Error Identification in Tourniquet Use : Error analysis of tourniquet use in trained and untrained populations

(1)

Linköping University | Department of Computer and Information Science Master thesis, 30 ECTS | Cognitive Science Spring term 2020 | LIU-IDA/KOGVET-A--20/016--SE

Error identification in tourniquet

use

Error analysis of tourniquet use in trained and

untrained populations

Molly Lundberg

Supervisor: Erik Prytz Examiner: Arne Jönsson

(2)

Upphovsrätt

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning.

Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art.

Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att

dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart. För ytterligare

information om Linköping University Electronic Press se förlagets hemsida

http://www.ep.liu.se/.

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances. The online availability of the document implies permanent permission for anyone to read, to

download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against

infringement. For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.

(3)

iii

Abstract

The number of prehospital deaths caused by large bleedings could be decreased if civilian people would act in time to help the injured patient. One way to help is to stop the bleeding with a tourniquet application. However, the tourniquet needs to be placed correctly in order to stop the bleeding. Therefore laypersons need to be educated in bleeding control to increase the rate of successful tourniquet application. This study used human error identification techniques such as Hierarchical Task Analysis and Systematic Human Error Reduction and Prediction Approach to identify possible errors of four commonly used tourniquet models: the CAT-7, Delfi-EMT, SAM-X and SWAT-T. The results show that many predicted errors are time-oriented and critical.

Video analysis of tourniquet application was performed to map occurred use errors from the videos with the predicted ones. The goal was to identify problems that could be solved by training or redesigns of the tourniquets. The results show that the most common errors for all participants during tourniquet application were of six error types. The errors were to not check time or write down time of application, to take too much time to place the tourniquet around the limb, to place the tourniquet upside down, to place the tourniquet band over the securing

mechanism instead of between and lastly to not secure the tourniquet correctly before

transporting the patient. The untrained laypersons made more errors than the trained laypersons and professional emergency personnel group. The trained laypersons also made fewer errors in a calm setting than in a stressed setting, comparing to the professional group who did the same error types in both settings.

The results indicate that untrained laypersons not only make more errors but also more critical errors than trained laypersons and professional emergency personnel. Future research should empirically test other tourniquet models than the CAT in the goal of finding use errors to be reduced.

Overall the results are in line with previous studies that show the need for education of bleeding control techniques in the civilian population.

(4)

(5)

v

Acknowledgement

Thank you to my supervisor Erik Prytz for insightful knowledge and experience, helpful advice and manageable solutions during my work to write this thesis. Thank you for always taking your time when I needed. I would also like to thank the Centre for Disaster Medicine and

Traumatology in Linköping and its committed employees for producing and sustaining

interesting projects in the goal of saving more lives and giving me the chance to be a part of this goal.

Molly Lundberg Linköping, 2020

(6)

(7)

vii

List of tables

Table 1: Error taxonomy for the SHERPA, provided by Stanton (2005). ... 19

Table 2: Number of participants in each group. ... 25

Table 3: Possible tasks for the untrained laypersons. ... 59

Table 4: Possible tasks for trained laypersons and professionals. ... 60

Table 5: Total of errors for untrained laypersons, trained laypersons, and professionals in both calm and stressful setting. ... 61

Table 6: Untrained and trained laypersons errors in both stressed and calm conditions. ... 62

Table 7: Untrained laypersons errors... 63

Table 8: Trained laypersons error in calm setting. ... 63

Table 9: Errors of the professional group in the calm setting. ... 64

Table 10: Error of the trained laypersons in the stress condition. ... 64

Table 11: Errors of the professional group in the stressed setting. ... 65

Table 13: Top most occurring errors for all groups and settings, in order. ... 67

List of figures

Figure 1: Picture of a CAT-7 tourniquet. ... 21

Figure 2: Picture of a Delfi-EMT tourniquet. ... 22

Figure 3: Picture of a SAM-X tourniquet. ... 22

Figure 4: Picture of a SWAT-T tourniquet. ... 23

Figure 5: Setup for the untrained layperson experiment. ... 25

Figure 6: Calm setting for trained laypersons and professionals. ... 26

Figure 7: Stressed setting for trained laypersons and professionals. ... 26

(9)

ix

Figure 9: CAT-7 HTA for goal 1... 29

Figure 10: CAT-7 HTA for goal 2. ... 30

Figure 11: CAT-7 HTA for goal 3, part 3.1 and 3.2. ... 32

Figure 12: CAT-7 HTA for goal 3, part 3.3. ... 32

Figure 13: CAT-7 HTA for goal 3, part 3.4. ... 33

Figure 14: CAT-7 HTA for goal 4. ... 33

Figure 15: SAM-X HTA for goal 3, part 3.3. ... 34

Figure 16: Delfi-EMT HTA for goal 2. ... 35

Figure 17: Delfi-EMT HTA for goal 3, part 3.1 and 3.2. ... 36

Figure 18: Delfi-EMT HTA for goal 3, part 3.3. ... 37

Figure 19: Delfi-EMT HTA for goal 3, part 3.4. ... 37

Figure 20: SWAT-T HTA for goal 2. ... 39

Figure 21: SWAT-T HTA for goal 3, part 3.1 and 3.2. ... 41

Figure 22: SWAT-T HTA for goal 3, part 3.3. ... 41

(10)

10

1. Introduction

Between the years of 1998-2004, there were 17 703 people who died before arriving at the hospital in Sweden (Gedeborg et al., 2012). This number of prehospital deaths could have been reduced if bystanders had acted in time (Ashour et al., 2007) and not waited for the emergency personnel to arrive (Scerbo et al., 2017). Mell et al. (2017) report that the time for emergency medical service to arrive at an injury scene can exceed 30 minutes (Mell et al., 2017). For some types of trauma, such as severe bleedings, death can occur within minutes. It is therefore important for laypersons to recognize that their help can make a difference when it comes to life-threatening bleedings (Mell et al., 2017).

There is a campaign called Stop the Bleed, and one of the goals of the campaign is to teach laypeople the skills necessary to stop life-threatening bleedings (Goolsby et al., 2018). The layperson could use direct pressure, wound packing, or tourniquet application for effective haemorrhage control in a civilian prehospital setting (Leonard et al., 2016). A tourniquet is an initial form of bleeding control done by restricting the blood flowing into the injured

extremity (Ave, 1982). The name tourniquet origins from the French word “tourner”, which means to turn (Mabry, 2006). Tourniquets are mainly used in the military field but they are becoming more commonly used in civilian medical emergency care and non-medical personnel in different settings (Kragh et al., 2008). If more laypersons were educated in haemorrhage control the chance for them to act and use a tourniquet in an emergency would increase (Ross et al., 2018).

