Can First Responders Be Sent to Selected 9-1-1
Emergency Medical Services Calls without
an Ambulance?
Craig B. Key, MD, Paul E. Pepe, MD, MPH, David E. Persse, MD,
Darrell Calderon, MD
Abstract
Objectives:To evaluate the feasibility and safety of initially dispatching only first responders (FRs) to selected low-risk 9-1-1 requests for emergency medical services. First respon-ders are rapidly-responding fire crews on apparatus without transport capabilities, with firefighters trained to at least a FR level and in most cases to the basic emergency medical technician (EMT) level. Low-risk 9-1-1 requests include automatic medical alerts (ALERTs), motor vehicle incidents (MVIs) for which the caller was unable to answer any medical dispatch questions designed to prioritize the call, and 9-1-1 call disconnects (D/Cs).Methods:A before-and-after study of patient dispositions was conducted using historical controls for comparison. During the historical control phase of six months, one year prior to the study phase, basic life support ambulances (staffed with two basic EMTs) were dispatched to selected low-risk 9-1-1 incidents. During the six-month study phase, a fire FR crew equipped with automated external defibrillators (AEDs) was sent initially without an ambulance to these incidents.Results:
For ALERTs (n ¼ 290 in historical group vs. 330 in study group), there was no statistical difference in the transport rate (7% vs 10%), but there was a statistically significant increase in the follow-up use of advanced life support (ALS) (1% vs 4%, p¼0.009). No patient in the ALERTs historical
group required airway management, while one patient in the study group received endotracheal intubation. No pa-tient required defibrillation in either group. Analysis of the MVIs showed a significant decrease (p\0.0001) in the pa-tient transport rate from 39% of controls to 33% of study patients, but no change in the follow-up use of ALS in-terventions (2% for each group). For both the ALERTs and MVIs, the FR’s mean response time was faster than ambu-lances (p\0.0001). Among the 9-1-1 D/Cs with FRs only (n¼1,028), 15% were transported and 43 (4%) received sub-sequent ALS care. Four of these patients (0.4%) received intubation and two (0.2%) required defibrillation. However, no patient was judged to have had adverse outcomes as a result of the dispatch protocol change.Conclusions:Fire apparatus crews trained in the use of AEDs can safely be used to initially respond alone (without ambulances) to selected, low-risk 9-1-1 calls. This tactic improves response intervals while reducing ambulance responses to these incidents. Key words: emergency medical services; EMS; ambulance; emergency medical technician; EMT; accidents, traffic; defibrillation; tiered; dispatch; medical priority; first responder; triage. ACADEMIC EMERGENCY MEDICINE 2003; 10:339–346.
Modern emergency medical services (EMS) systems were established in the United States during the late 1960s. Since then, EMS has evolved from a rudimen-tary hospital transport service with minimally trained
personnel to a sophisticated public health system delivering advanced medical care to patients with medical emergencies in the out-of-hospital environ-ment.1–6Today, more than 25 million EMS responses are made in the United States each year, including a significant number (5–15%) that may involve the advanced monitoring and invasive interventional skills of paramedics.6–9
In retrospect, however, between 30% and 50% of all EMS responses still are made to nonemergency situations, and an equal number of cases require only basic life support (BLS) skills such as splinting and spinal immobilization.7,10 Recognizing that only a small percentage of calls ever utilize advanced life support (ALS) skills, modern priority dispatch sys-tems have been developed that either triage the need for a rapid, ‘‘lights and siren’’ response (which may pose some additional traffic risk) or, in other cases, simply limit paramedic deployment to those emergen-cies that probably will require advanced skills.7,10–21 From the Department of Emergency Medicine, Ohio State
Univer-sity (CBK), and the Ohio State UniverUniver-sity Center for EMS (CBK), Columbus, OH; the Department of Emergency Medicine, University of Texas Medical School at Houston (CBK, DEP, DC), Houston, TX; the Department of Emergency Medicine, University of Texas Southwestern Medical Center at Dallas (PEP), and the Dallas Area Biotel (EMS) System (PEP), Dallas, TX; the Department of Surgery, Baylor College of Medicine (DEP), Houston, TX; and the City of Houston, EMS (CBK, DEP), Houston, TX.
Received June 28, 2002; revision received October 17, 2002; accepted October 28, 2002.
Presented at the SAEM annual meeting, San Francisco, CA, May 2000.
Address correspondence and reprint requests to: Craig B. Key, MD, Ohio State University, Department of Emergency Medicine, 016 Health Sciences Library, 376 West 10th Avenue, Columbus, OH 43210-1252. Fax: 614-293-3124; e-mail: key.30@osu.edu.
