Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Diese Studie geht den kritischen Aspekten des Anziehens nicht isolierter Tauchanzüge nach, die für das Überleben in maritimen Notfällen unverzichtbar sind. Untersucht werden die Faktoren, die Zeit und Korrektheit des Anziehens beeinflussen, einschließlich Alter, Geschlecht, Erfahrung und Unterrichtsmethoden. Die Studie zeigt, dass das Alter die Anziehzeit signifikant beeinflusst, mit einer Zunahme von 7,5% pro 10-jähriger Zunahme des Alters, und dass Videoanweisungen Anziehfehler im Vergleich zu schriftlichen Anweisungen signifikant verringern. Die Studie identifiziert auch zentrale Designprobleme mit dem Tauchanzug, wie Reißverschluss-Ausfälle und Sehstörungen, die sich negativ auf die Anziehleistung auswirken können. Außerdem vergleicht sie die Anzugleistung verschiedener Arten von Tauchanzügen und betont die Notwendigkeit anzugspezifischer Bewertungen. Die Ergebnisse unterstreichen die Bedeutung realistischer Leistungsdaten bei der Verbesserung von Evakuierungsmodellen und Regulierungsstandards, die letztlich darauf abzielen, die Sicherheit der Passagiere in den Polarregionen zu verbessern.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
The International Maritime Organization (IMO) regulates the provision of thermal protection, including Thermal Protective Immersion Suits (TPIS), for passenger vessels operating in Polar regions. These TPIS must be donned within 120 s. While IMO also requires evacuation modelling analysis to demonstrate timely abandonment of passenger vessels, the two regulatory requirements currently remain independent. TPIS usage is excluded from evacuation analysis partly due to insufficient data. This study's uniqueness and importance lies in its provision of a reliable and robust evidence base, quantifying donning times and correctness for a regulatory compliant, non-insulated TPIS, designed for use onboard passenger vessels. Donning time is defined as the duration taken to put on the TPIS, while donning correctness refers to the proper, watertight, and immersion-ready application of the TPIS. Experimental trials with 96 volunteers (67 males, 29 females) aged 18 to 72 revealed donning times from 55 to 186 s, with 39% donning correctly. Age affected donning performance (increase in time and errors with age), while gender and experience did not significantly impact it, which was inconsistent with prior donning studies of different types of TPIS. Surprisingly, watching a donning video did not significantly reduce time compared to written instructions, but it did significantly improve correctness. Clearly, donning times for this regulatory compliant TPIS exceeding 120 s question the appropriateness of suit testing protocols. Finally, donning time distributions for use in agent-based ship evacuation analysis is proposed and TPIS design improvements suggested.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
AG
Age group
DE
Donning error
DP
Donning phase
DT
Donning time
DTnew
DT for participants with new suit
DTused
DT for participants with used suit
E
Experienced
IMO
International Maritime Organization
ISO
International Organization for Standardization
LSA
Life-saving appliances
NDT
Net donning time
NE
Not experienced
NI
No instruction
NSD
Norwegian Centre for Research Data
PP
Preparation phase
PT
Preparation time
PXP
Combined preparation and extraction phase
PXT
Combined preparation and extraction time
SIKT
Norwegian Agency for Shared Services in Education and Research
SM
Supplementary material
SOLAS
International Convention for the Safety of Life at Sea
TDT
Total donning time
te
End trial time
TPIS
Thermal protective immersion suit
ts
Start trial time
txe
End time of the suit extraction process
VI
Video instruction
WI
Written instructions
XP
Extraction phase
XT
Extraction time
1 Introduction
In the field of safety science, understanding human behaviour during emergencies is critical for improving survival outcomes. Dr. Rita Fahy has been a leading figure in this area, particularly in her studies on human behaviour during fire evacuation scenarios (see for example [1‐4]). Her research has greatly contributed to shaping fire safety regulations by collecting and studying human factors data and showing how human factors data can be incorporated into safety standards [2]. Fahy’s work emphasizes the importance of realistic human factor data in emergency situations (see for example [2‐4]), a theme that resonates with the present study.
This paper focuses on human factors in the context of maritime emergencies, specifically on the collection of realistic donning performance data of immersion suits, which are critical for survival in evacuation scenarios in cold environments. Similar to Fahy’s research on evacuation behaviour, this study aims to gather empirical data on how quickly and correctly people can don immersion suits. This data is essential not only for understanding human performance in time-critical evacuation scenarios but also for feeding into evacuation modelling, a key area of Fahy’s work (see for example [5, 6]). Such models are used to simulate emergency scenarios, predict outcomes, and develop strategies to optimise evacuation procedures.
Anzeige
Moreover, this study’s findings have the potential to influence regulatory evacuation standards, much like Fahy’s work has done for fire safety. The primary objective of this research is to provide data that can be integrated into evacuation modelling frameworks, with the aim of informing and potentially supporting future changes in regulatory requirements related to passenger ship evacuations in polar regions.
In the maritime domain, the Polar Code of the International Maritime Organization (IMO) [7] regulates the provision of thermal protection for all passengers onboard vessels travelling in Arctic and Antarctic regions. Appropriate thermal protection enhances passenger survivability in an evacuation scenario. Evacuation of a vessel may be necessary in accident scenarios such as a major fire outbreak onboard the vessel, collision with another vessel or object, or grounding and subsequent capsizing or sinking. According to the Polar Code [7], thermal protection can range from woollen underwear to specialised waterproof and insulated garments, commonly known as immersion suits or survival suits. Key features of an immersion suit may include insulating materials, waterproof construction, provision of buoyancy, visibility, or integrated hood and gloves. In this paper, the term "thermal protective immersion suits” (TPIS) is used to describe these garments, acknowledging their diverse and specialised designs tailored to specific application scenarios. “Thermal protective" highlights the focus on providing protection against the cold, and "immersion suits" specifies the type of garment designed for use in situations that may involve immersion in water.
Allowable donning performances of immersion suits are regulated, amongst others, by the IMO [8] and the International Organization for Standardization (ISO) [9]. Both define a 2-min donning requirement, which is assessed using six human test subjects of varying defined height, weight, and gender. After receiving a demonstration all test subjects must be able to unpack and don the TPIS including a lifejacket if the immersion suit is to be worn in conjunction with a lifejacket. The donning should be performed without assistance within a time frame of under two minutes. Furthermore, while not explicitly stated in the IMO or ISO requirements, it is assumed that the TPIS must be donned correctly by the participants to be considered a pass. However, no description is provided as to the level of correctness required to be considered a pass.
IMO furthermore requires that the evacuation capabilities of passenger ships are assessed during the design phase of the vessel, with the assessment usually undertaken using computer simulation [10]. The purpose of the evacuation analysis is to assess the evacuation performance using prescribed parameters, such as passenger walking speeds, exit flow rates, passenger response times, and measure it against performance standards. As all ship evacuations are time-critical events, the main measurement of evacuation performance stipulated in the IMO standard is evacuation duration, which is currently based on an analysis of fire risk [10, paragraph 8]. Passenger ship evacuation time comprises two components, the time required to assemble the passengers and the time required to abandon the vessel. However, the current IMO guidelines for evacuation analysis do not include any consideration regarding thermal protective clothing, in particular TPIS, for cold environments. The use of TPIS during passenger ship evacuation can impact evacuation performance in several ways, e.g., the time required to correctly put on the TPIS (donning time), a likely reduction in walking speeds while wearing the TPIS, and additional time associated with procedural aspects regarding the distribution of the TPIS.
Anzeige
Advanced evacuation simulation can be a powerful tool to assess the impact of deploying TPIS on the evacuation process [11‐16]. Many evacuation simulation tools represent human behaviour using a stochastic approach and probability distributions [17‐19]. Furthermore, the current IMO guidelines for evacuation analysis have adopted the stochastic approach for the advanced evacuation simulation, which can be seen for example in the representation of the passenger response time distribution [20]. However, the stochastic approach requires a reliable evidence base describing and quantifying the identified behaviour, thus if the donning process is to be included within ship-based evacuation models, a reliable evidence base quantifying donning performance, including donning times, and walking speeds when wearing a TPIS, is essential.
Unfortunately, there are very few studies that assess the impact of deploying TPIS on the evacuation process. Most research to date concerning TPIS has a focus on the thermal protection performance [21‐23]. However, other aspects associated with TPIS that are likely to impact evacuation performance, such as donning time and donning correctness, the likely reduction in individual walking speeds, and procedural aspects associated with the use of TPIS, have received far less attention. Some procedural aspects may positively impact donning performance. For example, for ships on international voyages of more than 24 h, a muster and safety briefing must be conducted before or immediately on departure [24]. Safety demonstrations may improve passengers’ safety knowledge and familiarity of safety critical equipment and procedures. Especially, in vivo demonstrations were found to build more trust to crew members than video instructions, which is believed to be vital to ensure a safe and effective evacuation [25].
The donning time and donning correctness of a TPIS are likely dependent on the complexity of the suit design [26, 27]. Given that the test protocols [8, 9] only require six test subjects, as mentioned earlier, it is unlikely that the donning data produced for certification purposes accurately reflects the expected population performance and is also unreliable and inappropriate for use in the evaluation and planning of emergency procedures [28].
One of the few studies published in the academic literature that assess donning performance is the work performed by Azizpour et al. [26], as part of the ARCEVAC project (see below). Azizpour et al. [26] investigated factors that influence donning time and donning correctness of an immersion suit with fully integrated buoyancy and thermal insulation. The TPIS was supplied in a carry bag with a placard displaying donning instructions and satisfied the IMO requirements. In total 108 male and female volunteers between 18 and 72 years of age participated in the donning trials. A Net Donning Time (NDT) was defined as the time required to open the carry bag, extract the TPIS and don the suit. In addition, a Preparation Time (PT) was defined in which participants were anticipated to read the donning instructions provided on the carry bag. Some participants (N = 19) were shown a donning instruction video prior to the donning process. At the end of the donning process, the donning correctness of each participant was recorded. The influence of age, gender, experience, and method of instruction on NDT and number of donning errors was analysed and discussed.
Azizpour et al. [26] reported that 89% of participants had NDTs exceeding the 2-min regulatory requirement. Also Mallam et al. [27], who collected donning times of two suits of similar design and both similar to the suit used in the study by Azizpour et al. [26], reported 26% of donning times exceeding the IMO/ISO 2-min criteria. Furthermore, in the Azizpour et al. [26] study, 96% of the participants had at least one donning error and 75% had two, while in the Mallam et al. study 56.3% incorrectly donned the suit. These findings support the suggestion that donning data using the current certification protocol does not reflect the expected population performance and is unreliable and inappropriate for use in the evaluation and planning of emergency procedures. However, how generalisable are the research findings by Azizpour et al. [26] and Mallam et al. [27]? In particular, are the trends in donning times and donning error rates reported by Azizpour et al. [26] for an immersion suit with fully integrated buoyancy and thermal insulation applicable to different types of TPIS, such as lightweight survival suits [28, 29]?
To address this lack of data and amass an evidence base that can be used to assess the impact of TPIS on evacuation performance in cold environments, Western Norway University of Applied Sciences (HVL) and UiT The Arctic University of Applied Sciences embarked on the ARCEVAC (ARCtic EVACuation) project. As part of the ARCEVAC project, two different types of TPIS were used in a series of experiments to quantify donning times and the factors that influenced donning times and to assess their impact on walking speeds and stair ascent/descent speeds. The two TPIS differed substantially, one was a lightweight survival suit produced by Hansen Protection (Sea Pass passenger suit) [28] and the other was an immersion suit with fully integrated buoyancy and thermal insulation produced by Viking (Yousafe Blizzard PS5002) [26]. The impact of TPIS on walking speeds of individuals along a corridor at different angles of heel was recently reported in Azizpour et al. [30] while the impact of the TPIS on walking speeds on stairs is currently under investigation [31]. The aim of this paper is to build a reliable evidence base quantifying the donning performance of a specific type of TPIS (exemplified by the Hansen Protection TPIS), by typical passengers, with a focus on identifying key factors that influence donning time and correctness. The key research questions which are examined by corresponding hypothesis tests for each of the individual factors include:
1)
Is there a significant difference in the number of donning errors based on the factors age, gender, experience and/or instruction method?
2)
Is there a significant difference in the donning time based on the factors age, gender, experience and/or instruction method?
In addition, the study provides quantified donning time data for this type of suit, which can be applied in regulatory contexts, particularly in advanced computer simulations for passenger ship evacuation analysis.
2 Experimental Methodology
The experimental methodology employed in the collection of the TPIS donning data has been previously described and discussed in detail [28]. Here a brief outline of the methodology used in the study is provided.
