- Email: [email protected]
Incidence of Cardiac Arrest During Self-Transport for Chest Pain
GENERAL CLINICAL INVESTIGATION/BRIEF REPORT
Incidence of Cardiac Arrest During Self-Transport for Chest Pain From the King County Emergency Medical Services Division* and the Centerfor Evaluation of Emergency Medical ServiceS:, Seattle-King County Department of Public Health; and the Department of Medicine, University of Washington~, Seattle, Washington. Receivedfor publication November 20, 1995. Revisions receivedApri110, 1996, and July I5, 1996. Acceptedfor publicationJuly 25, 1996. Copyright © by the American College of Emergency Physicians.
Linda Becker, MA* Mary Pat Larsen, MS** Mickey S Eisenberg, MD, PhD*§
Study objective: To assess the incidence of cardiac arrest among patients who use self-transport to seek medical care for chest pain. Methods: This was a retrospective cohort study of patients admitted to a CCU for suspected acute myocardial infarction (AMI) and patients experiencing out-of-hospital cardiac arrest preceded by symptoms in King County,Washington, between January 1, 1992, and July 31, 1994. Participants were identified through use of the databases compiled by the Myocardial Infarction Triageand Intervention Trial, which reviewed medical records in all area hospitals, and the Cardiac Arrest Surveillance System, which tracks all incidences in which CPR is performed by EMS personnel in King County.Patientswhose sudden cardiac arrests were not preceded by symptoms were excluded. Hospital records were abstracted to find the means of transport for patients admitted to CCUs. For cardiac arrest patients, the medical history, presenceof symptoms, means of transport, and prehospital death information were abstracted from paramedic field reports. Outcome (admission, discharge, or in-hospital death) was obtained from hospital records. An event cause (cardiac or other) was determined from death certificates, hospital records, or consultation with private physicians. Results: During the 30-month study period, 13,187 patients sought help for cardiac symptoms and were either admitted to a CCU or died before admission after calling 911. A majority, 7,393 (59%), were transported by emergency medical services, and 5,182 (41%) used private transportation to obtain medical care; the means of transport could not be determined for 612 patients. Of the EMS group, 6,978 were admitted to the hospital without experiencing prehospital cardiac arrest, and 415 (5.6%)arrested before arriving at the hospital. Of the group using private transportation, 5,164 were admitted without arresting and 18 (.35%) arrested before arrival, after which 911 was called (P<.0Ol).
6 12
ANNALS OF EMERGENCY MEDICINE
28:6
DECEMBER 1996
SELF-TRANSPORT t~ec/zer et al
Conclusion: The incidence of cardiac arrest among patients
who attempted to reachthe hospital by private transportationwas very low compared with the incidence among those who chose the EMS system for transport. This suggests that patient selfselection occurs, with the more seriously ill patients more commonly calling 91 1 for transport. [Becker L, Larsen MP, Eisenberg MS: Incidence of cardiac arrest during self4ransport for chest pain. Ann ErnergMedDecember 1996;28:612-616]
INTRODUCTION Persons who seek medical care for chest pain must decide on a mode of transportation. Should they call 911, or should they drive to the hospital or have a friend or relative drive? The National Heart Attack Alert Program ~ and other educational campaigns ~,3 urge people with chest pain to call 911. The rationale is that paramedic personnel can start treatment with oxygen, provide analgesia and antiarrhythmic therapy, and perform defibrillation if needed. The treatment can begin within a few minutes of symptom onset. Despite this advice, many patients who experience chest pain choose to transport themselves. The purpose of this study was to determine the incidence of cardiac arrest during self-transport for chest pain.