A problem that stops laypersons from using tourniquets is that health care recommendations aimed to the public are to not block the blood flow with tourniquet like methods for more than 30 minutes, due to the risks with oxygen deprivation in the tissue (Martinez & Gunnarsson, 2019). However, scientific evidence strongly supports the notion that tourniquet application for up to two hours has a low risk of complications (Lee et al., 2007). The dated and opposing recommendations of tourniquet application could theoretically prevent civilians to act in a critical situation, along with the lack of education in haemorrhage control in the layperson populations. The tourniquet application process also includes many steps that could go wrong and untrained laypersons have a high rate of failure of tourniquet application (McCarty et al., 2019).

This report is part of the project at the Swedish Centre for Disaster Medicine and

(11)

11

Control. The project aims to develop the capacity of the general public to perform first aid actions in incidents that cause massive bleeding. The project explores subjects such as training effectiveness and instruction design for bleeding control techniques. The same project will also develop a new type of tourniquet for untrained laypersons. The project will improve the knowledge of the challenges that laypersons face during situations where life-threatening injuries occur and find out what is needed for laypersons to successfully contribute to correct bleeding control (Linköping University, 2017).

The goal of the current study is to create and compare task analyses of the application process of different tourniquet models. Thereafter, the task analyses will be used to identify potential use errors. Goolsby et al. (2015) recommend further studies to focus on instructions,

tourniquet design and brief education for laypersons, which is part of what this study is for. By identifying the errors for existing tourniquets such errors can be avoided when designing a new tourniquet.

1.1. Purpose and research statement

This thesis investigates different tourniquet models – more specifically the CAT-7, Delfi-EMT, SAM-X and SWAT-T – and what a person needs to do to apply a tourniquet to stop a bleeding. The purpose is to find risks and errors in tourniquet use.

The current study will highlight the limitations of tourniquet use. By knowing common use errors of tourniquet application new designs of tourniquets and education of tourniquet application can try to prevent the same errors.

The work has the following research questions:

• What are the potential use errors in the application process of four common tourniquet models?

(12)

12

1.2. Limitations

This work is limited to the investigation of the following versions of tourniquets due to availability:

• CAT-7 • DELFI-EMT • SAM-X • SWAT-T

The error analysis is also limited to the data that has previously been collected from studies using the CAT-7 in populations of untrained laypersons, trained laypersons and professional emergency personnel. However, CAT-7 is one of the most commonly used tourniquet models (Montgomery et al., 2019), which motivates the analysis of this model. The analysed

tourniquet models in the current study are also all listed as recommended from the Committee on Tactical Combat Casualty Care of the Joint Trauma System division of the Defense Health Agency (CoTCCC) (Montgomery et al., 2019). They recommend the following nonpneumatic limb tourniquets:

• Combat Application Tourniquet, generation 6 (CAT-6) • Combat Application Tourniquet, generation 7 (CAT-7) • SOF Tactical Tourniquet – Wide, generation 3 (SOFTT-W) • Tactical Mechanical Tourniquet (TMT)

• Ratcheting Medical Tourniquet – Tactical (RMT-T) / TX2 / TX3 Tourniquets • SAM Extremity Tourniquet (SAM-XT)

The CoTCCC recommend the following pneumatic limb tourniquets: • Emergency and Military Tourniquet (EMT)

(13)

13

2. Background

2.1. Prehospital haemorrhage control

17 703 people died before arriving at the hospital in Sweden during the years of 1998-2004 (Gedeborg et al., 2012). Trauma is the leading cause of death for people younger than 45 (Rhee et al., 2014) and injuries with large haemorrhage is one of the dominant causes of death around the world (Oyeniyi et al., 2017). Studies show that prehospital trauma deaths could have been prevented if bystanders acted in time (Ashour et al., 2007) and the implementation of haemorrhage control techniques and devices in civilian trauma patients can decrease the number of deaths from haemorrhage (Oyeniyi et al., 2017). This indicates that the number of prehospital deaths could be reduced if the prehospital emergency care improved.

A guideline for recognizing life-threatening bleeding is that when the blood loss volume coming from a wound is at half the volume of a soda can (half of 330 ml) and flowing steadily the bleeding is life-threatening and actions should to taken quickly to stop the bleeding

(Goolsby et al., 2018). There are several ways to stop a bleeding and the recommended methods differ depending on the situation and the volume of bleeding. Three of the methods are direct pressure, wound packing and tourniquet application. Wound packing is done by packing the wound with clean cloths and applying direct pressure to the injury (Zwislewski et al., 2019). QuikClot Gauze or other haemostatic dressing are good options for junctional zone haemorrhage (Shina et al., 2015) but tourniquet application should be a method of

haemorrhage control when the bleeding is life-threatening. Both the wound packing and tourniquet application are effective for haemorrhage control in a civilian prehospital setting for blunt and penetrating trauma and medical reasons for bleeding (Leonard et al., 2016). A rapid tourniquet application in less than two minutes could allow laypersons to prevent haemorrhage deaths and even if the tourniquet is just partially effective it can be better than no tourniquet at all (Goolsby et al., 2015). Scerbo et al. (2017) found that patients who had a tourniquet applied before arriving at the trauma centre had a less critical outcome than those who had a tourniquet applied after arriving at the hospital. Waiting to stop a bleeding with a tourniquet until arriving at the hospital can cause great negative consequences and sometimes death for the patient (Scerbo et al., 2017). Laypersons are willing to help in an emergency and just in time instructions increase the chance of successful tourniquet application (Goolsby et al., 2015). However, education is important to increase the chance of laypersons to act helpfully in the scene of an injury and prevent the risk of prehospital deaths (Ashour et al., 2007; Pelinka et al., 2004; Ross et al., 2018). Skills in haemorrhage control are drastically

(14)

14

improved after a short education program (Aekka et al., 2015). Goolsby et al. recognize three important parts in educating people in haemorrhage control: to motivate learners to act in an emergency, to teach learners the difference between life-threatening from non-life-threatening bleeding, and to teach learners to apply pressure (2018). Another method of increase the rate of successful tourniquet application is for the instructions to be improved by implementing more pictures and fewer words (B. R. Lowndes et al., 2017).

2.1.1. Tourniquets

For a long time bleeding control was performed through tight bandages, bands and cords proximal to wounds mainly as an aid in amputation (Mabry, 2006). The first recorded use of a tourniquet in the military was a haemorrhage control device is from 1674 in the French army. The surgeon placed a stick into the bandage on the thigh and twisted until the bleeding stopped (Mabry, 2006). An improvised tourniquet like device can be used to prevent blood loss, but there is a challenge for the user to make the device safe and effective (Kragh et al., 2015), it is, therefore, better to use one of the clinically approved tourniquet models

recommended in the report of Montgomery et al. (2019) to stop a bleeding.

