Chest pain is a very common complaint in Emergency Care. It is estimated that in the UK 20–30% of hospital admissions are for chest pain (Kendrick, 1997; Quinn, 2008). In 2011-2012 chest pain was the third most common reason to request an emergency ambulance in the North West, comprising 8.94% of all calls (North West Ambulance Service, 2013). As the prevalence of Coronary Artery Disease, a contributing factor to chest pain, rises (Quinn, 2008), those admissions and ambulance calls are likely to increase (Hamm et al 2011). Whether all of these patients need to be taken to Accident and Emergency can be debated.
At present, all patients in the North West of England complaining of non-traumatic chest pain should be taken to hospital. This is because the responsible ambulance trust has in place the ‘Paramedic Pathfinder’: a triage tool for paramedics (North West Ambulance Service, 2014; Newton et al, 2013). As the tool does not differentiate between pleuritic, cardiac or gastrointestinal origins of pain, many patients are, perhaps, being taken to hospital unnecessarily. The tool is therefore at odds with the current discourse within emergency care; namely to treat patients closer to home and avoid filling over-stretched Accident and Emergency Departments (Association of ambulance service chief executives (AASCE), 2011; NHS England Emergency and Urgent care review team, 2014).
It must be the goal of the ambulance clinician to differentiate between a high risk, potentially life-threatening condition, and everything else. Examples of potentially life threatening causes of chest pain are aortic dissection, large pneumothorax, an ST elevation myocardial infarction (STEMI) and Acute Coronary Syndrome (ACS). Common non-immediately life-threatening causes include gastro-oesophageal reflux, anxiety, a pulled muscle in the chest wall and pleurisy (NHS 2014). The pathfinder tool does not allow for clinicians to make this crucial differentiation between the two. For example if pathfinder was not in place, patients with pleurisy and reflux could be referred to a GP by paramedics for ongoing care and further investigation. A more challenging argument to make, but worth raising nonetheless and also discussed in this work, is for paramedics to find alternatives to A and E for patients presenting with chest pain that has a possible cardiac aetiology.
This work identifies enablers, which should remove the need for a non-traumatic chest pain discriminator. The enablers would allow paramedics to safely identify different aetiologies of chest pain and therefore refer patients more appropriately. Firstly however, ACS will be studied more closely. This is because there are variations in the disease process, which cause ACS. Variations cause different amounts of myocardium damage, some of which is reversible. Such variations can cause difficulty in finding these patients the most appropriate care. Finally, after identifying the enablers, this work will discuss risk associated with chest pain patients, before drawing conclusions. Current literature on various topics relating to chest pain is relayed as part of the discussion. Furthermore observations of the Finnish ambulance service undertaken by the author for a period of two 24 hour shifts are drawn upon.
Identifying a life-threatening cause
It is assumed paramedics can safely identify those chest pain patients who are critically ill: those that are hemodynamically unstable for example, as well as most of the life-threatening aetiologies of chest pain listed above. With regard to STEMI, it is encouraging to see paramedics now have a proven success in diagnosing STEMI for direct referral for PPCI (A de Belder, 2012). ACS is different from the other life-threatening causes as it is much more difficult for clinicians to accurately diagnose, assess the severity in primary care and therefore find an appropriate pathway. Difficulties arise for a number of reasons.
Firstly, ACS is potentially life-threatening as it is unstable - the patient may become critically ill quickly as there is potential to progress to Acute Myocardial Infarction (AMI) and/or sudden cardiac death (Braunwald et al, 2002). Secondly it can easily be confused with more minor ailments as there are a large number of differential diagnosis of chest pain (Jaffery and Grant, 2008; Albarren and Tagney, 2008). Thirdly atypical presentations of ACS are common (Jafferey and Grant, 2008; Hamm et al 2011). Patients do not always give the ‘classic’ description of cardiac chest pain. Fourthly there are no signs on physical examination that diagnose ACS (Jafferey and Grant, 2008; Hamm et al 2011). These difficulties do highlight a need for a non-traumatic chest pain discriminator to be included on the pathway tool. With a limited number of clinical diagnostic tests available to ambulance clinicians it is very difficult to exclude ACS in a differential diagnosis. However it is hoped the enablers identified shortly will help overcome these complexities.