2.1 Brief Description of Trial Participants, Location and the TPIS
The experimental programme made use of volunteer human test subjects and so followed appropriate ethical standards and protocols. The approach used in collecting and handling personal data was approved by the Norwegian Centre for Research Data (NSD, now called SIKT– Norwegian Agency for Shared Services in Education and Research). All appropriate measures were taken to ensure the safety and anonymity of participants (see [28] for details). In total 96 volunteers (67 males and 29 females) from the local communities were recruited through the local media, social media and word of mouth, with an emphasis on enrolling naïve test subjects. The test subjects were non-disabled, healthy participants aged between 18 and 72 years. Of the 96 volunteers, 91 participants were included in the donning time and donning error analysis (see sect. 2.6 and the Online Resource Supplementary Material (SM) sect. S3 for details). The trials took place at three shore-based facilities in Norway, the ARCOS safety centre in Tromsø, the Faculty of Science at UiT The Arctic University of Norway in Tromsø, and the ResQ safety centre in Haugesund (see the Online Resource SM sect. S1). The donning took place within a large room, with a smooth flat floor. Prior to commencing the trials, participants were requested to complete a pre-trial questionnaire (providing demographic information) and consent form, and were given a safety briefing.
The TPIS used in the trials was a lightweight, non-insulated immersion suit without integrated buoyancy. The TPIS is available in two sizes, adult and child, but only adult-sized suits were used in this study. The suit is complete with integrated hood, gloves, and foot coverings (see Fig. 1). It is designed for use onboard passenger vessels or recreational vessels, is approved as an uninsulated immersion suit in accordance with SOLAS’ LSA Code Med/1.5a [32] to be used with additional approved lifejacket. The suit has thus been tested to meet the 2-min donning requirement including the donning of an approved lifejacket. Note however, that the TPIS was donned without a lifejacket during the trials. According to the manufacturer [29], the TPIS is also compliant with the requirement of a thermal protective aid according to the IMO Polar Code [7], which is part of the survival equipment to be stored aboard life raft and -boats. General information and donning instructions are printed on the vacuum-packaging, but not on the suit itself. Furthermore, due to a limited number of new suits sealed in their vacuum packaging available for the trials (N = 24), unpacked used suits were re-used by the majority of the participants (N = 72). An essential component of the TPIS is the zipper located at the front of the suit that is used to close the suit once it is donned. During the trials, zippers on 10 suits broke, resulting in a reduction in the data available for donning time and donning correctness analysis (see sect. 2.6 and Table 1 for details).
Fig. 1
Picture of a unpacked immersion suit and b vacuum packed suit
A suit that has been removed from the vacuum packaging
68
19
61
49
Method of instruction
No instructions (NI)
Participants that were not provided with WI or VI
68
19
61
49
Written instruction (WI)
Participants that had access only to the written instructions printed on the sealed vacuum packaging (whether they read the instruction or not)
11
1
11
10
Video instruction (VI)
Participants that had access to a video demonstration of the correct procedure for donning the TPIS (whether they paid attention to the video or not). Note, those that had access to VI also had access to WI
12
1
11
11
Prior donning experience
No experience (NE)
Have never attempted to don a TPIS before, regardless of the type of TPIS. Based on response to post-trial questionnaire and previous participation in trials, if applicable
68
13
63
55
Experienced (E)
Participants that have donned a TPIS previously, regardless of when, how often or type of TPIS. Based on response to post-trial questionnaire and previous participation in trials, if applicable
23
8
20
15
Gender
Male
Participants identifying as “male”
62
19
54
43
Female
Participants identifying as “female”
29
2
29
27
Age (AG = Age Group)
AG1
18–29 years of age male/female
36
13
18/12*
12/11*
AG2
30–50 years of age male/female
33
4
18/13*
16/13*
AG3
51–72 years of age male/female
22
4
18/4*
15/3*
+10 zippers broke during the donning, reducing the number of data points that could be used in the analysis of the donning time and donning correctness (see sect. 2.6)
*Numbers in cells are shown male/female
2.2 Method of Participant Instruction
As part of the trial design, three methods of instruction were investigated: written instruction (WI), video instruction (VI), and no instruction (NI). The WI (as delivered by the manufacturer) described the donning procedure in eight steps, with each instruction expressed using a simple short sentence written in four languages (Norwegian, English, German, and French) and accompanied with an illustration (see Fig. 1). The VI consisted of a short instructional video demonstrating how to don the suit correctly and was prepared by the research team. The video was 2.3 min long and the soundtrack was in Norwegian, the first language of the participants that were exposed to the VI. Participants who experienced the VI also had access to the WI. As mentioned in sect. 2.1, donning instructions are printed on the vacuum-packaging, but not on the suit itself and thus only participants with new suits could be assigned to the WI or VI group. Thus, these groups consisted of a total of 24 participants. Prior to the start of the trials, a group of 12 participants were randomly selected to receive the VI and another group of 12 participants were randomly selected to receive the WI. As suggested by the title, NI meant that no instructions, neither VI nor WI were provided. Participants in the NI group donned previously used suits, and so the packaging and hence instructions were not available. A total of 72 participants were in the NI group.
2.3 Brief Description of Trial Protocol
Groups of up to 15 participants would attempt to don the suit at the same time. The amount of floor space available to each participant was minimum 0.6 m2 as required by the IMO regulations [33]. A TPIS was set at the feet of each participant, either an unused TPIS in the vacuum packaging or a folded used suit. The participants in the VI group were shown the instructional video just prior to positioning them behind their assigned suit. The trial manager would then set the scene by instructing the participants to imagine that they were at sea on board a passenger ship sailing in polar waters and the evacuation alarm had just been sounded. The participants were told that they had to don the suit as quickly and as correctly as possible so that they would be ready to safely evacuate the vessel. The trial would commence once the manager gave the “GO” command and the end point was defined as the time that the participant raised their arms above their head. On completing the trial, participants were required to complete the post-trial questionnaire that explored the participants’ opinions about their donning experience, and asked for suggestions and feedback to improve the donning task (see the Online Resource SM sect. S5 and/or [28] for the complete questionnaire).
The participants’ donning performance was recorded throughout using two GoPro Hero cameras (frame rate of 25 FPS). A range of quantitative and qualitative data was collected during the trials through video footage and post-trial questionnaires (see the Online Resource SM sect. S5 or [28] for details). Quantitative data concerning donning correctness and speed of donning was collected through analysis of the video footage using Adobe Premier Pro v15.0. Statistical analysis of the data was performed using IBM SPSS® Statistics version 28.0 using a significance level of 0.05 in all statistical tests.
2.4 Definition of Trial Control Variables and Key Events
Based on the experimental set-up, five control variables were defined: suit state, method of instruction, participant prior experience of donning a TPIS, gender, and age. Details of the definition of each control variable and the number of participants associated with each are presented in Table 1. Note that age is specified in three groups as defined in the current IMO guidelines for evacuation analysis [10].
The various phases and key events that occur during the donning process are schematically described in the time-line diagram presented in Fig. 2. The four key events that define the donning process are, event ‘a’, the start signal for the trial, event ‘b’, the start of the extraction phase, event ‘c’, the start of the donning phase and finally, event ‘d’, the end of the donning phase. As noted in Fig. 2, participants donning a new suit and used suit involve a different combination of key events as the used suits are not sealed within the vacuum packaging. The donning phases are the time intervals between the key events, and for donning a new suit this typically involves the preparation phase (PP), extraction phase (XP) and donning phase (DP), while donning a used suit only involves the DP. During the PP the participant may undertake a number of activities such as pick up the TPIS package, study the instructions on the package, etc. The XP is the period during which the TPIS is removed from the packaging and is ready to be donned. During the XP participants do not stop to study the instructions or undertake any other task. The PP and the XP is only applicable for participants with a new suit. Some participants began to open the packaging while still engaged in preparation activities. For these participants it is not possible to define a distinct PP and XP. However, a combined preparation and extraction phase (PXP) can be defined for all participants with new suits (see method 2 in Fig. 2 and Table 2).
Depiction of key events on timeline for donning of TPIS
The duration of each phase represents a key time interval during the donning process and are referred to as, preparation time (PT), extraction time (XT), combined preparation and extraction time (PXT) and donning time (DT).
To determine the key time intervals defining the donning process, the timing of the key events is derived from the video analysis. The process by which the timings of these events were reliably and consistently extracted from video footage relied on the specification of a data dictionary that precisely defined these key events. The precise definition of key timed events and parameters as defined in the data dictionary can be found in the Online Resource SM sect. S2. Here, the relevant key events and time intervals are described, and a more comprehensive description can be found in the Online Resource SM sect. S2.
The DT for a participant with a used suit (DTused) is defined as the time interval from the trial start time to their trial end time, i.e. from event ‘a’ to ‘d’ in Fig. 2 and Table 2, and is given by,
$${DT}_{used}= {t}_{e}-{t}_{s}$$
(1)
where ts is the start trial time and te is the end trial time.
The DT for a participant with a new suit (DTnew) is defined as the time interval from the end of their extraction phase to their trial end time, i.e. from event ‘c’ to ‘d’ in Fig. 2 and Table 2, and is given by,
$${DT}_{new}={t}_{e}-{t}_{xe}$$
(2)
where txe is the end time of the extraction process and te is the end trial time.
The DT is a measure of the time required to don the TPIS, and so is an inherent measure of the ease or difficulty associated with putting on the suit. This is a useful measure as it allows for the comparison of one TPIS design with another. Furthermore, the XT is a measure of the ease of opening the packaging and so enables the comparison of various packaging designs, while the PT is a measure of the time participants spend in preparing to don the TPIS, which can include studying the instructions.
For regulatory applications, the Total Donning Time (TDT) is required. The TDT is a measure of the time required for the entire process of preparation, extraction, and donning. The TDT is defined as,
$$TDT=PXT+ DT$$
(3)
The DT is given by Eqs. 1 and 2 for used and new suits, respectively. The duration of the combined preparation and extraction phase (PXT) is defined as the time interval from the start of the trial to the end of the extraction phase, i.e. from event ‘a’ to ‘c’ in Fig. 2 and Table 2, (i.e. PXT = PT + XT) and can be derived for the 23 participants donning a new TPIS (see Table 1 and sect. 2.6). However, a PXT could not be measured for participants donning a used TPIS, and so to determine the TDT for a used TPIS an approximation for PXT must be used. Depending on the nature of the intended application, PXT can be represented by the mean of the measured PXT distribution for the 23 new suits, or the maximum PXT or values can be randomly selected from the PXT distribution.
2.5 Definition of Donning Error
In addition to the timed events associated with donning the TPIS, behavioural data, such as donning errors, was quantified and recorded in the form of binary and nominal variables (see [28]). The number of donning errors incurred by a participant provides a Donning Error (DE) score. It is important to note, that DE measures the final state of key suit features, i.e., it does not measure if the participant incorrectly executes a donning task but corrects it before completing the donning process. An overview of key features and the correct final state is shown in Table 3. These are based on the features identified by the TPIS manufacturer as indicated in the instructions panel on the TPIS packaging (see Fig. 1). The scale for DE varies from a minimum of zero (no donning errors) to a maximum of five, indicating that the participant incurred all the donning errors.
Table 3
List of key donning features and their correct final state
2.6 Zipper Malfunction and Data Validation
A total of 25 new TPIS in vacuum sealed packaging were available at the start of the experiment, of which 23 resulted in valid data for DT and DE analysis. Another 72 participants donned a used TPIS, of which 68 resulted in valid data for DT and DE analysis (see the Online Resource SM sect. S3 for details).
During the trials, it became apparent that the zipper was of poor quality and that a substantial number of zippers malfunctioned and broke. However, given that the number of TPIS were limited, it was necessary to make use of the data from the TPIS with broken zippers. Thus, in order to not lose the data points, it was judged appropriate to include participants with a broken zipper in the DT analysis, that had performed all other donning tasks prior the attempt to close the zipper, and to add an appropriate correction factor to represent the time required to close the zip (see the Online Resource SM sect. S3 for details). In total 10 zippers (representing 40% of the available TPIS) broke during the donning process and a further 11 participants were required to don a TPIS with a broken zipper. Of the 21 participants with broken zippers a total of 13 (62%) were included in the DT analysis (see Table 1). All participants with broken zippers were excluded from the analysis of donning correctness as the zipper end-status for these participants could not be determined.
3 Descriptive Results
The main results from the data collections consists of data extracted from the video analysis supported by data extracted from trial questionnaires (see the Online Resource SM sect. S5 for details concerning the questionnaires).
A dataset comprising of 83 data points is available for the DT analysis while a dataset comprising of 70 data points is available for the DE analysis. Data from the participants excluded from the DT and DE analysis is included in the analysis of the post-trial questionnaires, where appropriate.