By sending nonparamedic ambulances staffed by BLS-providing emergency medical technicians (EMTs) to the lower-priority calls, EMS systems can function with fewer paramedics.1,6,7,22With fewer paramedics, each paramedic gains more experience with those true emergency situations for which he or she needs to be focused, thereby improving his or her clinical perfor-mance in these situations.6,7,13,16–18 In addition, by making the pool of paramedic ambulances (or other ALS response units) more available, one can actually improve response intervals to the more critical cases.7,22
In many systems, rapidly-responding fire suppres-sion apparatus or police vehicles, staffed with crews who can provide BLS, are used to supplement the initial ambulance response.1,6,21Although they do not transport patients, these ‘‘neighborhood’’ crews can routinely arrive much faster than ambulances (for myriad reasons) and thus provide more rapid basic medical care, ranging from oxygen administration and spinal immobilization to basic cardiopulmonary resuscitation (CPR) and automated external defibril-lator (AED) use.21
While the traditional use of these ‘‘first responders’’ (FRs) has been to provide a more reliable and rapid response to critical cases requiring time-dependent interventions (e.g., cardiac arrest, chokings), it may also be considered reasonable to dispatch well-trained FRs to provide faster triage and evaluation of those cases that may not require either urgent ambulance responses or even hospital transport altogether.
Based on prior research in our system,7 three categories of EMS incidents were identified as ‘‘low risk’’: automatic medical alerts (ALERTs), motor vehicle incidents (MVIs) for which the caller was unable to answer any medical dispatch questions designed to prioritize the call, and 9-1-1 call dis-connects (D/Cs). The purpose of this study was to evaluate the feasibility and safety of initially dis-patching only fire apparatus first responders (FAFRs, rapidly-responding fire apparatus without transport capabilities, with firefighters trained to at least a FR level and in most cases to the EMT-basic level) to these low-risk 9-1-1 requests for EMS.
METHODS
Study Design. The study was a before-and-after
investigation that used a six-month historical control period from the same six months one year prior to the observational period for comparison. The study was blinded in that dispatchers, administrators, and field personnel were not aware that a focused, formal study was being conducted. The ‘‘safety’’ of the strategy of initially only sending FAFRs to select low-risk dispatch types was evaluated by assessing the raw percentage of patients receiving transport to a hospital and/or subsequent ALS care as compared
with the historical control group to which only a BLS ambulance had been dispatched initially, using the same dispatch algorithm.
Institutional review board (IRB) approval was obtained from the Committee for Review of Human Subjects. Written consent was waived by the IRB because this was considered to be an observational study that was being conducted as a routine quality assurance activity of the EMS system medical direc-tors to independently evaluate the system changes that were being instituted by a governmental agency (municipal fire department). Since the type of activity involved to conduct this study was part of the routine quality assurance conducted for dispatch operations, only high-level fire department officials and the in-vestigators were prospectively involved in the study design. Dispatchers, field personnel, and quality as-surance personnel were unaware that such a study was being conducted.
Study Setting and Population. The study was an
EMS population-based investigation conducted in a large urban city with a daytime population of 3.6 million, a nighttime population of 1.7 million, and 620 square miles in geographic territory. The municipal fire department EMS division is the sole designated emergency out-of-hospital care responder for the city. By ordinance, the fire department is responsible for all aspects of the municipal EMS system, including dispatch, first response, patient care, and transport.
The EMS program used a tiered ALS/BLS ambu-lance system with a total of 54 ambuambu-lances at the time of study. In addition, there were 112 fire apparatus, staffed with four-person crews who provided BLS FR functions. Each fire apparatus had personnel trained to at least a FR level to perform CPR, use AEDs, perform basic airway maneuvers, control external bleeding, provide spinal immobilization, and com-plete a basic patient assessment. Most fire apparatus crews were trained to the EMT-basic level as part of a policy in place for 11 years prior to the study that all new fire department personnel were required to obtain and keep EMT-B certification. Personnel who staff FAFR crews also staff the basic ambulances, as basic ambulance and fire apparatus crews routinely rotate between ambulances and fire apparatus. The FAFR crews carried an AED, bag–valve–mask devices (BVMs), backboards, cervical collars, oxygen, and other basic equipment. None of the FAFR crews operated at the ALS level.
Around the clock, there were 32 ALS ambulances and seven EMS supervisors (in response cars) operating at the ALS level. The remaining 22 ambulances were BLS ambulances, staffed with two basic EMTs equipped with an AED, BVMs, oxygen, glucometers, oral glucose, and basic trauma equip-ment such as bandages, splints, cervical collars, and backboards. All ambulances (ALS and BLS) and FAFR
crews are dispatched ‘‘hot’’ (red lights and sirens) on all calls. There is no triage to ‘‘cold’’ response in the study system.
