Sepsis is a time-sensitive illness with high morbidity and mortality rates. It has received global recognition and prioritisation over the past 15 years. Medical and governmental bodies have recognised the serious consequences of sepsis (Dellinger et al, 2013). In Ireland, mortality, while improving, still stands at 17% for adults and is in a range of 2–4% in children aged under 16 years (Health Service Executive, 2017). Mortality rates in the United States are similar at 14.5% (Paoli et al, 2018). In the UK, an all-party parliamentary group established a confidential enquiry into sepsis, which concluded that the care of two out of three people with sepsis could be improved (National Confidential Enquiry into Patient Outcome and Death (NCEPOD), 2015).
Emergency medical services (EMS) personnel are often the first point of medical care for patients with severe sepsis (Wang et al, 2010), and they influence the type of care patients receive and when they receive it.
Estimates of the prehospital incidence of sepsis range from 1.8% of EMS calls in the UK (McClelland and Jones, 2015) to 3.3% in the United States (Seymour et al, 2012) and 3.8% in Switzerland (Tusgul et al, 2017), while patients with infection account for 6.9% of EMS transports in an American study (Guerra et al, 2013).
EMS staff being able to recognise time-sensitive conditions early and start patients on definitive care pathways has proved to be valuable for conditions such as ST-elevation myocardial infarction (STEMI) and stroke (Acker et al, 2007; Moyer et al, 2007). Prompt, accurate recognition of severe sepsis in the prehospital period could offer an invaluable opportunity to reduce time to IV fluids, antibiotics and hospital notification, thereby improving outcomes in patients with severe sepsis.
With no single definitive diagnostic test for sepsis (Carrigan et al, 2004; Singer et al, 2016), early recognition is often based on a variety of clinical indicators such as markers for systemic inflammatory response syndrome (SIRS) or signs of end-organ dysfunction, e.g. altered cognition, reduced urinary output and hypotension. Numerous studies have found that recognition of severe sepsis by EMS personnel varies a great deal (Guerra et al, 2013; Lane et al, 2016; Smyth et al, 2016).
Furthermore, much research on the recognition of severe sepsis in the prehospital setting focuses on identifying clinical signs of the condition, which is relevant only when clinicians reach the patient at the scene. This creates serious problems for call-takers and dispatchers when allocating a call priority and dispatching resources to such patients.
The aim of this audit was: to use the information on patient care reports (PCRs) to identify those with potential sepsis or severe sepsis; to examine the care provided and benchmark this against national standards; and, finally, to assess whether the medical dispatch system call prioritisation correlated with the clinical findings.
The study population were patients attended by Dublin Fire Brigade's (DFB) fire-based EMS service. It operates one of the two public, government-funded ambulance services in Ireland. It serves the people of Dublin city and county, which have a total population of 1.2 million. DFB employs both paramedic and advanced paramedic practitioners, all of whom are registered with the Pre-Hospital Emergency Care Council (PHECC), the national regulator for prehospital care.
In Ireland, there are national prehospital clinical practice guidelines (CPGs) that emergency medical technicians, paramedics and advanced paramedics for statutory, voluntary and auxiliary agencies should follow so that all patients are treated to the same standard irrespective of the organisation providing care.
PHECC-registered paramedics undergo a 2-year training programme and can administer 26 medications. PHECC-registered advanced paramedics can deliver advanced life support (including intubation, intravenous (IV) or intraosseous cannulation) and administer 45 medications (PHECC, 2017).
PHECC's medical advisory committee, which comprises prehospital practitioners and medical doctors, devise the CPGs. The sepsis CPG for paramedics and advanced paramedics was devised in conjunction with Ireland's Health Service Executive's (HSE) sepsis steering committee (HSE, 2018) as part of an integrated national approach to sepsis recognition and management. CPGs are periodically reviewed and updated.