Even if ACS is a possible diagnosis, it could be argued, that alternatives to Accident and Emergency could be found by paramedics for those patients experiencing milder forms of ACS. Before proceeding to identify enablers which could help in this regard, it is worth understanding a little more about ACS. Knowledge of the pathophysiology of ACS would help clinicians to understand ACS as a spectrum.
ACS is a common diagnosis, with 114,000 being admitted each year (Patient, 2014). Green and Tagney (2008: 72) define ACS as:
‘the clinical spectrum of the same disease process resulting in reduced coronary arterial blood flow which leads to unstable angina, ST segment elevation myocardial infarction (STEMI) or a non-ST segment elevation myocardial infarction (N-STEMI).’
Braunwald et al (2002) identify possible causes of ACS. Firstly thrombus is the main cause of STEMI but can also cause 35-75% of the cases of N-STEMI or Unstable angina. Other causes may be a spasming of the arteries, severe narrowing of the arteries by the atherosclerotic plaque or arterial inflammation. Patients with pre-existing coronary heart disease are likely to experience ACS due to myocardial hypoperfusion due to hypotension, anaemia, hypoxia, or fever demanding an increased cardiac output (collectively called secondary unstable angina). These processes can occur to greater or lesser degrees and can occur in different arteries of the heart creating a variety of presentations (Braunwald et al, 2002). By using the clinical tools identified below, it is hoped paramedics could safely identify patients where disease processes are occurring to a lesser extent and the risk of further myocardial damage is low. These patients could be referred directly to specialists for further investigation, avoiding Accident and Emergency.
Enablers to finding alternative pathways
To assist paramedics in identifying a non-cardiac cause of chest pain and, more controversially, a cardiac cause that could be dealt with without attending Accident and Emergency, a number of enablers have been identified.
Firstly, and perhaps the most beneficial enabler, is the university training of paramedics. Previously paramedics were trained over a few months by completing a vocational course. This has of course now changed to a 2-year diploma or a 3-year degree. More time spent learning about ECG, and greater understanding the pathophysiology of cardiac conditions and the knowledge of treatment options will all enable paramedics to find the best pathway for chest pain patients. For example, the ability to note changes on ECG such as T wave inversion and minor ST depression will aid assessment (Hamm et al, 2011) and help identify correct aetiologies of chest pain.
The 2005 ‘Taking Healthcare to the Patient’ report noted that:
‘Ambulance clinicians should be equipped with a greater range of competencies that enable them to assess, treat, refer, or discharge an increasing number of patients’
(Department of Health 2005: Page 44)
The report identifies that to gain these competencies Higher Education programmes are necessary for Paramedics (Department of Health 2005). Today there is a drive to stratify the content delivered by the various paramedic programmes delivered by various academic institutions (Lovegrove M and Davies J, 2013). A uniform approach to paramedic education would create a better understanding of the paramedic role by other healthcare professionals. If the paramedic role is better understood by GPs and specialist centres it becomes easier for paramedics to make referrals of chest pain patients. This is because potential alternatives to Accident and Emergency therefore have a better understanding of the level of clinical assessment that has taken place and the clinical investigations that have been carried out. Higher education is the first step away from a restrictive, protocol-driven practice for paramedics as exemplified by the paramedic pathfinder.
Secondly, ACS risk scoring is a tool that is used in hospital but has potential to be used by paramedics in the pre-hospital environment. A variety of tools already exist to aid clinical decision making. They work by combining features of the patient history, signs and symptoms and the results of clinical investigations to assign a score to a patient, indicating their level of risk. Goldberg et al (2010) highlight a difference between tools that stratify risk and and tools that classify likelihood. Likelihood classification refers to the likelihood the patient has ACS as opposed to a non-cardiac cause. Risk Stratification is establishing what a particular outcome would be over time by someone presenting with a cardiac aetiology (e.g mortality or risk of a major adverse coronary event). Paramedics could utilise a tool which covers the two areas. Likelihood classification would help identify a suitable pathway and risk stratification would identify how urgently the patent needed referral.