Formal ‘‘no transport’’ policies were in place during the control and experimental portions of the study that were the same for both the basic ambulance and FAFR crews. Many of the study calls were exempt from these policies when no patient was found at the scene. Basic EMT crews are required to obtain two sets of vital signs, establish the patient’s level of con-sciousness and competency for refusal, complete a primary and secondary physical examination, and obtain a complete list of medications before consid-ering nontransport. Contact must also be made with an EMS supervisor or physician to request permission for nontransport. If approved, a nontransport form must be completed for each patient not transported to an ED. Further, a specific comprehensive nontran-sport protocol is in place for victims of motor vehi-cle collisions. In addition to the requirements of the general nontransport guidelines, specific vital sign limitations, mechanism-of-injury limitations, and com-petency limitations are imposed along with specific documentation guidelines.
The public safety answering point (PSAP) in this system triages calls to either fire/EMS or police, and transfers the caller to the fire/EMS or police dispatch center. Some MVI calls are relayed from the police dispatch center when either the call is initially sent to the police department by the PSAP because the callers indicate no ambulance is required, or when a police officer on the scene requests an ambulance.
Fire/EMS dispatchers use a standard dispatch protocol with uniform training and oversight by dispatch supervisors. The dispatch system in use for this study is described in detail in a previous publication.7
Study Protocol.All data were collected prospectively in both the historical and study phases using the automated data collection system, and therefore, no data sets were incomplete. The three categories of emergencies studied were: 1) ALERTs; 2) MVIs for which the 9-1-1 caller was unable to answer any medical dispatch questions designed to prioritize the call; and 3) D/Cs in which the 9-1-1 caller provided no information to the call taker before disconnecting. The priority dispatch algorithms for these categories are relatively simple because there is no additional information, except for the automatic alarm signal or the 9-1-1 disconnect screen, that displays the location or the police dispatchers’ electronic relay information, regarding the address of an MVI. Dispatch procedures did not change in the study and control periods, except the deployment disposition (lone BLS ambu-lance vs. lone FAFR dispatch). However, historical data were not available for the 9-1-1 disconnect category, because, prior to the change in the dispatch
protocols, 9-1-1 disconnects were not identified as a separate category, but instead were grouped with other calls into a more general category called ‘‘unknown’’ calls. Except for this one new addition, no other dispatch changes were made and thus there was no expected effect on the other two categories (MVIs and ALERTs). As previously described, a com-prehensive electronic database system was used that automatically logs the dispatch type (e.g., automatic medical alert); the response, scene, and transport intervals; patient assessments; and medications and procedures.7,9,23Cases requiring subsequent ALS care are readily tabulated, and systems to provide follow-up on such cases, or identify cases that should have received ALS care, were in place as described in detail in previous publications.7,9,23All data were reviewed by two nonstudy emergency physicians and fire department quality assurance officers for appropri-ateness of care and safety.
Measurements. For each of the three categories
(ALERTs, MVIs, D/Cs), the numbers of incidents that eventually resulted in patient transport, including the number of BLS transports and the number of ALS transports, were determined from the database. ALS transports were defined as: 1) cases in which patients had at least one ALS intervention (electrocardio-graphic [ECG] monitoring, intravenous [IV] cannula-tion, drug administracannula-tion, endotracheal intubacannula-tion, or defibrillation/cardioversion) done by any EMS per-sonnel before hospital arrival; or 2) those historically rare cases in which quality assurance screens at area hospitals identified patients who should have had such out-of-hospital ALS interventions.7,23 The ALS calls were further stratified by type of ALS interven-tion, defining ‘‘critical’’ ALS intervention as endotra-cheal intubation or defibrillation, since precautionary IV access and cardiac monitoring may not actually indicate a need for ALS care.
Response intervals (recorded to the second) were measured for both phases of the study and were defined by the time between the dispatch of the EMS unit (ambulance or first responder) and arrival of each EMS unit at the scene, as previously defined by Spaite et al.24All AED rhythms and operational actions were recorded automatically and reviewed. As previously described, longitudinal patient status and patient care interventions were also recorded prospectively in electronic data logs (e.g., serial vital signs), as were narrative data describing patient conditions and progress throughout the incident up until trans-fer of patient care to emergency department (ED) physicians.