A tourniquet should be placed at least 5 cm proximal to wound and not on a joint directly onto the skin of the limb (Lee et al., 2007). The tourniquet should then be tightened and secured. The pressure needed to occlude blood flow in a leg is significantly larger than for an arm, because of its circumference. The effectiveness is measured by the external haemorrhage, not the pulse, and if the bleeding has not stopped the tourniquet should be tightened or

repositioned before consulting the need for a new tourniquet to be placed proximal to the previous. The time of application should always be documented and transferred to the emergency health care personnel. The tourniquet should stay in place until the patient is prepared for operation and surgical haemorrhage control can be made (Lee et al., 2007). However, there are rare cases when the tourniquet can be removed before the hospital. If the patient transport time is estimated to take longer than 1 hour and the vital signs are stable the removal of the tourniquet could be considered. Before careful release of the tourniquet, the user should have secured wound packing and applied direct pressure to the wound, along with windlass dressing. If the large bleeding returns the tourniquet should be reapplied until the patient is transported to a hospital (Lee et al., 2007). Ideally, the occlusion should not exceed 60 seconds and the securing of the tourniquet and time documentation should not proceed for longer than 90 seconds (Montgomery et al., 2019).

(15)

15

Research shows that tourniquet application can be used for a longer period than previously expected. A study confirmed significant functional recovery of a limb even after 16 hours of tourniquet wear (Kragh et al., 2007). The specific case was an arm and the patient had along with other things skilled medical personnel, methods such as lowered temperature and extensive rehabilitation training. The case can be inspirational for other civilian cases where patients are exposed to entrapment or other scenarios that elongates the transportation time of the patient (Kragh et al., 2007). The more common way of applying a tourniquet and its environment is not as controlled as in the reported case, thus it should not be used as an argument for wearing tourniquets for extensive periods without trying other methods. A limb with a tourniquet applied for longer than two hours risks permanent nerve injury, muscle injury, vascular injury and skin necrosis (Lee et al., 2007). By six hours of prevented blood flow, the muscle damage is almost complete with small chances of salvation (Lee et al., 2007).

The education of tourniquet application differs and should differ, among different populations. People are placed in different groups depending on their background with tourniquet education: laypersons, trained laypersons and professionals (Goolsby et al., 2018). The laypersons are the least likely to use emergency medical material than the other groups and have no medical knowledge. The trained laypersons have greater motivation to learn haemorrhage control techniques than the untrained. These are people who work in high-risk environments. The professionals are experienced medical personnel that are educated in emergency health care (Goolsby et al., 2018). By dividing learners into different populations the teaching can be more specific to increase the success rate of tourniquet application.

However, education could still be improved. Lowndes et al. (2019) suggest that the training of medical personnel could improve and be standardized regarding tourniquet application

because there are inconsistencies of knowledge, attitudes, and practice within singular populations (B. Lowndes et al., 2019).

2.2. Problems with TQ use

There are several problems with tourniquet use. Thierbach et al. (2004) report that many bystanders do not help or provide care in severe trauma, even though they are first at the scene. Bystanders provide care more frequently in less severe injuries than in severe traumas. There are many reasons as to why laypersons will not use a tourniquet. Studies show that civilians need to be more comfortable with using tourniquets without the fear of

(16)

16

Errors of tourniquet application include failure of applying enough pressure within reasonable time restrictions and failure to identify the severity of a situation and the need of tourniquet application (Flecha et al., 2018; Schroll et al., 2019). According to Baruch et al. (2017), the most common error in non-medical users is excess slack in tourniquet strap. Other errors are not enough rods with windlass rod, misunderstanding of buckle mechanism and incorrect placement (Baruch et al., 2017). Bauchwitz et al. (2019) report the following errors during tourniquet application: 1) error when achieving correct location, 2) the number of hand pumps to inflate tourniquet, 3) errors during the whole procedure of tourniquet application, 4) errors while preparing tourniquet, 5) errors of tightening and securing belt, 6) errors while inflating the tourniquet, 7) error for the time of application, 8) latency of preparing tourniquet, 9) latency of positioning belt and 10) latency to secure tourniquet. The above studies from Bauchwitz et al. (2019), Baruch et al. (2017), Flecha et al. (2018) and Schroll et al. (2019) report errors within tourniquet use, but they do not report errors detailed enough to be certain of the exact types of errors that can be made during tourniquet application.

2.3. Hierarchical Task Analysis

Task analysis identifies performance problems and solutions of a whole system, including the operators, tools, and environment (Annett & Stanton, 2006). The goal is to provide data about human functions that could result in determining working aids, training programmes, and the assessment of the performance of the system that is analysed. Data is collected to identify sources of error or knowledge that the operator needs to perform successfully and efficiently (Annett & Stanton, 2006).

Hierarchical Task Analysis (HTA) is one of many methods of cognitive task analysis (Adams et al., 2012). An HTA creates a hierarchy of goals, sub-goals, operations, and plans in a top-down approach. Because of its flexibility, the method becomes a recommended start of further investigations of cognitive tasks (Stanton et al., 2005). The method is popular within the Human Factors field due to its contribution to a greater understanding of cognitive tasks. The process typically includes interviews, domain knowledge, logic and knowledge of training approaches. By viewing tasks from different angles the design can be improved and create training programmes to increase safety, effectiveness and efficiency. The strength of the HTA is the focus on goals rather than actions and the fact that it is hierarchical and functional (Adams et al., 2012).

Doing an HTA starts by defining the task and the purpose of the analysis (Stanton et al., 2005). Next step is to collect data that will inform the HTA and be the knowledge base of the

(17)

17

analysis. The third step is to define the overall goal of the task. When the overall goal is determined it is time to break it down into sub-goals that together conclude the overall goal. Then even the sub-goals needs to be broken down into further sub-goals. The bottom level of this task is always an operation, meaning an action telling what needs to be done. After the sub-goals and operations have been defined plans need to be added. The plans decide how to achieve the goals (Stanton et al., 2005). An HTA can also result in a tabular diagram instead of a tree diagram. It is then called Tabular Task Analysis (TTA)(Stanton et al., 2005). The TTA is directly derived from an HTA, where only the bottom layer actions of the HTA is transported into the tabular columns.

The HTA data can be used for further analysis by investigating it from different perspectives. Crandall et al. (2006) write that the data from a task analysis needs to be treated separately from knowledge elicitation because it encourages different analysis processes, products, and representation formats to be more visible. The disadvantages are the risk of eliciting more general features of the task rather than deeper and more complex goals and situations within the task (Crandall et al., 2006). The result of an HTA can, for example, be summarised in a textbook as an instructions manual for an electric device, or it can be applied in teamwork as a way of determining the functionality or malfunctioning mechanisms of the team in efforts to organize new procedures (Annett & Stanton, 2006). An HTA supposedly seeks to finally propose solutions to possible problems by either training prescriptions or orientation of concepts or strategies. It can also propose guiding material or possible redesigns if helpful training cannot improve the system effectively enough (Annett & Stanton, 2006).

