Paramedics often have to make critical decisions based on several options and frequently make these alone (Bledsoe et al, 2007). They attend patients who do not present with classic textbook signs and symptoms. In order to make a clinical judgement, the paramedic is required to use their knowledge and experience to achieve a patient care focused critical decision (Bledsoe et al, 2007). Therefore, gaining a better understanding of a decision-making process will make a significant contribution to optimal patient care (Buckingham and Adams, 2000).
This article will use a case study to critically reflect and analyse the decision-making process used in an emergency situation. It will discuss a range of factors that influenced the clinical decision-making process and how this prompted immediate transport to a local hospital. A wide range of supporting evidence will be explored and the decision-making process will be questioned and challenged. Relevant professional, ethical and legal issues will be considered and discussed. The inclusion of other colleagues, patients, relatives and their involvement, within the decision-making process, will also be deliberated.
Case presentation
A relative made an emergency call requesting an ambulance for a 64-year-old female, who had collapsed at home and appeared to be having a stroke. Whilst responding on a Rapid Response Vehicle (RRV), vital information was passed including the patient's condition and the estimated time of arrival (ETA) of a conveying resource.
Table 1 shows the incident response times recorded on the Mobile Data Terminal (MDT). The safety, scene and situation, patient assessment triangle, catastrophic haemorrhage, airway, breathing, circulation, disability, and examination (CABCDE) approach was used, as it advocates a structured approach to all acute patient assessment (Halliwell et al, 2011).
| RRV incident response times | Ambulance incident response times | |
|---|---|---|
| Origin | 0830 | 0830 |
| Passed | 0831 | 0831 |
| Mobile | 0831 | 0831 |
| At scene | 0837 | 0850 |
| At patient | 0837 | 0850 |
| Left hospital | N/A | 0859 |
| At hospital | N/A | 0910 |
Primary survey
Safety, scene and situation
It appeared safe to enter and there were no immediate dangers. The scene was a patient collapsed on the bathroom floor. The situation was the patient appeared to be having a stroke. The mechanism of injury was not considered a potential for a c-spine injury.
Patient assessment triangle
The patient was responsive to pain, pale in colour and had a normal respiratory rate.
Catastrophic haemorrhage
There were no signs of catastrophic haemorrhage.
Airway
The patient's airway was partially obstructed by the tongue. There were no signs of c-spine injury. The airway was cleared by performing a head tilt, chin lift technique and the insertion of an oropharyngeal airway.
Breathing
The respiratory rate was 18 with an SPO2 reading of 96% on air. On auscultation, the chest was clear with equal air entry.
Circulation
The patient was hypertensive with a blood pressure of 158/112 mm/hg. A radial pulse rate of 60 bpm and a capillary refill of less than two seconds. The temperature was within a normal parameter of 36.9 °C.
Disability
The patient had a reduced Glasgow Coma Score (GCS) of 8, which Cifu and Caruso (2010) states is a 3–15 point score that assesses levels of consciousness, examining eye opening, verbalisation and motor response. The patient received a score of 2 for eye response, a score of 2 for verbal response and a score of 4 for motor response. According to Blaber and Harris (2011) the AVPU framework (see Table 2) is a tool used to assess a patient's level of consciousness, and in this case study the patient responded to painful stimuli. The pupils were equal, round and reactive to light. The patient had a blood glucose level of 5.2 mmol/L.
| A | The patient is alert, conscious and responds directly and appropriately to the clinicians questions. |
| V | The patient responds to vocal commands (verbally), which may be a grunt or groan. |
| P | The patient responds only to painful stimuli. |
| U | The patient is unresponsive. |
The Face, Arms, Speech and Time (FAST) test (see Table 3) was initially positive with slurred speech. Any difficulty in speaking should be viewed as significant, including expressive dysphasia and aphasia, not just the stereotypical slurred speech (London Ambulance Service NHS Trust, 2011).
| Facial movements |
| Ask patient to show teeth:
|
| Arm movements |
| Lift the patient's arms together to 90° if sitting, 45° if supine, and ask them to hold the position for 5 seconds before letting go:
|
| Speech |
|
|
| Time |
|
|
Expose, environment, evaluate, evacuate, reassess
A further examination revealed that there were no obvious injuries. A comprehensive history, as suggested by Gregory and Mursell (2010), was taken from the patient's daughter whilst waiting for the ambulance to arrive. The patient mentioned prior to collapsing she had woken up with a sudden onset of a severe headache and vomiting. This, combined with the eliciting of further information, led to a possible diagnosis of a subarachnoid haemorrhage. Therefore, the decision was made to immediately transfer the patient to the receiving stroke pathway at a local emergency department (ED). The rationale for the transport destination was based on a pre-hospital fast track pathway operating within the Ambulance Trust for acute stroke patients. A technician crew arrived on scene. The Health and Care Professions Council (HCPC) Standards of conduct, performance and ethics (2008) suggest that a registrant should not delegate a task to another person outside of their scope of practice. Therefore, due to the presenting complaint and the patient's condition, the decision was made by the paramedic to travel with the patient to hospital. Although the patient improved en route with a GCS of 15 and was FAST test negative (see Table 3), the patient still presented with a loss of power in her left leg. As the patient was eligible for the fast track pathway, she was referred and accepted by the stroke team direct and they were pre-alerted. This pathway aided the decision-making process and ensured the patient received the right care at the right place and the right time. Information provided included the patient's age, time of onset, medical complaint and history, investigations, vital signs and treatment (ATMIST). An ETA was also passed to the hospital.