MATERIALS AND METHODS Cases included in the study were those of persons who had experienced chest pain or other potentially serious cardiac symptoms and who had made a decision about transport to seek medical care. Patients included those admitted to a CCU with a diagnosis of suspected acute myocardial infarction (AMI) without arresting and those who experienced cardiac arrest before arrival at the hospital, either before or after arrival of emergency medical services. Patients whose cardiac arrests were not of primary cardiac origin, such as arrests resulting from suicide, trauma, respiratory disease, or cancer, were excluded, as were patients who arrested in the emergency department and died before admission. Also excluded were those patients who had no symptoms and collapsed without warning; these patients were not able to make a decision about transport, and 911 was caIled in every case. We retrospectively studied cases in King County, Washington (study population approximately 1,000,000 in 1992) that occurred between January 1, 1992, and July 31, 1994. Data were: drawn from two distinct, concurrent and
DECEMBER 1996
28:6
ANNALS OF EMERGENCY MEDICINE
complementary databases. The Cardiac Arrest Surveillance System (CASS) provided information on cardiac arrests. ~ Patients in this database had out-of-hospital cardiac arrest with CPR performed by EMS personnel. This data collection system began with the implementation of paramedic services in King County in 1976; definitions and coding procedures have remained the same since the beginning. Through a dual reporting system whereby a separate report of each CPR case is received from emergency medical technicians and paramedics, virtually 100% of relevant cases were identified. Data recorded were time of collapse; time CPR was begun; whether the event was wimessed; whether bystander CPR was performed; and times to the following: defibrillation, intubation, blood pressure, placement of IV tube, administration of medication, return of spontaneous circulation, and termination of efforts. The records of cardiac arrest patients admitted to a hospital were reviewed to determine whether they were discharged alive and the destination of discharge (home or nursing home). An event cause was assigned based on consideration of several factors: prior history, medications, symptoms, and circumstances of arrest. The discharge diagnosis or cause of death was obtained from the medical record or from the death certificate, respectively. If there were ambiguities regarding cause, the patient's private physician or family members were contacted by telephone for clarification. An event was considered to be of "cardiac" origin if one or more sources indicated such a cause. If there were multiple causes, or if it was unclear which cause was primary, the event was not considered to be of cardiac origin. Only cardiac cases were included in the stu@ Information on CCU admissions was derived from the Myocardial Infarction Triage and Intervention (MITI) Trial 5 database. This study identified all patients admitted to a CCU with a diagnosis of suspected AMI between November 1988 and July 1994. The study did not include chest pain patients who were not admitted (ie, who were sent home from the ED, were transported to another facility, or died in the ED). However, the total number of patients presenting to the ED with chest pain was obtained from a separate tally, the t~D Chest Pain Log. Data for the years 1992 and 1993 are complete in tallies from 15 area hospitals and indicate that 19,532 patients were seen in EDs for chest pain. They fall into four categories: those transported by paramedic services, admitted (n=4,886) or not admitted (n=842), and those using self-transport, admitted (n=~,216) or not admitted (n=7,728). Persons not admitted included those discharged home, those sent to another facility, and those who died in the ED.
613
SELF-TRANSPORT Becker et al
II
II
I
In the combined data from the CASS and MITI databases, four groups of patients were identified: (1) those who used self-transport and arrived at the hospital without prehospital cardiac arrest; (2) those who used self-transport and had a prehospital arrest; (3) those who called 911 and arrived at the hospital without arresting; and (4) those who called 911 and had a prehospital arrest. Probability values were determined with the use of Z2 statistics and Hotelling's multivariate test. RESULTS
During the study period, 13,187 patients with chest pain made a decision about transportation; there were an additional 612 patients for whom the means of transport could not be determined. Of the former group, 5,182 (41%) chose self-transport (by the patient or by a friend or relative), and 7,393 (59%) called 911 (Figure). Among those who chose self-transport, 18 (.35%) had a prehospital arrest (after which 911 was called). In contrast, among the group who initially called 911 for cardiac symptoms, 415 (5.6%) experienced prehospital cardiac arrest (P<.001). Using MITI data for CCU admissions, we compared a number of pre- and postevent factors between patients who
Figure, Means of transport among patients with suspected AMI and subsequent outcome. Total Cases: CCU Admissions + Prehospital Arrests 13,187
Unknown Means of Transport 612
Known Means of Transport
chose self-transport and those who chose EMS. Seventythree percent of patiehts in the EMS group were 65 years or older, compared with 60% in the self-transport group. The medical history for patients in the EMS group included AMI in 38%, congestive heart failure (CHF) in 24%, angina in 48%, and hypertension in 50%. The comparable figures for the self-transport group were AM[, 31%; CHE 18%; angina, 43%; and hypertension, 48% (Hotelling's multivariate test, 69.2; dr=5; P<.001). Likewise for the postevent factors: EMS patients had a 25% incidence of AMI, 53% had hospital stays longer than 4 days, and 8% died in the hospital, the comparable figures for self-transport patients being 19%, 44%, and 2%, respectively (Hotelling's multivariate test, 9.3; dr=3; P<.001). Results for both pre- and postevent factors show a fundamental difference between the population who were transported by EMS and those who attempted to drive themselves. We examined the records of the 18 patients who chose self-transport and then arrested before arriving at the hospital (Table). The symptoms listed were those obtained by emergency medical technicians and paramedic personnel at the scene of arrest and do not necessarily include alI the symptoms a patient may have had. A medical history was also obtained at the scene of collapse and was sometimes unknown or incomplete. All six survivors had ventricular fibrillation as the cause of cardiac arrest. It should be noted under circumstances of transport that some patients were attempting to drive to a clinic instead of a hospital; they were included because they were seeking medical care. We examined the MITI data to determine whether there was a difference in time elapsed before care was obtained between patients who chose EMS transport and those who chose self-transport. We found that the self-transport group had a median time from onset of chest pain to arrival at the hospital of 225 minutes; for those who called 911, the median time was 120 minutes (P<.001).
