Paramedics are responsible for treating and transporting patients in need of urgent care. In North America, paramedics have been referred to as the backbone of the out-of-hospital emergency care system (Institute of Medicine (IoM), 2006), and the safety net of healthcare (Ross, 2010). The decisions paramedics make while assessing and treating patients can have a major impact on the care delivered and the resultant clinical outcome (morbidity and mortality) and safety of the patient (IoM, 2006). This is especially true as the diagnostics and interventions paramedics administer become more complex and their scope of practice continues to evolve and expand (Paramedic Association of Canada, 2001; Emergency Medical Services Chiefs of Canada, 2006).
EMS in Canada
In most Canadian emergency medical services (EMS) systems, clinical protocols or medical directives direct paramedic care. EMS medical directors create these documents, based on existing norms and standard practice, using their judgment on what will be most successful in the local system. Protocols are often presented in algorithm format, and are analogous to ‘practice standards’, which are definitions of the correct practice, with few treatment options intended to be followed explicitly (Eddy, 1990). Adherence to protocols are often a main measure of the quality of care delivered.
Patients who require assistance from EMS are found in a variety of locations, making the practice of paramedicine more unpredictable than in hospital settings, including the emergency department. In some situations, paramedics have fewer clinical resources, including a lack of other skilled practitioners, incomplete patient medical histories, and even the events that precipitated the emergency call may be unclear. In combination with this, many EMS patients have high acuity, time-sensitive conditions, whether medical or trauma. These and other factors are why EMS is often called an ‘uncontrolled setting’ (Nelson, 1997). Given this context, it is essential to learn more about paramedic clinical decision-making (CDM).
This article will focus on paramedic clinical decision-making, specifically the judgments paramedics make that pertain to assessment, treatment and transport decisions. Two recent Canadian paramedic research studies will be discussed, along with the implications of this work on paramedic clinical practice, education, and future research. The goal of these projects was to learn more about what clinical decisions paramedics make that are most important for patient safety and clinical outcome, and how paramedics make clinical decisions. There has been little work done on paramedic CDM to date, so this work was intended to generate interest on this topic and be a catalyst for future research.
Emergency call scene management
Caring for a patient in the out-of-hospital setting can be challenging. What distinguishes paramedics from other health providers are not the diagnostics and interventions they use to assess and treat patients, but rather where they practice (Campeau, 2008). Campeau (2008), a Canadian paramedic who conducted research on paramedic scene management, commented:
‘Paramedics must ‘fit’ medical procedures into their work context; consequently, paramedic practice is a unique type of care. Paramedics achieve the remarkable objective of transforming everyday, uncontrolled locations where emergencies occur into settings that can be used to effectively deliver emergency care’
A schema, a concept from cognitive psychology, is the general information an individual acquires and organizes in their mind about an experience (Matlin, 2003). Schemas provide a cognitive template for what to expect when entering into a particular situation. Paramedic students quickly learn the schema of a typical emergency call: receive dispatch information, arrive on scene, conduct an assessment, perform initial treatment, move the patient to ambulance, perform repeated assessments and treatments en-route, arrive at destination, give report and transfer care of the patient.
These phases form the major events that occur in an emergency call. Process mapping allows for events that occur during a particular situation to be viewed in a linear fashion, which can increase understanding of the factors at play during a particular process. They are valuable not only to recognize areas where errors currently occur, but more importantly, to prospectively identify processes most vulnerable to adverse events (DeRosier et al, 2002). Process maps have been created to find areas susceptible to clinical error in the emergency department (Croskerry et al, 2006), and to outline the sub-processes required during out-of-hospital rapid sequence intubation (Blanchard et al, 2009).
Consensus on paramedic clinical decisions during high acuity emergency calls: results of a Canadian Delphi study
The objective of this study was to learn more about the most important decisions paramedics make during emergency calls, in terms of clinical outcome and patient safety, and to visualize those decisions on a process map of an emergency call (Jensen et al, 2009; Jensen et al, in press). Advanced care paramedics and EMS medical directors (emergency physicians who provide clinical oversight to paramedics working in the EMS setting) from across Canada participated in this multi-round online Delphi survey. The purpose of the survey was to achieve consensus among the group on the most important clinical decisions paramedics make.
In round I, participants listed all the clinical decisions made by paramedics during high acuity emergency calls they believed to be important for patient safety and clinical outcome. In Round II, participants scored each decision on a 1–5 scale on its importance for patient outcome and safety. In rounds III and IV, participants could revise their scores. If 80% or more of the panel scored a decision important or extremely important, it was included in the final list of important clinical decisions. Included decisions were categorized and plotted on a process map of a typical emergency call.
