EDUCATION AND PRACTICE

EDUCATION AND PRACTICE VALIDATION OF LOW-ACUITY EMERGENCY MEDICAL SERVICES DISPATCH CODES Glen E. Michael, BA, Karl A. Sporer, MD ABSTRACT Background. Computer-aided dispatch systems are used to assess the severity of a 9-1-1 caller s complaint and then assign an appropriate level of emergency medical services (EMS) response. Objective. To evaluate a group of low-acuity codes (defined as requiring advanced life support [ALS] intervention in fewer than 10% of cases) that has been derived and validated in one community. Methods. All of the 9-1-1 medical calls assigned to these predetermined emergency medical dispatch codes between January 1, 2004, and July 1, 2004, were analyzed. ALS care was defined as receiving one or more of the following: pulse oximetry measurement, blood glucose measurement, cardiac defibrillation, administration of any medication, airway maneuvers, or the placement of an intravenous (IV) catheter. A more restrictive definition of ALS care (use of IV fluid bolus, medication administration, intubation, or defibrillation) was also calculated. Results. A total of 1,799 calls were assigned low-acuity dispatch codes, and 1,597 met inclusion criteria. None of the 26 dispatch codes were found to be low-acuity by the study definition. Fifty-six percent of these patients received ALS care. Placement of an IV-catheter was the ALS intervention used most frequently (45% of cases), followed by pulse oximetry measurement (32%), glucose measurement (22%), medication administration (11%), intubation (0.13%), and defibrillation (0%). The medication administered most frequent was morphine. When using the more restrictive definition of acuity, patients in 19 of the 28 categories received ALS intervention less than 10% of the time. Patients in the other seven categories were considered high-acuity 13% to 36% of the time. Conclusion. Dispatch codes that had previously Received January 4, 2005, from the Department of Medicine, University of California, San Francisco (GEM, KAS), San Francisco California; and the Department of Emergency Services, San Francisco General Hospital (KAS), San Francisco, California. Revision received May 9, 2005; accepted for publication May 10, 2005. Address correspondence and reprint requests to: Karl A. Sporer, MD, UCSF/San Francisco General Hospital, Department of Emergency Medicine, 1001 Potrero Avenue, Room 1E21, San Francisco, CA 94110. e-mail: <ksporer@sfghed.ucsf.edu>. doi:10.1080/10903120500255875 been determined to be low-acuity were found not to be so in this community. The variation in clinical practice is likely explained by a more precautionary approach to care in this EMS system and the increased use of analgesics. This study demonstrates the need to define the optimal subset of prehospital patients who would benefit from these treatments. Key words: ambulances; cardiopulmonary resuscitation; emergencies; emergency medical service communication systems; emergency medical services; retrospective studies; risk assessment; triage; acuity; wounds and injuries; dispatch codes. PREHOSPITAL EMERGENCY CARE 2005;9:429 433 The Medical Priority Dispatch System (MPDS) is used by 9-1-1 call centers to assess the severity of a caller s complaint and then assign an appropriate level of emergency medical response. Callers are asked a series of scripted questions, and their answers are used to categorize each case into a standardized, complaint-based code. A group of these codes with low acuity (defined as requiring advanced life support [ALS] intervention in fewer than 10% of cases) has been derived and validated in one community. 1,2 The accurate identification of low-acuity calls at the time of dispatch would allow a lower level of medical support to be dispatched (i.e., a basic life support [BLS] ambulance rather than an ALS unit), resulting in a more appropriate allocation of emergency medical services (EMS) resources and potentially improving an EMS system s cost-efficiency. In addition, ambulances traveling to low-acuity dispatches could be instructed to reduce their response priority, thus significantly diminishing the risk of fatal motor vehicle accidents. 3 The objective of this study was to validate, in our own community, the ability of previously derived lowacuity dispatch codes to accurately identify 9-1-1 calls that require ALS interventions fewer than 10% of cases. Secondary objectives were to determine what ALS interventions were performed within the lowacuity dispatch codes and to examine whether these 429

