Fidelity in simulation is a term referring to the loyalty of a particular model to real life characteristics. Essentially, it refers to how ‘life-like’ a simulation may be. It is a concept that is often poorly defined and subject to various interpretations in the medical literature. Fidelity, or lack thereof, is an intuitively grasped property of any attempt at simulation. Maran and Glavin (2003) describe fidelity as ‘the extent to which the appearance and behaviour of the simulator/simulation match the appearance and behaviour of the simulated system’.
Fidelity is often attributed to a particular simulation tool, or simulator, as a fixed variable. At times this can be set arbitrarily by manufacturers as part of marketing. Technological sophistication can be confused with fidelity. Also, it is likely that incremental enhancements of a simulator to improve its fidelity will increase the cost. Beyond a certain, ill-defined, point these enhancements will add little to the role of the simulator as an educational tool. In reality, the fidelity of the simulation (rather than the simulator) is what is important, from the perspective of the learner. A ‘high-fidelity’ simulator used in a certain manner can result in a low-fidelity educational experience.
With a lack of a consensus of the definition of fidelity, interpreting the body of literature looking at simulation fidelity is difficult. It is also likely that the results of certain studies cannot be easily applied in other contexts. Many papers describe high-fidelity simulation as having occurred simply due to qualities of the simulator used.
Physical vs. Psychological
Miller (1953) first made the distinction between engineering, or physical, and psychological fidelity in the context of simulation for aviation. Engineering fidelity is the extent to which the training device replicates the physical characteristics of the real-world task. Psychological fidelity is how well the simulated task captures the skills necessary for that task in the real world.
These two aspects of fidelity do not necessarily equate. For example, a simple foam model provides relatively little physical fidelity when practising suturing, but excellent functional or psychological fidelity.
Spectrum of Simulators
Simulators do not need to be physical and include mental exercises, where learners engage in imaginary activities. See Table 5.1 for a tabulated spectrum of simulation.
Simulator characteristics | Fidelity | Cost | User feedback | Instructor input |
---|---|---|---|---|
Part task trainer | Low | Low to moderate | Nil | Low to moderate |
Mental simulation | Low | Low | Low to moderate | Low to moderate |
Computer-based system | Low to moderate | Moderate to high | Moderate to high | Nil to low |
Virtual reality and haptic simulator | Moderate to high | High to very high | High | Moderate to high |
Simulated patients | Low to moderate | Low | High | High |
Hybrid patients | Moderate to high | Low to moderate | High | High |
Integrated simulators – model-driven | High | Very high | High | Low to moderate |
Integrated simulators – instructor-driven | High | Moderate to very high | Moderate to high | High |
Part Task Trainers
These simulators seek to replicate only the salient portion of the environment relevant to the task at hand; for example, phlebotomy arms or static pelvic models for delivery simulation. As a consequence, they are relatively inexpensive to produce and maintain, with the most significant cost arising from the need for an instructor to teach and assess the clinical skills in question.
Computer Screen-Based Systems
Computers can be used to model different clinical scenarios, as well as faithfully replicate physiological and pharmacological data (de Wit-Zuurendonk and Oei, 2011). Students can then alter these models and the results can be observed, which enables real-time feedback to be supplied to the student while also allowing multiple attempts and therefore multiple outcomes. Moreover, as the only physical requirement for this type of simulation is an interface, multiple learners can be trained simultaneously in a semi-autonomous fashion.
Virtual Reality and Haptic Systems
Virtual reality (VR) represents the ultimate in computer-based simulation. The incorporation of a physical interface, often a part task trainer, that responds dynamically to the user and thereby attempts to replicate the physical feedback associated with a particular task allows for even further levels of interaction. This simulation method is used extensively in laparoscopic training models for surgical trainees. To date, the role for this modality has been limited in obstetric simulation.
Simulated Patients and Hybrid Simulators
Standard, or simulated, patients are already commonplace in medical education. Ranging from undergraduate history taking to communication assessment stations in objective structured clinical exams (OSCEs) at a postgraduate level, this type of simulation in isolation is almost exclusively aimed at imparting non-physical skills. Simulations involving patient actors were associated with improved perception of communication and safety compared with manikin based simulation (Crofts et al., 2008).
A standard patient can be combined with physical models such as part task trainers or even more complex integrated models (see below). These hybrid simulators allow for the integration of communication, interpersonal and psychomotor skills, as well as providing the overarching appropriate clinical context for each of these components. While lacking any specific pathology, they will potentially be able to portray symptoms accurately through verbal and non-verbal cues. This can readily change a simulation scenario from a low physical fidelity simulation into a high psychological fidelity simulation with the addition of communication skills and direct human interaction in addition to the psychomotor skills taught by a simple physical model. It is particularly significant in obstetric population, where a non-pregnant actor would likely challenge the immersion of participant involved in the simulated scenario.
