delays such as waiting for another family member, waiting for a ride, or waiting for a physician to call back.
Delay from ED arrival to CT scan was chosen as a relatively easily measurable point in the treatment pathway for rtPA. Although additional steps after completion of the CT scan (interpretation, physician consultation, and consent) may clearly need to be undertaken, getting the scan itself has been reported as a major obstacle.12 In the present study, the median time to CT scan completion was 1.1 hours. This is clearly above the 25 minutes to initiation of the CT scan and 45 minutes to interpretation of the CT scan, which have been set as a standard by a National Institute of Neurological Disorders and Stroke advisory committee.13
Median time to CT scan was also shorter for those patients arriving by EMS (1.0 versus 1.3 hours). This phenomenon has been cited in other studies.12 14 15 16 It may be that these patients have more severe strokes than those not arriving by EMS. What is likely more important is that these patients avoid triage, registration, and waiting for a room or bed. The physician may also perceive that these patients need to be seen sooner simply because they arrived by ambulance. In some centers, EMS may also alert the ED before arrival and allow the ED staff to prepare for the arrival and treatment of the patients.
The large number of subjects excluded because of missing data are clearly a potential problem in interpreting the results of the present study. It is reassuring, as shown in Table 1⇑, that the characteristics of those patients who were excluded are very similar to the characteristics of those patients who were included in the study. Another concern is that all of the EDs were involved in at least one clinical trial of acute stroke treatment. This may have altered their approach to stroke patients, and there may have been efforts within the communities to attract stroke patients and reduce delay time. Whereas this might affect the median times noted, it would likely not alter factors affecting delay.
Stroke severity was not measured in the present study; clearly, this may influence both prehospital delay times and time to CT scan. It is likely that greater proportions of the severe patients were transported by ambulance and would likely be seen sooner within the ED. However, the absolute number of severe patients is likely to be small. In another study of stroke delay, almost 95% of the patients presenting had Glasgow coma scales of ≥14.7
The present study offered an opportunity to examine prehospital and ED delays in a large and geographically diverse group of patients. As in other studies, a majority of patients would not arrive at the ED early enough to diagnose and treat acute stroke with rtPA or other therapies requiring a short time window. It also confirms, in a geographically diverse population with many different EMS systems, that arrival by ambulance is associated with markedly shorter prehospital and ED delay times. Because EMS use by stroke patients is a potentially modifiable behavior,17 efforts in this direction may be the most direct way to increase the number of patients eligible for acute stroke therapies. Intervention programs, which emphasize stroke recognition and appropriate response (calling 911), should be developed and critically evaluated.