3.1 Preparation and Extraction Time
The results presented in this section relate only to those participants with a new suit and no prior experience (NE-group). Participants with prior experience were excluded for two reasons: 1) it is believed that experience may impact the PT and PXT, and 2) since only two participants who donned a new TPIS had prior experience, the sample size was too small to assess the impact of experience on PT and PXT meaningfully. The PXT were identified for participants with written instructions (NE + WI group, N = 10) and video instructions (NE + VI group, N = 11) (see Table 4). Four of the total 21 participants extracted the TPIS while undertaking preparation tasks and so a distinct PT and XT could not be defined, resulting in 17 data points for PT and XT (see sect. 2.4, definition of PP and XP).
Table 4
Arithmetic mean, minimum, and maximum PT, XT, and PXT for participants with NE given their method of instruction
Preparation time (PT)
Method 1
Extraction time (XT)
Method 1
Preparation and extraction time (PXT)
Method 1 and Method 2
N
Mean (s)
Std Dev (s)
Min/Max (s)
N
Mean (s)
Std Dev (s)
Min/Max (s)
N
Mean (s)
Std Dev (s)
Min/Max (s)
NE + WI
7
8.8
8.0
2.0/20.3
7
8.5
3.2
4.8/13.1
10
17.4
6.6
6.9/28.0
NE + VI
10
2.2
0.7
1.3/3.6
10
5.6
2.4
2.8/11.7
11
8.3
3.1
5.5/14.4
The distribution of PT, XT, and PXT for the NE + VI and NE + WI groups are presented in Fig. 3. As can be seen from the distribution of PT, some participants in the NE + WI group display a longer PT than the participants in the NE + VI group. This is also reflected in the mean PT for the seven participants exposed to only WI, which is 8.8 s, while the 10 participants exposed to VI have a mean PT of 2.2 s (see Table 4).
Fig. 3
Distribution of a PT for NE + WI (N = 7) and NE + VI (N = 10) groups, b XT for NE + WI (N = 7) and NE + VI (N = 10) groups, and c PXT for NE + WI (N = 10) and NE + VI (N = 11) groups
Donning trial video footage suggests that only three of the seven participants in the WI group with a detectable PT were observed to look at the instructions during the PP while none of the participants in the VI group with a detectable PT were observed to look at the written instructions. The three participants that were observed to look at the instructions displayed a longer PT (between 13 and 21 s) than the other four participants that were not observed to look at the instructions (between 2 and 5 s). These observations are not surprising as participants that were exposed to the VI just prior to the commencement of the donning trials are unlikely to feel the need to study the donning instructions prior to attempting to don the TPIS, whereas those with only WI are likely to take time to study the instructions. Thus, the PT distribution for the WI group effectively consists of two distributions. One distribution representing those that appear to ignore the instructions, consisting of four participants resulting in very short PTs (between 2 and 5 s), and another representing those that appear to gaze at the instructions, consisting of three participants producing longer PTs (between 13 and 21 s).
In contrast to PT, the mean XT for participants with NE + WI and NE + VI are more similar, being 8.5 s and 5.6 s, respectively. Given that participants in the NE + WI group produced a longer PT than the NE + VI group, it follows naturally that the mean for the combined PXT is larger for the NE + WI group (17.4 s) than for the NE + VI group (8.3 s).
3.2 Donning Correctness
The number of donning errors (DEs) incurred by each participant is a measure of the donning correctness. A DE was identified for incorrect final states of shoes, hood, zipper, gusset strap, and ankle straps as defined in Table 3.
As shown in Table 5, the 70 participants incurred a total of 71 DEs, resulting in an average of 1.0 DE per person. The 43 males produced 43 DEs resulting in an average of 1.0 DE per male participant while the 27 females produced 28 DEs also resulting in an average of 1.0 DE per female participant. Thus, taken across all the age groups, experience and instruction categories, males and females produce the same number of DEs on average. The distribution of the number of DEs for male and female participants are also very similar and are presented in Fig. 4.
Table 5
Descriptive statistics for DEs according to different methods of instruction and experience category
Method of instruction experience (NE = No experience, E = Experienced)
All
No instructions (NI)
Written instructions (WI)
Video instructions (VI)
NE
E
NE
E
NE
E
(Total number of errors)/(total number of participants)
47/36
10/13
10/9
2/1
2/10
0/1
71/70
Average number of errors per person
1.3
0.8
1.1
2
0.2
0
1
Mode of errors per person/Frequency of mode
1/12
1/6
0,1/3
2/1
0/8
0/1
0/27
Min–Max number of errors per person
0–4
0–2
0–3
2–2
0–1
0–0
0–4
Fig. 4
Percentage of male (N = 43) and female (N = 27) participants incurring a given number of DEs across all experience and instruction categories
Experience tended to reduce the average number of DEs, from an average of 1.1 (59 out of N = 55) for the NE group to an average of 0.8 (12/15) for the E group, however, it is noted that the WI and VI groups only had one experienced participant each. For the NE group (N = 55), the method of instruction impacted the average number of DEs, with the NI (N = 36), WI (N = 9) and VI (N = 10) groups incurring an average of 1.3, 1.1 and 0.2 DEs, respectively.
Approximately 61% of all participants made at least one DE. However, this means that 39% of the participants made no DEs. Indeed, from Fig. 4 the most frequent observed outcome (i.e., the mode) was that a participant incurred zero DEs. Furthermore, the method of instruction impacted the proportion of the participants incurring at least one DE, with NI, WI and VI groups resulting in 69%, 70% and 18% of the group population respectively.
Presented in Fig. 5 is a pie chart showing the frequency of each type of DE across all experience and instruction categories. The least number of DEs was associated with the hood, representing only 3% (2) of the total number of errors, and– the same value– 3% (2) of the participants made this type of error (Note: The percentages, after rounding, are equal since the total number of participants, N = 70, and the total number of errors, N = 71, are so similar). The next highest number of DEs, representing 10% (7) of the total number of DEs (and 10% (7) of the population), involved the shoes. The shoes could be worn within the TPIS or externally, once the TPIS was donned. This DE involved the participants not wearing shoes once the donning was complete. The next highest number of DEs, representing 24% (17) of the total number of DEs, involved the gusset. This typically meant that the gusset was not closed. The neck cord at the top of the gusset must be tightened to seal the gusset and the gusset tucked into the suit. A total of 24% (17) of the population made an error associated with the gusset. Approximately 33% (14) of males and 11% of females (3) struggled with this task.
Fig. 5
Frequency of each type of DE taken across all experience and instruction categories (N = 71)
The second most common DE representing 25% (18) of the total DEs (and 26% (18) of the population) was associated with the zipper. When the zipper is correctly pulled up over the chin and the gusset correctly closed, a waterproof seal is created. However, participants struggled with pulling the zipper above the neck and chin. Approximately 26% (11) of males and 26% of females (7) struggled with this task.
The most frequent DE, representing 38% (27) of the total DEs (and 39% (27) of the population), was failure to close the ankle straps. The ankle straps are the only means, incorporated within the suit, to adjust its fit. Given the universal size (one size fits all) of the TPIS, the ankle straps are vital to ensure that the socks of the TPIS remain in place, especially if the shoes are worn inside the TPIS. Furthermore, the ankle straps prevent excess suit material from accumulating around the feet during movement. In total 33% (14) of men and 48% (13) of women failed to fasten the ankle straps.
3.3 Donning Time
Descriptive statistics for DT according to age group, gender, experience, and method of instruction are presented in Table 6. Across all categories, the DT for males (N = 54) varied from 55.2 to 186.3 s, with an overall mean of 104.9 s, while for females (N = 29) the DT varied from 60.9 to 183.6 s with an overall mean of 106.4 s. Taken across all categories, this suggests that on average, males and females were equally fast in donning the TPIS. The DT distributions for males and females, respectively, are presented in Fig. 6.
Table 6
Arithmetic mean, minimum, and maximum DT for male and female participants in AG1, AG2, AG3, given their level of experience and method of instruction
Method of Instruction*
Experience**
Mean, Min–Max DT and (Number of participants)
AG 1
18–29
Years of age
AG 2
30–50
Years of age
AG 3
51–72
Years of age
Overall
Male
Female
Male
Female
Male
Female
NI
NE
94.2
55.2–140.3
(8)
73.2
60.9–94.3
(4)
100.5
59.7–132.9
(8)
102.3
79.6–125.9
(7)
121.0
77.8–186.3
(14)
97.7
76.2–119.3
(2)
103.6
55.2–186.3
(43)
E
93.9
75.9–123.5
(6)
116.0
88.8–151.8
(4)
74.4
65.2–83.6
(2)
66.1
66.1–66.1
(1)
110.5
80.0–133.4
(4)
100.7
100.7–100.7
(1)
99.1
65.2–151.8
(18)
WI
NE
79.4
79.4–79.4
(1)
114.2
91.8–131.6
(3)
125.8
104.3–151.3
(3)
160.5
115.0–183.6
(3)
N/A
N/A
128.1
79.4–183.6
(10)
E
N/A
N/A
86.3
86.3–86.3
(1)
N/A
N/A
N/A
86.3
86.3–86.3
(1)
VI
NE
94.6
78.6–103.8
(3)
95.8
95.8–95.8
(1)
112.1
84.5–143.5
(3)
117.8
113.1–122.4
(2)
N/A
96.2
96.2–96.2
(1)
104.7
78.6–143.5
(10)
E
N/A
N/A
97.3
97.3–97.3
(1)
N/A
N/A
N/A
97.3
97.3–97.3
(1)
Total number
93.3
55.2–140.3
(18)
99.6
60.9–151.8
(12)
102.8
59.7–151.3
(18)
115.3
66.1–183.6
(13)
118.6
77.8–186.3
(18)
98.1
76.2–119.3
(4)
105.4
55.2–186.3
(83)
*NI No instruction, WI Written instruction, VI Video instruction
**NE No experience, E Experience
Fig. 6
Distribution of DT for males (N = 54) and females (N = 29) taken across all experience and instruction groups
By visual inspection, the DT distributions for males and females appear to follow either a normal or log-normal distribution. However, it could be argued that the characteristics of a log-normal distribution better represent the nature of the DT variable by, for example not allowing negative DT-values, and at the same time having non-negligible likelihood for observing larger DT values. To determine which provides a better fit to the data, Shapiro–Wilk tests were performed on the DT data and on log-transformed DT data. The results suggest that DT for male and female is not normally distributed (Male: W(54) = 0.955, p = 0.041; Female: W(29) = 0.924, p = 0.038), but fits a log-normal distribution (Male: W(54) = 0.986, p = 0.78, Female: W(29) = 0.971, p = 0.59). The DT distributions for male and female can be described by Eq. (4) and are depicted in Fig. 6.
where, PDFDT = probability density function for donning time, μ = location parameter of the log-normal distribution, Male: 4.619, Female: 4.630, σ = scale parameter of the log-normal distribution, Male: 0.264, Female: 0.278.
Across all categories, the overall mean DTs for AG1 (18–29 years, N = 30), AG2 (30–50 years, N = 31), and AG3 (51–72 years, N = 22), were found to be 96 s, 108 s, and 115 s, respectively. This suggests that age may impact DT, which is examined further in sects. 4.2 and 5.2.2. Similarly, to the DT distributions for males and females, the DT distribution for AG1, AG2, and AG3 can be described by a log-normal distribution (see the Online Resource SM sect. S4 for details) using Eq. 7, where:
Finally, for participants without previous donning experience, taken across all age and gender categories, the three methods of instruction, NI, WI, and VI produced average DTs of 104 s, 128 s, and 105 s respectively. These results suggest that the method of instruction may have an influence on the DT of those without prior experience, which is examined further in sects. 4.3 and 5.4.
3.4 Post-trial Questionnaires
A total of 95 participants completed the post-trial questionnaires, representing a 99% completion rate. The questionnaire consisted of nine questions (see the Online Resource SM sect. S5 or [28] for details) with questions 3–7 being closed questions asking participants to rate various aspects of their donning experience while questions 8 and 9 were open questions asking participants to suggest how to improve TPIS design to ease the donning process and general additional feedback. Detailed responses to these questions can be found in the Online Resource (see SM sect. S5) and in the discussion section (see sect. 5), however, it is informative to note the following responses, which are based on responses taken over all the instruction, experience, age, and gender categories. Percentages reported in the bullet points refer to the number of valid responses per question; as some participants skipped individual items, resulting in varying sample sizes, missing responses were treated as missing completely at random and excluded on a per-question basis:
Question 3 (SM sect. S5.2) asked participants to rate how easy the donning process was. In total, 49.5% (47) of the respondents thought it was easy/very easy, while 15.8% (15) thought it was difficult/very difficult to don the TPIS.