430 PREHOSPITAL EMERGENCY CARE OCTOBER /DECEMBER 2005 VOLUME 9/NUMBER 4 interventions were time-dependent or emergent. Our hypothesis was that those patients assigned a lowacuity MPDS code would require ALS intervention in fewer than 10% of cases, and that those ALS interventions occasionally needed would not be timedependent or emergent. METHODS San Mateo County is an urban/suburban region with a population of 700,000 and a size of 552 square miles. San Mateo County receives approximately 40,000 calls for emergency medical assistance annually. All 9-1-1 calls are received at a county-operated emergency communications center. This call center is currently in the process of becoming recognized as a Center of Excellence by the National Academy of Emergency Medical Dispatch. Each EMS call is assigned a dispatch code using the MPDS system (version 11), and a paramedic engine and a paramedic paramedic/emergency medical technician (EMT)-staffed ambulance are dispatched to the location in need of medical assistance. The computer-aided dispatch system records general information regarding each call, including date, time, and location of call, dispatch time, dispatch code, and disposition (transport, patient refusal, or cancellation). A single agency provides dispatch and emergency medical services to 95% of the county. An electronic prehospital care record is generated for each patient receiving medical attention. This record includes data regarding patient demographics, medical history, signs and symptoms, and clinical interventions. This retrospective cohort study analyzed all calls for EMS care in San Mateo County that were assigned previously derived low-acuity dispatch codes between January 1, 2004, and July 1, 2004. 1 These groups were defined as requiring ALS care in fewer than 10% of cases. For consistency, we maintained the same code groupings as used by the previous authors. Calls were excluded if the electronic prehospital care record could not be located (11.23%). This study was approved by the University of California, San Francisco, Committee on Human Research. We defined a low-acuity call (i.e., basic life support [BLS] care) as either a case in which the patient did not receive ALS care or a documented cancellation. ALS care was defined as any case in which the patient received one or more of the following: pulse oximetry measurement, blood glucose measurement, cardiac defibrillation, administration of any medication, airway maneuvers, or placement of an intravenous (IV) catheter. A secondary analysis was done using a more restrictive definition of ALS care that we believed more closely matched patient acuity. This definition included any case in which the patient received any IV fluid bolus, any medication administration, intubation, or defibrillation. Data from the computer-aided dispatch system were matched to the corresponding records in the electronic prehospital care database. The resulting Microsoft Access (Microsoft Corp., Redmond, WA) data set included dispatch code assigned, patient disposition, and all prehospital observations and interventions. RESULTS During the six-month study period, there were a total of 13,386 EMS medical calls, and 1,799 calls were assigned low-acuity dispatch codes. Medical records could not be located for 202 patients, and 1,597 patients met the inclusion criteria (1,167 transports and 430 cancellations). Two of the original 28 low-acuity MPDS codes had no calls remaining after application of the exclusion criteria. None of the remaining 26 dispatch codes were found to be low-acuity by this definition. Eight hundred, ninety-seven, or 56%, of the 1,597 analyzed subjects received ALS care. Table 1 shows the dispatch codes analyzed and the frequency of BLS care within each code. Figure 1 presents a comparison of the results of this study with those of the study used by Shah et al. to derive the low-acuity dispatch codes. 1 Figure 2 shows the ALS interventions that were provided to patients assigned to each low-acuity dispatch code. For the study population as a whole, IV placement was the ALS intervention used most frequently (45% of cases), followed by pulse oximetry measurement (32%), glucose measurement (22%), medication administration (11%), and intubation (0.13%). The following medications were administered to this group of patients: morphine, 136 patients (8.5%); albuterol, 14 (1%); nitroglycerin, 10 (<1%); aspirin, nine (<1%); glucose, six (<1%); normal saline bolus, five (<1%); epinephrine, one (<1%); and atropine, one (<1%). The majority of the morphine administration was for two codes, falls (17A) and traumatic injuries, specific (30A). Falls (17A) is an illustrative EMD code, with 346 patients, 68% of whom received ALS care. The treatments that this group received included IV catheter placement (46%), glucose measurement (19%), pulse oximetry (30%), medication administration (20%), normal saline infusion (0.6%), and intubation (0.3%). The medications that were given included morphine (67 patients), albuterol (1 patient), atropine (1 patient), and epinephrine (1 patient). When using the more restrictive definition of ALS care (use of IV fluid bolus, medication administration, intubation, or defibrillation), patients in 19 of the 28 categories received ALS intervention less than 10% of the time. Patients in the other seven categories received ALS care 13% to 36% of the time.