The panel (17 paramedics, 7 medical directors; mean 16.5 years experience) achieved consensus on 42 important clinical decisions, grouped into six categories: airway management (n=13 decisions); assessment (n=3); cardiac management (n=7); drug administration (n = 9); general treatment (n=6); and scene management (n=4). The airway management and cardiac management categories had the highest mean scores (4.49/5). Table 1 displays the list of included important clinical decisions.
| Code | Decision |
|---|---|
| S-1 | Recognize potential hazards (e.g., people, animals, environment, chemical/radiological/biological risks)—scene safety |
| S-2 | Decide to check for/triage patients at scene with several patients |
| A-1 | Initial assessment: is patient critical or not; level of distress/acuity, decide whether to start treatment right away, or complete assessment |
| A-2 | Recognize signs of life-threatening trauma |
| A-3 | Decide if patient has capacity to refuse or consent |
| S-3 | Decide when to leave scene vs manage on scene (load and go vs stay and play) |
| T-1 | Deciding on appropriate treatment |
| T-2 | Determine if patient requires immediate treatment or can wait til en route, arrival at ED |
| T-3 | Recognize contraindications/reason to withhold therapy |
| T-4 | Reassess patient after giving a treatment—decision on next action (stop drug, change, give another dose, etc) |
| T-5 | Decision to change care plan (switch protocol/med directive) based on patient changes |
| T-6 | Decide how to manage labour and delivery |
| D-1 | Provide ASA |
| D-2 | Give epinephrine for anaphylaxis |
| D-3 | Give epinephrine for severe asthma |
| D-4 | Give epinephrine for pediatric shock |
| D-5 | Decide to give TNK for STEMI |
| D-6 | Provide bronchodilators |
| D-7 | Decide to use drugs to facilitate intubation (sedation, opiates, paralytics) |
| D-8 | Decide on drug for tachycardia (amiodarone/lidocaine/adenosine) |
| D-9 | Decide whether to administer vasopressor |
| AW-1 | Decide on manual airway positioning—if necessary and how (head tilt, jaw thrust, etc) |
| AW-2 | Decide to insert airway adjuncts (OPA, NPA) |
| AW-3 | Decide to use supraglottic device (King LT, Combitube, LMA), ETI or BMV |
| AW-4 | Provide positive pressure ventilation with BVM in respiratory distress |
| AW-5 | Decide whether to attempt intubation in pediatric patient |
| AW-6 | Decide whether to attempt intubation in major trauma patient |
| AW-7 | Decide to use CPAP |
| AW-8 | Decide to perform chest needle decompression |
| AW-9 | How to clear obstructed airway (Heimlich maneuver, suction, forceps) |
| C-1 | Start CPR |
| C-2 | Begin chest compressions on decompensated child (shock) |
| C-3 | Remind/correct chest compressor on CPR quality; have chest compressors switch |
| C-4 | Decision to defibrillate |
| C-5 | Analyze cardiac rhythm (3 or 4 lead strip) |
| C-6 | Interpreting 12 lead ECG |
| C-7 | Decide on electrical cardioversion or medications for SVT |
| AW-10 | Decide how to confirm intubation |
| AW-11 | Decide to extubate if unsure of placement |
| AW-12 | Failed attempt at intubation—try again for ETI or switch to supraglottic device or BVM |
| AW-13 | Decide whether to perform cricothyroidotomy |
| S-4 | Decide most appropriate destination (trauma, heart, stroke centre, community ED, other) |
Decision density is the number of decisions that must be made simultaneously or over a short period of time (Croskerry, 2009a). The on-scene treatment phase of the process map of a typical emergency call appears to have the highest decision density (Appendix 1). For paramedics, scene management is an essential component of their duties.
Metz stated: ‘the measure of a man or woman doing paramedic work is always decided at the scene (1981: 93). Identification of areas of high decision density brings into focus point(s) of the call that have an increased susceptibility to near misses, adverse events and errors (Chisholm and Croskerry, 2009). The on-scene treatment phase is the period of a typical emergency call when paramedics deal with variable settings (outside the ambulance) and are likely to have the least amount of clinical support (e.g. when paramedics arrive on scene and then call for another crew for assistance). While not all of the decisions on the map happen during each emergency call, the map is valuable for increasing awareness of when paramedics are likely to be inundated with many decisions. This process map was not intended to be a representation of the thinking process that occurs in a paramedic's mind (consciously or subconsciously), as decision-making itself may not occur in a linear fashion. However, the main events of an emergency ambulance call (call dispatched, paramedics en route, patient contact, etc) are generally predictable, and in most instances an emergency call is a linear process through time.