Question 4 (SM sect. S5.3) asked if a visual (live or recorded) demonstration while donning would have made the donning process easier. In total, 80% (75) said yes while 19% (18) said no. When asked if physical assistance would have made the donning process easier, 30% (26) said yes while 56% (49) said no.
Question 5 (SM sect. S5.4) asked participants if they believed that they had donned the TPIS correctly. In total, 71% (50) said yes while 21% (15) said no.
Question 6 (SM sect. S5.5) asked if participants thought that attempting to don the TPIS in conditions of rough seas would impact the time required to don the TPIS. In total, 34% (33) thought it would significantly increase donning time, 55% (53) thought it would slightly increase donning times while only 5% (5) believed it would have no impact.
Question 7 (SM sect. S5.6) asked if participants thought that wearing the TPIS would impact their walking speeds (in corridors or ascending/descending stairs). In total, taken across all types of walking terrain, 53% (147) thought it would impact their walking speed while 38% (105) said it would not.
4 Statistical Analysis
The main results concerning key trial parameters relating to PT, XT, DT, and DE extracted from the experimental data were presented in sect. 2.4. In this section the interrelationships between these parameters and their dependence on population demographics is explored.
The statistical analysis focuses on a plausible worst-case scenario, referring to participants that have no prior donning experience and no access to donning instructions. The rationale behind the plausible worst-case scenario is based on observations from the trials involving participants with new suits who were provided only with written instructions. In these trials, only 50% (5 out of 10) appeared to ‘view’ the written instructions, suggesting that the remaining 50% attempted to don the TPIS without first studying them. In a time-critical emergency situation, it is possible that even fewer people would take the time to read the instructions. Furthermore, previous studies have shown that experience impacts donning performance, whereas one of the main aims of the study is to collect donning performance data for naïve (i.e. non-experienced) people, representing typical passengers. Thus, the donning performance for participants that had no written instructions is explored as a possible plausible worst-case scenario. The dataset used for the worst-case scenario is the NE + NI dataset consisting of 36 data points for the DE analysis (see Table 5) and 43 data points for the DT analysis (see Table 6). Restricting the analysis to the NE + NI group also provides an indication as to how intuitive the TPIS used in this study is to don.
The statistical analysis makes use of the non-parametric Mann–Whitney U test to compare two or more groups. In statistical analysis, the outcome is defined as the dependent variable (e.g. PT, XT, PXT, DT, and DE), while the factor that is manipulated or controlled to observe its effect on the dependent variable is defined as the independent variable (e.g. age, gender, instruction category, and experience). Potential confounding variables are accounted for by isolating the primary variable through dividing the data set in sub-groups based on relevant subcategories as defined in sect. 2.4. Analysis related to instruction categories WI and/or VI were performed excluding AG3, since only one participant belonged into the sub-category AG3. In addition, due to the small number of male and female participants in some of the age groups (see Table 1), to strengthen the statistical power for some of the DT and DE analysis relating to age and gender, the subcategory “Age Groups” was re-defined into two rather than the three original subgroups. These modified subgroups were defined as follows; AG A (18–45 years of age), representing a younger sample of participants, with 14 and 11 in the male groups for DT and DE respectively and 11 and 11 in the female groups, and AG B (46–72 years of age), representing an older sample, with 16 and 12 in the male groups for DT and DE respectively and 2 and 2 in the female groups. More information concerning the sub-groups can be found in the Online Resource (see SM sect. S4).
The subsections of this section are organised as follows: First, the PT, XT and PXT are analysed with respect to the instruction categories. Then the impact of gender and age on DE and DT are explored focusing on the worst-case scenario. In addition to the worst-case scenario, the impact of method of instruction on DE and DT is explored in sect. 4.3, and the impact of experience on the donning performance (i.e. on PXT, DT and DE) is analysed in sect. 4.4. Unfortunately, there is insufficient data to perform a meaningful statistical analysis of the impact of DE on DT.
Table 7 provides an overview of the dependent and independent variables, along with their respective descriptive statistics for each group, and the probability values (p-value) from the Mann–Whitney U tests to indicate the significance of the differences between groups for each variable. Note, a Mann–Whitney U test was used for all analyses, except for the analysis of the relationship between age and DT, where a linear regression was applied (see annotation * in Table 7 and refer to sect. 4.2).
Table 7
Mann–Whitney U Test Results for dependent and independent variables and descriptive statistics
Sections
Dependent variable
Independent variable
G1: Group 1
G2: Group 2
Descriptive values
p-value
Group 1
Group 2
Mean
Min–max
(N)
Mean
Min–max
(N)
Sections 4.1–4.3 includes participants with no experience (NE) only:
4.1
Preparation time (PT)
Instruction category
G1: Written instruction (WI)
G2: Video instruction (VI)
WI
VI
0.045
8.8 s
2.0 s-20.3 s
(7)
2.2 s
1.3 s-3.6 s
(10)
Extraction time (XT)
Instruction category
G1: Written instruction (WI)
G2: Video instruction (VI)
8.5 s
4.8 s -13.1 s
(7)
5.6 s
2.8 s-11.7 s
(10)
0.051
Preparation and extraction time (PXT)
Instruction category
G1: Written instruction (WI)
G2: Video instruction (VI)
17.4 s
6.9 s-28.0 s
(10)
8.3 s
5.5 s-14.4 s
(11)
0.002
4.2
Donning errors (DE)
Age Group
G1: Males in AG A (18–45 yrs)
G2: Males in AG B (46–72 yrs)
Males in AG A
Males in AG B
0.020
0.8
0–3
(11)
1.8
0–4
(12)
Gender
G1: Males in AG A (18–45 yrs)
G2: Females in AG A (18–45 yrs)
Males in AG A
Females in AG A
0.387
0.8
0–3
(11)
1.3
0–3
(11)
Donning time (DT)
Gender
G1: Males
G2: Females
Males
Females
0.113
108.4 s
55.2 s-186.3 s
(30)
92.6 s
60.9 s-125.9 s
(13)
Gender
G1: Males in AG 2 (30–50 yrs)
G2: Females in AG 2 (30–50 yrs)
Males in AG 2
Females in AG 2
0.817
100.5 s
59.7 s-132.9 s
(8)
102.3 s
79.6 s- 125.9 s
(7)
Gender
G1: Males in AG A (18–45 yrs)
G2: Females in AG A (18–45 yrs)
Males in AG A
Females in AG A
0.381
100.3 s
55.2 s-140.3 s
(14)
91.7 s
60.9 s-125.9 s
(11)
Donning time (DT)*
Age*
N/A*
N/A*
0.004*
4.3
Donning errors (DE)
Instruction category
G1: NI in AG 1 and AG 2 (18–50 yrs)
G2: WI in AG 1 and AG 2 (18–50 yrs)
NI in AG1 + AG2
WI in AG1 + AG2
0.843
1.04
0–3
(23)
1.11
0–3
(9)
Instruction category
G1: WI in AG 1 and AG 2 (18–50 yrs)
G2: VI in AG 1 and AG 2 (18–50 yrs)
WI in AG1 + AG2
VI in AG1 + AG2
0.015
1.11
0–3
(9)
0.11
0–1
(9)
Donning time (DT)
Instruction category
G1: NI in AG 1 and AG 2 (18–50 yrs)
G2: WI in AG 1 and AG 2 (18–50 yrs)
NI in AG1 + AG2
WI in AG1 + AG2
0.018
95.1 s
55.2 s-140.3 s
(27)
128.1 s
79.4 s -183.6 s
(10)
Instruction category
G1: WI in AG 1 and AG 2 (18–50 yrs)
G2: VI in AG 1 and AG 2 (18–50 yrs)
WI in AG1 + AG2
VI in AG1 + AG2
0.121
128.1 s
79.4 s -183.6 s
(10)
105.6 s
78.6 s– 143.5 s
(9)
Section 4.4 includes participants with experience (E) and no experience (NE) from the instruction category no instruction (NI):
4.4
Donning errors (DE)
Experience category
G1: No experience (NE) in AG 1 and AG 2
G2: Experienced (E) in AG 1 and AG 2
NE in AG1 + AG2
E in AG1 + AG2
0.638
1.04
0–3
(23)
0.78
0–2
(9)
Donning time (DT)
Experience category
G1: No experience (NE) in AG 1
G2: Experienced (E) in AG 1
NE in AG 1
E in AG 1
0.166
87.2 s
55.2 s-140.3 s
(12)
102.7 s
75.9 s-151.8 s
(10)
Experience category
G1: No experience (NE) in AG 3
G2: Experienced (E) in AG 3
NE in AG 3
E in AG 3
0.680
118.1 s
76.2 s-186.3 s
(16)
108.5 s
80.0 s-133.4 s
(5)
*Performed statistical analysis: linear regression on log-transformed DT
4.1 Preparation and Extraction Time
The PXT and PT distributions for the NE + WI and NE + VI groups show significant differences, with a Mann–Whitney test indicating that the method of instruction affects both PT and PXT (p < 0.045, refer to Table 7). Specifically, VI is more effective than WI in reducing the time participants spend preparing to don the TPIS. In contrast, the XT distribution is similar between the two groups, and no significant dependence on the method of instruction was detected (p = 0.051). This is expected, as opening the TPIS packaging is intuitive and requires only tearing it open.
4.2 Impact of Gender and Age on Donning Performance
4.2.1 Donning Correctness
The descriptive results (see sect. 3.2, Fig. 4 and Table 5) suggest that there is no difference between the DE rate for males and females across all instruction and experience categories. Restricting the analysis to the plausible worst-case (NE + NI) suggests that there could be gender differences in the DE rate as shown in Fig. 7. From the descriptive data, females made slightly fewer DE (N = 13, Mean = 1.23, Mode = 0) than the male participants (N = 23, Mean = 1.35, Mode = 1). However, it is also noted that the female group is on average 8 years younger than the male group (average age males/females = 46.8 years / 38.5 years) and this age imbalance between the genders may also exert an influence on the DE rate.
Fig. 7
Percentage of male (N = 23) and female (N = 13) participants in NE + NI group incurring a given number of donning errors
The potential impact of age on the number of DEs cannot be reliably assessed using the original three age groups due to small numbers in the female age groups and the small number of males in AG3 (see Table 8 and the Online Resource SM sect. 4.2 for details). Thus, here the potential impact of age on DE rate is assessed using male data only in the modified age sub-groups, AG A (18–45 years) and AG B (46–72 years). Males in AG A (N = 11) made statistically significantly fewer DEs than males in AG B (N = 12) (p = 0.020, see Table 7). To explore the impact of age and gender on the number of DEs, participants in one age group, i.e., AG A are considered (note there are insufficient females in AG B for analysis, N = 2). No statistically significantly difference between male and females was found (p > 0.38). That no statistical difference in number of DEs was found between males and females in the younger population confirms that the imbalance of age between the genders in the overall data set could impact the DEs analysis.
Table 8
Descriptive statistics for donning errors for participants in group NE + NI according to different AGs and gender
Age Group
Gender
AG 1 18–29 years of age
AG 2 30–50 years of age
AG 3 51–72 years of age
AG A 18–45 years of age
AG B 46–72 years of age
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
(Total number of errors)/(total number of participants)
5/5
7/4
5/7
7/7
21/11
2/2
9/11
14/11
22/12
2/2
Average number of errors per person
1.0
1.8
0.7
1.0
1.9
1.0
0.8
1.3
1.8
1.0
Mode of errors per person
0,1
2
1
0
2
0,2
1
0
1,2
0,2
Frequency of mode
2
2
5
3
4
1
6
4
4
1
Min number of errors per person
0
0
0
0
0
0
0
0
0
0
Max number of errors per person
3
3
1
3
4
2
3
3
4
2
Furthermore, based on the findings from the male population, it is considered reasonable that older females would also make more DEs than younger females. It is also reasonable to assume that there is unlikely to be a difference in the DE rate between older males and older females, based on the findings of the younger male–female analysis.
4.2.2 Donning Time
The possible impact of gender on DT for the worst-case scenario (NE + NI group) is assessed for various age groupings. As there is insufficient gender-based data in each of the original three age sub-groups the analysis is undertaken for the following age groupings: the entire sample (Age: 18–72 years, Nmale = 30, Nfemale = 13), for AG 2 (Age: 30–50 years, Nmale = 8, Nfemale = 7), and for AG A (Age: 18–45 years, Nmale = 14, Nfemale = 11) (see Table 7). Unfortunately, AG 1, AG 3, and AG B, did not have sufficient data points in the female population for a meaningful statistical analysis. The results of the statistical tests for these three age groupings showed no statistically significant difference in DT between males and female (p > 0.11).
