Should we add smells in simulation training? A systematic review of smells in healthcare-related simulation training

Abstract

Olfactory stimuli are rarely used in healthcare-related simulation training. Their addition may improve simulator validity, biographical memory formation and coping mechanisms for exposure to strong malodours. Some military simulators already employ smells in simulation training, and the technology and principles may be used to cross over into medical simulation training. We set out to determine if there is evidence to suggest that smell should be routinely incorporated into medical simulation training. We carried out a systematic review of the literature relating to use of olfactory stimuli in medical simulation training, and identified 5 relevant papers. 3 were experimental studies and 2 were observational studies. The results of the experimental studies were mixed, though there were some indications that the use of a clinically relevant smell instead of a congruent background smell may be more effective. We discuss the benefits of the inclusion of smell in simulation training and identify that there are currently few high-quality studies addressing the use of smell in medical simulation training.

Introduction

Olfactory stimuli are rarely employed in healthcare-related simulation training. Yet simulation validity is an important factor in the educational value of these simulations, and efforts are constantly being made to improve this.^{1 2} While sight, touch and sound are routinely used, the sense of smell, possibly because of historical technological reasons or lack of awareness of its importance, is usually neglected.³ Some researchers have hypothesised that addition of a further sensory modality to simulations may enhance the simulator validity and therefore improve the educational value of simulator training.³

The importance of smell in diagnosis has been recognised since ancient times, and is still taught in some forms of traditional medicine, yet its role in current medical education is unknown and also largely unexplored.⁴ Current medical smell-related technologies include chemosensory devices and sniffer dogs which can diagnose cancers.⁵ Yet the human faculty of smell often remains naive until confronted for the first time with particular olfactory stimuli during real-life clinical scenarios. Exposing healthcare professionals earlier in their training to smells which may aid in diagnosis (eg, ketone) and to malodours which could complicate management (eg, gangrene, melena) might improve diagnostic gestalt and allow more timely diagnosis, as well as the possibility of developing coping mechanisms for dealing with strong malodours.

Olfactory stimuli are sometimes used in military and law enforcement simulation training in several countries.⁶ The benefits of using smells in such training are partly to desensitise participants to strong unpleasant odours, as a teaching tool to warn of danger (eg, mustard gas), and also to induce a state of arousal in learners.⁷ These smells may be transmitted over a larger area than medical smells would require, and are produced either by specialist companies or simple measures such as setting tires alight.⁸ The integration of experience gained in non-medical areas, such as the airline industry, into medical training has already been well proven, and the authors hypothesise that inclusion of smell, as used in military training, may enhance medical simulation training.

The addition of smell to high fidelity medical simulation may improve the validity of simulators, could desensitise trainees to strong malodours and potentially assist in biographical memory formation.^{9 10} Yet little is known about what types of smell would be most effective, how these smells should be presented, and what barriers prevent smells from being used routinely in medical simulation. Though there is not a large volume of evidence about the incorporation of smell into medical simulation training, there is a need to consolidate what is known on the subject, and to stimulate further study into this novel area, which has potential to add a new sensory dimension to simulation training. We carried out a systematic review of literature relating to the use of olfactory cues in medical simulation to determine if there is evidence that smell should be routinely incorporated into medical simulation training.

Methods

A systematic review of published work was conducted using the protocol specified by the Cochrane collaboration and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for the conduct of meta-analyses.¹¹ The following sources were searched: MEDLINE via PubMed (from inception to June 2015); EMBASE (same date restriction), and generic internet search engines. There was no limitation on publication type or language. The following medical subject headings (MeSH) were used: ‘Patient simulation’, ‘Olfactory perception’, ‘Smell’ and ‘Training’. Articles were also identified by hand searching of references and use of the related articles function in PubMed.

Data were extracted independently by two researchers (SJWK and FHK). From the MEDLINE search, 325 studies were screened using title and abstract by two of the researchers. Studies pertaining to testing or treatment of anosmia using smell training were excluded.

Eleven full texts were retrieved and six of these studies examined smell in other settings such as smoking cessation training, localisation of smell, cognitive memory or virtual restorative environments. Five studies were identified which focused on smell in simulation training. Three of these were experimental and two were observational discussions. Studies were appraised independently by two of the authors (SJWK and CWB; figure 1).

Figure 1.

Identification of eligible studies.