12,575
DISCUSSION
614
Sel#Transp0rt
911
5,182
7,393
Arrived at Hospital
Arrested Before Hospital Arrival
Arrived at Hospital
Arrested Before Hospital Arrival
5,164
18
6,978
415
Patients who chose self-transport had a much lower incidence of prehospital cardiac arrest than those who called 911. Only 18 patients among the thousands who used self-transport had a cardiac arrest, and 6 of these (33%) survived to hospital discharge. We cannot predict with accuracy the outcome of these 18 events had 911 been called initially Conceivably~patients may have avoided cardiac arrest had they been able to remain at rest and obtain rapid access to oxygen and medication for pain; on the other hand, the event in each of these cases may have progressed to cardiac arrest after para-
ANNALS OF EMERGENCY MEDICINE 28:6 DECEMBER1996
SELF-TRANSPORT l~ecker et al
I
medic personnel were called or after their arrival. The study measured only cardiac arrest as the outcome; it was not possible to measure morbidity or the extent to which paramedic intervention may have reduced the extent of cardiac damage during AMI. Because this was not an intervention study, a true relative risk cannot be determined. It may seem counterintuitive that patients choosing EMS transport have a much higher incidence of cardiac arrest than those who transport themselves. We believe: this result reflects self-selection or, to use an epidemiologic concept, confounding by condition. Our data show that patients who called 911 were older, had a higher incidence of prior AMI, had longer hospital stays, and had a higher in-hospital mortality rate than patients who transported themselves. The two populations are distinct in terms of mortality rate. Other studies
eliminated). Such an assumption could explain the lower rate of cardiac arrest among those who transport themselves. However, we found Lhat transport time was much longer for such patients, and the data indicate that this group is at less risk of cardiac arrest despite the longer time to treatment. Because community-based surveillance systems were used, it is conceivable that some cases may have been missed. For example, it was not possible to include cases in which a cardiac arrest precipitated an auto accident and subsequent death from trauma. Another category of potentially missed cases is that of chest pain patients who died in the ED before admission. Although information on these patients was not collected during the MITI study, ED data from one of the area's receiving hospitals indicate that, for the year 1995, there were no patients with chest pain who used self-transport and died in the ED. Therefore, we believe the number of these patients to be small. It is possible that communities in which a higher percentage of patients choose self-transport in a cardiac emergency may experience a higher number of cardiac-caused deaths in the ED. This study does not address the reasons why people delay or do not call 911 for chest pain; these reasons have been described elsewhere. 8 During the 2-year period during which data from ED visits were collected by ED Chest Pain Log, there were a total of 46,663 paramedic calls. Of Lhese, approximately 10,759 were calls for chest pain. During the same period,
Table,
Description of 18 patients who used self-transport to seek care for cardiac symptoms and arrested before arrival at the hospital. Age (Years)
Cardiac Rhythm
Sex
History
Symptoms and Circumstances of Transport
Outcome
Dyspnea, gas pains 3 days; collapsed in car while husband preparing to transport Died F Asystole HTN Acute dyspnea; went to clinic and collapsed Died F IDle HTN, IDDM Chest pain; wife drove to clinic; collapsed in car during evaluation by MD Discharged M VF Unknown Chest pain; collapsed in car while wife driving to hospital Died M VF MI Dyspnea; collapsed while husband was putting her in car to go to hospital Died F iDle Unknown Dyspnea; collapsed after wife put him in car to go to hospital Died M Heart block Unknown Vomiting; collapsed while wife was putting him in car Died M Asystole ASHD Chest pain; drove to clinic and