Introduction
Emergency Medical Services (EMS) are designed to respond to and manage patients experiencing life-threatening emergencies; however, not all responses by EMS are for emergent or life-threatening conditions. Emergency Medical Services have the potential to respond more efficiently to the higher acuity patient, which is hindered by the availability of vehicles and resources transporting the low-acuity patient. Reference Dami, Golay, Pasquier, Fuchs, Carron and Hugli1 High volume response to non-emergency calls with lights and sirens not only increases the risk of motor vehicle accidents, Reference Eastwood, Morgans, Smith, Hodgkinson, Becker and Stoelwinder2 but also predisposes EMS staff to factors such as poor staff morale, poor job satisfaction, skills decay, and fatigue. Reference Eastwood, Morgans, Smith, Hodgkinson, Becker and Stoelwinder2,Reference Sofianopoulos, Williams, Archer and Thompson3
With regards to the emergency department (ED), there exists a substantial amount of research on the low-acuity (non-emergent) patient Reference Bayati, Kwasnick, Luo and Plambeck4,Reference Munjal, Belachew and Tan5 and the negative consequences these patients have on the higher acuity patients. Some of the research includes overcrowding and the time dedicated to the low-acuity patient in the ED. It has been shown that the arrival of one low-acuity patient to an ED significantly impacts the waiting times for treatment provided to a patient of high acuity. Reference Bayati, Kwasnick, Luo and Plambeck4 It is also noted that one-third of ED patients are low acuity and a major contributor to this proportion is ambulance arrivals. Reference Alpert, Morganti, Margolis, Wasserman and Kellermann6 One study has shown that 41% of patients transported to hospital were deemed to be low acuity and 79% of these patients could have been safely treated in a primary care center. Reference Munjal, Belachew and Tan5
Much research has been published around prehospital discharge or treat-and-refer procedures internationally. There is also indication that telephonic assessment of emergency calls identified patients that were less likely to require treatment in the ED and therefore had the ability to reduce unwarranted response of emergency vehicles to some calls. Reference Dale, Williams and Foster7 A further study showed that telephonic advice for these calls had proven to be safe and recommended full implementation in a clinical trial setting. Reference Cameron, Gabbe, Smith and Mitra8
In the US alone, it is estimated that US$560 million could be saved annually through the appropriate management of patients with lower acuity. Reference Alpert, Morganti, Margolis, Wasserman and Kellermann6 Consideration may be provided for emergency care practitioners in local EMS systems to provide on-scene discharge of patients and that the exact operational procedures for this requires further consideration and research prior to implementation. Reference Goldstein, Sibanyoni and Vincent-Lambert9 Although prehospital discharge is a useful mechanism in reducing low-acuity patient transport to hospitals, it does not alleviate the impact of the initial ambulance response to these calls. Dispatching of EMS resources timeously and appropriately based on patient acuity would be the ideal dispatching system, balancing patient safety and adequate resource allocation and utilization. Reference Cameron, Gabbe, Smith and Mitra8
In South Africa, the problem of lack in emergency resources is well-established with only 2,000 of the required 5,700 ambulances being in operation, 10 therefore compounding the highlighted predisposing factors with emergent response to low-acuity calls. It was evidenced that 39% of patients presenting to an ED were transported by ambulance, Reference Hodkinson and Wallis11 and 47% of these patients did not require admission. Reference Hodkinson and Wallis11 Another study showed that 58% of all patients attended to by EMS required no prehospital clinical intervention (only transport) and concluded that EMS systems are unable to meet the demand as a result of such responses. Reference Newton12
Accuracy in dispatching resources according to triage is a key issue to bridge the gap between patient medical needs and EMS resources dispatched. Reference Stassen, Larsson, Wood and Kurland13 This extends to existing dispatching protocols to be efficient and accurate in categorizing and triaging patients. International research all indicate a problem with over-triage resulting in the unnecessary response and transport of the low-acuity patient to hospital. Reference Stassen, Larsson, Wood and Kurland13-Reference Feldman, Verbeek, Lyons, Chad, Craig and Schwartz15 Over-triage internationally ranges from 23% to as much as 78%. Over-triage in South Africa has been reported to be as high as approximately 94%, Reference Newton12 which is staggering in comparison to the rest of the world.