Results

Birnbach et al randomised 165 medical students and medical interns to two groups, and asked them to examine a simulated patient with palpitations. A commercially available scent distribution device was used to emit a fresh citrus smell, similar to that found in some cleaning products, in 79 of the simulations. The groups which were exposed to the smell showed more frequent hand washing behaviour than the control group, who were exposed to the background smell in the simulation centre. The results of this experiment were statistically significant and continued to occur in subgroup analysis of medical students and interns separately. The study is well conducted with the only potential source of bias being identified that all of the ‘with smell’ groups were tested in the afternoon, though it is not known whether this is likely to have an effect on hand washing behaviour. No further information was given about the conduct of the simulation aside from the addition or not of a smell.¹²

Nanji et al randomised 103 anaesthesiologists to two operating theatre-based anaesthesia scenarios, identical except for the addition of smoke produced using a diathermy device on bovine tissue (n=52 intervention group). The authors describe how they developed their scenarios to maximise physical fidelity. Outcomes were measured using a seven-question survey assessing participant perceptions of physical, conceptual and emotional/experiential fidelity; and whether realism enabled learning engagement. Results of this study did not show any statistically significant difference in how real the participants felt the scenario was or how emotionally involved they were. The authors hypothesised that this may be due to the pre-existing high level of fidelity in the control group, indicating a high perception of baseline realism and making any incremental benefit negligible. The authors also consider that their results may be affected by the lack of sensitivity of the survey to detect changes in realism between the two groups. No risks of bias were identified.¹³

Robertson et al used a quasi-experimental design to study the effects of the addition of a malodour to a simulated wound care scenario. One hundred and thirty-seven student nurses carried out a wound care scenario with no added smell and were invited to carry out the scenario again. Forty-nine of them volunteered to return and complete the scenario again, this time with a malodour included in the simulation. Outcomes were measured using Likert scale questionnaires enquiring about the perceived realism of the simulation, student's ability to determine the care needs of the patient, and confidence in managing the situation again. Control questionnaires were randomly drawn from the first group, though no information on randomisation was given. Results indicated that the group who repeated the simulation with the addition of a malodour rated the realism of the situation more highly, and felt more prepared to manage the situation in the future. Limitations of this study include the self-selection of the experimental group and the undisclosed method of random selection of responses from the ‘control group’.¹⁴

Spencer's and Krueger's papers were observational discussions, with no experimental element. The first is a report funded by the US military, which presents investigations into the viability of using smells within virtual reality settings to train surgeons. The project was at a preliminary stage, and discussed issues surrounding production of smells and infusion of these smells in appropriate quantities and timings. They identified that the major barriers to using smells within medical virtual reality settings included realistic smell production and smell delivery.¹⁵ The second of the observational papers discussed the technological aspects of using computer-controlled smell-generating devices to enhance healthcare-related simulations, both in simulation laboratories and also using personal computers and laparoscopic simulators in distance learning. They highlighted the growing number of commercially available smell generators, which can blend smells generated by oils which are heated and dispersed by fan. They did not discuss how smells might be effectively delivered throughout a simulation room, removed after simulation, or how specific medically related smells might be produced. No source of bias was identified.³

With only five studies identified, and different outcome measures used in each, no further analysis was possible (table 1).

Table 1.

Study characteristics

Reference (first author)	Type of study	Size of cohort	Smell used	Outcome
Birnbach12	Randomised, controlled	n=165	Fresh citrus smell	Increased hand washing seen in group exposed to the smell
Nanji13	Randomised, controlled	n=103	Smoke from diathermy	No improvement in trainee perceptions of fidelity of simulation
Robertson14	Cross over, partially controlled	n=49	Gangrenous wound smells, created using cheeses	Improvement in perceptions of realism and educational value
Spencer3	Non-experimental	NA	Various, including garlic to represent arsenic poisoning	Discussed potential use of currently available smell producing devices in surgical simulation training
Krueger15	Non-experimental	NA	Various, including iodine scrub, chlorhexidine scrub	Discussed issues relating to smell production and delivery, and the need for further work in these areas

NA, not available.