Thinking about clinical decision-making
In many EMS systems, paramedics use clinical protocols to help guide the care they deliver. A traditional assumption has been that paramedics make most of their clinical decisions by choosing the most appropriate protocol and following it from memory (i.e. algorithmic thinking), but this is probably not a sufficient explanation for how paramedics actually make decisions in practice (Bigham et al, 2010).
Clinical decision-making has been explored by other health professionals and disciplines of study. The dual process theory is the predominant decision-making theory in cognitive psychology (Table 2). The theory divides decision-making into two processes. Type I thinking, often referred to as subconscious thinking, is reflexively employed when mental short cuts are used to make decisions, without conscious thought, and is essential for minimizing thinking effort (Evans, 2008). Type 2 thinking underlies the decisions that require purposeful contemplation and analytic thought.
| Characteristic | Type I | Type II |
|---|---|---|
| Cognitive style | Intuitive/heuristic | Analytical/systematic |
| Awareness | Low | High |
| Conscious control | Low | High |
| Automaticity | High | Low |
| Cost/effort | Low | High |
| Rate | Fast | Slow |
| Reliability Low | High | |
| Errors | Vulnerable to error | Few but large |
| Predictive power | Low | High |
| Emotional valence | High | Low |
| Detail on judgment process | Low | High |
| Scientific rigour | Low | High |
Although Type 1 processes are more prone to error, they are effective in minimizing reaction time, avoiding ‘paralysis by analysis’ (Croskerry, 2009a). For example, a novice paramedic may feel unsure about how to approach an unresponsive trauma patient, but as they gain experience, the cervical spine is immediately held still without conscious deliberation.
Type 2 processes are engaged in more complex situations, when a decision cannot be made quickly. In this mode of thinking, the thinker must weigh the pros and cons of each option and make a conscious decision on the best fit. A paramedic may decide a patient's chest pain is cardiac in origin, rather than musculoskeletal, after they have conducted a physical exam, a patient interview and collected diagnostic data such as electrocardiogram and vital signs. The paramedic may make the decision with purposeful deliberation between the competing possibilities (using Type 2 processes), or it may be instantly made by recognizing the combination of presenting symptoms as likely cardiac (such as crushing retrosternal pain, nausea, sweating). Subsequent actions are based on this pivotal decision.
Another category of thinking strategies exists, which do not readily fall into either Type I or Type 2 processes. The term ‘Type 2-by proxy’ refers to situations in which clinicians use thinking ‘tools’ to speed decision-making. These tools, such as algorithms, clinical prediction tools, and differential diagnosis lists, have been developed by experts using Type 2 processes, typically by heavily referring to research evidence. Clinicians memorize the tool and during the patient encounter, the steps or list is recalled. An example of this would be working through a resuscitation guideline while managing a cardiac arrest (i.e. algorithmic thinking). Several thinking strategies have been proposed on how emergency physicians make decisions (Sandhu and Carpenter, 2006) (Table 3). Each of these strategies can be categorized as employing Type 1, Type 2 or Type 2-by proxy processes.
| Name | Type | Details |
|---|---|---|
| Event driven therapy | I | Treat symptoms and then re-evaluate with further evaluation, depending on response to |
| Intuition | I | Decisions made without conscious thought |
| Pattern recognition | I | Combination of salient features establish likely diagnosis with corresponding evaluation and management plan |
| Exhaustive | II | Accumulate facts indiscriminately and then sift through them for diagnosis |
| Hypotheticodeductive | II | Inference based on preliminary findings, idea modification based on subsequent findings, response to therapy and exclusion of competing possibilities |
| Algorithmic | II by proxy | Preset diagnosis or treatment pathway, based on pre-established criteria |
| Rule out worst scenario | II by proxy | Consideration of pre-existing ‘can't miss’ list of diagnosis for presenting condition |
Clinical decision-making by Canadian advanced care paramedics: a think aloud study
In this study, Canadian advanced care paramedics (ACP) verbalized their reasoning while working their way through two scenarios of emergency calls (one trauma and one medical) (Jensen et al, unpublished observations, 2010). The think aloud technique was used to identify the decisions made and thinking strategies used by the participants during the verbal ‘paper patient’ scenarios.
Purposeful sampling was used to create an equally mixed sample of novice (less than two years experience at the ACP level) and experienced participants (greater than two years experience at this level). Following the think aloud technique, participants were encouraged to stop and explain why they made each assessment, treatment and transport decision during the audio-recorded scenarios (Fonteyn et al, 1993). The clinical decisions made and thinking strategies used were identified in the interview transcripts.