Data Analysis. For cases resulting in transport, the historical control and study groups were compared using two-tailed Fisher’s exact tests in each of the three transport categories (total, BLS, and ALS
transports). Statistical significance was interpreted at a reduced alpha of 0.017 (0.05/3). For the MVI group, four transport categories were compared (total, BLS, ALS, and critical ALS intervention [either intubation or defibrillation]) and interpreted at a reduced alpha of 0.0125 (0.05/4). Mean time intervals, such as response intervals for BLS ambulances and FRs, were compared using two-sample t-tests with Sat-terthwaite’s correction for unequal variances.
RESULTS
ALERT Group.The BLS and ALS transport rates for the control and study ALERTs groups are found in Table 1. Of the 290 requests for EMS response by automatic medical alert device in the six-month period one yearpriorto lone FAFR crew deployment (BLS ambulance phase), 19 (7%) resulted in patient transport. Two patients (\1%) eventually received an ALS intervention, with one receiving ECG rhythm monitoring alone and the other receiving an IV line and nitroglycerin for ‘‘sharp’’ chest pain (without relief). No patient deteriorated, required airway management, or required defibrillation.
In the six months after lone FAFR deployment was initiated, there were 335 requests for EMS response by automatic medical alert, and 35 (10%) resulted in patient transport (no significant difference). Thirteen (3.9%) of the cases did involve some type of ALS care, a statistically significant increase over the historical control (p¼0.009). However, eight patients received rhythm monitoring only (three complaining of chest
discomfort, two with weakness, two with abdominal pain, and one complaining of fever, cough, and difficulty breathing). One patient with altered mental status received IV access without any drug adminis-tration. Three others were given self-administered type drugs (one received nitroglycerin for chest pain and two received albuterol for difficulty breathing). One patient did receive a critical ALS intervention, intubation, for severe difficulty breathing. No patient required defibrillation.
On average, FAFR crews took 5.8 minutes (95% confidence interval [95% CI]¼5.41 to 6.14,n¼335) to arrive at these incidents, while the average historical BLS ambulance response time of 10.0 minutes (95% CI
¼9.25 to 10.81) was significantly higher (p\0.0001). In subsequent follow-up and review of cases, by the routine (nonstudy) EMS quality assurance officers, there was no adverse outcome that could be attributed to delays in the dispatch of a transport or ALS unit during either phase of the study. Theoretically, calls for ALS backup (approximately another 7-minute interval to respond) were initiated, on the average, 4.2 minutes earlier due to the earlier arrival of FAFR crews. For the patients who were transported, there was no significant difference in the total out-of-hospital time (9-1-1 call to out-of-hospital arrival) when comparing the two study phases.
MVI Group. Rates for the transport categories of MVIs when the caller could not answer any of the priority dispatch questions are shown in Table 2. In the basic ambulance group, there were 6,463 requests for EMS response; and 2,498 (39%) resulted in trans-ports (2,346 BLS, 152 ALS). Eleven of the 6,463 cases (0.2%) eventually involved a critical ALS intervention such as endotracheal intubation or defibrillation, but seven of those 11 patients presented with trauma-associated cardiac arrest and an initial AED rhythm of asystole. The remaining four patients included one with an opiate overdose who responded to naloxone and survived to hospital discharge, and three with primary cardiac arrests who each achieved return of spontaneous circulation, with one surviving to hospi-tal discharge. This patient had ventricular fibrillation and received prompt defibrillation with an AED by the BLS ambulance crew. In contrast, for the FAFR phase, there were 3,372 MVIs, and 1,103 (33%) resulted in transport (p\0.0001) (1,021 BLS, 82 ALS). One person (0.03%) eventually received a critical ALS intervention (endotracheal intubation). However, that patient presented with asystolic cardiac arrest associ-ated with blunt trauma. Therefore, no true adverse outcome could be attributed to delays in the dispatch of a transport or ALS unit. BLS ambulances required an average of 9.1 minutes (95% CI ¼ 8.98 to 9.31,
n¼5,690) to arrive at these incidents, compared with the FAFR crews, who took 6.3 (95% CI¼6.04 to 6.48,
n ¼ 2,272) minutes to arrive at these incidents,
TABLE 1. Number of Responses Made to Automatic Medical Alerts Coming into a Municipal 9-1-1 Center*
Ambulance
Only FAFR Only No. Patients (%) No. Patients (%) p-value Number of patients n¼290 n¼335 No transportation 271 (93%) 300 (90%) NSy BLS transportation 17 (6%) 22 (7%) NS ALS transportation 2 (1%) 13 (4%) 0.009 ECG monitor 1 (0.3%) 8 (2%) Intravenous access 1 (0.3%) 1 (0.3%) Drugs 1 (0.3%) 3 (1%) Endotracheal intubation 0 1 (0.3%) Defibrillation 0 0
*Also shown are the resulting rates of basic life support (BLS) and advanced life support (ALS) transportation to a hospital, and rates of ALS interventions. Comparison is shown between a six-month period using BLS ambulances as initial first-responders and the same six-month period one year later when fire apparatus first responder crews (FAFRs) were initially deployed without an ambulance. ECG¼electrocardiogram. yNS¼not significant.
presumably, diminishing time to ALS backup by nearly 3 minutes.