collapsed Discharged M VF Unknown M VF CHF Acutely short of breath; collapsed while family was helping him to car Died F VF Angioplasty Chest pain; called friends who came and found her collapsed Died F IDle CHF Short of breath; collapsed while neighbor was preparing to transport Died Chest pain; collapsed in car during transport to hospital Died M Asystole Heart problem F Asystole Heart disease Dyspnea; collapsed while family was preparing to transport Died M VF Unknown Crushing chest pain; refused to call 911; drove to doctor's office and collapsed Discharged F EMD Unknown Back pain, nausea; went to clinic and collapsed Died M VF HTN, IDDM Chest pain; was driven to fire station by son and collapsed Discharged M VF ASHD Chest pain; collapsed while pulling into doctor's parking lot Discharged F VF CHF, MI Not feeling well; collapsed after being put in car to go to hospital Discharged ASHD, arterioscleroticheart,disease;CHF,congestiveheartfailure; EMD, electromechanicaldissociation;HTN, hypertension;IDle, idiovantricular;IDDM, insulin-dependentdiabetesme[titus; ML myocardialinfarction;VI:, ventricular. 70 65 56 73 82 80 63 65 84 79 79 67 57 54 62 65 76 86
DECEMBER 1996
28:6
ANNALS OF EMERGENCY MEDICINE
aI 5
SELF-TRANSPORT
Be&er et al
there were 11,94~r patients who arrived at EDs without calling 911 (4,21 d admitted and 7,728 not admitted). Transport of all 22,703 patients by EMS would increase the number of EMS responses by 25%, requiring a considerable addition of units, staff, and administrative support. We found the incidence of cardiac arrest among patients choosing self-transport after experiencing chest pain to be very low (less than 1%), suggesting that the current level of awareness of the benefits of using 911 is sufficient to bring most of the highest-risk patients to the hospital by EMS transport. If everyone who had chest pain had called 911, a maximum of 18 cardiac arrests might have been prevented, and 12 additional lives might have been saved, in a 30-month period. We believe that other communities can benefit by assessing the risk to their populations of use of self-transport as a means of seeking medical help for a cardiac emergency. An estimate of this incidence can serve as a basis for policy and public education planning.
Stefanie Ostergard provided assistance in case identification. Steve Call provided admfnistrative support. The authors appreciate the assistance of King County paramedics and emergency medical technicians in conducting this study.
Reprint no. 47/1/77733 Address for reprints: Linda Decker, MA King County Emergency Medical Services Division Seattle-King County Department of Public Health 900 Fourth Avenue, Suite 850 Seattle, Washington 98164
REFERENCES 1. National Heart, Lung and Blood Institute: Patient/BystanderRecognitionandAcrion:Rapid Identificationand Treatmentof Acute MyocardialInfarction. Publication#93-3303. Bethesda, Maryland: US Departmentef Health and Human Sewices, Public Health Service,National institutes of Health, 1993. 2. EpplerE, EisenbergM, Schaeffer S, et al: 911 and emergencydeportmentuse for chest pain: Results of a media campaign.Ann EmergMed 1994;24:202-208. 3. He M, EisenbergM, Litwin P, at el: Delay between onset of chest pain and seeking medical care: The effect of public education.Ann EmergMed1989;18:727-731. 4. EisenbergM, BergnerL, HallstromA: Paramedicprogramsand out-of-hospital cardiacarrest: I. Factors associatedwith successful resuscitation.Am JPublic Health1979;69:30-38. 5. Maynard C, Weaver WD, Litwin PE,et al: Hospital mortality in acute myocardialinfarction in the era of reperfusiontherapy (the MyocardialTriage and intervention Project).Am J Cardiol 1993;72:877-882. 6. Weaver WE;: Acute myocardialinfarction: Seattle, 1990. Journalof Applied Cardiopulrnonary Physiology1991;4:93-41. 7, Meischke H, EisanbergM, Schaeffer S, et ah Utilization of emergencymedical servicesfor symptoms of acute myocardialinfarction. HeartLung1995;24:11-18. 8, Meischke H, Ho MT, EisenbergMS, et el: Reasonspatients with chest pain delay or do not call 9-1-1. Ann EmergMad 1995;25:193- 197.