During analysis, investigators matched the participants' explanation of their decision-making and the context of each decision to one of seven predefined thinking strategies (Table 3)—hypotheticodeductive, exhaustive, algorithmic, rule out worst scenario (ROWS), event-driven, pattern recognition and intuition) (Sandhu and Carpenter, 2006).
Eight ACPs with a mean 9.6 years of overall paramedic experience (SD 6.7) participated. Twenty-nine decisions were made in the trauma scenario. Eighteen decisions were made in the medical scenario. In the trauma scenario, participants used event-driven and algorithmic thinking most frequently. In the medical scenario, algorithmic and ROWS were employed the most. Event-driven thinking was used more often in the trauma scenario compared to the medical scenario (45 decisions vs 0, t-test, P<0.001), otherwise no differences in thinking strategy used by scenario types were found. Decisions in both scenarios were made most by using Type 2-by proxy thinking strategies (Table 4).
| Type | Total decisions | Trauma decisions | Medical decisions |
|---|---|---|---|
| I (event-driven, pattern recognition, intuition) | 78 | 60 | 18 |
| II (hypotheticodeductive, exhaustive) | 60 | 34 | 26 |
| II-by Proxy (algorithmic, ROWS) | 137 | 80 | 57 |
The novice paramedics failed to verbalize a significantly larger number of decisions than the experienced paramedics in both scenarios (mean 8.50 decisions not made per participant versus mean 4.12 decisions not made per participant, t-test, P<0.05). Differences in thinking strategies used were not found between the novice and experienced ACPs. It is possible that the thinking strategies paramedics use do not change significantly with more experience.
A major limitation of the think aloud study is that intuition cannot be directly measured. The requirement to think aloud and verbally report on decisions may bear no direct relationship to any intuitive thinking that might have occurred, and may even change it (Nisbett and DeCamp Wilson, 1977; Hogarth, 2005). Therefore, it is possible paramedics rely on Type 1 processes to a greater extent than could be detected in this study.
Some health professionals do not believe algorithms and clinical prediction rules can outperform clinical judgment (Paley, 2007), and may feel these are a threat to their decision-making autonomy and lead to inflexible care plans. However, several reviews and one meta-analysis comparing clinical judgment to clinical predication rules found that these tools are almost always more accurate, and often require less clinical information than individual clinician judgment (Grove et al, 2000).
Therefore, in chaotic or time-sensitive situations, it is preferable for paramedics to use decision tools, rather than rely on Type 1 processes, which can be influenced by bias, the emotional state of the thinker, and inappropriate use of mental short cuts (Croskerry, 2005; 2010). It is also preferable to taking too much time deliberating each competing hypothesis or decision option using Type 2 processes before acting (Croskerry, 2009b).
Although small and exploratory, this study contributed evidence to dispel assumptions that may exist that paramedics solely rely on algorithmic thinking to make clinical decisions. This finding provides us with new challenges: if paramedics use more than one thinking strategy, which are the best for which situations? Also, how can we teach paramedics to recognize their own thinking strategies?
Implications of clinical decision-making research
Implications for future paramedic research
The two studies have generated further research questions about paramedic CDM. This is a complex topic, and many studies are required to build a cohesive body of knowledge in this area, as it applies to paramedic practice. Future paramedic research questions that can be asked with the think aloud technique include: the impact of paramedic variables on thinking strategy, such as paramedic level, call volume (rural compared to urban service), work setting (ground ambulance versus air), and type of paramedic education (full-time diploma, part-time diploma, undergraduate degree).
Think aloud studies could be conducted in a simulation lab, to give participants a more realistic sense of an emergency call. Further, it would be possible to have paramedics think aloud while they are working the ambulance setting, in order to determine thinking strategies in real time (Fonteyn and Fisher, 1995; Aitken and Mardegan, 2000). Other aspects of clinical reasoning can be explored, including the use of mental short cuts by paramedics.
Finally, paramedics make decisions during emergency calls that are unique to that profession. These factors may include timing (how long to spend on scene, how many interventions can be done in the time it takes to get to the hospital) and clinical support (do I have the time or hands to call the medical director for advice, should I call for another paramedic crew or medical first responders to help?).
Studies like these are important for developing the field of clinical decision-making further, in order to understand more about how CDM varies, and the impact on patient outcome and safety. The field of decision-making is multi-faceted, and includes theory and research from psychology, philosophy, neuroscience, statistics, computer science and others (Croskerry, 2000). The academic work of these scientific disciplines needs to be interpreted and applied to the real time setting of paramedics.