The following institutions, principal investigators, and study coordinators participated in the GSPS study: Erfan A. Albakri, MD, and Liliana Vinci, RN, Columbia Brandon Regional Medical Center, Brandon, Fla; Gregory W. Albers, MD, and Nanette Hock, RN, MS, Stanford University Medical Center, Stanford, Calif; Richard P. Atkinson, MD, Deidre Wentworth, MSN, RN, and Christi Delemos, RN, Mercy General Hospital, Sacramento, Calif; Rodney D. Bell, MD, Jamie Strause, RN, BSN, Patricia Lynch, RN, and Toby R. Mazer, MPH, Thomas Jefferson University Hospital, Philadelphia, Pa; Andrew M. Blumenfeld, MD, and Carol Elliott, RNP, Southern California Kaiser Permanente Medical Center, San Diego; William H. Brooks, MD, and Patricia K. Howard, RN, MSN, CCRN, CEN, Central Baptist Hospital, Lexington, Ky; Thomas G. Brott, MD, and Judith Spilker, RN, The University of Cincinnati Hospital, Cincinnati, Ohio; David Brown, MD, and Anne Courtney-Eighmy, Washington Regional Medical Center, Fayetteville, Ark; Askiel Bruno, MD, Linda Chadwick, BSN, RN, and Carol Kempf, RN, Indiana University Medical Center, Indianapolis; John C. Brust, MD, and Norman Pflaster, MD, Harlem Hospital Center, New York, NY; James Bumgartner, MD, and Patty Brown, RN, CNRN, Roper Hospital, Charleston, SC; Paul Casale, MD, Lawrence Rodichok, MD, Lisa Hollywood, RN, Lou Anne Frey, RN, Kay Knepper, RN, and Joann Tuze, RN, Lancaster General Hospital, Lancaster, Pa; Craig Coccia, MD, Robert Glanzman, MD, Dagmar Prout, RN, MSN, CCRN, and Terry Durley, RN, BSN, CCRN, Marquette General Hospital, Marquette, Mich; John Comber, MD, and Trish Bankes, RN, BSN, Abington Memorial Hospital, Abington, Pa; Antonio Culebras, MD, and Mary Therese Dean, US Veterans Medical Center, Syracuse, NY; T.J. Curfman, MD, and Charlotte Howells, RN, Home Hospital, Lafayette, Ind; James DeMatties, MD, and Ruth Fries, BSM, ASCP, Hamot Medical Center, Erie, Pa; Dennis W. Dietrich, MD, Janet Horton, RN, and Debbie Wilson, RN, Benefits Health Care, Great Falls, Mont; Jonathan Dissin, MD, and Connie Borschell, Albert Einstein Medical Center, Philadelphia, Pa; Jay M. Ellis, DO, and Patricia Rosier, RN, MS, CS, Berkshire Medical Center, Pittsfield, Mass; William M. Feinberg, MD, and Diane Rose Taylor, University Medical Center, Tucson, Ariz; C. Stephen Ford, MD, and Jill Wright, RN, High Point Regional Hospital, High Point, NC; Michael Frankel, MD, Grady Health System, Atlanta, Ga; Thomas Giancarlo, DO, and Alice Schuster, RN, St. John Hospital and Medical Center, Detroit, Mich; Camilo R. Gomez, MD, Nancy Lemons, RN, and Keith W. Brooks, RN, University of Alabama Hospitals, Birmingham; Brad Grayum, MD, and Kathie Dermott, Crozer Chester Medical Center, Chester, Pa; Alan B. Grindal, MD, and Michael R. Halbert, PharmD, Sarasota Memorial Hospital, Sarasota, Fla; Rizwan Hassan, MD, and Edwina Hand, RN, Columbia Wesley Medical Center, Wichita, Kan; David C. Hess, MD, and Angela D. Touhey, BS, US Veterans Medical Center, Augusta, Ga; Andrea Hidalgo, MD, and R. Doreen Monks, RN, CS, MSN, CNRN, University of Medicine and Dentistry, New Jersey Medical School, Newark; Daniel Homer, MD, and Barbara Small, MBA, RN, Evanston Hospital, Evanston, Ill; Deborah Ronnie Horowitz, MD, Sandra Augustine, RN, and Jameelah Ali, RN, Mt. Sinai Medical Center, New York, NY; Steven Horowitz, MD, Joan Estrem, RN, and Anne L. Bonnett, RN, University of Missouri Hospital and Clinic, Columbia; Kinan K. Hreib, MD, PhD, Lauri Dion, RN, and Elizabeth Kennedy, Lahey Hitchcock Medical Center, Burlington, Mass; Chung Y. Hsu, MD, PhD, and Ellen Balestreri, RN, Barnes-Jewish Hospital, St. Louis, Mo; Richard L. Hughes, MD, and Clark A. Anderson, MD, University Hospital, Denver, Colo; Te-Long Hwang, MD, and Robert L. Frank, RN, EdD, Richland Memorial Hospital, Columbia, SC; Michael R.K. Jacoby, MD, and Lori Parsons, Mercy Hospital Medical Center, Des Moines, Iowa; Dara G. Jamieson, MD, and Michelle Hellstern, RN, BSN, CCRN, Pennsylvania Hospital, Philadelphia; Ronald Karlsberg, MD, David Reisbord, MD, Susan Jackman, and Tracey S. Gerez, BS, MA, Brotman Medical Center, Culver City, Calif; Scott E. Kasner, MD, Jeanne A. Dora, MSN, RN, CNRN, and Susan Unipan, BSN, RN, University of Pennsylvania Medical Center, Philadelphia; Roger E. Kelley, MD, and Colette Hilliard, Louisiana State University Medical Center, Shreveport; Teri