A popular system used in dispatching of EMS is the Medical Priority Dispatch System (MPDS). This dispatching system has reduced EMS response to calls of low acuity, however still has some problems in over-triage. Reference Sporer, Craig, Johnson and Yeh16,Reference Clawson, Olola, Heward, Patterson and Scott17 Major aspects to be considered with this system are the cost, which is in excess of $10M, 18 and the lack of validation of this system in a low- and middle-income country’s (LMIC’s) EMS environment, like South Africa. With eleven official languages, the diverse language profile of South Africa is another contributor to the accuracy of dispatch. This is particularly pronounced during emergency call taking, as the telephonic words and descriptors of the caller really provide the only clues for Emergency Medical Dispatch (EMD) staff to recognize and accurately assign acuity of patients. Reference Mann, Schmidt and Cone19
Research in the appropriate use of EMS resources has tried to define greater efficiencies in protocols used for the dispatching of resources, however this has been difficult. Organizations and countries operate independently and protocols coupled with outcome criteria for each EMS service vary, therefore studies have been difficult to evaluate. Reference Munjal, Belachew and Tan5 However, a consistent fact is that if the dispatch protocol is more accurate, this will reduce the amount of unnecessary resources being dispatched to lower acuity calls. Problems associated with low acuity are compounded by the fact that there is no clear evidence-based criteria to the concept of “low acuity in EMS,” Reference Cone, Galante and MacMillan20 and those available are vague and vary subjectively between health care providers. Reference Hasson, Keeney and Mckenna21
The aim of this study is to develop consensus-based descriptors for low acuity in EMS within South Africa.
Methods
Design
A Modified Delphi study was conducted to obtain a set of consensus-based descriptors of low-acuity patients in EMS through experts in the field of prehospital and emergency medicine. These experts are well-experienced in EMS 22 and from various geographical locations around the country.
Setting
Emergency Medical Services in South Africa service patients through public and private systems. The public system is funded by the Department of Health, while private EMS service patients through reimbursement models such as medical insurance. In South Africa, 82% of the population does not have medical insurance and therefore relies on the public health care system. From an EMS perspective, there should be approximately 5,700 ambulances (1/10,000 population) Reference Ludwig23 in South Africa, but there are less than 2,000 ambulances in operation. 10
Sample and Sampling
Participants recruited for this study were professional medical providers with tertiary qualifications within prehospital and emergency medicine. Participants were also required to be registered with the relevant professional regulating body in South Africa. Further, only participants with a minimum of two years operational experience within South Africa were eligible for participation.
The initial group of participants were nominated by the researcher via purposive sampling to ensure quality of the candidates. Reference Oh24,Reference Jünger, Payne, Brine, Radbruch and Brearley25 Participants were then provided the opportunity to nominate more candidates for the study through snowball sampling. A list of interested candidates with contact details was gathered and invited to participate in the study. 22 These participants were evaluated by the researcher prior to invitation and were excluded if minimum recruitment criteria were not attained. The online survey tool Lime Survey (LimeSurvey GmbH; Hamburg, Germany) was used to conduct the study.
Procedure
This survey was conducted in three rounds. All participants were requested to provide informed consent for their participation.
The first round was developed through information gathered from literature available. Reference Sporer, Craig, Johnson and Yeh16,Reference Clawson, Olola, Heward, Patterson and Scott17 Participants were asked for their agreement on three initial binary questions that should not be considered as low acuity (respiratory distress, unconsciousness, and severe hemorrhage). The balance of questions posed to the expert panel were open-ended questions under groups of potential emergency response categories as identified in the literature. Participants were asked to provide their expert opinion on criteria that could denote a patient as low acuity under each of the emergency response categories presented to them.
After the first round, open-ended questions within each emergency response category resulted in descriptors of low acuity being developed. Following content analysis, these descriptors were presented to the panel in round two through binary outcomes of “agree” or “disagree.” Participants were also provided free-text options to justify their responses. These comments were condensed and subjected to content analysis and presented to participants in round three which offered clarification and information for further consideration. Round three of the survey had binary outcomes only with no provision for free text in order to finalize consensus. Figure 1 depicts the process used.
Figure 1. Process Flow of Delphi Study.
Consensus for each descriptor was obtained when 75% of the participants either agreed or disagreed to the binary question. Consensus obtained on a particular descriptor would result in it either being excluded or included in the final list and was not presented to the participants in a subsequent round.