As there was no indication from the statistical tests that gender has an impact on DT, the effect of age was analysed using the combined gender sample (N = 43). Simple log-linear regression was used to test whether age significantly predicted the average DT. This means that the regression analysis was based on the log-transformed DT data. The overall regression was statistically significant (R.2 = 0.18, F(1,41) = 9.44, p = 0.004), indicating that DT significantly depends on age. The fitted regression model is displayed in Fig. 8 and can be represented by Eq. (5)
Fig. 8
Regression of DT against age in the NE + NI group (N = 43)
In Eq. (5), \(Y\) is the Donning Time (DT) in seconds. The model is derived from the corresponding linear regression model, \(\text{ln}\left(Y\right)={\beta }_{0}+{\beta }_{1}{x}_{1}+\varepsilon \), where \({\beta }_{0}\) is the intercept and\({A}_{0}={e}^{{\beta }_{0}}=72.2\). The variable \({x}_{1}\) is the explanatory variable, representing age in years, and \({A}_{1}= {e}^{{\beta }_{1}}=1.0073\) is the growth factor corresponding to\({x}_{1}\). The error term \(\varepsilon \) is assumed to follow a normal distribution,\(\varepsilon \sim Normal(\text{0,0.26})\), which corresponds to \(u={e}^{\varepsilon }\) being log-normally distributed.
The appropriateness of the model was corroborated by residual analysis of the data-fitted values of \(u\). A Shapiro–Wilk test was used to assess the fit of the standardised residuals, which suggested a strong fit to the normal distribution (N = 43, p = 0.66). Moreover, this log-linear model excludes the possibility of negative DT values, as opposed to the common linear model.
From Eq. (5), the growth factor for each 10 years increase in age can be found by:
$${{A}_{1}}^{10 years}={1.0073}^{10 }=1.075$$
(6)
From Eq. (6), the DT increases by 7.5% for each 10 years increase in age.
4.3 Impact of Method of Instruction on Donning Performance
4.3.1 Donning Correctness
A comparison of the number of DEs made by the participants without experience and in AG1 and AG2 between the various instruction categories is shown in Fig. 9. Focusing on the NE + NI (N = 23) and NE + WI (N = 9) groups, the participants in NE + NI (AG1 + AG2) (Mean = 1.04, Mode = 1) and NE + WI (AG1 + AG2) (Mean = 1.11, Mode = 0,1) incurred a similar number of DEs. It is noted that the WI group made slightly more errors on average than the NI group, which is counterintuitive, as it is expected that providing instructions would reduce the number of DEs compared to having no instructions available at all. However, no statistically significant impact of instruction category on the number of DEs was found (p > 0.84, see Table 7).
Fig. 9
Percentage of participants in AG1 + AG2 with E + NI (N = 9), NE + NI (N = 23), NE + WI (N = 9), and NE + VI (N = 9) incurring a given number of donning errors
Comparing the number of DEs incurred by the NE + WI (AG1 + AG2) group (N = 9, 1.1 DE/person) to that of the NE + VI group (AG1 + AG2) (N = 9, 0.11 DE/person), indicates that providing video instruction may significantly reduce the number of DEs compared to providing written instruction only. Table 7 shows that, indeed, a statistically significant impact of VI on the number of DEs compared to WI was found (p = 0.015).
4.3.2 Donning Time
To compare the effect of different instructional methods on DT, statistical analyses were conducted across various groups, as outlined in Table 7. Excluding AG3 from the NE + NI group (N = 27) results in an average DT of 95 s, which is 33 s shorter than the average DT for the NE + WI group (N = 10, Mean DT = 128 s), with this difference being statistically significant (p = 0.018, see Table 7). The mean DT for the NE + VI group, again including only participants from AG1 + AG2 (N = 9), was 106 s, about 22 s faster than the NE + WI group. However, this difference was not found to be statistically significant (p > 0.12, see Table 7).
4.4 Impact of Experience on Donning Performance
The number of participants with experience in the NI group is small (N = 13) and not all sub-groups provide sufficient data points for a meaningful analysis thus, only variables that were found to significantly impact DT and DE are considered. In sect. 4.2, age was found to significantly impact DT and DE, but gender was not. Therefore, only age was considered by testing the impact of experience on DT and DE (see Table 7 for the tests and results and the Online Resource SM sect. 4.3 for more analysis). Note that for the DE analysis, AG1 and AG2 were collapsed into a single grouping. Combining these data sets is considered appropriate as no significant difference was observed between participants below the age of 50 years for the NE group.
The statistical analysis presented in Table 7 shows that no statistically significant difference in DT or DE between experienced and inexperienced participants was found for any of the performed tests (p > 0.16). Thus, for the TPIS investigated in this study, experience was not a statistically significant factor in determining average DT nor DE.
5 Discussion
The discussion section addresses the factors influencing donning performance of TPIS, beginning with an examination of preparation and extraction times. This is followed by an analysis of the impact of gender and age on donning time and correctness. The effect of different instructional methods and prior experience on performance is also considered. A summary of the key characteristics affecting donning performance is then provided. Finally, the discussion covers the influence of design features on donning performance, focusing on how these elements may affect both the speed and accuracy of the process in emergency situations.
5.1 Preparation and Extraction Time
The PXT is a combination of PT and XT and it was found that the PT and thus, PXT is statistically significantly shorter for the NE + VI group than the NE + WI group. It is noted that the trial involved an ideal situation where participants in the VI group were shown the donning instruction video immediately prior to attempting to don the TPIS– as in the IMO/ISO testing requirements [8, 9]. However, this is unlikely to be the case in an actual emergency situation where passengers are likely to only have access to WI. While it is possible that passengers may have seen an instructional donning video as part of a safety drill or briefing, prior to or immediately following departure as required by IMO SOLAS [24], during an actual emergency there is unlikely to be sufficient time for passengers to watch an instructional donning video prior to attempting to don the TPIS. Nevertheless, these results demonstrate that VI can significantly reduce the PT. What remains to be demonstrated is whether VI administered several days prior to attempting to don the TPIS, will have a beneficial impact on PT. And if it does, how long can the gap between the VI and the donning event be before the beneficial effect is diminished?
It is noted that the NE + WI group displayed an unexpectedly short PT, varying from 2 to 20 s with a mean of 8 s. This is likely due to following three reasons: 1) an intuitive and simple donning process, 2) short and clear donning instructions, and 3) the fact that while it may be reasonable to expect participants to study the donning instructions prior to attempting to don the TPIS, this may not be the case. The first reason is supported by the responses of the participants collected through the post-trial questionnaires, where 64.8% (8) of participants with WI thought that the TPIS was easy/very easy to don, while only 16.7% (2) thought it was difficult/very difficult (see question 3 of the post-trial questionnaire in the Online Resource SM sect. S5.2). Even among participants with NI, 40.8% (20) rated the TPIS as easy/ very easy to don, which is approximately 50% more than those who rated it as difficult/ very difficult (18.3% (13)). For a full breakdown of responses, please refer to the Online Resource SM sect. S5.2 The second reason is supported by a visual inspection of the donning instruction, which appears simple and clear with pictograms supported by short, clear text (see Fig. 1). The third reason is supported by analysis of the donning trial video footage, which suggests that only three of the seven participants in the WI group with a detectable PT were observed to look at the instructions during the PP while none of the participants in the VI group with a detectable PT were observed to look at the written instructions. The three participants that were observed to look at the instructions displayed a longer PT (between 13 and 21 s) than the other four participants that were not observed to look at the instructions (between 2 and 5 s). Thus, most of the participants ignored the WI prior to attempting to don the TPIS.
The short PT measured in these trials is not considered a weakness of the experimental methodology. In a real emergency situation, it is likely that not everyone will make the effort to study the written instructions prior to attempting to don the TPIS, or that the conditions of the emergency situation are not conducive to studying the instructions due to, e.g., poor lighting, time pressure, or ship movement etc. Additionally, if passengers do attend to the instructions, they may only give them a passing glance.
5.2 Impact of Gender and Age on Donning Performance
5.2.1 Donning Correctness
For a TPIS to function as intended and enhance the safety of the person wearing it, it is essential that the TPIS is worn correctly. Generally, DEs can detrimentally impact the safety of the person wearing the TPIS by compromising features associated with survival and escape, in particular features associated with thermal protection, buoyancy, or mobility.
Key donning features for the TPIS in this study that are important for survival and escape, concern the final state of the shoes, ankle straps, gusset strap, zipper, and hood. For example, an incorrect final state of the gusset strap or zipper will compromise the watertight integrity of the suit enabling cold water to enter upon immersion in water, reducing the thermal protection and buoyancy capability of the TPIS. Neglecting to fasten the ankle straps compromises the fit of the TPIS allowing excess material to accumulate around the feet. This may make it more difficult to walk, increasing the time required to reach a place of safety or potentially causing trips and falls. Furthermore, in an attempt to counter-act this DE during movement, a person may attempt to physically gather the excess material with their hands. However, during evacuation it may be essential for passengers to hold onto handrails, for example during conditions of bad weather or due to adverse vessel orientation. Not holding onto handrails may negatively impact overall evacuation performance and may lead to an increase in injuries or fatalities. Failing to wear the hood correctly and close the zipper all the way may also reduce the effectiveness of the thermal protection provided by the TPIS, reducing the survival time offered by the TPIS. Finally, neglecting to wear shoes may impact escape capabilities by making it difficult to walk in some situations, in particular where there may be debris on the deck. Clearly, multiple DE may have a cumulative effect, resulting in multiple compromises to survival and escape capabilities.
Thus, successful evacuation and enhanced survivability is not only dependent on the speed at which the TPIS can be donned, but also the degree of donning correctness that is achieved. It is thus essential to identify and understand the factors that contribute to DEs so that the number of DEs incurred by each passenger can be reduced.
It is noteworthy that more than a quarter of the participants (i.e., 28% comprising 5 males and 5 females) with no experience and no instructions were able to don the TPIS without error and a further 33% (10 males and 2 females) only incurred a single DE (see Fig. 7). Compare these results with the arguably more complex TPIS (heavier and with buoyancy) investigated by Azizpour et al. [26], where only 6% (3 out of 47) of those with NE + WI donned without errors and only 3.7% (4 out of 108) of those across all experience and instruction groups donned without errors. Thus, while 72% of the participants in the current study incurred at least one DE, given that so many could don the suit correctly, suggests that the TPIS investigated in this study is somewhat intuitive to don. This assertion is supported by the response of the participants to the post-trial questionnaire, with 41% (29) of the respondents with NI suggesting that it was ‘easy/very easy’ to don and a further 41% (29) stating it was ‘not difficult nor easy’. In addition, 67% (34) suggesting that they believed that they had donned the TPIS correctly (see the Online Resource SM sect. S5, Question 3 and Question 5, respectively).
The statistical analysis for the worst-case scenario in sect. 4.2 suggests that the number of DEs is not influenced by gender but is significantly influenced by age for completely naïve participants (i.e., the NE + NI group). Older participants (above 45 years of age) appear to make more DEs than younger participants. This observation is in contrast to the finding of Azizpour et al. [26] who reported for the TPIS investigated in their study, that neither age nor gender significantly contributed to the number of DEs, and none of the interaction terms in the regression analysis (such as method of instruction and age) were significant. It has already been noted that the TPIS in the Azizpour et al. [26] study was arguably a more complex suit to don than the one used in the current study, and these factors likely contributed to the observed performance differences. The noted differences in performance for the two TPIS support the view that factors influencing the number of DEs are likely to be suit specific. However, it is noted that the test conditions in the Azizpour et al. [26] study were not identical to those of this study; in particular, all participants in that study had some form of instructions (WI or VI), which would arguably reduce the DE rate. In addition only 10% (11 out of 108) of the participants in the study by Azizpour et al. [26] were over 50 years of age, which could have made it less likely to detect any differences due to age.
5.2.2 Donning Time
Passenger ship evacuation is a time-critical process and so all elements of the evacuation process must be thoroughly understood, quantified, and efficient, enabling robust planning to optimise the evacuation process. If donning a TPIS is a necessary part of the evacuation process, then quantifying DTs and understanding the factors that influence DTs are vital to enable appropriate planning to enhance the evacuation process.