9-1-1 Disconnect Group. Results from the 9-1-1
disconnect group are in Table 3. During the six-month study period, there were a total of 1,028 consecutive callers to 9-1-1 who hung up without providing any emergency information, and 159 (15%) of these incidents resulted in patient transport (116 BLS, 43 ALS). Of the 1,028 total, 21 (2%) of the patients received only ECG rhythm monitoring, while an additional 13 (1%) received ‘‘precautionary’’ IV access
without administration of fluids or medications. Five
(0.5%) of the patients eventually received a medication such as nitroglycerin (with no other advanced care), but two others (0.2%) were eventually intubated (one head injury, one septic child), and two (0.2%) were defibrillated (by the FAFR crew) and later intubated for primary cardiac arrest. With the exception of delayed intubation in two cases (septic child and primary arrest), there was no adverse outcome that could be distinctly attributed by the routine quality assurance officers to the delays in the dispatch of either a transport or ALS unit.
DISCUSSION
Consistent with police and fire dispatch procedures, EMS priority dispatch systems were implemented in the mid to late 1970s.7,11,12The ability of dispatchers to identify patients in cardiac arrest and provide CPR instructions over the phone is well documented,7,19 and priority dispatch concepts have also been expanded to include discrimination between BLS and ALS situations.7,10–15,25 Earlier research efforts from this institution and others have demonstrated that a computer-aided priority dispatch algorithm can safely and reliably help dispatchers distinguish between ALS and BLS incidents.7,10,14,15,20,25 Such systems can be of value, particularly in busy urban EMS systems.22 In addition to improving skills, such systems can paradoxically improve paramedic (ALS) response intervals because they help to spare para-medic response units for the more critical calls such as cardiac arrest.7,22
As addressed in this study, one might also consider a secondary triage system to spare other transport vehicles (BLS units) from low-risk calls that often do not result in patient transport. High rates of false alarms and responses to minor problems can strain the EMS system by preoccupying those EMS crews who must respond to investigate these incidents. The control group data of this study demonstrated that many cases in certain dispatch categories do not result in patient transport to a hospital, let alone patient care, and a previous study in this system demon-strated that some categories are not likely to need ALS care.7 Therefore, this study took an additional step and looked at the feasibility of simply sending a rapidly-responding BLS first responder unit equipped with an AED to a select group of low-risk dispatch categories, including automated medical alerts, MVIs for which the caller could not provide information regarding a patient’s condition, and those incidents in which the 9-1-1 caller hung up.
Overall, the study confirmed the feasibility of sending a FAFR crew alone versus the lone response of a BLS ambulance. Sending a FAFR crew alone to these study calls allowed faster arrival of a defibrilla-tor and, theoretically, faster paramedic backup in the rare cases that might need ALS attention, because the
TABLE 3. Number of Responses Made to Incidents Involving 9-1-1 Telephone Call Disconnects at a Municipal 9-1-1 Center*
Total number of patients n¼1,028 No transportation 869 (85%) BLS transportation 116 (11%) ALS transportation 43 (4%) ECG monitor 40 (4%) Intravenous access 20 (2%) Drugs 7 (0.7%) Endotracheal intubation 4 (0.4%) Defibrillation 2 (0.2%) *Also shown are the resulting rates of basic life support (BLS) and advanced life support (ALS) transportation to a hospital, and rates of ALS interventions. Results are from the six-month period when fire apparatus first responder crews (FAFRs) were initially deployed without an ambulance. ECG¼ electrocar-diogram.
TABLE 2. Number of Responses Made to Requests Coming into a Municipal 9-1-1 Center for Motor Vehicle Incidents for Which No Patient Information Could Be Provided*
Ambulance
Only FAFR Only No. Patients (%) No. Patients (%) p-value Total number of patients n¼6,463 n¼3,372 No transportation 3,965 (61%) 2,269 (67%) \0.0001 BLS transportation 2,346 (36%) 1,021 (30%) \0.0001 ALS transportation 152 (2%) 82 (2%) NSy Critical ALS intervention (intubation/ defibrillation) 11 (0.2%) 1 (0.03%) NS (0.069) *Also shown are the resulting rates of basic life support (BLS) and advanced life support (ALS) transportation to a hospital, and rates of ALS interventions. Comparison is shown between a six-month period using BLS ambulances as initial first-responders and the same six-month period one year later when fire apparatus first responder crews (FAFRs) were initially deployed without an ambulance.