Michael and Sporer Low-Acuity EMD PREHOSPITAL EMERGENCY CARE OVTOBER/DECEMBER 2005 VOLUME 9/NUMBER 4 431 TABLE 1. Frequency of Basic Life Support (BLS)-only (Low-acuity) Care in Low-acuity Medical Priority Dispatch System Codes in San Mateo County Dispatch Included Percent Code Descriptor Remarks Calls BLS 1A Abdominal pain/problems Abdominal pain 132 23.5 1C2 Abdominal pain/problems Females with fainting or near fainting, aged 12 50 yr 28 21.4 1C3 Abdominal pain/problems Males with pain above navel aged 35 yr 41 9.8 1C4 Abdominal pain/problems Females with pain above navel aged 45 yr 43 11.6 4A Assault/sexual assault 6 66.7 4B1A Assault/sexual assault Possibly dangerous body area 23 78.3 4B3A Assault/sexual assault Unknown status 4 75.0 4D3A Assault/sexual assault Abnormal breathing 8 62.5 4D4A Assault/sexual assault Dangerous body area 1 0.0 5A Back pain 98 45.9 7A Burns (scalds)/explosion 5 80.0 16A Eye problems/injuries 6 83.3 17A Falls 346 41.6 18A Headache 10 40.0 18C4 Headache Sudden onset of severe pain (<3 hours) 25 40.0 21A Hemorrhage/lacerations 38 81.6 24A Pregnancy/childbirth/miscarriage First-trimester hemorrhage or miscarriage 3 0.0 24B1 Pregnancy/childbirth/miscarriage Labor (delivery not imminent, >5 months) 0 24B2 Pregnancy/childbirth/miscarriage Unknown status 0 24D3 Pregnancy/childbirth/miscarriage Imminent delivery (>5 months) 17 17.7 24D5 Pregnancy/childbirth/miscarriage High-risk complications 5 0.0 25A Psychiatric/abnormal behavior Nonviolent and nonsuicidal (alert) 31 61.3 25B1 Psychiatric/abnormal behavior Violent 41 68.3 25B2 Psychiatric/abnormal behavior Threatening suicide 13 53.9 25B4 Psychiatric/abnormal behavior Unknown status 12 66.7 26A Sick person No priority symptoms 436 44.7 30A Traumatic injuries, specific 122 45.9 30B1 Traumatic injuries, specific Possibly dangerous body area 103 63.1 Total 1,597 43.8 DISCUSSION Dispatch protocols tend to err on the side of sensitivity at the expense of specificity, resulting in overtriage and potential overtreatment by prehospital providers. The propensity for overtriage and overuse of ALS stemming from the MPDS system specifically has been documented in several studies. 4 6 The validation and proof of safety of low-acuity dispatch codes would allow a more nuanced approach to dispatching an appropriate level of prehospital resources to each 9-1-1 call, and could produce significant cost savings FIGURE 1. Advanced life support (ALS) rates in two communities, by dispatch code. Rochester data from: Shah MN, Bishop P, Lerner EB, Czapranski T, Davis EA. Derivation of emergency medical services dispatch codes associated with low-acuity patients. Prehosp Emerg Care. 2003;7:434 9.

432 FIGURE 2. PREHOSPITAL EMERGENCY CARE OCTOBER / DECEMBER 2005 VOLUME 9 / NUMBER 4 Advanced life support (ALS) interventions performed, by dispatch code, in San Mateo County. IV-intravenous. in many EMS systems. However, our results clearly show that the dispatch codes determined to be lowacuity in one community are not low-acuity in our community. The significant variance between our results and those of Shah and colleagues (Figure 1) could be attributable to numerous factors. One likely explanation is regional variation in prehospital protocols and practices. For example, 45% of all subjects in our study had IV catheters placed. This could be a result of a precautionary approach to prehospital care or simply the technical imperative of having a paramedic evaluate and transport all medical patients. Similarly, the use of morphine sulfate in subjects with painful injuries and illnesses was more prevalent in our system. In our study, subjects assigned to code 5A (back pain) received morphine in 12% of cases, those in code 17A (falls) in 18% of cases, and those in code 30A (traumatic injuries) in 26% of cases. Contrasting with these numbers, the proportions of subjects within these codes receiving any ALS interven- tions in the Shah study were 4.1%, 6.5%, and 3.1%, respectively. Other studies have also demonstrated wide regional variation in the prehospital provision of analgesia7 11 Placement of an IV catheter and measurement of oxygen saturation via pulse oximetry accounted for the vast majority (68.2%) of ALS interventions in our study population. It is possible that in most cases these procedures were precautionary only and, therefore, not by themselves indicative of high-acuity illness. For example, while 45% of all subjects had an IV catheter placed, only 1.6% of these actually received an infusion of fluids, and only approximately 10% received IV medication. The more restrictive definition of ALS care demonstrated that patients in 19 of the 28 categories received one of these interventions less than 10% of the time and are possible candidates for low-acuity calls (Table 2). Patients in the other seven categories received these interventions 13% to 36% of the time.