Discussion

High fidelity simulation aims to provide a realistic immersive setting in which participants can practice, experience and learn without potentially harmful consequences. To improve the immersive nature of the scenario being practiced, realism is improved by mimicry of the sights, sounds and tactile sensations participants might experience in a real-world scenario, and scenarios are designed to be emotionally realistic. Smell is rarely used during medical simulations, though it has the potential to alter the behaviour of participants,^{12 16} to further improve simulator validity,^{1 3 13} to assist in biographical memory formation^{9 17 18} and to desensitise trainees to strong malodours.^{7 8 19} We carried out a systematic review of literature relating to the use of smell in medical simulation training.

The memories formed in connection with smells are thought to be strongly autobiographical, as opposed to procedural memories which are formed when a task is repeated and learnt or declarative memories which are facts which can be consciously recalled. Biographical memories are closely linked to emotional states and also to fine details of when the smell was first encountered.^{9 17 20} This makes the memories formed in connection with smell ideal for use in a simulation setting, where attempts are made to recreate emotionally aroused states, and the ability to diagnose and manage a problem is practiced. By adding a clinically relevant smell to the scenario, the biographical memory formed may be recalled on future exposure to the smell, and details relating to diagnosis and management may potentially be recalled more quickly and accurately.

A second possible benefit of the addition of olfactory stimuli is the improved fidelity of the simulation, making it closer to the real-life scenario. The benefits of improved fidelity are well documented, and may improve knowledge retention and trainee performance.¹

A third potential benefit of using smell in simulation is to desensitise trainees to malodours. Learning to inhibit a natural, involuntary gag reflex on exposure to certain smells is important for all healthcare providers, as it allows the relationship with a patient to progress without reinforcement of low self-image which might ensue given a strong negative reaction to patient odour. It may also help practitioners to complete tasks efficiently in a malodorous environment such as the airway management of burns patients or the wound dressings of ulcerated skin. Stress inoculation training in combat medics is thought to improve battlefield performance, and use of realistic smell simulators is a key in inducing appropriate sympathetic responses in trainees.^{7 21} Indeed we hypothesise that smell may become an important element of situational awareness which is rarely investigated in clinical situations.¹⁰

Recent studies have examined whether certain background smells have an effect on improved retention of memories, improved immune status and emotional state, with promising initial results. Smells have also been used with success in virtual restorative environments, which use smell to enhance the validity of virtually generated environments, though the therapeutic effects of smells are out with the scope of this review.^{10 16 18 22} The use of smells to prompt certain behaviours via unconscious mechanisms, such as those shown by Birnbach et al,¹² also has much potential in the clinical setting.¹⁶

We identified three experimental studies which assessed various impacts of using smell in simulation training, and two observational studies which discussed strategies and barriers to incorporation of smell into a surgical simulator. Two of the experimental studies were randomised controlled trials, and one was a quasi-experimental observational study design. Study sizes were both between 49 and 165 participants, and each of the studies involved addition of only one smell.

Results of the experimental studies identified were mixed—Nanji et al showed no improvement to simulator validity while Robertson et al showed improvements in both simulator validity and trainee perceptions of preparedness. Birnbach showed a change in behaviour in the smell exposed group, but did not examine simulator validity or any educational outcomes. The experimental studies can be broadly divided into those that employ a clinically relevant smell (ie, one which suggests a particular diagnosis) and those which added an incidental background smell (smoke or clean citrus smells). In the first study which added an ‘incidental’ background smell in order to improve the simulator validity (smoke from a diathermy in an anaesthesia simulation), there was little difference in the participants perception of the realism of the situation, and no measurable benefit to adding to smell; however, the addition of a background citrus smell did impact positively on hand washing behaviour. In the quasi-experimental study which used a clinically relevant smell (mouldy cheese which suggested a diagnosis of wound infection), participants reported improved outcomes from the simulation.

None of the studies used a clinically relevant smell with a randomised controlled design to measure a reproducible, objective outcome related to the educational value of medical education. The study which used a clinically relevant smell was not robustly designed and contained significant sources of bias, and the well-designed trial which used a citrus smell to demonstrate change in hand washing behaviour does not have a clear educational application. Future studies using smell in medical simulation training, which measure objective outcomes related to educational value of the simulation, are required to establish the role of smell in medical simulation training.

As both medical simulation technology and olfactory delivery systems improve, the authors hypothesise that smell may eventually have a role in modern medical simulations. While the technology exists to add olfactory stimuli to medical simulations, there is currently little evidence about whether this would be beneficial.^{3 23} This systematic review identifies that few studies have been carried out in this area, and suggests that further research is required to determine the importance of incorporation of smells into medical simulation training.