From the statistical analysis referring to the worst-case scenario, gender was not found to statistically significantly impact the DT for the NE + NI group. As with the number of DEs (see sect. 4.2), this observation is in contrast to the findings of Azizpour et al. [26] who reported for the TPIS investigated in their study, that the net donning time was dependent on gender. In their analysis, they report that females take about 33% longer to don the TPIS than males. They also report dependence on age along with experience and method of instruction (see sect. 5.3 for discussion concerning method of instruction and sect. 5.4 for discussion concerning experience). In contrast to gender, for the TPIS examined in this study, age, was found to increase DT by 7.5% for each 10 years increase in age. This observation is very similar to the findings of Azizpour et al. [26] who reported for the TPIS investigated in their study, that age exerted a slightly smaller effect of about 6.6% increase in net donning time per every 10-year increase in age. While the impact of age is similar for both TPIS, that gender is not an influential factor for this suit, but was for Azizpour et al. [26] supports the suggestion that the influencing variables for both DEs and DT are suit specific.
At first thought, it would be reasonable to assume that the average DT decreases as the number of DEs increases as essential tasks are not completed. However, achieving a correct final donning state not only requires that key activities are completed but they are completed correctly. Depending on their nature, DEs may increase or decrease the DT. For example, the correct final state for the footwear is that they are worn at the end of the donning process– inside or outside the TPIS. If the footwear is not worn, this means that the footwear was removed during the donning process, which takes time, and was not worn at the end of the process. Thus, this DE could add time to the DT compared to someone who kept their footwear on during the donning process but would decrease the DT compared to someone who removed their footwear and then correctly wore the footwear at the end of the process. Whereas an incorrect final state of other key parameters typically refers to an omission of an activity, and thus a decrease in DT might be expected. For example, not pulling the zipper all the way up will decrease the DT. However, this non-executed task might have been preceded by a struggle to perform the task followed by a failure to complete the task, resulting in an increase in the DT. Thus, the relationship between the number of DEs and the DT is complex and is dependent, not simply on the number of DEs, but also the nature of the DEs and how they were incurred.
5.3 Impact of Method of Instruction on Donning Performance
5.3.1 Donning Correctness
While analysis of participant performance in the NI group is important as it provides insight into a plausible worst-case scenario, it is also important to explore the impact of instruction method on performance. From the statistical analysis, it was found that providing WI did not significantly reduce the number of DE made. This might be attributed to three factors, 1) the WI needs improvement, 2) participants did not read the WI and/or 3) the TPIS is somewhat intuitive to don. Regarding the former, it is noted that the WI appears simple and clear with pictograms supported by short, clear text (see Fig. 1). Nevertheless, participants made suggestions on how these could be improved (see the Online Resource SM sect. S5.7 (a)) and further suggestions on improving the WI are provided in the Online Resource SM sect. S6. These relate to issues associated with the ankle strap, which incurred the greatest number of DE’s in the WI group (see Fig. 10). Referring to factor 2, it was noted from the video analysis that a large proportion, i.e. eight of the ten participants in the NE + WI group (80%), glanced at the WI on several occasions during the donning trial. However, the amount of information the participants actually received from glancing at the WI, was not measured. The third factor is supported by the fact that over a quarter of the NE + NI group manged to don the TPIS without error, suggesting that donning the TPIS is somewhat intuitive, and so perhaps a passing glance at the donning instructions is unlikely to significantly impact the number of DEs. However, from the post-trial questionnaire, more participants with WI perceived it was easy/very easy (65% or 8) compared to those with NI (41% or 29) (see the Online Resource SM S5.2). Thus, by providing WI, participants perceived that it was easier to don the TPIS, although it did not result in a significant improvement in the actual donning correctness.
Fig. 10
Frequency of each type of donning error taken across a NE + NI group (Number of DE = 47), b NE + WI group (Number of DE = 10), and c NE + VI group (Number of DE = 2)
Results from comparing the WI and VI group suggest that providing VI immediately prior to the donning process can significantly reduce the number of DE incurred, compared to simply providing WI. This is supported by the study performed by Azizpour et al. [26], who also reported a similar finding for their more complex TPIS. This finding is particularly relevant for performance standards used for evacuation analysis purposes or emergency procedures planning. Certification standards for donning of TPIS [8, 9] require a visual donning demonstration immediately prior to the donning performance testing. As demonstrated in both this work and Azizpour et al. [26], providing VI (or a live demonstration) immediately prior to attempting to don the TPIS significantly reduces the average number of DEs. However, it is questionable how relevant this testing approach is to real world situations, since during an emergency it is unlikely that passengers will have time to watch a video or live demonstration describing the donning procedures. Thus, the performance derived from certification testing is considered extremely optimistic and as such, can have detrimental consequences for the survivability of passengers. A relevant question not explored in this work relates to how effective video instruction is in reducing DEs, when the video instruction occurs several days prior to the donning event.
Finally, it is noted that DEs were made by participants in all instruction category groups (see Fig. 10). While participants in the WI group (see Fig. 10b) incurred almost every type of possible error, those in the VI group incurred only one error type, involving incorrectly closing the zipper (see Fig. 10c). And as can be expected, those in the NI group incurred every possible type of error (see Fig. 10a).
5.3.2 Donning Time
Looking at the impact of method of instruction on DT, the statistical analysis found that participants in the NE + WI group took significantly longer to don the TPIS than the NE + NI group (see sect. 4.3.2). However, there was no corresponding significant improvement in the donning correctness (see sect. 4.3.1). This may be attributed to the TPIS used in the trial being somewhat intuitive to don (as already discussed in sects. 5.2.1 and 5.3.1) and thus not having instructions would not be a major hindrance, and without instructions, participants would not spend time studying the instructions during the donning process. This is supported by the fact that from the video analysis it was noted that eight of the participants in the NE + WI group (N = 10) glanced at the donning instructions on several occasions during the donning trial. The number of times a participant gazed at the instructions varied from 1 to 6 times, with a mean of 3.3 times. Thus, having access to WI may encourage participants to refer to the instructions for reassurance. Each time a participant gazes towards the instructions they are extending their DT. This may account for the increase in DT for the NE + WI group.
The result regarding reduced average DT for participants without WI, while important, has limited general applicability to ship evacuation planning as it relates to the specific TPIS used in this study. However, it is informative to note that if the TPIS is sufficiently intuitive to don, passengers may be able to don it within a reasonable time without recourse to studying written instructions. This is clearly desirable as in an emergency, it is likely to be impractical for passengers to follow printed instructions while donning the TPIS, especially if there is poor lighting, or adverse vessel orientation and limited space.
Notably, previous research by Azizpour et al. [26] found that for the TPIS in their study, the method of instruction had a significant impact on net donning time, with VI increasing net donning time by as much as 21% compared to WI. The increase in net donning time resulting from VI, while counterintuitive, was most likely a result of participants performing fewer DEs and so undertaking all the donning tasks correctly, without missing out important steps in the donning process. Unlike the research by Azizpour et al. [26], where VI significantly increased the net donning time, the data in this study, while not being statistically significant, showed an opposite impact of VI with tendency towards a reduction in average DT. These results further support the suggestion that the influencing variables for both DE and DT are suit specific and cannot be generalised to any and all TPIS.
Again, for the TPIS assessed in this study, providing VI immediately prior to attempting to don the TPIS tends to reduce DT and significantly reduces the number of DEs (see sect. 4.3). However, this advantage is unlikely to be experienced in a real situation– as it is unlikely that there will be sufficient time to show a donning video to passengers during an emergency - and so the performance achieved under these conditions is unrepresentative of what is likely to be achieved in practice. As mentioned previously with regards to the impact of VI on DEs (see sect. 5.3.1), these results bring into question the efficacy of current IMO/ISO certification standards for donning of TPIS [8, 9] which require a visual donning demonstration immediately prior to the donning performance testing. A relevant question not explored in this work relates to how effective video instruction is in reducing DTs, when the video instruction occurs several days prior to the donning event.
5.4 Impact of Experience on Donning Performance
The preceding sections focused on a plausible worst-case scenario and participants with experience were excluded. However, previous research has suggested that experience will result in shorter DT and number of DE [26, 27]. Indeed, Azizpour et al. [26] found that for the TPIS in their study, previous experience could reduce the net donning time by as much as 17% and donning errors by 27%. However, for the TPIS used in this study, statistical analysis suggests that experience did not significantly impact DT nor DE (see sect. 4.4).
It is noted that the quality and nature of the experience of this group of participants was quite diverse. Participants were asked to identify whether they had previously donned a similar TPIS. However, it was not specified when they last donned the similar TPIS and how often they had donned it. Thus, the effectiveness of the gained experience may vary substantially, which is likely to have impacted the quality of their donning performance. Additionally, the TPIS used in this study is an emergency suit, that is, according to the manufacturer, specifically designed for the use onboard passenger ships or as an additional protection onboard leisure craft. As such, this particular TPIS is unlikely to be used in a professional work setting, such as anti-exposure suits used in the oil and gas industry, immersion suits for fishing vessels, or helicopter transport suits. Thus, it is likely that few participants would have had an opportunity to have gained donning experience in exactly this type of TPIS and it can be assumed that most participants refer to donning experience with a professional type of TPIS. This is supported by the observation that experienced participants tended to remove their footwear and kept their footwear off as generally required when using a professional TPIS (see Fig. 11). Finally, given that the TPIS was relatively intuitive to don, it is unlikely that experience in donning a TPIS of any kind would have provided significant advantage in donning performance.
Fig. 11
Percentage of experienced (N = 13) and not experienced participants (N = 36) in the NI group incurring each type of donning error (all age groups)
5.5 Summary of Donning Performance Characteristics
A key finding of this work is that TPIS donning performance, in terms of number of donning errors and time required to don is strongly related to the design of the TPIS, the quality of the instructions and the means by which the instructions are provided. This is clearly demonstrated in Table 9 by comparing the performance, under similar experimental conditions (see the Online Resource SM sect. S7, Table SM 7 for details), of two very different types of TPIS. Trends in performance cannot be generalised from one type of TPIS to another, in particular when the TPIS are significantly different in form. This suggests that each type or class of TPIS, if not each TPIS, needs to be assessed separately to determine its likely donning performance. A more complete summary of the differences and similarities between the two types of TPIS can be found in the Online Resource SM (see sect. S7, Table SM 8).
Table 9
Comparison of key performance parameters for two different types of TPIS
Males < 50 years of age made significantly fewer donning errors than males > 50 years of age
All: Not significant
Experience+
NI group (< 50 years): Not significant
All: Significant
Reduces donning errors by 27%
Method of instruction
WI vs VI (NE group, < 50 years): Significant
WI vs NI (NE group, < 50 years): Not significant
WI vs VI (All): Significant
Receiving VI significantly reduces the number of donning errors compared to WI
Factors influencing DT/NDT
Gender
NE + NI group: Not significant
All: Significant*
Females take about 33% longer to don than males
Age
NE + NI group: Significant
Increase in DT of 7.5% per every 10-years increase in age
All: Significant*
Increase in NDT of 6.6% per every 10-years increase in age
Experience+
NI group: Not significant
All: Significant*
About 17% reduction in NDT with having previous experience
Method of instruction
WI vs VI (NE group, < 50 years): Not significant
However, mean DT for VI is 22 s (∼18%) lower than for WI
WI vs NI (NE group, < 50 years): Significant
NI mean DT is 33 s (~ 26%) lower than for WI
WI vs VI: Significant*
About 21% increase when receiving VI
*Based on Model 2
+Results considered to be only indicative due to the lack of accounting for the quality, frequency and how the donning experience was gained
5.6 TPIS Design Impacting Donning Performance
Clearly, the design of the TPIS will impact donning performance, both in terms of average DT and the number of DEs, which in turn impacts survival and escape. Differences in TPIS design will not only directly impact thermal and flotation performance but will also impact the speed at which TPIS can be donned, and hence evacuation performance, and the number of DEs likely to be incurred, and hence the effectiveness of the thermal protection offered and flotation performance.
For the particular TPIS explored in this study, several design issues were highlighted that adversely impacted performance. The most important issues concern the zipper. This is a key component of the TPIS as it assists in sealing the suit (along with the gusset) providing both thermal and buoyancy protection. During the trials a substantial number of zippers broke, either during the first donning attempt (3 out of 25 or 12%) or subsequent donnings (8 out of 59 or 13.6%). This represents a failure rate of 13.1% (11 out of 84), which for a life critical component of a life critical system is unacceptable. Clearly, a stronger zipper is required.
Furthermore, the zipper received most attention by the participants in the post-trial questionnaires. Many comments related to the size of the zipper and the ability for the wearer to grab and pull the zipper while wearing the oversized, one-size-fits-all gloves (see the Online Resource SM sect. S5.7 for details). Several participants suggested that the zipper should include a long ‘grab tag’ to make it easier to clasp and pull the zipper while wearing the oversized gloves.