2.4. Human Error Analysis

Today human error is seen as the consequence of failure within a larger system (Stanton et al., 2009). Changing the name from a human error to human factors also changes the perspective of human error. The new perspective is understanding why people do what they do, to

systematically tweak and change the work environment with the goal of forming human activity that functions well (Dekker, 2002). The field of human factors views error with a systematic approach, where human error is seen as a consequence rather than a cause.

The field of human factors reviews system safety. However, safety is not inherent in systems, but rather something that people create (Dekker, 2002). The tools, tasks and environment all have features that result in human error that could be understood by focusing on their

relations. By understanding the cooperation within the system and recognize patterns and mechanisms safety can be increased (Dekker, 2002).

(18)

18

An opposing argument to the human error field is that there is a problem of defining “error” (Dekker, 2002). An error can either be the cause of the failure or the failure itself. An error can also be the process of departing from the norm. Depending on what standard to use, errors can result in different types of conclusions (Dekker, 2002). The problem of determining what is a cause and a consequence is a long-gone discussion, but according to Dekker (2002), the problem is insignificant when an error is stripped from its context and unavailable to deeper analysis due to meaningless categorisation of fragments.

Kirwan (1998) writes that there are many methods of error identification. The most common one is the taxonomic approach, and more specifically the THERP’s error taxonomy. Kirwan also writes about three different components of error: external error mode, performance shaping factors and psychological error mechanisms. According to Kirwan (Kirwan, 1998), there are seven major error types:

• Slips and lapses

• Cognitive errors: diagnostic and decision-making errors • Maintenance errors and latent failures

• Errors of commission • Rule violations • Idiosyncratic errors

• Software programming errors

Kirwan (1994) writes about human reliability assessment and how to analyse human error within a system. Human error identification (HEI) is what follows a task analysis. The types of errors that are considered are error of omission, error of commission, extraneous act and error-recovery opportunities. The errors that might be identified to have great negative consequences on the system alone or together with other errors must be managed within a following risk analysis. It is important to find all possible errors because if a certain error is not initially identified it will not be considered in the risk analysis and the risk might be underestimated. Human error probability (HEP) is a formula as follows: the number of errors occurred divided by the number of opportunities for error. The HEP is a quantification of human error and is considered in the risk analysis (Kirwan, 1994).

(19)

19

Stanton (2005) states that in terms of performance the Systematic Human Error Reduction and Prediction Approach (SHERPA) is the best HEI method in combination with field expert knowledge. SHERPA was originally created to be applied in the nuclear reprocessing industry but is commonly used in various domains that include human activity. It is supposed to

predict potential human or design induced error in a task or scenario. SHERPA is done by firstly collecting data and concluding a task analysis. Then the bottom layers are categorised into taxonomies such as action, retrieval, checking, selection and information communication (see Table 1). After categorising the tasks it is time for the actual human error identification; writing down all possible errors and translate them into external taxonomy error modes. Next step is to assign consequences and recovery potential to the errors. Probability rate is also included in the analysis, using low, medium and high risk of occurrence. Step seven is an error criticality analysis, also using low, medium and high criticality. The last step is a remedy analysis where error reduction solutions are proposed. The solutions are typically categorised into equipment, training, procedures or organisational remedies (Stanton et al., 2005). Table 1: Error taxonomy for the SHERPA, provided by Stanton (2005).

Action errors

A1 Operation too long/short A2 Operation mistimed A3 Operation too little/much A5 Misalign

A6 Right operation on wrong object A7 Wrong operation on right object A8 Operation omitted

A9 Operation incomplete

A10 Wrong operation on wrong object Checking errors

C1 Check omitted C2 Check incomplete

C3 Right check on wrong object C4 Wrong check on right object C5 Check mistimed

C6 Wrong check on wrong object Retrieval errors

R1 Information not obtained R2 Wrong information obtained R3 Information retrieval incomplete Communication errors

(20)

20

I2 Wrong information communicated I3 Information communication incomplete Selection errors

S1 Selection omitted S2 Wrong selection made

SHERPA provides a structured and thorough method of error prediction and provides a detailed analysis of potential errors. The disadvantages of this error identification method are that it is repetitive and overall time consuming to do correctly (Harris et al., 2005). Sharit (2006) also writes that an error analysis might find different errors, but the reasons for one error mode can differ.

(21)

21

3. Method

This report is the result of error analyses of different models of tourniquets. The analyses are Hierarchical Task Analyses (HTAs), Tabular task analyses (TTAs) as well as Systematic Human Error Reduction and Prediction Approach (SHERPA). The current study also reviewed videos of tourniquet usage to count numbers of errors to compare with the error analysis output.

3.1. Error analysis

The error analyses were based on instructional videos of tourniquet application and four different models of physical tourniquets.

3.1.1. Material for error analysis

The material used in the error analysis was already filmed videos of tourniquet usage from two separate studies as well as instructional videos of tourniquet application. Other than films the actual tourniquets themselves have been tested and reviewed for this part of the current study. The tourniquet models that were reviewed are CAT-7, DELFI-EMT, SAM-X and SWAT-T.

The CAT-7 is one of the most commonly used tourniquets and is commercially available. It consists of a constricting band for placement on the limb, a windlass rod for further tightening of the tourniquet, clips to secure the windlass rod and a velcro strap to ensure the windlass rod stays secure and allow to note the time of application (B. R. Lowndes et al., 2017). See Figure 1 for reference.

(22)

22

The Delfi-EMT consists of a wider band, a clamp and pump to fill the band with air that ensures pressure on limbs. See Figure 2 for reference.

Figure 2: Picture of a Delfi-EMT tourniquet.

The SAM-X is a tourniquet similar to the CAT-7. The difference is that the buckle makes a clicking sound when the strap is pulled hard enough and locks in place by the holes in the strap. Other differences are that the windlass rod is made of metal instead of plastic material. See Figure 3 for reference.

(23)

23

The SWAT-T is a lightweight tourniquet that consists of one wide elastic band that while stretching it out is supposed to be wrapped around the limb to stop blood flow. The band has images on it to help the user to know how hard to stretch the band. See Figure 4 for reference.

Figure 4: Picture of a SWAT-T tourniquet.

3.1.2. Task analysis

The current study reviewed videos of tourniquet application as well as empirically tested them to create a Hierarchical Task Analysis (HTA) of the application process of the four tourniquet models. The task of applicating the tourniquets were broken down into smaller parts of goals and plans on how to perform the actions. The HTA created a large tree diagram of goals, plans and actions. After placing only the bottom layer of actions from the HTA in a tabular format it created a tabular task analysis (TTA), with columns and rows for task numbers and task descriptions. One HTA and one TTA were made for each tourniquet model. The four created TTAs could then be further used for the error analysis.