When a patient or bystander in Dublin city or county calls 999, 112 or 911, their call is answered by the DFB eastern regional control centre (ERCC) call-takers and dispatchers using a scripted medical priority dispatch system that codes the calls. DFB has been using the International Academy of Emergency Medical Dispatch (IAEMD) Advanced Medical Priority Dispatch System (AMPDS) as its call-management system since 2004. The ERCC has been accredited by the IAEMD since 2006 (IAEMD, 2020). AMPDS is supported by the PHECC's priority dispatch committee.
This call-management system both classifies the chief complaint of the call and allocates a call priority. The call is categorised by urgency of response required or call priority as an echo, delta, charlie, bravo, alpha or omega call. Echo and delta calls have the highest priority and alpha and omega calls are the lowest priority. In Ireland, the EMS response (which includes vehicle type, practitioner level and response time) is determined by the PHECC priority dispatch committee (Table 1). The nature of the patient's complaint is categorised into chief complaint call codes (Warner et al, 2014).
Pre-Hospital Emergency Care Council priority dispatch standard
|Clinical||Code||Description||Essential response||Response to scene||Vehicle type||Additional response||Other resources||Further response|
|1 Life-threatening||Echo||Life-threatening: cardiac or respiratory arrest||Ambulance with a minimum paramedic (two paramedics)||Lights and siren||Ambulance||
||Ambulance officer according to operational requirements||Helicopter emergency medical service (HEMS), fire service, Garda, coast guard, utility services as required|
|Delta||Life-threatening other than cardiac or respiratory arrest||Ambulance with minimum paramedic||Lights and siren||Ambulance||
||Ambulance officer according to operational requirements||HEMS, fire service, Garda, coast guard, utility services as required|
|2 Serious not life-threatening||Charlie||Serious not life threatening — immediate||Ambulance with minimum paramedic||Lights and siren||Ambulance||Advanced paramedic for specified DCR codes||Ambulance officer according to operational requirements||HEMS, fire service, Garda, coast guard, utility services as required|
|Bravo||Serious not life-threatening — urgent||Ambulance with minimum paramedic||Lights and siren||Ambulance||Ambulance officer according to operational requirements||HEMS, fire service, Garda, coast guard, utility services as required|
|3 Non-serious or non-life-threatening||Alpha||Non-serious or non-life-threatening||Ambulance with minimum EMT||Lights and/or siren discretion||Ambulance or intermediate care vehicle||Ambulance officer according to operational requirements||HEMS, fire service, Garda, coast guard, utility services as required|
|Omega||Minor illness of injury||Ambulance with minimum EMT||Lights and/or siren discretion||Ambulance or intermediate care vehicle||Ambulance officer according to operational requirements||HEMS, fire service, Garda, coast guard, utility services as required|
CFR: communitiy first responder; EFR: emergency first responder; EMT: emergency medical technician; HEMS: helicopter emergency medical service;
Source: Power, 2014
All patients consenting to transport are brought to an emergency department. There are six adult and three paediatric emergency departments in Dublin.
All PCRs from patients assessed by DFB prehospital practitioners in the week starting 27 March 2017 were manually examined by an independent researcher to identify patients with potential sepsis or potential severe sepsis. All patients who had a manual temperature recording of <36° C or >38.3° C were included in the study and were classed as potential sepsis. Patients were excluded from the study if they had a temperature of <36° C or >38.3° C but no other SIRS signs, signs of sepsis or clinical features of sepsis; for example, they may have had a history suggestive of acute minor trauma or cardiac arrest.
The cases included were further classified as potential sepsis or potential severe sepsis based on the PHECC sepsis CPG 2017 (PHECC, 2017). Characteristics of potential severe sepsis include: being a member of certain ‘at-risk’ populations (e.g. patients with cancer or taking immunosuppressant medication); signs of poor perfusion; practitioner suspicion of severe infection; and whether antibiotics or IV fluids were administered or a pre-alert was given to the emergency department.
This information was extracted by an independent researcher from the vital signs, past medical history, clinical impression and free-text boxes on the PCR (PHECC, 2011).