FAFRs arrived faster than BLS ambulances and could thus summon ALS sooner. Also, all FAFR units in-cluded at least one EMT-B, as well as four-person crews. Therefore, they provided the same level of BLS care, but with additional help compared with their basic ambulance crew counterparts (two EMT-Bs).
It might be assumed that this study is relevant only to a tiered ALS–BLS ambulance system. However, it could conceivably also apply to an all-paramedic ambulance system as well, in that it also delineates a number of situations in which paramedics’ responses can be spared. Truly, the main counterargument to this approach in any EMS system would be the concern over the two or three cases (of the many thousands studied here) in which an advanced airway might have been provided earlier. Emotionally, such con-cerns would warrant a universal paramedic dispatch. However, one can also argue that there is a much stronger relative risk in unnecessarily sending an additional emergency response vehicle (staffed with paramedics with lights and sirens) through traffic for the thousands of cases in which no transport or even basic medical care is required. In addition, the universal paramedic response approach mandates many more paramedics in the system, diminishing individual skills utilization and intubation success rates.6,7,22It could be further argued that there is more of a risk of delayed, failed, or unattempted intubations for the thousands of other patients (in all categories of response) in those EMS systems that do not provide for these triage mechanisms.7,22 In the particular EMS system under study, it has been demonstrated pre-viously that by creating diminished paramedic staffing needs through a dispatch triage system, ALS response times improved for hundreds of critical cases and intubation success rates also dramatically improved, thus directly affecting many dozens of patients.7 In addition, intubations are attempted more often and are accomplished earlier. Thus, ‘‘delays in intubation’’ in two cases (which may or may not have had any effect on outcome) may pose amuchsmaller risk to patients than multiple failed, or even unattempted intubations for hundreds of other patients in a lesser skilled cadre of paramedics.6,7,22 In fact, studies now indicate that survival rates are improved when such priority dispatch approaches are used.18,26,27
Naturally, the current study results probably are affected by certain biases. For example, when a BLS ambulance crew responds, presumably they are more apt to transport since they are operating a transport vehicle. In addition, they may not choose to wait for ALS backup in cases where this might be detrimental, particularly in MVIs (trauma cases). In turn, this may increase transport rates and diminish ALS use in such cases. Conversely, if a FAFR crew is unsure about the need for ALS, they may err on the side of an ALS response. In turn, an arriving ALS unit may be more predisposed to use the simple skills that they have
(ECG monitoring, IV access). Thus, the number of ALS interventions might be slightly increased and trans-ports reduced when switching from BLS to FAFR. In fact, these considerations may explain the findings in the ALERTs group for which a statistically significant increase in the number of ALS calls was noticed in the FAFR phase. The majority of these ALS calls involved the use of an ECG monitor only, while one involved intubation. Nevertheless, such biases, if they truly existed, really did not affect the overall outcomes.
Most automatic medical alerts did not result in patient transport either before (93%) or after (90%) the dispatch protocol change. Because many of the au-tomatic medical alert devices work by notifying 9-1-1 after a period of inactivity by the patient, approx-imately 60% of the time there was no one home upon the arrival of EMS units. In some cases, patients had traveled out of the city and had forgotten to turn off their devices. Also, when people were home, rea-sons given for activation of the automatic medical alert included ‘‘the grandchildren were playing with the buttons,’’ ‘‘I needed the police, not EMS,’’ and ‘‘I wanted to have my blood pressure checked.’’ One person even stated, ‘‘I just need my blinds adjusted and I can’t reach them.’’
In addition, there actually was no difference in the rate of ALS care after implementation of the FAFR-only response for MVIs without caller information. However, there was a statistically significant decrease in the rate of MVIs involving patient transportation to the hospital, including a corresponding decrease in the rate of BLS transport. Again, this outcome did not result in any known adverse affects, complaints, or quality assurance concerns.
In the 9-1-1 disconnect category, 85% resulted in no patient transportation to a hospital. There was a slightly higher rate of persons (4%) who did require ALS intervention, and four patients (0.4%) required critical ALS care (intubation or defibrillation); two of these were of potential concern in terms of delay. Nevertheless, as discussed previously, this risk is probably far outweighed by the benefits afforded by such a system.