Another area requiring improvement for this particular TPIS concerns the gusset and gusset strap, which received the second highest number of comments in the post-trial questionnaires. The gusset is a key component of the TPIS as it is the final barrier sealing the suit providing both thermal and buoyancy protection (along with the zipper). The three main issues suggested by the participants requiring improvement were:
The gusset was too oversized, creating a vast surplus of fabric which made it difficult when stepping into the suit and attempting to close the zipper.
The gusset strap was considered too long and interfered with the zipper when the suit was being closed.
The gusset tightening mechanism was considered to be too small, its function unclear and not very intuitive.
The amount of gusset fabric may be difficult to reduce as it is required to be generous since it determines the size of the opening to enter the TPIS. Nevertheless, replacing the current toggle and drawstring mechanism with a larger and more intuitive closing mechanism, that can be easily operated while wearing the oversized gloves, is recommended.
Finally, an issue raised by some participants concerned vision impairment. One participant observed that their glasses tended to fog up, while another participant experienced impaired vision when the TPIS was fully zipped up. Moreover, wearing the TPIS with the hood up and the zipper fully closed caused discomfort to the bridge of their nose due to their glasses.
6 Quantification of Donning Time for Regulatory Purposes
In this section, the experimental donning data is compared with the regulatory process to explore issues with the current compliance procedures. In addition, donning time distributions are proposed for use in evacuation simulation analyses intended to demonstrate that proposed vessel designs meet IMO requirements.
6.1 Regulatory Compliance
It is an IMO regulatory requirement that, after receiving a donning demonstration, the TPIS can be unpacked and donned without assistance within 120 s, including a lifejacket [8, 9] (see sect. 1). For the TPIS investigated the TDT (derived from Eq. 3) for participants in the NE + VI group (N = 10) (see Fig. 12a) and the NE + NI group (N = 43) (see Fig. 12b) are compared with the regulatory requirement. For the PXT component in Eq. 3, the mean PXT derived for the NE + NI group (N = 10), 17.4 s, and the PXT derived for the NE + VI group (N = 11), 8.3 s is used.
Fig. 12
TDT distribution highlighting the 120 s regulatory requirement (dashed line) based on the a (NE + VI) distribution (N = 10) and (b) (NE + NI) distribution (N = 43)
The TDT for the NE + VI group varies from 86.9 to 151.8 s and the TDT frequency distribution (see Fig. 12a) suggests that 30.0% (3 out of 10) of the participants have a TDT in excess of the maximum permitted donning time of 120 s. For the NE + NI group, the TDT varies from 72.6 to 203.7 s (see Fig. 12b), suggesting that 44.2% (19 out of 43) of the participants exceed the maximum permitted donning time. For regulatory purposes, to demonstrate that the donning performance of the TPIS satisfies regulatory requirements it is essential to follow the testing protocols set out in the guidance documentation [8, 9]. This requires that the test subjects are first shown a visual donning demonstration. Thus, the NE + WI results cannot be used for exploring TPIS compliance with regulatory requirements.
However, the trial conditions for the NE + VI group closely follow the test standards required by IMO/ISO [8, 9], including showing the participants a visual donning demonstration in form of a video immediately prior to undertaking the donning task, but excluding the time required to don a lifejacket, as is required for this TPIS. As the measured TDT does not include the time required to don a lifejacket, the measured TDT should be well below the 120 s criteria. As can be seen from Fig. 12a, 30% of the TDTs exceed the regulatory requirement, even without donning the lifejacket. The mean DT for the NE + VI group is 104.7 s (see Table 6), and with a mean PXT of 8.3 s (see Table 4) this produces a mean TDT of 113.1 s, leaving 6.9 s to don the lifejacket. While donning lifejackets was not considered in this study, other studies have reported lifejacket donning times. Brown et al. [20] measured lifejacket donning times of between 15.6 and 83.9 s, with a mean of 38.5 s for test participants wearing normal clothing. It is noted that lifejacket donning times will vary depending on the nature of the lifejacket, and that wearing the TPIS while attempting to don the lifejacket is likely to be a more difficult task, especially due to the over-sized rubber gloves integral to the TPIS. If the mean lifejacket donning time from the Brown et al. [20] study is adopted, this results in a mean TDT for the NE + VI group of 151.6 s, well in excess of the regulatory permitted maximum. Even if the minimum lifejacket donning time is used, this results in a mean TDT of 128.7 s, again exceeding the regulatory permitted maximum. Furthermore, if the mean lifejacket donning time is incorporated within the minimum TDT for the NE + VI dataset (86.9 s), this produces a time of 125.4 s, suggesting that 100% of the TDTs exceed the regulatory permitted maximum.
As already noted, exposing test participants to an instructional video immediately prior to donning the TPIS substantially reduces the number of DEs and the DT compared to other situations (see sect. 5.3). However, it is unlikely that passengers would be shown an instructional video on how to don a TPIS during an actual time-critical emergency situation aboard a passenger vessel, primarily due to time constraints. It is more likely that passengers will have to rely on memory (from the initial pre-sailing demonstration), intuition, or written instructions, should they have time and the care to read them. This consideration highlights the practical relevance of the NE + NI condition, in which participants received no prior demonstration and were not provided with written instructions. The impact this has on regulatory compliance testing is demonstrated by the TDT data presented in Fig. 12b for the arguably more realistic NE + NI case. As can be seen from Fig. 12b, 44.2% of the TDTs exceed the regulatory requirement, even without donning the lifejacket. This is a greater proportion than in the group that received video instruction. If the mean lifejacket donning time (38.5 s) is added to the TDT for the NE + NI group, this produces a TDT of between 111.1 and 241.2 s with a mean of 159.5 s, resulting in 90.7% (39 out of 43) exceeding the regulatory permitted value. Additionally, regulatory compliance is not based solely on time; the TPIS must also be correctly donned. Assuming this implies zero DE, it is notable that 72% of participants in the NE + NI group incurred at least one error (see Table 9). Furthermore, of the four participants who did meet the time requirement, all had at least one donning error, suggesting that they may not have satisfied the criterion for correct donning.
6.2 Evacuation Modelling
A main aim of this work was to provide a reliable evidence base to quantify the time required to don a TPIS that could be used in passenger ship evacuation analysis as specified in the IMO evacuation modelling guidelines (MSC.1/Circ 1533 [10]). Here, a TDT distribution for the specific TPIS examined in this work is proposed for use in agent-based evacuation modelling. The proposed TDT distribution could be incorporated into agent-based evacuation modelling much the same way as passenger response time data is currently used. A probability distribution, based on real-world data for the TPIS, can be used to randomly allocate a TDT to agents within the simulation based on their age group. In this way, the age-appropriate total donning time allocated to an agent will contribute to the determination of the total evacuation time for the vessel. The TDT for each unique passenger is defined by, TDT = PXT + DT (see Eq. 3 in sect. 2.4).
This contribution could be quite important as DTs for TPIS that have satisfied the IMO donning time requirement can be up to 120 s (as required by IMO regulation) or more (see Fig. 12). This will contribute to ensuring that the passenger ship evacuation analysis required by IMO will be adapted to accommodate one of the issues associated with passenger ship evacuation in polar waters.
As discussed in sects. 4 and 5 a plausible worst-case scenario is considered to specify the DT. This consists of the DT data for passengers without previous experience (NE) and without written instructions (NI). This consists of a dataset comprising 43 data points (see Table 6). As demonstrated in sect. 5.2 the trial data indicate that DT for this type of TPIS is independent of gender but is dependent on age. To be consistent with the IMO evacuation modelling guidelines [10] three AGs were defined, AG1 < 30 years, AG2 30–50 years, AG3 50 + years however, given the nature of the TPIS data collected, the lower limit for AG1 was set to 18 years and the upper limit for AG3 was set to 72 years.
As discussed in sect. 3.3, a log-normal distribution was found to best fit the DT data. The distribution curves presented in Fig. 6 are for the entire dataset while here distributions for the NE + NI dataset are required. The mean DTs, \({DT}_{mean}\), for each AG were calculated based on mean ages, \({Age}_{mean}\) (see Eq. 7), and the mean DT regression curve (see Eq. 8) found from the collected data (see Eq. 5 in sect. 4.2.2):
where \({Age}_{upper}\) is the upper age limit, and \({Age}_{lower}\) is the lower age limit of AG1, AG2, and AG3, respectively.
$${DT}_{mean}=72.2*{1.0073}^{{Age}_{mean}}$$
(8)
Mean ages equate to 23.5 years, 40 years, 61.5 years for AG1, AG2, and AG3, respectively, and mean DTs to 85.7 s, 96.6 s, and 112.9 s, respectively.
The variances for the log-transformed DTs (NE + NI group) across all three age groups were assessed using Levene’s test, which showed no statistically significant difference, F(2,40) = 0.897, p = 0.42. Thus, an average standard deviation on the log-normal scale is used for all three age groups. And finally, the probability density function (PDF) is suggested to be truncated after the 99.5% quantile for each AG, with the reasoning that crewmembers or other passengers may assist passengers that seem to struggle with the donning process. For a cruise ship of 3000 passengers, this would equate to 15 passengers being assigned a shorter DT. The PDFs are described by Eqs. (9), (10), and (11) for AG1, AG2, and AG3, respectively.
To complete the definition of TDT a PXT is required (see Eq. 3), where PXT = PT + XT. However, the PXT was only measured for the new suits, not the used suits that were utilised to define the DT distribution. Given the complex nature of the behaviours during the PP and the relatively small number of data points in the Preparation Phase (PP) and Extraction Phase (XP), rather than attempt to specify separate distributions for each phase, a combined distribution is defined for the PXP. This combines the behaviours and times associated with preparing to don the TPIS and opening the packaging. As the PXT for the NE + NI group was not measured, here the PXT for the NE + WI group is adopted. This is based on the assumption that passengers will have been given new suits (still in the packaging) and passengers are unlikely to have been shown a training video immediately prior to donning the TPIS. For the NE + WI data, this consists of 10 data points (see Table 4 and Fig. 3). To determine whether a normal or log-normal distribution better represents the PXT distribution of the NE + WI group, a Shapiro–Wilk test was used. This suggested that the collected data could fit both, a normal distribution (W(10) = 0.96, p = 0.73) and a log-normal distribution (W(10) = 0.93, p = 0.48). However, since PXT is a non-negative stochastic variable, a log-normal distribution was considered a more appropriate fit as its values cannot become negative. The log-normal distribution (see Eq. 4) describing the NE + WI PXT data is defined with μ = 2.79 and σ = 0.402 as shown in Fig. 14.
The TDT for the specific TPIS used in the trials described in this paper is defined according to Eq. 6 using Eqs. 9–11 to determine the DT and Eq. 12 to determine the PXT. It is suggested that until additional data for different TPIS are collected, these equations can be used to represent this type of TPIS, i.e., lightweight TPIS without integrated buoyancy. Finally, for completeness, as the TPIS explored in this work did not include integrated buoyancy device, it is necessary to include a time required to don a lifejacket. As discussed in sect. 6.1, based on the work of Brown et al. [20], a mean lifejacket donning time of 38.5 s should be added to the total donning time derived using Eqs. 9–12.
An example of generating DTs for evacuation modelling utilising the suggested DT and PXT distributions for modelling purposes is provided in the Online Resource SM sect. S8.
7 Limitations
As with any experimental study involving human test subjects, there are limitations associated with this work which should be considered when reviewing the results. The limitations of the current study are identified as follows:
The donning data was collected in land-based facilities in controlled environments without the simulation of hazardous conditions that may arise on a vessel necessitating its evacuation. Thus, it is important to realise that the data represents donning times under ideal conditions and that donning undertaken in a hazardous situation may require longer time.
Participants were instructed to remove excessive clothing such as jackets or heavy jumpers prior to the start of the trial, and participants came appropriately dressed wearing trousers and flat shoes. Onboard a vessel, passengers may be dressed in skirts or dresses, wearing high heels, may be wearing additional warm clothing, which may impact the PT and/or DT. The DT and TDT measured in the trials can be expected to be optimistic.
The physical space available to the participants during the trials was representative of the minimum floor area per passenger required by international regulation. However, as not all participants scheduled for a specific time slot always showed up, the space available for each participant varied from cohort to cohort. Generally, participants had ample space available. Nevertheless, it is possible that in actual emergency situations, passengers may be in environments with less physical space which may make donning more challenging.
For ethical reasons, all trial participants were in good health, non-disabled and below the age of 72 years to minimise the risk of injury. Some 24% were over the age of 50, with 8% being over the age of 65 years [28]. Thus, most participants were standing upright during the donning process. However, elderly people, or people with mobility impairments, may need to sit down to don the suit. As IMO require that 40% of the passengers simulated in evacuation modelling analysis for evacuation certification applications are assumed to have mobility impairments [10], it is essential that TPIS donning data for mobility impaired passengers is collected. Furthermore, a dynamic environment would add to the need for support during donning. This will have an impact on the donning performance and on the space required, as well as the necessity for seating options all of which may impact the DT, suggesting that donning times measured in these trials are optimistic. Furthermore, of those aged over 50 years of age, only 4 out of 22 (18%) were female. Thus, females were under-represented in the higher age range, which may have impacted the gender independence noted in the trials.