Patients who met the inclusion criteria but did not meet the potential severe sepsis criteria were deemed to have potential sepsis.
The data collected on patients included anonymised demographic details, clinical information as detailed above (vital signs, medical history, clinical impression and free-text box) and the treatment provided, including oxygen therapy, paracetamol, antibiotics and IV fluid therapy. Call-taking AMPDS data (call priority and call code) were recorded. Practitioner clinical level was extrapolated using practitioner personal identity numbers. Pre-alert notification given to the receiving emergency department was captured.
Data were inputted and analysed using Microsoft Excel 2013. No specific funding was allocated for this research project. Ethical approval was not required for audit.
The incidence of potential sepsis or potential severe sepsis over a 1-week period was 3.7%. That is, 59 of the 1582 patients attended by DFB that week met the study's inclusion and exclusion criteria (Figure 1); there were 51 adults and eight children.
Figure 1. Consort diagram of study population selection
The incidence of seizure, cardiac arrest, STEMI and FAST-positive (face, arm, speech test) stroke were 5.4%, 0.5%, 0.2% and 0.6% respectively, internal audit showed. Potential sepsis was more common at the extremes of age; 8.5% of patients were aged less than 1 year and 58% were aged over 65 years. The median age of paediatric patients (≤15 years) was 13 months and adults (≥16 years) was 72 years. Male (49%) and female (51%) patients were almost equally represented.
The paramedic PCR records five key vital signs (temperature, pulse rate, respiratory rate, blood pressure and Glasgow Coma Scale; 45 (76%) patients had all five recorded, three (5%) patients had three recorded (two of them were infants) and 11 (19%) had four recorded.
Two (3.4%) patients were recognised as ‘at-risk’ patients on the PCR, 12 (20%) had ‘SIRS response’ documented in the free text and 16 (27%) had ‘severe infection’ recorded and 22 (37%) had a potential source identified on the PCR.
Using the criteria detailed above, the investigator determined that there were 17 patients (29%) with potential severe sepsis, and 42 (71%) with potential sepsis (Figure 1).
Table 3 and Table 4 outline the care given to adult and paediatric patients, practitioner clinical level and pre-alert information. Where all the time points were accurately recorded (69% n=41), the median time from arrival on scene to hospital arrival was 22 minutes (range 9–44 minutes).
Sepsis care provided: adults
|Potential sepsis n=36||Potential severe sepsis n=15||All patients n=51|
|Oxygen therapy||7 (19%)||4 (26.6%)||11 (21.5%)|
|Paracetamol||4 (11%)||3 (20%)||7 (14%)|
|IV/IO antibiotics||0 (0%)||0 (0%)||0 (0)%|
|IV/IO fluids||1 (3%)||0 (0%)||1 (2%)|
|Level of practitioner|
|Paramedic||35 (97%)||15 (100%)||50 (98%)|
|% of patients with pre-alert notification||6 (16.6%)||1 (6.6%)||7 (14%)|
IV: intravenous; IO: intraosseous
Sepsis care provided: paediatrics
|Potential sepsis n=6||Potential severe sepsis n=2||All patients n=8|
|Oxygen therapy||0 (0%)||1 (50%)||1 (12.5%)|
|Paracetamol||2 (33.3%)||1 (50%)||3 (37.5%)|
|IV/IO antibiotics||0 (0)%||0 (0%)||0 (0%)|
|IV/IO fluids||0 (0%)||0 (0%)||0 (0%)|
|Level of practitioner|
|Paramedic||6 (100%)||2 (100%)||8 (100%)|
|% of patients with pre-alert notification||0 (0%)||0 (0%)||0 (0%)|
IV: intravenous; IO: intraosseous
AMPDS call prioritisation
Interrogation of dispatch data showed that 28 patients (48%) were prioritised as delta (high priority), 19 (32%) were prioritised as alpha or omega (lowest priority) and, of the latter, 37% were classified as having ‘potential severe sepsis’ (Table 5).