6I 6
ANNALS OF EMERGENCY MEDICINE
28:6
DECEMBER 1996
Incidence of Cardiac Arrest During Self-Transport for Chest Pain From the King County Emergency Medical Services Division* and the Centerfor Evaluation of Emergency Medical ServiceS:, Seattle-King County Department of Public Health; and the Department of Medicine, University of Washington~, Seattle, Washington. Receivedfor publication November 20, 1995. Revisions receivedApri110, 1996, and July I5, 1996. Acceptedfor publicationJuly 25, 1996. Copyright © by the American College of Emergency Physicians.
Linda Becker, MA* Mary Pat Larsen, MS** Mickey S Eisenberg, MD, PhD*§
Study objective: To assess the incidence of cardiac arrest among patients who use self-transport to seek medical care for chest pain. Methods: This was a retrospective cohort study of patients admitted to a CCU for suspected acute myocardial infarction (AMI) and patients experiencing out-of-hospital cardiac arrest preceded by symptoms in King County,Washington, between January 1, 1992, and July 31, 1994. Participants were identified through use of the databases compiled by the Myocardial Infarction Triageand Intervention Trial, which reviewed medical records in all area hospitals, and the Cardiac Arrest Surveillance System, which tracks all incidences in which CPR is performed by EMS personnel in King County.Patientswhose sudden cardiac arrests were not preceded by symptoms were excluded. Hospital records were abstracted to find the means of transport for patients admitted to CCUs. For cardiac arrest patients, the medical history, presenceof symptoms, means of transport, and prehospital death information were abstracted from paramedic field reports. Outcome (admission, discharge, or in-hospital death) was obtained from hospital records. An event cause (cardiac or other) was determined from death certificates, hospital records, or consultation with private physicians. Results: During the 30-month study period, 13,187 patients sought help for cardiac symptoms and were either admitted to a CCU or died before admission after calling 911. A majority, 7,393 (59%), were transported by emergency medical services, and 5,182 (41%) used private transportation to obtain medical care; the means of transport could not be determined for 612 patients. Of the EMS group, 6,978 were admitted to the hospital without experiencing prehospital cardiac arrest, and 415 (5.6%)arrested before arriving at the hospital. Of the group using private transportation, 5,164 were admitted without arresting and 18 (.35%) arrested before arrival, after which 911 was called (P<.0Ol).
6 12
ANNALS OF EMERGENCY MEDICINE
28:6
DECEMBER 1996
SELF-TRANSPORT t~ec/zer et al
Conclusion: The incidence of cardiac arrest among patients
who attempted to reachthe hospital by private transportationwas very low compared with the incidence among those who chose the EMS system for transport. This suggests that patient selfselection occurs, with the more seriously ill patients more commonly calling 91 1 for transport. [Becker L, Larsen MP, Eisenberg MS: Incidence of cardiac arrest during self4ransport for chest pain. Ann ErnergMedDecember 1996;28:612-616]
INTRODUCTION Persons who seek medical care for chest pain must decide on a mode of transportation. Should they call 911, or should they drive to the hospital or have a friend or relative drive? The National Heart Attack Alert Program ~ and other educational campaigns ~,3 urge people with chest pain to call 911. The rationale is that paramedic personnel can start treatment with oxygen, provide analgesia and antiarrhythmic therapy, and perform defibrillation if needed. The treatment can begin within a few minutes of symptom onset. Despite this advice, many patients who experience chest pain choose to transport themselves. The purpose of this study was to determine the incidence of cardiac arrest during self-transport for chest pain.