Some observers have expressed concern about sending fire apparatus to EMS calls because of increased fuel and maintenance costs. An in-house audit of the costs associated with operating fire apparatus as FRs revealed that increased mainte-nance, fuel, and supply costs were minimal. The cost of operating all fire apparatus as FRs was approxi-mately equal to operating a single ALS ambulance (this number includes salary costs for the ambulance crews as well as fuel, maintenance, and supplies).21
LIMITATIONS
The use of historical control groups is an obvious limitation of this study. However, using a
randomiza-tion process to dispatch ambulances would have been confusing to the dispatchers as well as the field personnel, and would have required special pro-gramming of the computer-aided dispatch system. It also might have increased biases because of the dispatchers’ and field personnel’s knowledge of the study. Interestingly, compared with the historical pe-riod, there was a significant decrease in the tabulated number of requests for 9-1-1 service in which the caller could not provide information about MVIs. One consideration explaining this decrease in the number of callers unable to provide information may have been due to dispatchers’ being more aggressive about interrogating callers (who are sometimes police dispatchers when their field officers request ambu-lances, or the call was initially sent to the police department because the caller indicated no need for EMS), because they realized that a FR would be sent without an ambulance if the caller could not provide any information. In addition, more aggressive dis-patcher interrogation of the caller was conceivably responsible for the decrease in critical ALS incidents (intubation/defibrillation) and a decrease in the overall transport rate found in the dispatch of FAFR units only. This would be difficult to detect on routine quality-assurance review of dispatchers. Meanwhile, closer interrogation of the caller would not affect the automatic medical alert and 9-1-1 disconnect catego-ries because these categocatego-ries do not involve any caller interrogation. It is therefore conceivable that a more focused review of the MVIs might yield improved discrimination. In the end, however, the results do show that ambulances were spared from more than 6,700 emergency responses, including nearly 5,000 nontransport cases on an annualized basis, just for MVIs alone. The lack of ramifications (no complaints, no missed problems identified by community hospi-tals) reinforces the resource utilization tactic as being cost-effective. Indeed, this very tactic has been used for three decades without reported sequelae in some of the successful pioneer EMS systems.1
Another limitation is a lack of follow-up data on patients who were not transported. This study, however, examined sending a crew of basic EMTs to a scene either by ambulance in the control phase or a nontransport vehicle in the experimental phase. Patients are being triaged by personnel with essen-tially the same level of training, the only difference being that they arrive in a different type of vehicle. During the experimental phase, not only are basic EMTs present, but a fire officer (captain) is a part of the crew doing the assessment. In the system being studied, EMTs that staff the ambulance one day are on a fire apparatus the next day, further reinforcing that the only real difference between the two phases is the vehicle in which the personnel arrive at the scene. Routine quality-assurance surveillance is in place within the EMS system to detect inappropriate
non-transports and patient care. No adverse event associated with the study runs was reported during the study period.
CONCLUSIONS
A first-responder EMS system composed of fire engines and ladder trucks staffed by firefighter/ EMT-Bs with AEDs can safely and reliably respond alone to triage 9-1-1 hang-up calls (disconnects), automated medical alarms, and motor vehicle inci-dents for which the caller cannot provide patient information. Sending such first responders may allow the system to operate more efficiently because most of these calls do not result in patient transportation and ambulances are spared from responding to these calls. In almost every case, patients who eventually do receive ALS care have no adverse outcomes. Patients requiring ALS care received initial basic interventions, including defibrillation, in a more timely manner. The authors thank Kay T. Kimble, PhD, Baylor College of Medicine, Department of Medicine, for her assistance with statistical analysis and study design. The authors thank the firefighters of the City of Houston for their outstanding public service and readiness to research new ways to better serve their fellow citizens. They also thank Carol W. Smith for helping to prepare the manuscript.
References
1. Cobb LA, Alvarez H, Copass MK. A rapid response system for out-of-hospital cardiac emergencies. Med Clin North Am. 1976; 60:283–90.
2. Crampton RS, Aldrich RF, Gascho JA, Miles JR Jr, Stillerman R. Reduction of prehospital, ambulance and community coronary death rates by the community-wide emergency cardiac care system. Am J Med. 1975; 58:151–65.
3. Durham LA, Richardson RJ, Wall MJ Jr, Pepe PE, Mattox KL. Emergency center thoracotomy: impact of prehospital resusci-tation. J Trauma. 1992; 32:775–9.
4. Eisenberg M, Bergner L, Hallstrom A. Paramedic programs and out-of-hospital cardiac arrest: I. Factors associated with successful resuscitation. Am J Public Health. 1979; 69:30–8. 5. Lewis RP, Stang JM, Fulkerson PK, Sampson KL, Scoles A,
Warren JV. Effectiveness of advanced paramedics in a mobile coronary care system. JAMA. 1979; 241:1902–4.