Kintzer, MD, and Donna Post, RN, Columbia St. Petersburg Medical Center, St. Petersberg, Fla; Howard S. Kirshner, MD, and Sharon O’Connell, RN, Vanderbilt University Medical Center, Nashville, Tenn; Richard L. Koller, MD, and Pam Andrews, Abbott Northwestern, Center City, Minneapolis Minn; Dexter Koons, MD, James A. Farrier, MD, and Debbie Gobin, Parkview Episcopal Medical Center, Pueblo, Colo; Derk W. Krieger, MD, Patti Bratina, RN, and Mary Ann Young, RN, Hermann Hospital, Houston, Tex; E. Francis LaFranchise, MD, and Trixy Schrimpf, RN, Candler General Hospital, Savannah, Ga; Steven R. Levine, MD, Christopher A. Lewandowski, MD, and Patricia Marchese, RN, Henry Ford Hospital and Health Sciences Center, Detroit, Mich; David A. Lombardi, MD, and Yvonne Schooley, RN, MS, Western Medical Center Santa Ana, Santa Ana, Calif; Patrick Lyden, MD, Karen Rapp, RN, and Terri McLean, RN, UCSD Medical Center/Hillcrest, San Diego, Calif; Kenneth Madden, MD, and Lynda Stephani, Marshfield Clinic, Marshfield, Wis; Jonathan Maisel, MD, K.N. Sena, MD, and Donna Yasick, RN, Bridgeport Hospital, Bridgeport, Conn; Robert Margolis, MD, and Therese Gaertner, Christian Hospital North East, St. Louis, Mo; Lee C. Marmion, MD, and James Cwengros, PharmD, Butterworth Hospital, Grand Rapids, Mich; Thomas G. Mattio, MD, PhD, and Kathy Gendron, RN, MSN, FNPC, Theda Clark Regional Medical Center, Neenah, Wis; Jerold Mikszewski, MD, and Maureen Burke, RN, Fairfax Hospital, Falls Church, Va; Dexter L. Morris, MD, PhD, and Susan N. Pusek, RN, University of North Carolina Hospitals, Chapel Hill; Richard Mortara, MD, and Les Gardina, St. Joseph’s Hospital, Lexington, Ky; Frederick Munschauer, MD, and Sandra S. Starr, RN, Buffalo General Hospital, Buffalo, NY; Susan Nadis, MD, and Terri Flegel, RN, BSN, MacNeal Hospital, Berwin, Ill; George C. Newman, MD, University Hospital at Stony Brook, Stony Brook, NY; Aurora K. Pajeau, MD, and Annette R. Kinsella, RN, Mt. Sinai Hospital Medical Center, Chicago, Ill; Stefan M. Pulst, MD, and Brenda Williams, RN, Cedars-Sinai Medical Center, Los Angeles, Calif; Louis Rosa III, MD, Jane Gaeta, RN, and Jacky Saeger, RN, Morton Plant Hospital, Clearwater, Fla; Daniel M. Rosenbaum, MD, and J.P. Noonan, RN, Montefiore Medical Center, Bronx, NY; Vernon D. Rowe, MD, and Tricia Gentzler, RN, North Kansas City Hospital, North Kansas City, Mo; Michael O. Sauter, MD, Robert Hickey, and Phyllis Parker, RN, St. Charles Hospital and Rehabilitation Center, Port Jefferson, NY; Michael K. Sauter, MD, MSc, and Carol P. Clayton, RN, Westmoreland Hospital, Greensburg, Pa; Michael J. Schneck, MD, and Cynthia Motton, RN, BSN, Cook County Hospital, Chicago, Ill; Carol Schultz, MD, and Shirley Frederiksen, MSN, RN, University of Michigan Medical Center, Ann Arbor; Christopher Shields, MD, and Marlot Wigginton, Norton Hospital, Louisville, Ky; Richard H. Smith, MD, Vivian Noonan, RN, and Margaret Kelly, RN, Provenant St. Anthony Hospital Central, Denver, Colo; Rodney Smith, St. Joseph Mercy Hospital, Ann Arbor, Mich; Wade Smith, MD, PhD, Carol Applebury, and Nel O’Neil, RN, MSN, FNP, San Francisco General Hospital, San Francisco, Calif; Raymond M. Stephens, MD, and Kacey Hansen, RN, John Muir Medical Center, Walnut Creek, Calif; Gary A. Stobbe, MD, and Kathy E. Dalton, CMA, Stevens Memorial Hospital, Edmonds, Wash; Lynne P. Taylor, MD, and Cheryl C. Weaver, CCRC, Virginia Mason Medical Center, Seattle, Wash; Gretchen Tietjen, MD, and Andrea Korsnack, RN, Medical College of Ohio, Toledo; Anthony Turel, MD, and Edie Szoke, Geisinger Medical Center, Danville, Pa; Janet L. Wilterdink, MD, and Jo-Ann Sarafin, RN, BSN, Rhode Island Hospital, Providence; Stanley D. Wissman, MD, and Mary Ann Wissman, RN, MSN, Lutheran Hospital of Indiana, Fort Wayne, Ind; Mark L. Young, MD, and Gail Wallace, RN, St. Mary’s Medical Center, Duluth, Minn.