Michael and Sporer Low-Acuity EMD PREHOSPITAL EMERGENCY CARE OVTOBER/DECEMBER 2005 VOLUME 9/NUMBER 4 433 TABLE 2. Patients Receiving Emergent Advanced Life Support Interventions No. Subjects Percent Receiving Receiving Medication, Medication, No. Intubation, Fluid Intubation, Dispatch Included Infusion, or Fluid Infusion, Code Calls Defibrillation or Defibrillation 1A 132 18 13.6 1C2 28 5 17.9 1C3 41 2 4.9 1C4 43 2 4.7 4A 6 0 0.0 4B1A 23 2 8.7 4B3A 4 0 0.0 4D3A 8 0 0.0 4D4A 1 0 0.0 5A 98 16 16.3 7A 5 1 20.0 16A 6 0 0.0 17A 346 75 21.7 18A 10 0 0.0 18C4 25 2 8.0 21A 38 0 0.0 24A 3 0 0.0 24B1 0 0 24B2 0 0 24D3 17 0 0.0 24D5 5 1 20.0 25A 31 1 3.2 25B1 41 0 0.0 25B2 13 0 0.0 25B4 12 1 8.3 26A 436 34 7.8 30A 122 39 32.0 30B1 103 8 7.8 Total 1597 207 13.0 LIMITATIONS This retrospective study was subject to a number of limitations that should be noted. The sample size used for some dispatch codes was quite small. We did not include hospital outcomes and we were not able to determine which ALS interventions were medically necessary. It is also important to consider a distinction between the methods used in our study and those used in the Shah study. While Shah and colleagues excluded from their study population calls that were cancelled, we included such calls and assigned them to the low-acuity pool. It is likely that most cancelled calls involve patients who are not in need of ALS care, and such calls should be taken into account in any study seeking to create pragmatic dispatch triage guidelines for EMS systems. If anything, this approach led to an overestimation in our study of the numbers of BLS-level subjects. Dispatched calls that did not have a matched patient care record or a documented disposition were excluded (11%). Experience in our system suggests that these patients were commonly undocumented cancellations, and their inclusion would have similarly increased the possible BLS population. CONCLUSIONS The low-acuity dispatch codes derived in a previous study could not be validated in our system. On the contrary, our results demonstrate that patients assigned a low-acuity dispatch code in fact received ALS interventions in the majority of cases. Another measure of acuity demonstrated that patients in 19 of the 28 categories received ALS interventions less than 10% of the time, while patients in the other seven categories received ALS care 13% to 36% of the time. The variation in clinical practice is likely to be explained by a more precautionary approach to care in our EMS system and the more frequent use of analgesics. The authors thank the employees of the San Mateo County EMS Agency and the San Mateo Public Safety Communications Division for their support in this study. References 1. Shah MN, Bishop P, Lerner EB, Czapranski T, Davis EA. Derivation of emergency medical services dispatch codes associated with low-acuity patients. Prehosp Emerg Care. 2003;7:434 9. 2. Shah MN, Bishop P, Lerner EB, Fairbanks RJ, Davis EA. Validation of using EMS dispatch codes to identify low-acuity patients. Prehosp Emerg Care. 2005;9:24 31. 3. Kahn CA, Pirrallo RG, Kuhn EM. Characteristics of fatal ambulance crashes in the United States: an 11-year retrospective analysis. Prehosp Emerg Care. 2001;5:261 9. 4. Kennedy JD, Sweeney TA, Roberts D, O Connor RE. Effectiveness of a medical priority dispatch protocol for abdominal pain. Prehosp Emerg Care. 2003;7:89 93. 5. Cone DC, Wydro GC. Can basic life support personnel safely determine that advanced life support is not needed? Prehosp Emerg Care. 2001;5:360 5. 6. Schmidt T, Atcheson R, Federiuk C, et al. Evaluation of protocols allowing emergency medical technicians to determine need for treatment and transport. Acad Emerg Med. 2000;7:663 9. 7. Abbuhl FB, Reed DB. Time to analgesia for patients with painful extremity injuries transported to the emergency department by ambulance. Prehosp Emerg Care. 2003;7:445 7. 8. Fullerton-Gleason L, Crandall C, Sklar DP. Prehospital administration of morphine for isolated extremity injuries: a change in protocol reduces time to medication. Prehosp Emerg Care. 2002;6:411 6. 9. McEachin CC, McDermott JT, Swor R. Few emergency medical services patients with lower-extremity fractures receive prehospital analgesia. Prehosp Emerg Care. 2002;6:406 10. 10. Vassiliadis J, Hitos K, Hill CT. Factors influencing prehospital and emergency department analgesia administration to patients with femoral neck fractures. Emerg Med (Fremantle). 2002;14:261 6. 11. White LJ, Cooper JD, Chambers RM, Gradisek RE. Prehospital use of analgesia for suspected extremity fractures. Prehosp Emerg Care. 2000;4:205 8.