The concept of hybrid simulators was first reported in 2002 for procedural skills (Kneebone et al., 2002) and has expanded to include a wide range of procedural and operative skills with simulation manufacturers designing products for this specific purpose. Higham et al. (2007) report the concept for vaginal examination and Pap smear. Cooper et al. (2012, 2016) describe a birthing suit that is worn by an SP and enables midwifery students to learn effective teamwork during delivery. Similarly, Kumar et al. (2016) describe a hybrid model of an SP and birthing model ‘in situ’ for supporting the development of interprofessional collaborative practice for homebirths. An alternative form of hybrid simulation developed by Pauline Lyon simply uses a life-size photograph of a woman in labour that is placed alongside the birthing model (Lyon, personal communication). Together with sound effects, the visual representation engages simulation participants differently to simplify learning on the model.
Figure 5.1 shows use of a wearable part task trainer as a hybrid solution.
Figure 5.1 A hybrid postpartum haemorrhage simulation. A faculty member playing the part of a patient (standard patient), wearing the low-cost part task trainer MamaNatalie to facilitate a high-fidelity simulation.
Integrated Simulators
Integrated simulators combine many of the above features to be as fully immersive a simulation experience as possible. Model-driven simulators, controlled by sophisticated physiological and pharmacological algorithms that enable autonomous responses independent of instructor input, are hugely expensive. These incorporate full-body manikins capable of replicating a wide variety of clinical parameters, including speech, ECG rhythms, central venous pressure traces and chest sounds. However, they often lack non-verbal cues (e.g. changes in facial expression, muscle tone or skin colour).
An ‘intermediate’-type simulator might have significant computer modelling built into it, but is still dependent upon instructor input to modify ‘physiological’ responses to learner interaction.
Used in combination with faithful location replication and simulated patients, and possibly even clinical staff, these simulators allow for both excellent physical and psychological fidelity in the hands of expert instructors, who have sufficient experience with the models to minimise delays in response and inappropriate cueing to trainees.
High vs. Low Fidelity
The fidelity of simulation (and simulators) is often arbitrarily defined along a continuum of low-, intermediate- and high-fidelity. The utility of the delineation is questionable. There is no standard which describes how close to reality a simulation has to be to be termed high-fidelity. It is also likely that certain current simulations considered high-fidelity will be deemed lower fidelity in the future, with developments in the science of healthcare simulation. The terms are used mostly in a relative sense in that one particular simulator may be more ‘realistic’ and hence of higher fidelity than another. For example, a low-fidelity simulator in epidural anaesthesia may be a plastic model of the lumbar spine, while a high-fidelity model is a life-size replica of a patient’s back with multiple fluid-filled compartments capable of replicating loss-of-resistance during Tuohy needle insertion.
Low-fidelity simulation is most often used for demonstration and practice of manual technical skills, e.g. suturing, urinary catheter insertion or intravenous cannulation. In this respect, it is frequently, although not exclusively, used to impart a one-dimensional skill set. The learning objective is to acquire competency in a defined set of technical skills. Moreover, instructor interaction is necessary to enable the skill acquisition – low-fidelity simulators, typically, do not respond dynamically to trainee input.
High-fidelity simulation, by comparison, is deemed a more accurate attempt at replicating ‘real life’, such as full clinical case scenario rather than a single skill. An example of this would be a resuscitation training scenario involving an advanced, computer-driven, patient simulator (e.g. SimMan, Laerdal Medical, Stavanger, Norway). High-fidelity simulation seeks to provide context in which a skill is not only deployed, but also examines the appropriateness of the decision to use said skill. In that regard, high-fidelity simulation also involves non-technical skills such as communication, leadership and stress management. High-fidelity simulators are characterised by their ability to respond to trainee input, either through indirect instructor input via a computer or through automatic scripted responses built into the simulation software.
At present, the evidence to favour high-fidelity over low-fidelity simulation is scant. In a small study of nursing students undergoing training in congestive cardiac failure, Kardong-Edgren et al. (2007) failed to show an improvement in post-training test scores between high- and low-fidelity arms. DeStephano et al. (2015) found that a low-fidelity hybrid spontaneous vaginal delivery model was as effective at teaching normal delivery to medical students as a high-fidelity computer-controlled manikin. Interestingly, the students felt more confident in their abilities after training on the low-fidelity hybrid model, but this may be confounded by the human interaction component this included that the remote controlled manikin lacked. Friedman et al. (2009) found no difference in the performance of anaesthesia students trained to perform epidural insertion on a high- or a low-fidelity model, and no apparent change in steepness of their learning curve over subsequent cases.
This is in contrast to a randomised controlled trial conducted by Crofts et al. (2006), which demonstrated that a high-fidelity model of shoulder dystocia, which incorporated force feedback, offered additional training benefit in terms of more successful deliveries with less force applied.
Scholz et al. (2012) found that although two groups of medical students made equivalent obstetric decisions regardless of whether they underwent training on low- or high-fidelity simulators, those that underwent high-fidelity training felt more confident in their own abilities, as well as demonstrating improved obstetric skills.
Studies need to include not only indicators that training has transferred into clinical practice, but also that alterations in clinical outcome have improved patient outcomes. However, at this point there is little evidence that high-fidelity simulation training impacts clinical outcomes. A multicentre randomised control trial by Fransen et al. (2017) failed to show a difference in a composite endpoint of obstetric complications following high-fidelity simulation team training.