Call prioritisation for potential sepsis
|Potential sepsis n=42||Potential severe sepsis n=17||Total n=59|
|Echo/delta||20 (48%)||8 (47%)||28 (48%)|
|Charlie/bravo||10 (24%)||2 (12%)||12 (20%)|
|Alpha/omega||12 (28%)||7 (41%)||19 (32%)|
AMPDS chief complaint
Further interrogation showed that the chief complaints most common in adults were ‘breathing problems’ and ‘sick person’ (33% and 26% respectively) and the incidence was higher in those with potential severe sepsis (40% and 33%) (Table 6). In children, convulsion/fitting was the predominant chief complaint (Table 7).
Chief complaint: adults
|Potential sepsis n=36||Potential severe sepsis n=15||Total n=51|
|Breathing problem||11 (31%)||6 (40%)||17 (33%)|
|Sick person||8 (22%)||5 (33%)||13 (26%)|
|Fall||4 (11%)||1 (7%)||5 (10%)|
|Other||13 (36%)||3 (20%)||16 (31%)|
Complaint ‘other’ includes abdominal pain (n=3), back pain (n=1), haemorrhage (n=4), stroke (n=2), heart problem (n=1) and faint (n=2)
Chief complaint: paediatric
|Potential sepsis n =6||Potental severe sepsis n =2||Total n = 8|
|Breathing problem||1 (17%)||0 (0%)||1 (12.5%)|
|Sick person||1 (17%)||0 (0%)||1 (12.5%)|
|Convulsion/fitting||4(66%)||1 (50%)||5 (62.5%)|
|Other||0 (0%)||1 (50%)||1 (12.5%)|
Complaint ‘other’ includes stroke (n=1)
Defining sepsis is complex. Sepsis is not a specific illness but a syndrome with a still-uncertain pathobiology. At present, it can be identified by a constellation of clinical signs and symptoms in a patient with suspected infection (Singer et al, 2016). The International Task Force on Sepsis recognises that sepsis is a syndrome without a validated criterion standard diagnostic test, and this key deficit that has led to major variations in reported incidence and mortality rates (Singer et al, 2016). The definitions of potential sepsis and potential severe sepsis used in this audit reflect the arbitrary nature of defining the sepsis spectrum.
The authors had to consider several variables to allow the accurate capture of potential sepsis from the PCR as, at the time of the audit, there was no option to record ‘sepsis’ as a ‘clinical impression’ on the PCR. This meant that the investigators had to consider data from several sources in deciding whether the patient had potential sepsis or potential severe sepsis. Patient follow-up to confirm the diagnosis was not practical as the patients were brought to nine hospitals and patients do not have individual health identifier numbers in Ireland. This was also outside the scope of an audit.
The recording of key vital signs on the PCR was high, with 95% of patients having four or more vital signs documented. This was in keeping with similar paramedic data from the UK and exceeded vital sign recording by GPs in the same study (NCEPOD, 2015). It differed from Dutch data, which showed that vital signs were not routinely assessed (Groenwoudt et al, 2014).
The free-text comments on the PCR (such as ‘SIRS response’, ‘severe sepsis’ and ‘at risk’) reflect the effectiveness of having specific clinical practice guidelines for sepsis and sepsis-specific paramedic training. This contrasts with the previously cited opinion that paramedic education programmes have not focused sufficient attention on sepsis (Smyth et al, 2016).
The difficulties in identifying patients with potential sepsis from the PCR was fed back to the PHECC by the study authors via the medical advisory committee. As a result of this, a revised edition of the PHECC's PCR clinical practice guidelines now includes sepsis as a clinical impression (PHECC, 2019). These changes may assist future research and audit.