3.1.3. Error analysis

The rows of action processes from the TTAs based the SHERPAs. The error analysis added columns in the tabular diagram for error mode, error description, consequence, recovery, probability, criticality and remedial strategy in addition to task number. As many errors as possible were estimated for each action and placed in the table. All possible errors were then

(24)

24

given an error mode, see Table 1. Some errors were identified by starting with an error mode and placing them on the current task. For example, placing the error mode A3 (Operation too little/much) on the task 3.3.1.2 Pull strap (from the HTA of CAT-7) to predict that the pulling action can be performed both too hard and too loose. The column of remedial strategy in the SHERPA was omitted.

The probability rates and criticality rates in the CAT-7 SHERPA are based on the results of the video analysis in the current study, as well as literature results. The probability rate of the identified errors from the video analysis is based only on the errors of untrained and trained laypersons results (excluding the professional score). When an error was present in less than 20% of trials the probability was rated as “low”. When an error was present in 20-80% of trials the probability rate was rated as “medium” and when an error was present in more than 80% of trials, it was rated as probability “high”. If an error was not found in the trials its probability rate was considered within the literature, the researcher’s expert review, or not estimated at all. The error analysis with probability rates and criticality rates of Delfi-EMT, SAM-XT, and the SWAT-T was based on the researcher’s expert review.

3.2. Video analysis

The data for the video analysis was collected in two different studies. The first study was executed during an exhibit close to Washington DC in the USA were layperson volunteers applied to participate in a study to apply a CAT-7 on a simulated arm. The second study was conducted during a masters’ thesis project where laypersons and emergency professionals (fire rescue, EMS, health care) were asked to apply a CAT-7 during a calm setting and a stressed setting. In the stressed setting, the tourniquet application was conducted before and after several other tasks. In one of those tasks, the participants were exposed to paintball fire. Read more of this study and its method in the report of Marc Friberg (2019).

3.2.1. Participants

The participants for the video analyses were novices that had not used a tourniquet before as well as trained professionals. The novices were divided into two categories: trained and untrained laypersons. Because the laypersons in the second study also participated in a small education programme they were in the current study placed in the trained laypersons' group. There were 17 (4 male, 13 female) untrained laypersons participating in the first study. There were 59 participants (48 male, 11 female) participating in the second experiment. Among these participants, there were 31 professionals (28 male, 3 female) and 28 trained laypersons

(25)

25

(20 male, 8 female). The professionals were 20 fire rescuers and 11 emergency medical service (EMS) professionals. See Table 2 below for a visual of frequency and gender of the participants.

Table 2: Number of participants in each group.

Laypersons Trained laypersons

Professional EMS Total

Total 17 28 31 76

Female 13 8 3 24

Male 4 20 28 52

3.2.2. Setup of the experiments

The videos from the first study show when untrained people were put in a hypothetical scenario where there was an explosion that started a life-threatening bleeding from the arm of another person. They were given an open CAT-7 tourniquet and were asked to apply the tourniquet on a simulated arm to stop the bleeding. The untrained laypersons were not given any instructions on how to apply a tourniquet. See Figure 5 for the setup of this study, where the tourniquet is laying in front of the participant in the background of the image and the simulated arm is in the front of the image, further from the participant. The sides of the table are blocked off by walls.

Figure 5: Setup for the untrained layperson experiment.

The videos from the second study show how laypersons and professional emergency workers apply a tourniquet in a calm and a stressful situation. All participants in this study participated

(26)

26

in both settings. Both groups in this study received a short instruction of what a tourniquet is aimed to do and how to apply it correctly before participating in the experiment. Both the calm setting and the stressed setting included a secondary task to solve a mathematical problem while applying the tourniquet. See Figure 6 for a visual of the calm setting from this study, where the doll has simulated bleeding on its lower leg and the tourniquet is laying closed on top of the thigh. Figure 7 shows the stressed setting for the participants where the simulated wound is in the upper thigh of the doll.

Figure 6: Calm setting for trained laypersons and professionals.

(27)

27

3.2.3. Procedure of video analysis

The procedure of error identification during the video analysis was performed by marking the occurred errors in the SHERPA table for each participant during reviewing the videos, as well as noting if the error was recovered or not. This showed how many times an error occurred and simplified the process of calculating the occured errors for each group. All marked errors were summed and calculated to identify frequencies for all groups as well as within the groups or settings. The identified errors were mapped to the tasks from the HTA to identify which steps in the application process were the most error-prone. The error frequencies for the groups were compared with each other to find differences and similarities between the groups.

3.3. Ethics

………

The current study analysed past study data and did not collect data on its own. The material was not shown to third parties that did not have reason to see the results and videos were not saved for personal use. Names were anonymised to prevent results to be matched with real people.

The study with trained laypersons and professionals was approved by the Regional Ethical Board in Linköping at 2018-08-15, reference number 2018/305-31. All participants gave written consent before participating in the experiment. See more details in the report of Marc Friberg (2019).

(28)

28

4. Results

4.1. Hierarchical task analysis

The top layer of all of the task analyses is the goal of stopping life-threatening bleeding using the current (CAT-7/Delfi-EMT/SAM-X/SWAT-T) tourniquet. The four subgoals are 1) Determine if a tourniquet is necessary, 2) Prepare tourniquet application, 3) Apply tourniquet and 4) Verify bleeding control. Following are the HTA’s for the different models of

tourniquets.

4.1.1. HTA for CAT-7

This first figure (Figure 8) is the top layer of the HTA of the CAT-7 tourniquet.

Figure 8: First layer of the CAT-7 HTA.

For the user to apply the CAT-7 (0.0) the first step is to determine if the tourniquet is needed (1.0). For visual, see Figure 9. The user needs to estimate the bleeding simultaneously as assessing the severity of the situation (1.2). The investigation of the wound can be done both visually (1.1.1) and tactile (1.1.2) depending on how much the bleeding is visible.

(29)

29

Figure 9: CAT-7 HTA for goal 1.

The next step is to prepare the tourniquet (2.0), see Figure 10. To understand the application process (2.1) the user needs to either read the instructions (2.1.1) or recall prior training (2.1.2). Next goal is preparing the application site (2.2). The user needs to remove any clothing (2.2.1) or other items (2.2.2) that are blocking the application site. The goal of preparing the tourniquet itself (2.3) has two subgoals; to open the packaging (2.3.1) and to open the tourniquet (2.3.2). To open the packaging the user will need to tear open the plastic (2.3.1.1), remove plastic packaging (2.3.1.2) and set the instructions to the side (2.3.1.3). To open the tourniquet the user first needs to open the velcro straps (2.3.2.1) and then retract the strap through the buckle mechanism (2.3.2.2) until the loop is large enough to slide on to the extremity or until the strap is completely removed from the buckle mechanism.