MATERIALS AND METHODS Cases included in the study were those of persons who had experienced chest pain or other potentially serious cardiac symptoms and who had made a decision about transport to seek medical care. Patients included those admitted to a CCU with a diagnosis of suspected acute myocardial infarction (AMI) without arresting and those who experienced cardiac arrest before arrival at the hospital, either before or after arrival of emergency medical services. Patients whose cardiac arrests were not of primary cardiac origin, such as arrests resulting from suicide, trauma, respiratory disease, or cancer, were excluded, as were patients who arrested in the emergency department and died before admission. Also excluded were those patients who had no symptoms and collapsed without warning; these patients were not able to make a decision about transport, and 911 was caIled in every case. We retrospectively studied cases in King County, Washington (study population approximately 1,000,000 in 1992) that occurred between January 1, 1992, and July 31, 1994. Data were: drawn from two distinct, concurrent and
DECEMBER 1996
28:6
ANNALS OF EMERGENCY MEDICINE
complementary databases. The Cardiac Arrest Surveillance System (CASS) provided information on cardiac arrests. ~ Patients in this database had out-of-hospital cardiac arrest with CPR performed by EMS personnel. This data collection system began with the implementation of paramedic services in King County in 1976; definitions and coding procedures have remained the same since the beginning. Through a dual reporting system whereby a separate report of each CPR case is received from emergency medical technicians and paramedics, virtually 100% of relevant cases were identified. Data recorded were time of collapse; time CPR was begun; whether the event was wimessed; whether bystander CPR was performed; and times to the following: defibrillation, intubation, blood pressure, placement of IV tube, administration of medication, return of spontaneous circulation, and termination of efforts. The records of cardiac arrest patients admitted to a hospital were reviewed to determine whether they were discharged alive and the destination of discharge (home or nursing home). An event cause was assigned based on consideration of several factors: prior history, medications, symptoms, and circumstances of arrest. The discharge diagnosis or cause of death was obtained from the medical record or from the death certificate, respectively. If there were ambiguities regarding cause, the patient's private physician or family members were contacted by telephone for clarification. An event was considered to be of "cardiac" origin if one or more sources indicated such a cause. If there were multiple causes, or if it was unclear which cause was primary, the event was not considered to be of cardiac origin. Only cardiac cases were included in the stu@ Information on CCU admissions was derived from the Myocardial Infarction Triage and Intervention (MITI) Trial 5 database. This study identified all patients admitted to a CCU with a diagnosis of suspected AMI between November 1988 and July 1994. The study did not include chest pain patients who were not admitted (ie, who were sent home from the ED, were transported to another facility, or died in the ED). However, the total number of patients presenting to the ED with chest pain was obtained from a separate tally, the t~D Chest Pain Log. Data for the years 1992 and 1993 are complete in tallies from 15 area hospitals and indicate that 19,532 patients were seen in EDs for chest pain. They fall into four categories: those transported by paramedic services, admitted (n=4,886) or not admitted (n=842), and those using self-transport, admitted (n=~,216) or not admitted (n=7,728). Persons not admitted included those discharged home, those sent to another facility, and those who died in the ED.
613
SELF-TRANSPORT Becker et al
II
II
I
In the combined data from the CASS and MITI databases, four groups of patients were identified: (1) those who used self-transport and arrived at the hospital without prehospital cardiac arrest; (2) those who used self-transport and had a prehospital arrest; (3) those who called 911 and arrived at the hospital without arresting; and (4) those who called 911 and had a prehospital arrest. Probability values were determined with the use of Z2 statistics and Hotelling's multivariate test. RESULTS
During the study period, 13,187 patients with chest pain made a decision about transportation; there were an additional 612 patients for whom the means of transport could not be determined. Of the former group, 5,182 (41%) chose self-transport (by the patient or by a friend or relative), and 7,393 (59%) called 911 (Figure). Among those who chose self-transport, 18 (.35%) had a prehospital arrest (after which 911 was called). In contrast, among the group who initially called 911 for cardiac symptoms, 415 (5.6%) experienced prehospital cardiac arrest (P<.001). Using MITI data for CCU admissions, we compared a number of pre- and postevent factors between patients who
Figure, Means of transport among patients with suspected AMI and subsequent outcome. Total Cases: CCU Admissions + Prehospital Arrests 13,187
Unknown Means of Transport 612
Known Means of Transport
chose self-transport and those who chose EMS. Seventythree percent of patiehts in the EMS group were 65 years or older, compared with 60% in the self-transport group. The medical history for patients in the EMS group included AMI in 38%, congestive heart failure (CHF) in 24%, angina in 48%, and hypertension in 50%. The comparable figures for the self-transport group were AM[, 31%; CHE 18%; angina, 43%; and hypertension, 48% (Hotelling's multivariate test, 69.2; dr=5; P<.001). Likewise for the postevent factors: EMS patients had a 25% incidence of AMI, 53% had hospital stays longer than 4 days, and 8% died in the hospital, the comparable figures for self-transport patients being 19%, 44%, and 2%, respectively (Hotelling's multivariate test, 9.3; dr=3; P<.001). Results for both pre- and postevent factors show a fundamental difference between the population who were transported by EMS and those who attempted to drive themselves. We examined the records of the 18 patients who chose self-transport and then arrested before arriving at the hospital (Table). The symptoms listed were those obtained by emergency medical technicians and paramedic personnel at the scene of arrest and do not necessarily include alI the symptoms a patient may have had. A medical history was also obtained at the scene of collapse and was sometimes unknown or incomplete. All six survivors had ventricular fibrillation as the cause of cardiac arrest. It should be noted under circumstances of transport that some patients were attempting to drive to a clinic instead of a hospital; they were included because they were seeking medical care. We examined the MITI data to determine whether there was a difference in time elapsed before care was obtained between patients who chose EMS transport and those who chose self-transport. We found that the self-transport group had a median time from onset of chest pain to arrival at the hospital of 225 minutes; for those who called 911, the median time was 120 minutes (P<.001).