Once sepsis is recognised, the clinical practice guideline in Ireland focuses on the Give 3 element of the Sepsis Six bundle, that is, oxygen, antibiotics and IV fluids (Department of Health, 2014). This study demonstrated that 21% of adult patients received oxygen, none received antibiotics and 2% received IV fluids. Seymour et al's (2012) retrospective study of patients with severe sepsis showed that 37% had IV access and 72% were given oxygen. There was no reference to IV antibiotics. Casu and Häske's (2016) survey of medical directors of a physician-delivered EMS in Germany showed that 25.6% of respondents had antibiotics on EMS vehicles (most commonly ceftriaxone) and 10.3% of rescue districts had an algorithm for patients with sepsis. Groenewoudt et al's (2014) study of patients with sepsis arriving by EMS in the Netherlands showed treatment (oxygen and or IV fluids) was started in 43.6% of patients with no reference to antibiotic use.
For patients with septic shock, hospital evidence from large clinical trials shows that mortality increases with every hour delay in administering appropriate antibiotics (Kumar et al, 2006; Ferrer et al, 2014). If the goal of treatment in patients with septic shock is early antibiotics, then prehospital administration of antibiotics may reduce the time to antibiotics if prehospital on-scene times are prolonged. This would require that paramedics trained in IV cannulation and IV antibiotic administration be tasked to look after patients with potential sepsis.
Many of the patients with potentially severe sepsis in this audit did not receive IV fluids or antibiotics, although these were permitted by the clinical practice guidelines. Further interrogation of the data demonstrated that the majority of patients were treated by paramedics, not advanced paramedics, so these interventions were outside of their scope of practice. It can be argued that, because of the urban nature of work of DFB firefighter paramedics, prehospital IV antibiotics and fluids are less important because on-scene and travel times are short (the median total prehospital time is 22 minutes) which means sepsis treatment is prompt, usually starting within 60 minutes.
Seymour et al's (2012) study in King County in Washington State showed the average prehospital time was 45 minutes, including 12.6 minutes average travel time. Ideally, EMS systems would ensure the resources needed for patients with sepsis reflect the local organisation of EMS care and the communities they serve, to ensure timely administration of IV fluids and antibiotics whether in a prehospital or hospital setting.
One key step in this process is pre-alerting the receiving hospital to the incoming patient and this audit demonstrated that 14% of adults and no children had pre-alert messages. All emergency departments served by DFB can already accept pre-alert information. Since this audit was performed, national pre-alert criteria have been agreed and these include sepsis for pre-arrival notification.
Medical dispatch call prioritisation and classification
If IV fluids and antibiotics are to be administered prehospitally, medical dispatch systems should ideally recognise potential sepsis, allocate a high call priority and deploy practitioners who have the scope of practice to administer both of these.
UK data demonstrate that 51.9% of critically ill patients with sepsis are brought to hospital by ambulance (NCEPOD, 2015). Seymour et al (2012) calculated that 40% of all patients hospitalised with severe sepsis arrived in the emergency department by EMS transport. This audit of more than 1500 patients suggests an incidence of potential sepsis/severe sepsis of 3.7%, demonstrating that this illness forms a considerable volume of emergency ambulance calls in Dublin.
It is concerning that one-third of patients in this study had a low alpha or omega call priority and, within this cohort, 37% (7/19) were classified as having ‘potential severe sepsis’. This finding is in keeping with international practice; Groenewoudt (2014) found that half of EMS transports for sepsis were not considered urgent, even in the 34.6% of patients with septic shock and Seymour et al (2012) found that 30% of patients with severe sepsis were given a non-urgent call priority.
It is imperative that EMS call prioritisation systems consider sepsis if further improvements in sepsis morbidity and mortality are to be made. In the UK, NICE (2016) guidelines recommend that ‘all people with suspected sepsis outside acute hospital settings be referred to emergency medical care by the most appropriate means (usually 999 ambulance) if they have any high-risk criteria’. However, many international sepsis guidelines overlook the prehospital arena (NICE, 2016; Singer et al, 2016; HSE, 2017), and so are neglecting an area where considerable improvements in direct patient care could be made. It is incumbent on the prehospital community now to drive these changes.