6. McManus WF, Tresch DD, Darin JC. An effective prehospital emergency system. J Trauma. 1977; 17:304–10.
7. Curka PA, Pepe PE, Ginger VF, Sherrard RC, Ivy MV, Zachariah BS. Emergency medical services priority dispatch. Ann Emerg Med. 1993; 22:1688–95.
8. Emergency Cardiovascular Care Committee of the American Heart Association. 2000 Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation (Special Supplement). Aug 22, 2000; 102:1–384.
9. Pepe PE, Mattox KL, Fischer RP, Matsumoto CM. Geographic patterns of urban trauma according to mechanism and severity of injury. J Trauma. 1990; 30:1125–31; discussion 1131–2. 10. Slovis CM, Carruth TB, Seitz WJ, Thomas CM, Elsea WR. A
priority dispatch system for emergency medical services. Ann Emerg Med. 1985; 14:1055–60.
11. Clawson JJ. Emergency medical dispatch. In: Kuehl AE (ed). EMS Medical Director’s Handbook. St. Louis: C. V. Mosby, 1989, pp 59–90.
12. Clawson JJ. Emergency medical dispatching. In: Roush WR, Aranosian RD, Blair TMH, et al. (eds). Principles of EMS System: A Comprehensive Text for Physicians. Dallas: Amer-ican College of Emergency Physicians, 1989, pp 119–33. 13. Braun O, McCallion R, Fazackerley J. Characteristics of
mid-sized urban EMS systems. Ann Emerg Med. 1990; 19:536–46. 14. Kallsen G, Nabors MD. The use of priority medical dispatch to distinguish between high and low-risk patients [abstract]. Ann Emerg Med. 1990; 19:29.
15. Stratton S. Triage by emergency medical dispatchers. Prehosp Disaster Med. 1992; 7:263–9.
16. Pepe PE, Bonnin MJ, Almaguer DA, Prentice FD, Mattox KL. The effect of tiered system implementation on sudden death survival rates [abstract]. Prehosp Disaster Med. 1989; 4:71. 17. Sammel NL, Taylor K, Selig M, O’Rourke MF. New South Wales intensive care ambulance system: outcome of patients with ventricular fibrillation. Med J Aust. 1981; 2:546–50. 18. Persse DE, Key CB, Bradley RN, Rodriguez DJ, Dhingra A,
Baldwin JB. Impact of paramedic deployment strategy on cardiac arrest survival in a large urban EMS system [abstract]. Acad Emerg Med. 2000; 7:476.
19. Eisenberg MS, Hallstrom AP, Carter WB, Cummins RO, Bergner L, Pierce J. Emergency CPR instruction via telephone. Am J Public Health. 1985; 75:47–50.
20. Saunders CE, Braun O, Amick P. Accuracy of paramedic dispatchers at prospectively determining call priority and level of service (ALS vs BLS) in San Francisco [abstract]. Prehosp Disaster Med. 1990; 5:297.
21. Pepe PE, Kelly JE, Ivy MV, Rodriguez DJ. Resource utilization and impact of using fire apparatus for a fully integrated EMS first-responder program [abstract]. Ann Emerg Med. 1991; 20:488.
22. Stout J, Pepe PE, Mosesso VN Jr. All-advanced life support versus tiered-response ambulance systems. Prehosp Emerg Care. 2000; 4:1–6.
23. Sirbaugh PE, Pepe PE, Shook JE, et al. A prospective, population-based study of the demographics, epidemiology, management and outcome of out-of-hospital pediatric cardio-pulmonary arrest. Ann Emerg Med. 1999; 33:174–84.
24. Spaite DW, Valenzuela TD, Meislin HW, Criss EA, Hinsberg P. Prospective validation of a new model for evaluating emer-gency medical services systems by in-field observation of specific time intervals in prehospital care. Ann Emerg Med. 1993; 22:638–45.
25. Bailey ED, O’Connor RE, Ross RW. The use of emergency medical dispatch protocols to reduce the number of inappro-priate scene responses made by advanced life support personnel. Prehosp Emerg Care. 2000; 4:186–9.
26. Eisenberg MS, Horwood BT, Cummins RO, Reynolds-Haertle R, Hearne TR. Cardiac arrest and resuscitation: A tale of 29 cities. Ann Emerg Med. 1990; 19:179–86.
27. Nichol G, Detsky AS, Stiell IG, O’Rourke K, Wells G, Laupacis A. Effectiveness of emergency medical services for victims of out-of-hospital cardiac arrest: A metaanalysis. Ann Emerg Med. 1996; 27:700–10.