The ATMIST algorithm is used within a local ambulance service and the local hospitals are familiar with this process (South Western NHS Foundation Trust, 2013). En route, observations were recorded every 10 minutes and the patient was continuously observed for any change. There were no changes en route and the patient remained stable throughout. The patient's history, examination and treatment were handed over to the stroke nurse and the patient was left in the care of the hospital. The patient was later transferred to a specialist hospital, following the result of a computerised tomography (CT) scan showing a subarachnoid haemorrhage.
Critical reflection
Thomson et al (2000) investigated the management of subarachnoid haemorrhages in emergency care environments and found only 49% of patients were correctly diagnosed before a CT scan was performed. Therefore, it is important as a paramedic to comprehend the clinical decision-making process, in order to analyse data effectively and devise a practical management plan that optimises patient care (Bledsoe et al, 2007). However, it is essential to understand that decision making is a complex process that encompasses cognitive, intuitive and experiential processes (Pugh, 2002). Muir (2004) suggests there are many models that have been devised to attempt to describe decision making processes. However, the two competing conceptual frameworks are the hypothetico-deductive model and the intuitive model. It is suggested that all decisions are made either hypothetico-deductively or intuitively (Buckingham and Adams, 2000). However, both models were used in this case study.
The hypothetico-deductive model
As highlighted by Cioffi (2002: 53), the hypothetico-deductive model may be useful in clinical situations where certainty is high. However, the model may have limitations where conditions of uncertainty exist, e.g. when patients have multiple pre-existing conditions. Keller (2009) suggests the hypothetico-deductive approach is based on certain presuppositions. Furthermore, Thompson and Dowding (2009) describe the hypothetico-deductive approach as individuals who go through different stages of reasoning processes when making clinical judgements and decisions. Thompson and Dowding (2009) highlight that there are four stages in this process and these are cue acquisition, hypothesis generation, cue interpretation and hypothesis evaluation.
Cue acquisition
The first stage, cue acquisition, encompasses gathering vital information about the patient. Thompson and Dowding (2009) suggest information is often obtained prior to seeing the patient and in this case study information was passed from the daughter who made the call to the ambulance control centre. However, as suggested by Deakin et al (2009), the accuracy of the possible diagnosis of a stroke remained particularly debatable due to the limited clinical knowledge of the ambulance call taker. However, NHS Pathways enables a specially designed clinical assessment to be carried out by triaging telephone calls from the public, based on the symptoms they report when they call (NHS Connecting for Health, 2011). Therefore, the call takers are not allowed to deviate from the script and are not able to illicit further or deeper responses. In this case cues were gained from the history given by the daughter, the environment and clinical findings gained from examining the patient. Gregory and Mursell (2010) suggest that the reliability of third party data may not always be accurate. However, in this case the patient was unconscious at the time of the paramedic's arrival and it was pertinent to gain further information from the daughter. The signs and symptoms were a sudden onset of a severe headache and vomiting, followed by a period of reduced consciousness, a loss of power in the left leg and initial slurred speech.
Hypothesis generation
The second stage, hypothesis generation, involved generating some initial possible explanations from the information gathered. The hypotheses immediately created were an ischaemic or haemorrhagic stroke, of which a subarachnoid haemorrhage or intracerebral haemorrhage was suspected. According to Rae et al (2011), 80% of all strokes are due to ischaemia and 15% are due to a haemorrhage. A subarachnoid haemorrhage represents about 5% of all strokes (Alway and Cole, 2009. According to the Stroke Association (2012a), only 10% of strokes are intracerebral haemorrhages. However, other potential causes could have been a migraine, affecting 17% of the population (Thompson and Dowding, 2009). According to The Stroke Association (2012b), studies have shown that the risk of a haemorrhagic stroke may be increased in women aged less than 45 years who have migraine with aura. Other explanations are hypoglycaemia, a seizure or an infection. These differential diagnoses show similar symptoms to a stroke, and therefore although considered, were ruled out individually.