The average Body Mass Index (BMI) for male and female participants was 26 (SD = 3) and 24 (SD = 5), respectively. The majority of participants in the trial were within the normal BMI range with only 8 (9%) of the participants in the obese category. It is noted that in the UK and the US 27% and 38%, respectively, of the population are classified as obese [34]. Thus, the sample population used in the trials may not be considered fully representative of the target population. While further research is required to include a wider cross-section of the public, the donning times measured in these trials may be considered to be optimistic.
Concerning the validity of the statistical analyses, some of the sub-groups suffered from small numbers of participants, e.g., only 10 participants in the NE + WI group, for the WI group, there were no participants in AG3, and only one participant in the VI group. With such small numbers it is often difficult to generalise, and some caution should therefore be taken when considering the analysis of the factors that influence donning errors and donning times.
While the number of participants is high considering the time and effort needed to conduct the experiments, the strength of several of the statistical tests could have been greatly improved with higher number of participants. More influencing factors affecting DE and DT could potentially have been revealed if the data set had been larger.
The relative high number of statistical tests that are performed in the analysis increase the risk of having false positive results. Test with p-values in the same order as the chosen significance level of 0.05, should be treated with caution, and it is recommended that more experimental research is performed to validate those results in particular.
Additional data concerning the impact of prior experience of donning a TPIS is required to explore the impact of experience on donning performance. The datasets presented in this paper that are recommended for regulatory application are based on naïve test subjects with no prior donning experience and so are considered conservative.
8 Conclusion
The aim of this work was twofold, first to build a reliable evidence base describing Thermal Protection Immersion Suit (TPIS) donning performance by typical passengers and identifying factors that impact donning time and donning correctness. Secondly, to suggest a TPIS dataset that could be used in regulatory applications such as advanced computer simulation of passenger ship evacuation. To address these aims a series of experimental donning trials were conducted involving 96 volunteers (67 males and 29 females) aged between 18 to 72 years. The particular type of TPIS used in the trials was a lightweight, one-size fits all, non-insulated suit without buoyancy.
The uniqueness and importance of this study is that it identifies and quantifies, for the first time, the factors that influence donning time and correctness for the type of TPIS used in this study. Good performance in both is essential for life safety; if the TPIS takes too long to don, this could reduce the likelihood of a safe evacuation, and if donned incorrectly, this could adversely impact the protection offered by the suit, reducing survivability probability. The insight that this information provides is of benefit to both ship operators and TPIS designers. For ship operators it assists them to develop procedures to minimise the time required to correctly don the TPIS and for TPIS designers, it identifies design issues that make it difficult to quickly and correctly don the TPIS.
A key finding of this work is that, while the particular TPIS used in these trials was somewhat intuitive to don quickly and correctly, various factors were found to impact the donning performance, in particular age. Older participants made significantly more donning errors than younger participants and took significantly longer to don the TPIS. While the results were not conclusive, experience was not found to significantly impact speed and/or correctness of donning. Furthermore, and unlike other studies, gender was not found to significantly impact donning time. It is suggested that this finding could be dependent on the nature of the TPIS used, which was both light and relatively simple to don, compared to TPIS used in other studies.
Three methods of donning instruction were assessed to determine their impact on both donning time and correctness, including provision of, no instructions, written instructions (provided by the manufacturer), and video (plus manufacturers written) instructions. The video instruction was provided immediately prior to the donning trial. An interesting finding was that providing no instruction resulted in significantly lower donning times than providing written instructions. This counterintuitive result is thought to be related to the intuitive nature of donning this particular TPIS, resulting in participants not needing to read instructions to don. It is thus important not to generalise this finding to other types of TPIS. Another unexpected result was that watching a video prior to donning, while resulting in shorter donning times on average compared to the written instruction group, did not statistically significantly reduce donning time. However, video instruction did significantly reduce the number of donning errors compared to the other methods of instruction.
A significant observation resulting from this work concerns the appropriateness of testing protocols utilised by the IMO/ISO to approve TPIS for use in cold environments. Current testing only requires the use of six test subjects to determine whether they can correctly don the TPIS within a specified period of time. In addition, test subjects are shown an instructional video immediately prior to the donning trial. The analysis presented in this paper demonstrates that, given the complexity of correctly donning a TPIS, even a relatively intuitive TPIS as used in this trial, significantly more than six test subjects are required to quantify donning performance. Indeed, data from this work demonstrate that, given the power of the current study (53 participants in the NI and VI groups) more than 91% of the sample would fail to meet the IMO/ISO donning time requirement and 100% would have failed if donning errors were considered. Furthermore, in a real emergency incident requiring the abandonment of a vessel, it is unlikely that there will be sufficient time to show all the passengers an instructional video demonstrating how to don the TPIS correctly. Results from this work demonstrate that showing participants an instructional video on how to correctly don the TPIS immediately prior to donning can reduce donning times and significantly reduce donning errors, compared to simply providing written instructions or no instructions. It is thus suggested that IMO/ISO need to review their test protocols to better represent the diversity of the user population and the means of instruction likely available in an emergency.
Resulting from this work are several suggestions to improve the donning performance of the TPIS used in this study. The main suggested design change concerns the undersized suit closing mechanisms (zipper and gusset strap), which are challenging to operate when wearing the provided oversized rubber gloves. Also, the zipper was found to be of poor quality and a significant number broke during the donning trials. In a real emergency incident, such failures could compromise passenger safety.
Finally, addressing the second aim of the paper, a donning time distribution that can be used in agent-based passenger ship evacuation analysis was developed. The suggested donning time distribution reliably represents typical donning times of passengers donning the type of TPIS used in this study. However, donning times are believed to be suit specific, or at least specific to various types or classes of suit design. Thus, the suggested donning time distribution should only be used for TPIS of the type described in this paper. Using the proposed donning time distribution in ship evacuation models will enable engineers to assess whether the time required to don the TPIS critically impacts the evacuation process, and if it does, enables them to refine procedures to reduce the impact.
In reviewing these results, it is important to note the limitations associated with the study. In particular, about 76% (N = 69) of the study were under the age of 50 years, with just about 8% (N = 7) being over the age of 65 years and only 9% (N = 8) being classed as obese (BMI ≥ 30). Furthermore, the trials were conducted in ideal environmental conditions. These limitations were a combination of practical and ethical considerations, the latter being intended to reduce the risk of injury to the participants.
Further research involves quantifying the impact wearing a TPIS has on walking speeds of passengers ascending and descending stairs, and determining the impact of TPIS on ship evacuation through incorporating TPIS performance data within an agent-based passenger ship evacuation model.
Acknowledgements
The authors would like to express their great appreciation to MARKOM-2020 (T92) for providing funding, Hansen Protection for providing the TPIS used in this study, and UiT The Arctic University of Norway, ARCOS safety centre in Tromsø, Norway, and ResQ safety centre (now Maersk Training) in Haugesund, Norway, for providing access to the data collection facilities. We would like to offer our special thanks to the staff at UiT and HVL that supported the data collection, and finally, we are particularly grateful for all the trial participants that freely provided their time without which this project would have been impossible.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Fahy RF, Proulx G (1997) Human behavior in the world trade center evacuation. In: International association for fire safety science, fifth international symposium. pp 713–724
2.
Proulx G, Fahy RF (1997) The time delay to start evacuation: review of five case studies. Fire Saf Sci 5:783–794CrossRef
3.
Fahy RF, Proulx G, Aiman L (2012) Panic or not in fire: clarifying the misconception. Fire Mater 36:328–338CrossRef
4.
Proulx G, Fahy RF (2008) Human behavior and evacuation movement in smoke. ASHRAE Trans 114:159
5.
Fahy RF, Proulx G (2001) Toward creating a database on delay times to start evacuation and walking speeds for use in evacuation modeling. In: 2nd international symposium on human behaviour in fire. Boston, pp 175–183
6.
Fahy RF (1994) Exit 89-an evacuation model for high-rise buildings-model description and example applications. In: Fire safety science–proceedings of the 4th international symposium. pp 657–668
7.
International Maritime Organization (IMO) (2016) International code for ships operating in Polar waters (Polar Code). London
8.
IMO (1998) Resolution MSC.81(70) Revised recommendation on testing of life-saving appliances
9.
International Organization for Standardization (ISO) (2012) 15027-2:2012 Immersion suits - part 2: abandonment suits, requirements including safety
10.
IMO MSC.1/Circ 1533 (2016) Revised guidelines for evacuation analysis for new and existing passenger ships. 6 June 2016
11.
Galea ER, Owen M (1994) Predicting the evacuation performance of mass transport vehicles using computer models. IMarE Conf 106:
12.
Galea ER (2000) Safer by design: using computer simulation to predict the evacuation performance of passenger ships. Inst Mar Eng 112:7–16
13.
Glen IF, Galea ER (2001) Ship evacuation simulation: challenges and solutions. In: Annual meeting society of naval architects and marine engineers. SNAME transactions, pp 121–139
14.
Vassalos D, Kim HS, Christiansen G et al (2002) A mesoscopic model for passenger evacuation in a virtual ship-sea environment and performance-based evaluation. Pedestrian and evacuation dynamics. Springer, Netherlands, pp 369–391
15.
Gwynne S, Galea ER, Lyster C, Glen I (2003) Analysing the evacuation procedures employed on a Thames passenger boat using the maritimeEXODUS evacuation model. Fire Technol 39:225–246CrossRef
16.
Galea ER, Deere S, Brown R, Filippidis L (2013) Two evacuation model validation data-sets for large passenger ships. J Speech Lang Hear Res 57:155–170
17.
Averill JD (2011) Five grand challenges in pedestrian and evacuation dynamics. Pedestrian and evacuation dynamics. Springer, Cham, pp 1–11
18.
Ronchi E, Kuligowski ED, Peacock RD, Reneke PA (2014) A probabilistic approach for the analysis of evacuation movement data. Fire Saf J 63:69–78CrossRef
19.
Grandison A, Deere S, Lawrence P, Galea ER (2017) The use of confidence intervals to determine convergence of the total evacuation time for stochastic evacuation models. Ocean Eng 146:234–245CrossRef
20.
Brown R, Boone J, Small G, et al (2008) Understanding passenger ship evacuation through full-scale human performance trials. In: International conference on offshore mechanics and arctic engineering. pp 645–650
21.
Zhang H, Song G (2014) Performance of immersion suits: a literature review. J Ind Text 44:288–306CrossRef
22.
Ducharme MB, Brooks CJ (1998) The effect of wave motion on dry suit insulation and the responses to cold water immersion. Aviat Space Environ Med 69:957–964
23.
Power J, Ré AS, Barwood M et al (2015) Reduction in predicted survival times in cold water due to wind and waves. Appl Ergon 49:18–24CrossRef
24.
International Maritime Organization (IMO) (2013) Amendments to the international convention for the safety of life at sea (SOLAS), 1974, Resolution MSC.350(92)
25.
Hystad SW, Olaniyan OS, Eid J (2016) Safe travel: passenger assessment of trust and safety during seafaring. Transp Res Part F Traffic Psychol Behav 38:29–36CrossRef
26.
Azizpour H, Galea ER, Erland S et al (2023) Factors influencing the time required to don thermal protective immersion suits correctly. Saf Sci 164:106064CrossRef
27.
Mallam SC, Small GR, MacKinnon SN (2012) Donning time of marine abandonment immersion suits under simulated evacuation conditions. J Ocean Technol 7:45–59
28.
Brünig R, Galea ER, Batalden B-M, Oltedal HA (2021) A methodology for collecting donning times of thermal protective immersion suits intended to be worn by passengers on vessels operating in cold environments. In: IOP conference series: materials science and engineering. IOP Publishing, p 12056
Azizpour H, Galea ER, Erland S et al (2022) An experimental analysis of the impact of thermal protective immersion suit and angle of heel on individual walking speeds. Saf Sci 152:105621CrossRef
31.
Brünig R, Galea ER, Erland S, et al The impact of thermal protective immersion suits on walking speeds while ascending and descending stairs (in preparation)
32.
IMO LSA Code (2017) International life-saving appliance (LSA) code. London
33.
International Maritime Organization (IMO) (1974) International convention for the safety of life at sea (SOLAS), as amended
34.
Gallagher D, Heymsfield SB, Heo M et al (2000) Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index. Am J Clin Nutr 72:694–701CrossRef