(30)

30

(31)

31

The goal of applying the tourniquet (3.0) parted into four steps, see Figure 11. The first step is to inform the patient about the application process (3.1) to calm them down. The second step is to place the tourniquet on the correct spot (3.2). If the tourniquet is in a loop the first task in placing the tourniquet is to slide the tourniquet to at least 5 cm proximal to the wound and not on a joint (3.2.1). If the tourniquet is fully unfolded it needs to be looped around the limb (3.2.2) by wrapping the strap around the limb (3.2.2.1) and inserting the strap through the buckle mechanism (3.2.2.2). If the buckle mechanism is difficult to reach and manage the user should rotate the tourniquet (3.2.3). If the situation is very critical or it is not clear exactly where the bleeding comes from the user should instead slide the tourniquet “high and tight” (3.2.4) to stop the blood flow as quickly as possible and then rotate if needed.

Next subgoal is to tighten the tourniquet (3.3), see Figure 12. This step is divided into two subgoals: tighten the strap (3.3.1) and rotate the windlass rod (3.3.2). To tighten the strap the user needs to hold the tourniquet in a stable position (3.3.1.1), pull the strap (3.3.1.2) and secure the Velcro strap to itself around the tourniquet (3.3.1.3). To tighten the strap further by rotating the windlass rod (3.3.2.1) the user should rotate until the bleeding has stopped and then secure the rod between the clips (3.3.2.2). To secure the tourniquet for patient transport (3.4) the user should do either task 3.4.1 to push the remaining strap between the clips or task 3.4.2 wrap the remaining strap over the windlass rod. Then the user should secure the time strap over the clips (task 3.4.3), ensuring the strap and the rod will stay in place.

The last subgoal is to note the time of application (goal 3.4.4), see Figure 13. The first task is to check the current time (task 3.4.4.1) and then write down the time on the time strap of the tourniquet (3.4.4.2). If there is no way for the user to know the current time or if the user has no pen to write down time with, the user should inform other caretakers of the time of application in other ways (3.4.4.3).

(32)

32

Figure 11: CAT-7 HTA for goal 3, part 3.1 and 3.2.

(33)

33

Figure 13: CAT-7 HTA for goal 3, part 3.4.

The last goal is to verify bleeding control (goal 4.0), see Figure 14. The user should look for further blood flow from the wound (task 4.1) at the same time as checking for pale skin further down on the extremity that the tourniquet is applied to (task 4.2). If the situation is unclear whether the blood flow is stopped or not the user should measure the pulse on the extremity (task 4.3).

Figure 14: CAT-7 HTA for goal 4.

4.1.2. HTA for SAM-X

The only difference in the HTA for the CAT-7 and the SAM-X is in the plan 3.3.1 where the user should hear a clicking sound if the tourniquet strap is pulled hard enough. Otherwise, all other steps of the tourniquet application are the same as with the CAT-7. See Figure 15 for visual of goal 3, part 3.3. of the SAM-X.

(34)

34

Figure 15: SAM-X HTA for goal 3, part 3.3.

4.1.3. HTA for Delfi-EMT

One difference compared to the CAT-7 is that the Delfi-EMT has instructions both in a booklet and on the tourniquet itself. To understand the application process the user can do either one or both task 2.1.1 and task 2.1.2 to read the instructions. Then during the

preparation of the tourniquet (2.3), the user has to open a clamp (2.3.2.1) instead of Velcro straps to open the tourniquet (2.3.2). See Figure 16 for a visual task analysis of goal 2.

(35)

35

(36)

36

During the application process (3.0) the user uses a different method for tightening (3.3). To tighten the strap the clamp needs to be secured (3.3.1.3). To further tighten the tourniquet (3.3.2) the user needs to pump the bulb (3.3.2.1) until strap is filled with air and bleeding has stopped. If no air is filling the band or the process is slow the user needs to do task 3.3.2.2 to close the air ventilation. Because the Delfi-EMT has no Velcro strap to secure the last step of preparing the patient for transport is to note the time (3.4.1), although not writing it down on the band but somewhere else. See Figure 17, 18 and 19 below for goal 3 overview.

(37)

37

Figure 18: Delfi-EMT HTA for goal 3, part 3.3.

Figure 19: Delfi-EMT HTA for goal 3, part 3.4.

4.1.4. HTA for SWAT-T

The SWAT-T needs some carefully used strength to provide enough pressure to the extremity to stop a bleeding. The packaging also comes with a treatment band to place as a bracelet on the patient.

The packaging has full instructions on it, which means that the user needs to read those before opening the packaging (2.1.1). During the opening of the packaging, the user also needs to be careful not to rip up any of the instructions when tearing along one of the indents of the

(38)

38

packaging (2.3.1.1). Because the treatment band is also in the packaging the user not only needs to remove the tourniquet from the packaging (2.3.1.2) but also remove the bracelet (2.3.1.3) and read the instructions on the bracelet. To open the tourniquet (2.3.2) the user will have to unfold the band (2.3.2.1) before also perceiving the visuals on the band (2.3.2.2). See Figure 20 for an overview of goal 2 for the SWAT-T.

(39)

39

(40)

40

When applying the tourniquet (3.0) to the extremity the user needs to stretch out the first part of the tourniquet band (3.2.1) until the circles on the band look like ovals. To place the

tourniquet on the correct spot (3.2.2) the user needs to place that first stretch part of the band 5 cm proximate to wound (3.2.2.1) and holding it to place with one hand while wrapping more of the band around the limb (3.2.2.2). To tighten the band (3.3.1) the user needs to hold the band in position with one hand (3.3.1.1) simultaneously as pulling the band to stretch (3.3.1.2) and wrapping the band around the limb on top of the previous layer (3.3.1.3). All of these three steps should be done for a few laps until the bleeding has stopped. To secure the tourniquet (3.3.2) the user will hold a part of the latest lap of the band in one hand (3.3.2.1) while simultaneously tucking the remaining band under that created loop (3.3.2.2). When the end is tucked the band should be released (3.3.2.3). When preparing the patient for transport (3.4) the user needs to update the bracelet (3.4.2) by checking the time (3.4.2.1), write down the time on the bracelet (3.4.2.2). If available the user should check for the patient’s heart rate and blood type (3.4.2.3) and then document the heart rate and blood type on the bracelet (3.4.2.4). If unavailable the user can skip those two steps. Lastly, the user should apply the bracelet to the arm of the patient (3.4.3). See Figure 21, 22 and 23 for an overview of goal 3.

(41)

41

Figure 21: SWAT-T HTA for goal 3, part 3.1 and 3.2.

(42)

42

Figure 23: SWAT-T HTA for goal 3, part 3.4.