This study was approved by the Human Research Ethics Committee (HREC) of the University of Cape Town (Cape Town, South Africa; HREC Ref 491/2019). The results are presented in accordance with the proposed Conducting and REporting DElphi Studies (CREDES) checklist. Reference Alshehri, Pigoga and Wallis26
Results
Invitations were sent to 60 candidates, 31 of whom completed round one of the survey providing a response rate of 52%. In round two, 30 responses were received with an attrition rate of three percent. A further attrition rate of three percent was noted with 29 respondents in round three, providing a total response rate of 94% for the study. Candidates that completed all three rounds of the survey (n = 29) were either Advanced Life Support practitioners in EMS (n = 26) or a medical doctor employed in the field of prehospital or emergency medicine (n = 3). Demographics of the participants that completed all three rounds in the Delphi are presented in Table 1.
Table 1. Demographics of Participants
Table 2 provides a summary of the levels of consensus achieved through each round of the survey. Consensus was achieved in round one on all three binary disqualifier questions posed to participants, which were then excluded in the following rounds. Round one also yielded (n = 152) descriptors for all emergency response categories provided, which were then posed to participants in round two for consensus and opinion.
Table 2. Summary of Consensus
Round two of the survey achieved 44% (n = 67) consensus on all binary questions, which were then excluded in round three. For the remaining questions in round two, the free text was analyzed, which resulted in questions being excluded (n = 29) and questions being clarified (n = 85) which were then asked again in round three. The final round achieved 54% (n = 46) consensus on all questions, which resulted in an overall 75% consensus received on all binary questions through all three rounds.
Table 3 presents the descriptors that reached consensus on conclusion of the Delphi survey and therefore provide criteria for low acuity within each emergency response category.
Table 3. Descriptor Consensus for Low Acuity Calls in South Africa
Abbreviations: BSA, body surface area; CNS, central nervous system; GCS, Glasgow Coma Scale; RTA, road traffic accident.
Discussion
The aim of this study was to develop descriptors for low-acuity emergency calls for the South African context. “Low acuity” is a term that categorizes patients subjectively and is vague in its definition Reference Hasson, Keeney and Mckenna21 making it difficult to apply operationally. Emergency calls in the EMD are also over-triaged as there are concerns that patients may die or staff may be subject to medicolegal litigation for incorrect dispatch. 27
Determining which patients require an emergency response is a challenge and often results in disagreements between EMD staff and EMS operational staff. 27 New and safe dispatching protocols in the EMD are required for the reduction in unnecessary responses and transport of patients. 27 Their development may be guided through the implementation of algorithms based on the criteria for low-acuity denotation identified in this study. However, the complexity of developing these algorithms is exemplified in these results as each emergency response category had (mostly) unique consensus-based criteria that should be applied.
The panel identified three universal criteria that disqualified cases from being deemed as low acuity. The panel felt that instances where loss of consciousness, respiratory distress, or severe hemorrhage are reported cannot be considered as low acuity. Respiratory distress is of particular interest as this was highlighted as one of the emergency response categories with a high risk for over-triage, but was also ranked as the third highest emergency category in the Western Cape, South Africa. 27 Respiratory pathologies are ranked in the top ten causes of death in South Africa 28,Reference Health29 and were also shown to have a reduced discharge rate from the ED in a similar study conducted in Gauteng, South Africa. Reference Goldstein, Sibanyoni and Vincent-Lambert9 Comments received from the expert panel indicated that respiratory complaints are difficult to denote as low acuity telephonically because such a decision would be entirely dependent on an unknown caller’s ability to assess a patient in respiratory distress. The risk in dispatching an emergency case with a patient in respiratory distress is evident in this study and therefore recommend that these cases be distinguished as high acuity with the development of an algorithm to enhance these dispatches.