Call-taking and dispatch systems can be medical priority dispatch system (MPDS), used mainly in Anglo-Saxon countries, or a criteria-based dispatch (CBD) used in Nordic and European countries (Bohm and Kurland, 2018). Recognition of sepsis and call prioritisation are challenges in both call-taking systems. Pedersen et al's (2017) observational study of dispatch and prehospital transport of acutely septic patients found that the Danish EMS CBD system did not have sepsis-specific descriptors available and patients with non-specific complaints were twice as likely as others to get a low-priority dispatch code. Polito et al (2015) showed that 40% of patients with sepsis had a chief complaint of sick person and 23% had breathing difficulties, similar to these audit findings. Smyth et al (2016) propose that validated criteria are needed at the call-taking stage of the patient journey to identify potential sepsis if a timely response with appropriately skilled practitioners is to be deployed.
A starting point may be a review of the determinant descriptors and priority symptoms. Frequent call codes in adults are breathing problems and sick person, and convulsion/fitting in children. The NICE (2016) sepsis guideline states that those particularly at risk of sepsis are infants and young children, immunocompromised people (including those being treated for cancer), people who have recently had surgery, those with indwelling medical lines or devices and women following childbirth. Polito et al's (2015) study on the prehospital recognition of sepsis recognised six EMS characteristics, four of which could be determined by call-takers.
The authors of this study propose that additional discriminators, such as those in the NICE (2016) guideline, may assist in recognising patients with sepsis (Box 1) and patients at the extremes of age should be allocated a higher call priority.
Box 1.Call-taker determinants that indicate a high risk of sepsis
- Hot to touch
- Current use of prescribed antibiotics
- Possible meningitis
- Risk of neutropenia (active cancer treatment, organ transplant or medications)
- Recent surgery
- Indwelling medical lines or devices
- Recent childbirth
- Transfer from a nursing home
- Chief complaint of ‘sick person’
This audit was undertaken in an urban environment and may not reflect the population of other EMS systems. Also, as an audit, it was designed to measure performance and not to show causality.
The definition of sepsis is arbitrary and, while there were merits to a manual review of all PCRs to identify patients with suspected sepsis, having only one researcher review these PCRs could result in a lack of inter-rater reliability.
This audit shows that potential sepsis forms a considerable volume of calls to an urban, fire service-based EMS system. Patients at the extremes of age are most affected. Capturing the clinical impression of sepsis has improved since the audit as the audit findings led to the national PCR being reconfigured.
The ability of the current call-taking system is limited in its ability to identify potential sepsis. The consequence of this is that patients may be allocated a lower call priority with longer waiting times and being attended by a paramedic rather than an advanced paramedic. Both these factors can delay time to fluid therapy and IV antibiotics.
The MPDS is not designed to specifically identify sepsis but, given the high mortality associated with the condition, research is required to identify discriminating questions to include in call-taking systems to better recognise sepsis and to allocate a call the appropriate priority within current coding systems.
- Severe sepsis is a life-threatening, time-sensitive condition; early interventions with oxygen, intravenous fluids and antibiotics reduce mortality and morbidity
- The initial prehospital recognition of sepsis by both call-takers and practitioners is critical to the provision of timely interventions
- Patient care reports should be configured to allow a practitioner to document a clinical impression of sepsis
- Education should focus on the higher incidence of sepsis at the extremes of age, having criteria on pre-alerting the emergency department and being aware that dispatch information may not capture the possibility or severity of potential sepsis
- Research is required to identify discriminating questions to include in call-taking systems to more effectively recognise potential sepsis and to allocate a call the appropriate priority
CPD Reflection Questions
- Does your patient care report allow the clinical impression of ‘sepsis’ to be recorded? How does it do this and what are its advantages and disadvantages?
- What prehospital care do you provide to patients with potential sepsis?
- Are your call-taking and dispatch system able to recognise and respond to potential sepsis? If not, what would make you suspect the condition?