4.2. SHERPA

This result section of the SHERPA analysis will go through the CAT-7 tourniquet in more detail before proceeding with the other models by describing how they differ from the SHERPA output of the CAT-7.

One error is the same for application of all tourniquet models and that is to take too long when applying it. Time is a critical factor when dealing with blood loss and the only recovery is to hurry up. This error could occur during any step of the application process and will not be described further.

Because most steps in the HTA were actions they resulted in mainly action error modes. However, some errors were also given checking errors, retrieving errors, and communications errors. The predicted errors for the CAT-7 were in total 17 checking errors, 14 retrieval errors, 70 action errors, 3 selection errors, 6 communication errors. Many of the actions can be recovered if the user does the same task again, however, to do the task again extends the application time which also is a critical error.

(43)

43

4.2.1. SHERPA output for CAT-7

For the user to apply the CAT-7 the first step is to estimate the bleeding. This can be done both visually and tactile. The visual estimation can be incomplete (error mode C2). This will long term result in making the patient bleed for longer than needed and prevent quick bleeding control. A solution for not finding the wound is to do the tactile search by feeling with the hands over the extremity. If the user can still not find exactly where the bleeding is coming from they can place the tourniquet “high and tight” instead, which means to place the tourniquet as high up on the limb as possible. This solution helps the user by not needing to find the exact location of the wound. Task 1.2 is to assess the situation simultaneously as estimating the wound severity. An error would be to not apply a tourniquet even though the bleeding is large and no other caretakers are nearby (error mode S2). A less critical error would be to apply a tourniquet even though the bleeding is small (also error mode S2). This might harm healthy tissue and be unnecessarily painful to the patient.

Task Error mode

Error description Consequence Recovery

1.1.1 R2 underestimate bleeding patient loses more blood task 1.1.2 R2 overestimate bleeding may damage healthy tissue

unnecessarily

task 1.1.2 C5 take too much time to

estimate bleeding

patient loses more blood task 1.1.2 C2 guess the bleeding is from

wrong limb

will place TQ on wrong limb task 1.1.2 C2 guess the bleeding is on

wrong place

may apply TQ incorrectly task 1.1.2 C2 can't find source of bleeding cannot place TQ task 1.1.2

or task 3.2.4 1.1.2 R2 underestimate bleeding patient loses more blood task 1.1.1 R2 overestimate bleeding may damage healthy tissue

unnecessarily

task 1.1.1 C5 take too much time to

estimate bleeding

patient loses more blood task 1.1.1 C2 guess the bleeding is from

wrong limb

will place TQ on wrong limb task 1.1.1 A7 make the bleeding larger

from being too rough

patient loses more blood C2 guess the bleeding is on

wrong place

may apply TQ incorrectly task 3.2.4 C2 can't find source of bleeding cannot place TQ task 3.2.4 1.2 S2 use TQ when not needed may damage healthy tissue

unnecessarily

task 1.1.1 and 1.1.2 again

(44)

44

S2 not use TQ when needed patient loses more blood task 1.1.1 and 1.1.2 again

The task 2.1.1 is to read the instructions, and the reading can be done incompletely or not at all, which may result in an incorrect application of the tourniquet. The application will also be incorrect if the user recalls wrongly of its previous experience. The solution would be to read the instructions. If there are no instructions available the user will have to guess and try to apply the tourniquet as best as possible, ask for help or try another method for bleeding control.

The tourniquet needs to be placed tightly to the limb, therefore all blocking clothing or other items needs to be removed (2.2.1 and 2.2.2). If none or not enough clothing or items are removed (error mode A1, A8) the tourniquet will not be placed tight enough to stop blood flow. This error could be prevented by making sure no clothing or items are left between the tourniquet and the limb.

When opening the packaging (2.3.1.1) an error would be if the instructions were ripped, resulting in unreadable text. When the instructions need to be read they have to quickly be puzzled back together as best as possible, spending time that could be used for applying the tourniquet. If the user does not remove the tourniquet from the plastic packaging (2.3.1.2) the tourniquet also is not usable, which is why it is a critical error. The instructions should also be removed from the packaging (2.3.1.3) and placed near to prevent them from getting lost and needing to spend time on searching for them.

Task 2.3.2.1 is to open the straps in preparation of applying the tourniquet. If the straps are not opened the application process cannot proceed. It is less critical if the user opened the time strap instead, then the user only needs to open all straps that are visible on the tourniquet. If the tourniquet breaks, the user will have to find a new tourniquet to apply. After opening the straps the user needs to remove the strap from the buckle mechanism (2.3.2.2). If no strap is removed from the buckle the application process cannot proceed because there is no opening in the strap.

2.1.1 R2 interpret wrongly may not stop bleeding after applying TQ incorrectly

R1 not find instructions may not stop bleeding after applying TQ incorrectly

try another method for bleeding control

(45)

45

R3 not read all instructions may not stop bleeding after applying TQ incorrectly

R1 don't read instructions may not stop bleeding after applying TQ incorrectly

2.1.2 R2 recall wrongly may not stop bleeding after applying TQ incorrectly

task 2.1.1 R1 can't remember may not stop bleeding after applying TQ

incorrectly

task 2.1.1 2.2.1 A9 don't remove enough

clothing

TQ may not be effective if clothing is blocking

make sure limb is reachable

A8 don't remove clothing at all

TQ may not be effective if clothing is blocking

2.2.2 A9 don't remove enough items

TQ may not stop blood flow if there are items blocking

A8 don't remove blocking items at all

TQ may not stop blood flow if there are items blocking

2.3.1.1 A3 ripping the instructions manual

if instructions are unreadable user may apply TQ incorrectly

quickly puzzle the parts together if needed

A3 ripping the TQ itself cannot proceed find a new TQ to

apply

2.3.1.2 A8 don't remove plastic cannot proceed remove plastic before proceeding 2.3.1.3 A8 don't put away

instructions

instructions may get lost and not be readable

make sure all parts of the packaging is near

A3 put instructions too far away

patient may lose more blood place instructions close if needed 2.3.2.1 A8 fail to open straps cannot proceed

A3 break TQ cannot proceed find a new TQ to

apply A6 open other straps TQ cannot be used if velcro straps are not

open

open all velcro straps

2.3.2.2 A8 don't remove strap from buckle

cannot proceed if there is no opening in the strap

remove some of the strap from the buckle

Task 3.1 is to calm the patient down by explaining what is going on. If this fails the user may need to explain the process in other ways or omit this step. When placing the tourniquet (3.2.1 and 3.2.4) the user may place the tourniquet wrongly: on a joint, on the wound, too close to the wound or on the wrong limb. The recovery is to slide the tourniquet further from the wound and not on a joint. To rip up the wound further when sliding past the wound is also an error that could be recovered when placing and tightening the tourniquet correctly and

stopping the blood flow. If the placement is too far from the wound it may not be wrong as far at the placement is covered in the “high and tight” method, which is high up on the correct limb. When the strap is placed it needs to be wrapped around the limb (3.2.2.1) and if the

(46)

46

wrapping is omitted or not done properly the tourniquet will not work. If the strap is wrapped around the wrong limb it is also critical, and the remedy is to locate the bleeding again (1.1.1 and 1.1.2) to ensure correct placement. The wrapping of the strap can also be done upside down, which is a critical error that needs to be recovered by turning the strap over to the correct side with the windlass rod mechanism on the outside. Then the strap should be inserted into the buckle mechanism (3.2.2.2) to be able to pull and secure the strap later. If omitting to insert the strap it will not be able to be secured. Another error would be if the user inserted the strap from the wrong way when it is upside down or inserted the strap through the clips instead. Neither of these errors would lead to correct placement and the action should be redone. If the user has difficulties with understanding the buckle mechanism the solution is to read the instruction again (2.1.1).