12,575
DISCUSSION
614
Sel#Transp0rt
911
5,182
7,393
Arrived at Hospital
Arrested Before Hospital Arrival
Arrived at Hospital
Arrested Before Hospital Arrival
5,164
18
6,978
415
Patients who chose self-transport had a much lower incidence of prehospital cardiac arrest than those who called 911. Only 18 patients among the thousands who used self-transport had a cardiac arrest, and 6 of these (33%) survived to hospital discharge. We cannot predict with accuracy the outcome of these 18 events had 911 been called initially Conceivably~patients may have avoided cardiac arrest had they been able to remain at rest and obtain rapid access to oxygen and medication for pain; on the other hand, the event in each of these cases may have progressed to cardiac arrest after para-
ANNALS OF EMERGENCY MEDICINE 28:6 DECEMBER1996
SELF-TRANSPORT l~ecker et al
I
medic personnel were called or after their arrival. The study measured only cardiac arrest as the outcome; it was not possible to measure morbidity or the extent to which paramedic intervention may have reduced the extent of cardiac damage during AMI. Because this was not an intervention study, a true relative risk cannot be determined. It may seem counterintuitive that patients choosing EMS transport have a much higher incidence of cardiac arrest than those who transport themselves. We believe: this result reflects self-selection or, to use an epidemiologic concept, confounding by condition. Our data show that patients who called 911 were older, had a higher incidence of prior AMI, had longer hospital stays, and had a higher in-hospital mortality rate than patients who transported themselves. The two populations are distinct in terms of mortality rate. Other studies
eliminated). Such an assumption could explain the lower rate of cardiac arrest among those who transport themselves. However, we found Lhat transport time was much longer for such patients, and the data indicate that this group is at less risk of cardiac arrest despite the longer time to treatment. Because community-based surveillance systems were used, it is conceivable that some cases may have been missed. For example, it was not possible to include cases in which a cardiac arrest precipitated an auto accident and subsequent death from trauma. Another category of potentially missed cases is that of chest pain patients who died in the ED before admission. Although information on these patients was not collected during the MITI study, ED data from one of the area's receiving hospitals indicate that, for the year 1995, there were no patients with chest pain who used self-transport and died in the ED. Therefore, we believe the number of these patients to be small. It is possible that communities in which a higher percentage of patients choose self-transport in a cardiac emergency may experience a higher number of cardiac-caused deaths in the ED. This study does not address the reasons why people delay or do not call 911 for chest pain; these reasons have been described elsewhere. 8 During the 2-year period during which data from ED visits were collected by ED Chest Pain Log, there were a total of 46,663 paramedic calls. Of Lhese, approximately 10,759 were calls for chest pain. During the same period,
Table,
Description of 18 patients who used self-transport to seek care for cardiac symptoms and arrested before arrival at the hospital. Age (Years)
Cardiac Rhythm
Sex
History
Symptoms and Circumstances of Transport
Outcome
Dyspnea, gas pains 3 days; collapsed in car while husband preparing to transport Died F Asystole HTN Acute dyspnea; went to clinic and collapsed Died F IDle HTN, IDDM Chest pain; wife drove to clinic; collapsed in car during evaluation by MD Discharged M VF Unknown Chest pain; collapsed in car while wife driving to hospital Died M VF MI Dyspnea; collapsed while husband was putting her in car to go to hospital Died F iDle Unknown Dyspnea; collapsed after wife put him in car to go to hospital Died M Heart block Unknown Vomiting; collapsed while wife was putting him in car Died M Asystole ASHD Chest pain; drove to clinic and collapsed Discharged M VF