The most prevalent cause of death in South Africa after HIV is cardiovascular disease, which has shown a remarkable increase from 2007 to 2017. 28 The expert panel in this study felt very strongly that patients exhibiting chest pain cannot be categorized as low acuity. Statistics of call-outs from the EMD in the Western Cape for chest pain are slightly above two percent of the total emergency calls received, however all staff indicated that these emergency response categories carried a high risk of over-triage. 27 The potential of over-triage for these chest pain patients are evident as common symptoms of a heart attack can be linked to other forms of pathologies, such as respiratory problems, heartburn, or even anxiety attacks. Reference Buma, Saunders, Watermeyer and Stassen30
In the case then of respiratory distress and chest pain, it is perhaps not which calls to denote as low acuity, but instead to accept a priority response from the outset and develop algorithms that may instead distinguish those patients with higher risk from the others. Reference Mould-Millman, Dixon and Sefa31 This may ensure that those prehospital providers with the requisite scope of practice and equipment to manage patients with asthma and myocardial infarction be dispatched preferentially. These algorithms will be heavily reliant on the caller’s description of the patient. Reference Buma, Saunders, Watermeyer and Stassen30
The highest discharge rate for medical patients in Gauteng brought in by ambulance was among those with gastrointestinal complaints (36%). Reference Goldstein, Sibanyoni and Vincent-Lambert9 Gastrointestinal emergencies also constitute almost 11% of emergency responses in the Western Cape. 27 The high rates of call-outs, together with discharge rates, indicate that these patients may not have warranted transport to hospital but may have benefitted from other medical assistance such as referral to primary health care facilities or telephonic advice. Consideration must be provided for confounding factors in these cases such as the prevalence of gastrointestinal complaints found in patients with sepsis; a South African study found that 40% of patients with the prehospital diagnosis of sepsis were described as having gastrointestinal symptoms at initial emergency telephone call. Reference Mann, Schmidt and Cone19 The consequent descriptors from the expert panel may provide distinguishing criteria for low-acuity cases in reducing the high call-out rates and therefore contribute to the development of algorithms in reducing unnecessary dispatch of abdominal cases.
Obstetric emergencies in Africa have been the second highest call-out volume for EMS Reference Thaddeus and Maine32 and make up six percent of call-outs in the Western Cape. 27 Although delays in these responses contribute to maternal mortality through obstetric complications, Reference Pacagnella, Cecatti and Parpinelli33 implementation of any dispatch protocol in this respect requires meticulous consideration in order to mitigate maternal deaths, which are known to be as high as 84% in LMICs. Reference Gilboy and Tanabe34
Road traffic accidents (RTAs; 47%) and patient assaults (37%) accounted for 83% of trauma-related discharges brought in by ambulance. Reference Goldstein, Sibanyoni and Vincent-Lambert9 These statistics are significant, especially if there exist mechanisms to redirect these patients from the ED, providing efficiency to the health care system. Not dispatching emergency vehicles to RTAs received via telephone may be a challenge as injuries at the time of the call may not be known; however, the descriptors may provide value in supporting other systems like on-scene discharge to limit the unnecessary transportation of patients to the ED. Many of the patients in these cases with minor medical requirements, like alleviation of pain or tetanus toxoid injections, may be transported or preferably referred to primary care clinics after evaluation by emergency medical staff. Reference Gilboy and Tanabe34
The development of algorithms using these descriptors in each call-out category may not only terminate responses to calls of low acuity, but may also provide distinguishing criteria between high-acuity and low-acuity dispatch models and contribute to alternative dispatch systems that can influence policy of response to calls. This may be in the form of a time critical response approach to some calls where lights and sirens are required, or other cases where an acceptable time frame for response may be determined through evidence-based decisions. The implementation of algorithms in dispatching emergency resources to calls will also support other EMS systems such as on-scene discharge or telephonic advice systems. The challenge of receiving emergency calls in the EMD in a multilingual country with varying levels of literacy and education may also be alleviated through protocol-driven algorithms which can be informed through the low-acuity descriptors of this study.
Limitations
The results and recommendations of this study are based on expert opinion, which represents a low level of evidence-based medicine. For this reason, further development of these criteria is recommended as well as future research to refine and test these results.
Participants for this study were all South African registered emergency care practitioners and medical doctors to ensure that descriptors were contextual to the setting. This might limit the external validity of the study and results should be applied to other settings with caution.
Conclusion
The results of this study have provided a set of descriptors in order to define low acuity within the EMS environment of South Africa. A significant level of responsibility lies with the EMD center in reducing the dispatch of resources to low-acuity calls. These descriptors may be developed into algorithms that will affect an appropriate response to emergency calls. Further research is required in validating the application of these definitions in the EMS environment.
Conflicts of interest
The authors declared no conflicts of interest.
Open access