This project was funded by Genentech, Inc. The authors would like to acknowledge Stewart Hallett for managing the study.
Reprint requests to Dexter L. Morris, PhD, MD, Department of Emergency Medicine, CB 7594, University of North Carolina, Chapel Hill, NC 27599-7594.
Drs Morris and Rosamond have been consultants to Genentech, Inc, in the past but have no current conflict of interest. Dr Hamilton is employed by Genentech, Inc.
- Received June 8, 2000.
- Revision received June 8, 2000.
- Accepted July 14, 2000.
- Copyright © 2000 by American Heart Association
Chiu D, Krieger D, Villar-Cordova C, Kasner S, Morgenstern L, Bratina P, Yatsu F, Grotta J. Intravenous tissue plasminogen activator for acute ischemic stroke: feasibility, safety, and efficacy in the first year of clinical practice. Stroke. 1998;29:18–22.
Grotta JC. Acute stroke therapy at the millennium: consummating the marriage between the laboratory and bedside: the Feinberg lecture. Stroke. 1999;30:1722–1728.
Kwiatkowski T, Silverman R, Paiano R, Libman R. Delayed hospital arrival in patients with acute stroke. Acad Emerg Med.. 1998;3:538. Abstract.
Menon SC, Pandey DK, Morganstern LB. Critical factors in determining access to acute stroke care. Neurology. 1998;51:427–432.
Clark WM, Wissman S, Albers GW, Jhamandas JH, Madden KP, Hamilton S. Recombinant tissue-type plasminogen activator (Alteplase) for ischemic stroke 3 to 5 hours after symptom onset: the ATLANTIS Study: a randomized controlled trial: Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke. JAMA. 1999;282:2019–2026.
Feldmann E, Gordon N, Brooks J, Brass L, Fayad P, Sawaya K, Nazareno F, Levine S. Factors associated with early presentation of acute stroke. Stroke. 1993;24:1805–1810.
Bratina P, Greenberg L, Pasteur W, Grotta J. Current emergency department management of stroke in Houston, Texas. Stroke. 1995;26:409–414.
Marler J, Winters Jones P, Emr M, eds. Proceedings of a National Symposium on Rapid Identification and Treatment of Acute Stroke. Washington, DC: National Institute of Neurological Disorders and Treatment of Acute Stroke; 1997. Publication No. (NIH) 97-4239.
Williams L, Bruno A, Rouch D, Marriott D. Stroke patients’ knowledge of stroke: influence on time to presentation. Stroke. 1997;28:912–915.
Rosamond W, Morris D, Evenson K, Vallee J, Schmidt K, Cook S. Time of onset, awakening with symptoms, and delay in seeking care for acute stroke. Circulation. 1999;99:1119. Abstract.
Alberts M, Perry A, Dawson D, Bertels C. Effects of public and professional education on reducing the delay in presentation and referral of stroke patients. Stroke. 1992;23:352–356.