Implications for paramedic education
Through didactic learning, and even more so, during their clinical preceptorship, student paramedics learn how to manage an emergency call in a routine fashion. This occurs with the establishment of an emergency call schema in their memory. This seems to happen naturally over the course of preceptorship, which involves bridging the theory to practice gap through observation, repetition, and following the actions of experts (their paramedic preceptors) (Boyle et al, 2008).
While students may quickly learn the process of an emergency call, it is more difficult to develop an understanding of how decisions are made. Much has been written about medical education and the importance of teaching quality clinical reasoning through examples (Kassirer and Kopelman, 1989). In typical paramedic education, students learn and are tested extensively with simulation. These scenarios should test thinking strategies, not just clinical conditions and treatment paths (Kassirer and Kopelman, 1989).
Paramedic educators should feed information (such as past medical histories) to the student slowly, to replicate how it is uncovered in real life (Kassirer, 1983). Every time a student asks a question, requests more information, or performs an assessment or treatment task, the scenario should pause and the student explain why they are making the decision, recognizing the benefits and pitfalls of the process used. Through this type of exercise, students can learn how to use different thinking strategies (Banning, 2008b). For instance, they can increase their ability to tap into Type 2 processes by pausing, developing hypotheses, and ruling them in or out as new information becomes available. This would be a departure from the tradition of teaching decisionmaking by following algorithms.
Similarly, when paramedic students are in the clinical phase of their training, their preceptors should probe them about why they are making each decision, and discuss different thinking strategies. If this is not possible to do in real-time during the call, the questioning and discussion should ensue immediately after the call is complete.
In addition to teaching paramedics how to appreciate different thinking strategies, paramedic educators should specifically discuss which clinical decisions are important, and require deliberate, conscious decision-making, versus those decisions safely and effectively made with intuitive or unconscious thinking strategies. Given that the decisions paramedics make can have a major impact on some patients' outcomes, it is important for all students to learn about how decisions are made. A rigorous study should be conducted, comparing student decision-making between a group that has received a module on CDM and those who have not. The outcome might determine if paramedic CDM should become a mandatory competency for paramedic education.
Implications for paramedic practice
Croskerry et al (2000) proposed that each health discipline should identify meaningful patterns in their own practice that are prone to error. The onscene treatment phase of emergency calls was found to have the highest decision density, and therefore increased vulnerability to error and subsequent adverse events (Croskerry and Sinclair, 2001). This knowledge has important repercussions. Continuous quality improvement (CQI) paramedics and medical directors are tasked with ensuring the quality of care is high and risk of error is low. It would seem important that they work closely with paramedics who have made clinical errors, and encourage them to reflect on their thinking and try new strategies, instead of only focusing on the clinical aspect of the error.
It is imperative for CQI paramedics to be aware of the decisions that were found to be the most important for patient outcome and safety, and seek these out while conducting clinical audits of emergency calls.
Morbidity and mortality (M and M) rounds have a long-standing tradition in EMS. In these sessions, paramedics and medical directors gather to discuss emergency calls that were challenging or resulted in an adverse event (Cosby, 2009). The operational and clinical aspects of the call are discussed, and consensus is reached between the presenting paramedic and his or her colleagues on what the most ideal actions would have been. It is rare for a presenter to discuss the thinking strategies they used. Presenters should be encouraged to conduct a cognitive autopsy as soon as possible after the call, in order to maximize recall. Cognitive autopsies are ‘a form of cognitive and affective root cause analysis’ (Croskerry, 2005: 10).
During M and M sessions, the paramedic should focus on the events of the call and the decisions made, and also what they were thinking and feeling at the time. This metacognitive exercise would inevitably lead to improvements in clinical practice as paramedics learn more about how they make decisions during emergency situations. This information is at least as important to share as the clinical details. Further to this, paramedics should be encouraged to write up case reports of calls that required challenging decision-making.
As an example, Campbell et al (2007) published a case report that included a detailed analysis of cognitive biases that resulted in an important diagnosis being missed in an emergency department patient. Perhaps through the incentive of continuing education credits, paramedics should be encouraged to submit and share case reports of this nature.
Conclusion
High quality care is achieved when practitioners make clinical decisions that are safe and effective. Paramedics often make decisions while working in variable settings, unique from other healthcare providers. The on-scene phase of an emergency call has the highest important decision density; it is important for paramedics to be aware of this in order to minimize adverse events. In their work, paramedics engage a spectrum of decision-making strategies in their work that include Type 1 processes, Type 2 processes, and Type 2 –by proxy tools, that typically involve the use of algorithms, clinical decision rules, and practice guidelines. CDM is a topic of utmost importance to the development of the paramedic profession.