The European Society of Cardiology and NICE recommend the Global Registry of Acute Coronary Events (GRACE) score (Hamm, 2011; NICE, 2010). This may be problematic in pre-hospital care as it requires computing power to provide an estimate due to its complexity (Hamm, 2011). Mobile computing is not commonly found in the pre-hospital environment. More useful may be Christenson et al (2006) Vancouver Chest Pain Rule. Out of a possible 123 predictors 48 were chosen to include in an algorithm. Predictors include ECG findings, descriptions of pain and biomarker results. The aim of the tool is to provide safe discharge after only two hours at Accident and Emergency. The tool was tested by following up patients included in the study 30 days later. The tool was found to be 98.8% sensitive and 32.5% specific. Unfortunately the study requires validation as it was only conducted in one centre but certainly has potential.
Goldberg et al (2010) reviewed many of the tools currently developed. They concluded that:
‘there is no single scoring instrument that can identify all the patients with ACS and at the same time avoid hospitalising patients without ACS’. Goldberg et al (2010: 55)
Since 2010 tools are still being developed (such as Than, 2012 (2)) which have good results. An opportunity exists for research and testing existing or new ACS risk scoring tools in the pre-hospital environment. In order to avoid more protocol-driven practice any tools utilised by paramedics should not include a threshold that necessitates transport to hospital. Rather it should be a useful tool that a paramedic could use in order to justify identifying a safe pathway for the patient.
Most ACS scoring tools necessitate the inclusion of cardiac biomarkers. Elevated cardiac troponins reflect myocardial cellular damage (Hamm, 2011) and are widely accepted by the like of the American Heart Association and the European Society of Cardiology as crucial in risk stratification and NSTEMI identification (Hamm, 2011;Braunswald, 2002). Troponin assays are usually carried out in the laboratory. However bedside or point of care testing exists. In Finland paramedics have bedside troponin measuring equipment. Depending on patient presentation paramedics can take a measurement, and return 6 hours later to repeat the test. This is necessary as a troponin rise is not seen until 4 hours after symptom onset (Albarren and Tagney, 2008). Troponin bedside monitoring is something not commonly done in UK ambulance services. In busy UK ambulance services returning some hours later to repeat troponin would be a challenge but is possible. Single manned rapid response vehicles could be an efficient way of carrying out repeat testing.
Unfortunately studies of pre-hospital troponin measurement have not yet proven the benefits of prehospital testing; however, they have not shown it to be worthless (Erdhart et al, 2002, A de Belder, 2012). In the ADAPT trial bedside troponin measurement (in hospital), along with ECG results and use of risk-scoring tools, was shown to safely discharge patients within 2 hours. (Than et al, 2012 (2)). As previously alluded to a similar study to the ADAPT trial could take place in the pre-hospital environment and begin to demonstrate that a combination of tools, used by paramedics, could safely find alternative pathways for patients complaining of chest pain.
A fourth enabler involves using current evidence-based guidance for investigating chest pain. Clear guidance on the treatment of chest pain is available from the National Institute for Clinical Excellence (NICE, 2010). NICE present two treatment pathways: one is for patients presenting with stable chest pain; the other, for patients presenting with possible ACS. For patients with stable chest pain (anginal or a non cardiac cause) paramedics could refer patients to GP's for follow-up according to NICE recommendations. In the second pathway (for suspected ACS) all patients are advised to be referred to hospital. The guidelines suggest, however, the degree of urgency is a matter of clinical judgement. It is advised that a same day appointment is made at hospital for anyone who has had pain in the last 72 hours, even if they are pain free at the time of assessment. For patients that currently have cardiac pain and/or and abnormal ECG NICE recommends transfer to hospital should be an emergency. The advice goes some way to recognise the diverse nature of chest pain as opposed to the simplistic non-traumatic chest pain discriminator. The question remains, however whether paramedics can accurately determine whether a chest pain is non-cardiac, stable or identify the speed which an acute episode of chest pain requires investigation. If proved that paramedics can, NICE guidance provides a useful pre-existing pathway.