Cue interpretation
The third stage, cue interpretation, comprised of interpreting the cues gathered during the data-gathering stage and classifying them as confirming, contesting, or not contributing to the initial hypothesis generated. In helping to make the final decision between the potential diagnoses, a set of clinical guidelines were used: Stroke: diagnosis and initial management of acute stroke and transient ischaemic attack (National Institute of Health and Care Excellence, 2008). The NICE guidelines (2008) suggest in order to diagnose a stroke, a validated tool (FAST) should be used pre-hospital for patients with sudden onset of neurological symptoms. The FAST test is a reliable and valid tool for identifying stroke patients in the community. However, Harbison et al (2003) allude that FAST was shown to identify only 79% of cases, which suggests that it is not a reliable tool in the remaining 21% of cases. This could have been due to a misdiagnosis and as a result of symptoms that mimicked a stroke. Furthermore, it could therefore be argued that this is unreliable if the diagnosis of a stroke cannot be made with a GCS less than or equal to 7 (Harbison et al, 2003). Therefore, in this case, where the patient presented with an initial GCS of 8, the initial FAST test could be deemed as accurate. Using this tool the patient was initially FAST test positive with slurred speech, indicating a possible stroke. However, once the patient improved with a GCS of 15, a further assessment showed the patient was FAST test negative, but later reported a loss of power in her left leg. Harbison et al (2003) suggest the FAST test was designed for the assessment of a seated subject and therefore the assessment of leg weakness was omitted. Furthermore, increasing the complexity of the FAST test would lengthen paramedic assessment time and could increase the proportion of false positive diagnoses. However, as a single responder, an extensive assessment could be carried out while waiting for a transporting crew.
Cioffi (2002: 52) highlighted studies that have revealed too much reliance on a sign and symptom that has little or no validity. Therefore, it was vital to rule out other medical conditions that may mimic stroke symptoms. Thompson and Dowding (2002) describe this as a process where information is used to assess alternatives and select a course of action. Cicala (1999) highlights other conditions presenting with similar symptoms. However, from taking the patients temperature there was no sign of pyrexia and therefore this patient did not appear to have an infection. The patient's blood glucose was within normal parameters, which Gregory and Mursell (2010) suggest is between 5–7 mmol/L. However, according to Diabetes UK (2009), the recommended levels for a non-diabetic patient is 3.5–5.5 mmol/L before meals and less than 3 mmol/L two hours after meals. Furthermore, a blood glucose below 4 mmol/L would indicate hypoglycaemia, and therefore prompt treatment (Joint Royal Colleges Ambulance Liaison Committee (JRCALC), 2013). It is important that blood glucose is checked in suspected stroke patients as hypoglycaemia may mimic stroke symptoms. A thorough history was taken and there was no previous history of migraines. In this case a migraine could be ruled out, as according to Detsky et al (2006), a migraine is 24 times more likely if the person's headache has four or more features of the following presentation: pulsating, duration of 4–72 hours, unilateral, nausea and disabling. In this case there was no evidence of a seizure. However, according to Pinto et al (1996), between 4 and 10% of all patients with subarachnoid haemorrhage have seizures at the onset of bleeding.
Hypothesis evaluation
In helping to reach the hypothesis evaluation, each possible explanation was explored in order to achieve a diagnosis. It was important as an experienced clinician to process the information accurately to achieve a diagnosis (Thompson and Dowding, 2002). Schoenberg and Scott (2011) suggest an ischaemic stroke will present with neurological deficits which develop abruptly and are typically painless without reduced consciousness and evolve over the course of seconds to hours. However, a haemorrhagic stroke will present with a headache and altered consciousness (Schoenberg and Scott, 2011). From the history and examination of the patient, an ischaemic stroke could be excluded as the patient did not present with any of these symptoms. There are two types of haemorrhagic strokes, a subarachnoid haemorrhage and intracerebral haemorrhage. Alway and Cole (2009) suggest the signs and symptoms of an intracerebral haemorrhage are sudden neurological dysfunction, such as weakness, numbness, vision loss, diplopia, dysarthria, gait disorder, vertigo, aphasia; or disturbed level of consciousness, hypertension, headache, nausea and vomiting. A sudden onset of severe headache followed by a diminished level of consciousness is a typical presentation of a subarachnoid haemorrhage (Alway and Cole, 2009). Therefore, this evidence suggests an intracerebral haemorrhage was unlikely. Furthermore, the patient presented with a severe generalised headache with vomiting and loss of consciousness, which is a presentation of a subarachnoid haemorrhage (Schoenberg and Scott, 2011). However, the paresis in the left leg is a more localising feature. JRCALC guidelines (2013) suggest a subarachnoid haemorrhage may present with an altered level of consciousness and a headache of severe, sudden (thunderclap) onset. They go on to say that a subarachnoid haemorrhage is time critical and the patient must be transported to the nearest available receiving hospital and not left at home (JRCALC, 2013).