3.1 I2 fails to calm patient down takes time to apply TQ, patient loses more blood

explain the process differently or omit 3.2.1 A5 puts TQ on joint TQ is not useful slide TQ to a better

place

A5 puts TQ on wound TQ is not useful slide TQ further from

wound

A5 puts TQ too close to wound TQ is not useful slide TQ further from wound

A5 puts TQ too far from wound may damage tissue unnecessarily

slide TQ closer to wound

A6 puts TQ on wrong limb patient loses more blood place TQ on correct limb A7 roughly rips up wound more

when sliding past bleeding

patient loses more blood 3.2.2.1 A9 lays strap on limb TQ won't work, patient loses

more blood

pull strap tighter to limb A8 don't wrap strap around limb TQ won't work, patient loses

more blood

A6 wrap around wrong limb may damage tissue unnecessarily

task 1.1.1 and 1.1.2 A5 wrap strap upside down TQ won't adhere to itself turn over the strap 3.2.2.2 A8 omit to insert strap cannot proceed task 2.1.1

R3 don't understand buckle mechanism

patient loses more blood task 2.1.1 A6 insert strap between clips

instead

TQ will loosen up remove and insert through buckle A5 insert strap from the wrong

way

velcro strap may not work remove and insert the correct way

(47)

47

If the buckle is placed in a difficult position it should be rotated to a more manageable angle (3.2.4). If omitting this step when needed or rotating it too much or too less it will be difficult to tighten and secure the tourniquet.

To pull the strap the user needs to hold the tourniquet in a stable position (3.3.1.1). If the tourniquet is not stable enough it may slip and be placed on the wrong location. Another error would be if the user when holding the tourniquet stable also puts a knee or other weight on the wound. This may cause unnecessary pain to the patient or harm the wound further. The focus should be to place the tourniquet as quickly as possible instead. When pulling the strap (3.3.1.2) an error would be to omit to pull the strap or pull too loosely. This would result in a not tight enough tourniquet that does not stop blood flow. The solution is to pull tighter. If the strap is pulled too hard it may break and the user needs to find another tourniquet. After pulling the strap the user should secure the strap to itself along the tourniquet band (3.3.1.3). If the strap is secured crookedly, not enough or not at all the strap may unwrap and loosen up. If the strap is secured over the buckle mechanism or the windlass rod the tourniquet will not be able to be tightened. If the user secures the strap before inserting the strap through the buckle it will also not be pulled hard enough to stop the blood flow. The user, therefore, needs to make sure that it does task 3.2.2.2 before securing the strap and does it properly.

3.2.4 A5 puts TQ on wound TQ is not useful task 3.2.1 A7 rips up wound even further

when sliding past bleeding

patient loses more blood

A5 puts TQ on joint TQ is not useful task 3.2.1

A5 puts TQ too close to wound TQ is not useful task 3.2.1

A5 puts TQ too far from wound may damage tissue unnecessarily

task 3.2.1 3.3.1.1 A1 not holding enough TQ slips and secures on wrong

place

position one hand at TQ band and pull strap

S2 put weight on wound may cause more damage to

wound

focus on applying TQ 3.3.1.2 A3 pulls too loosely TQ won't be pulled enough to

strangle blood flow

pull strap tighter

A3 pulls too hard unnecessary tissue damage or

TQ break

pull just until strap is tightly applied, if broken find new TQ

A8 don't pull TQ won't be pulled enough to

strangle blood flow

pull strap before proceeding 3.3.1.3 A7 secure strap crookedly TQ may lose grip rearrange A5 secure strap over the buckle

mechanism

strap cover other parts of TQ unwrap to manage TQ properly

(48)

48

A8 omit securing the strap TQ won't be pulled enough to stop blood flow

secure strap before proceeding

A2 secure strap before inserting strap through buckle

TQ won't be pulled enough to stop blood flow

task 3.2.2.2

A9 not securing strap enough TQ might loosen up secure strap before proceeding

To further tighten the tourniquet the user should rotate the windlass rod (3.3.2.1). If omitted or not rotated enough the tourniquet will not be tight enough to stop blood flow. The solution is to rotate the rod further. If the user rotates the windlass rod too much the tourniquet may break or cause unnecessary damage to the patient. The solution is to wound back the windlass rod until the next task is possible or find a new tourniquet to apply if the first one breaks. If the user rotates the windlass rod before securing the strap the user should secure the strap first and if the hands lose their grip of the windlass rod the user should just try again and rotate the rod further. The next step is to secure the windlass rod between the clips (3.3.2.2) to make sure the rod stays in place after tightening the tourniquet. If omitted, not secured enough or secured too soon the tourniquet will loosen up or be too loose to stop blood flow.

After securing the rod the excess strap should also be secured and pushed between the clips (3.4.1). This will make sure the tourniquet will not unwrap during transport. If omitted there is a higher risk of unwrapping the strap. Therefore the user should be careful to do this step before transport. If the strap is pushed between the clips after securing the time strap it may be difficult and the tourniquet may loosen up. The strap could also be inserted between the clips before securing it to itself or placed between clips before securing the rod. Both of these errors would result in a not secured tourniquet or the rod will not have enough room. The strap could also be placed over the clips instead of between, which could make the tourniquet unwrap. The user needs to be careful during this step.

Task 3.4.2 is to wrap the strap over the windlass rod instead of pushing it between the clips. If done incorrectly the tourniquet may unwrap. A solution could be to push the strap between the clips instead because it is less risky. Then the user should secure the time strap over the

windlass rod clips (3.4.3). If omitted the tourniquet can loosen up during transport and not stop the blood flow, which means this is a very critical step. If the time strap is secured before securing the windlass rod the user needs to open the strap before securing the rod. If the strap is secured crookedly there is a risk of the tourniquet loosening up.

Error Identification in Tourniquet Use : Error analysis of tourniquet use in trained and untrained populations