A fifth enabler exists for paramedics to refer patients who have resolved symptoms of possible ACS. In South Tees, Paramedics can take patients, or make referrals for patients, to Rapid Assessment Chest Pain Clinics (RACPCs) (South Tees Hospitals, 2014). The National Service framework for coronary heart disease, published in 2000 recommended to establish theses clinics across the country, following the success a model established in South-East London, in an attempt to modernise cardiac care across England (Boyle, 2007). Sometimes led by nurses these centres allow testing by specialists to establish if the cause of chest pain is cardiac and facilitate ongoing treatments (Boyle, 2007). RACPCs can however have quite restrictive criteria for referral: only excepting patients above a certain age, patients without pre-existing cardiac history and those with symptoms of stable angina (Marshall, 2011). This therefore limits the scope of referral of many patients that present to paramedics. In Finland, paramedics often discuss patient cases with doctors to determine the best patient outcomes. This is true also of patients presenting with chest pain, often speaking directly with specialists. Where facilities exist a similar model could be put in place here. RACPC could be the specialists paramedics refer to, gaining crucial expert advice to maintain the safety for patients.
A discussion of risk
The first version of the pathfinder tool excluded patients suffering from cardiac chest pain (Newton et al, 2013). This was changed to non-traumatic chest pain upon testing to improve sensitivity with an accepted decrease in specificity (Newton et al, 2013). The change is understandable given the aforementioned difficulty of diagnosing a patient presenting with a chest pain. Even once investigated in hospital, according to Albarren and Tagney (2008), up to 30% of patients presenting with chest pain receive a misdiagnosis. If hospitals and GPs have difficulty finding the correct aetiology for chest pain, many would argue it would be illogical to allow alternative pathways to be found by paramedics.
The risk adverse nature in dealing with chest pain is not unique to the paramedic path-finder. Studies show a mismatch between numbers of patients initially thought to have cardiac chest pain and the number of diagnosis of ACS. Deakin et al (2006) showed that only 1 in 18 patients that call 999 with acute chest pain will have ACS as a diagnosis. Cooper et ale (2008) study of the performance of Emergency Care Practitioners (ECP's) showed that of 50 patients reporting a cardiac problem only 2 had a cardiac diagnosis after 24 hours. Furthermore Fox (2005) demonstrated that two thirds of patients referred to rapid access chest pain clinics will not receive a cardiac diagnosis. Than et al (2012)(2) makes reference to physicians in America admitting patients for prolonged observation, due to concerns of missing a diagnosis of MI or the potential for major adverse coronary event. These statistics appear to present a picture of over-referral and risk aversion when clinicians are presented with a patient complaining of chest pain.
Three problems are apparent with such risk aversion and question the validity of the non-traumatic chest pain discriminator. Firstly there are obvious cost and resourcing implications as low risk patients are investigated without need. Secondly there is a concern that the risk of investigating the complaint outweighs the risk of the disease process. Than et al (2012) (1) highlight the test threshold. There becomes a point when testing false positive patients causes more harm than the disease itself, given risks associated with further cardiac investigations. Even something as simple as giving Asprin, as commonly given by paramedics to patients with suspected cardiac chest pain, is not without risk (. Finally, taking all non-traumatic chest pain patients to hospital is a way of ensuring risk is avoided by the ambulance service. Conversely it could be argued that taking all non-traumatic chest pain patients to hospital causes increased risk as it leads to long delays in the emergency department resulting in patients who desperately need the A and E services not being able to access the resource promptly. This, not surprisingly appears to be an increasingly frequent feature of media reports (Smith, 2014; Johnston, 2014). Therefore some element of risk is avoided by using the pathfinder however further risks are created elsewhere within the healthcare system.
Conclusions
The discriminator of non-traumatic chest pain within the paramedic pathfinder does have justification. However, it does appear to clash with current general policy direction of UK ambulance services, namely more patients to be treated closer to home and avoid unnecessary Accident and Emergency attendance. Giving an increased number of tools to paramedics could help clinical decision making in the field, removing the need for the discriminator. Most of the enablers identified do however need development and research in the pre-hospital environment. Given the quantity of calls to ambulance services that relate to chest pain, the possibility that the number of these calls will rise as the prevalence of CHD rises and the current pressures on Accident and Emergency there does appear to be an urgency to change this discriminator. Removing the discriminator would change paramedic practice by not restricting paramedics to transporting all chest pain patients to Accident and Emergency, and instead, allow a higher education background and potentially other clinical tools to safely find alternative outcomes for patients complaining of chest pain.