Eccles and Grimshaw (2000) suggest clinicians should be vigilant, as value judgements made by guideline development groups may represent the wrong choices for individual patients. However, Thomas et al (1999) describe clinical guidelines as methodically developed statements to assist practitioner decisions. This is evident with the JRCALC guidelines (2013), which suggest it is often difficult to differentiate between a simple headache and a potentially more serious condition. Therefore, the symptom of a sudden onset of a severe thunderclap headache requires the patient to undergo hospital assessment. However, clinical guidelines have potential limitations in that recommendations might be wrong and can result in flawed suggestions, which may harm the patient (Rycroft-Malone, 2002: 150–1). This may ultimately lead to bias and misleading findings. However, Rycroft-Malone (2002: 151) goes on to suggest that clinical guidelines have a part to play in combining research evidence into clinical decision-making and therefore improving the quality of care delivered. Furthermore, they enable standardisation and encourage accountability (Kirkwood, 2000).
The intuitive model
The intuitive model, or the heuristic model, was also used in conjunction with the hypothetico-deductive model as part of the clinical decision-making process. Thompson and Dowding (2002) define intuition as immediate knowing of something without conscious thought. This experience could also be referred to as a sixth sense or a hunch (Nyatanga and Vocht, 2008). However, Thompson and Dowding (2002) argue the model attracts competing definitions.
Riley (2003) defines heuristics as the ‘rule of thumb’, which entails shortcuts. Cioffi (2001) suggests that these shortcuts are based on past experiences. As an experienced paramedic, the recollection of a similar presentation created a hunch that this patient was presenting with a subarachnoid haemorrhage. Riley (2003) suggests heuristics initiate hypothesis gathering, which over time can become well structured. However, there remains some degree of bias that may reduce its rationality. Riley (2003) highlights there are three important heuristics which are open to misinterpretation. One important heuristic is representativeness, i.e. how similar it is to other cases. This patient displayed signs and symptoms which were similar to encountering previous patients having a subarachnoid haemorrhage. Availability is another important heuristic, i.e. how easy it is to remember similar problems. As highlighted earlier by Alway and Cole (2009), a subarachnoid haemorrhage only represents about 5% of all strokes, and could have made it difficult to remember previous cases. Anchoring and adjustment involves the way information is presented. As suggested by Bost et al (2010), it was important accurate clinical information was not missed during the handover to the stroke nurse in order to avoid misinterpretation.
Benner (1984) defines a continuum of practice development from novice to expert and that expertise should be something to strive for. It could therefore be argued that using the intuitive approach to make clinical decisions could prove to be a challenge as a paramedic striving to be an expert within the field. However, as it can be seen in this case study, the expert is able to focus on important cues, asking the right questions in order to generate a hypothesis (Cioffi, 2002: 57). However, this can take some time to accrue (Benner, 1984; Crow et al, 1995; Lamond and Farnell, 1998). As suggested by Riley (2003), it can be argued that although the final clinical decision in this case study may have appeared to be intuitive, it was preceded by consideration of cues and the generation of hypotheses.
As the lead clinician and as suggested by the HCPC Standards of conduct, performance and ethics (2008), the decision was made to travel with the crew. This may have prevented the patient from coming to harm. However, beneficence states that there is a duty to do good to others and to maintain a balance between benefits and harms (Daniels, 2004). In this case the patient had the potential to deteriorate and ultimately the responsibility would lie with the paramedic.
Conclusions
A paramedic applies their clinical experience and exercises independent decisions as they develop and implement a management plan (Bledsoe et al, 2007). The decision-making process has been critically analysed throughout this case study and the overall conclusion is that a combination of hypothetico-deductive and intuitive strategies should be used in the decision making process (Harbison, 2001; Thompson and Dowding, 2002; Muir, 2004). It is evident that the hypothetico-deductive reasoning was utilised in a logical manner to identify the patient's condition. Clinical guidelines and up-to-date evidence was utilised in order to reach a clinical decision. Levels of expertise and intuitive judgements were considered. However, the decision was not solely based on pure intuition but on prior knowledge and experience of patients with similar symptoms.
Throughout this case study a broad understanding of clinical decision-making has been attained. This will enable a more thorough approach to making good clinical decisions on a daily basis. A good understanding of the different types of strokes and in particular subarachnoid haemorrhage has been acquired. This has been achieved through in-depth research and revision of the subject. This knowledge has enabled a more thorough decision-making process to take place, and as a result of recognising the associated signs and symptoms, could in the future ensure the patient is taken to the appropriate hospital. On-going development of critical decision-making skills should be continued by the paramedic, and according to Beebe and Myers (2009), good clinical decision-making depends on a willingness to accept new ideas and to practice creative thinking in the field.