Unknown M VF CHF Acutely short of breath; collapsed while family was helping him to car Died F VF Angioplasty Chest pain; called friends who came and found her collapsed Died F IDle CHF Short of breath; collapsed while neighbor was preparing to transport Died Chest pain; collapsed in car during transport to hospital Died M Asystole Heart problem F Asystole Heart disease Dyspnea; collapsed while family was preparing to transport Died M VF Unknown Crushing chest pain; refused to call 911; drove to doctor's office and collapsed Discharged F EMD Unknown Back pain, nausea; went to clinic and collapsed Died M VF HTN, IDDM Chest pain; was driven to fire station by son and collapsed Discharged M VF ASHD Chest pain; collapsed while pulling into doctor's parking lot Discharged F VF CHF, MI Not feeling well; collapsed after being put in car to go to hospital Discharged ASHD, arterioscleroticheart,disease;CHF,congestiveheartfailure; EMD, electromechanicaldissociation;HTN, hypertension;IDle, idiovantricular;IDDM, insulin-dependentdiabetesme[titus; ML myocardialinfarction;VI:, ventricular. 70 65 56 73 82 80 63 65 84 79 79 67 57 54 62 65 76 86
DECEMBER 1996
28:6
ANNALS OF EMERGENCY MEDICINE
aI 5
SELF-TRANSPORT
Be&er et al
there were 11,94~r patients who arrived at EDs without calling 911 (4,21 d admitted and 7,728 not admitted). Transport of all 22,703 patients by EMS would increase the number of EMS responses by 25%, requiring a considerable addition of units, staff, and administrative support. We found the incidence of cardiac arrest among patients choosing self-transport after experiencing chest pain to be very low (less than 1%), suggesting that the current level of awareness of the benefits of using 911 is sufficient to bring most of the highest-risk patients to the hospital by EMS transport. If everyone who had chest pain had called 911, a maximum of 18 cardiac arrests might have been prevented, and 12 additional lives might have been saved, in a 30-month period. We believe that other communities can benefit by assessing the risk to their populations of use of self-transport as a means of seeking medical help for a cardiac emergency. An estimate of this incidence can serve as a basis for policy and public education planning.
Stefanie Ostergard provided assistance in case identification. Steve Call provided admfnistrative support. The authors appreciate the assistance of King County paramedics and emergency medical technicians in conducting this study.
Reprint no. 47/1/77733 Address for reprints: Linda Decker, MA King County Emergency Medical Services Division Seattle-King County Department of Public Health 900 Fourth Avenue, Suite 850 Seattle, Washington 98164
REFERENCES 1. National Heart, Lung and Blood Institute: Patient/BystanderRecognitionandAcrion:Rapid Identificationand Treatmentof Acute MyocardialInfarction. Publication#93-3303. Bethesda, Maryland: US Departmentef Health and Human Sewices, Public Health Service,National institutes of Health, 1993. 2. EpplerE, EisenbergM, Schaeffer S, et al: 911 and emergencydeportmentuse for chest pain: Results of a media campaign.Ann EmergMed 1994;24:202-208. 3. He M, EisenbergM, Litwin P, at el: Delay between onset of chest pain and seeking medical care: The effect of public education.Ann EmergMed1989;18:727-731. 4. EisenbergM, BergnerL, HallstromA: Paramedicprogramsand out-of-hospital cardiacarrest: I. Factors associatedwith successful resuscitation.Am JPublic Health1979;69:30-38. 5. Maynard C, Weaver WD, Litwin PE,et al: Hospital mortality in acute myocardialinfarction in the era of reperfusiontherapy (the MyocardialTriage and intervention Project).Am J Cardiol 1993;72:877-882. 6. Weaver WE;: Acute myocardialinfarction: Seattle, 1990. Journalof Applied Cardiopulrnonary Physiology1991;4:93-41. 7, Meischke H, EisanbergM, Schaeffer S, et ah Utilization of emergencymedical servicesfor symptoms of acute myocardialinfarction. HeartLung1995;24:11-18. 8, Meischke H, Ho MT, EisenbergMS, et el: Reasonspatients with chest pain delay or do not call 9-1-1. Ann EmergMad 1995;25:193- 197.
6I 6
ANNALS OF EMERGENCY MEDICINE
28:6
DECEMBER 1996