Source-linked AI summary
An Immersive Virtual Reality Serious Game to Enhance Earthquake Behavioral Responses and Post-earthquake Evacuation Preparedness in Buildings
Zhenan Feng, Vicente A. González, Robert Amor, Michael Spearpoint, Jared Thomas, Rafael Sacks, Ruggiero Lovreglio, Guillermo Cabrera-Guerrero
TL;DR
Conventional earthquake-preparedness training can be limited in engagement and realistic hazard simulation. This study tests an IVR serious game at Auckland City Hospital, built around best-practice evacuation scenarios and assessed through knowledge and self-efficacy. Participants’ knowledge and self-efficacy increased significantly after training, and they found the system engaging and easy to use.
Problem
Conventional training approaches may have low engagement and limited ability to reproduce realistic earthquake hazards, while evidence on IVR serious games for earthquake responses and evacuation preparedness remains limited.
Source: §1 · p. 3Method
The study developed an IVR serious game for a hypothetical Auckland City Hospital earthquake, embedding best-practice scenarios and measuring knowledge and self-efficacy before and after training.
Source: §1 · p. 4, §6 · p. 25Results
Participants’ knowledge of earthquake responses and post-earthquake evacuation, and their self-efficacy, increased significantly after training.
Source: §6 · pp. 24–26Takeaways & Limitations
Participants reported that the IVR serious game was engaging and made learning best-practice knowledge easy.
Source: §6 · pp. 25, 27, §7 · p. 27Most participants were young adults, leaving the fit of IVR training for children and seniors unclear; around 6% stopped the experiment because of motion sickness.
Source: §7 · p. 28
Abstract
from arXiv · showhide
Enhancing the earthquake behavioral responses and post-earthquake evacuation preparedness of building occupants is beneficial to increasing their chances of survival and reducing casualties after the main shock of an earthquake. Traditionally, training approaches such as seminars, posters, videos or drills are applied to enhance preparedness. However, they are not highly engaging and have limited sensory capabilities to mimic life-threatening scenarios for the purpose of training potential participants. Immersive Virtual Reality (IVR) and Serious Games (SG) as innovative digital technologies can be used to create training tools to overcome these limitations. In this study, we propose an IVR SG-based training system to improve earthquake behavioral responses and post-earthquake evacuation preparedness. Auckland City Hospital was chosen as a case study to test our IVR SG training system. A set of learning outcomes based on best evacuation practice has been identified and embedded into several training scenarios of the IVR SG. Hospital staff (healthcare and administrative professionals) and visitors were recruited as participants to be exposed to these training scenarios. Participants' preparedness has been measured along two dimensions: 1) Knowledge about best evacuation practice; 2) Self-efficacy in dealing with earthquake emergencies. Assessment results showed that there was a significant knowledge and self-efficacy increase after the training. And participants acknowledged that it was easy and engaging to learn best evacuation practice knowledge through the IVR SG training system.
1. Introduction
Conventional earthquake-preparedness training can lack engagement, realistic hazards, and feedback, while IVR and serious games offer ways to simulate hazards and support interaction. This study investigates an IVR serious game for earthquake responses and post-earthquake evacuation preparedness, assessing knowledge and self-efficacy.
1. Introduction
- Seminars, posters, videos, and drills have pedagogical limitations, including low engagement, unrealistic hazards, and limited feedback on whether behaviors meet best practice.These limitations may affect learning outcomes and behavioral changes.Source: §1 · p. 3
- IVR can present simulated life-threatening scenarios, while serious games support interaction and feedback for teaching and training.IVR immerses participants in computer-generated environments; serious games have pedagogical purposes.Source: §1 · p. 3
- IVR serious-game evacuation training remains rare, especially for earthquake emergencies, despite applications in healthcare, driving, and workplace safety.A recent systematic literature review indicated this gap (Feng, González, Amor, Lovreglio, & Cabrera-Guerrero, 2018).Source: §1 · p. 4
- The study proposes an IVR serious-game framework for a hypothetical Auckland City Hospital earthquake and assesses preparedness through knowledge, self-efficacy, training efficacy, and engagement measures.Training content follows New Zealand Civil Defence and Auckland District Health Board guidance.Source: §1 · p. 4
2. Background
Prior work describes serious games and IVR as tools for emergency training and behavioral study, but earthquake evacuation education remains limited. New Zealand and Auckland hospital procedures provide the guidance and practical context for the study’s training scenarios.
2.1 IVR SGs for emergency preparedness
- Serious games have been used for emergency training and behavioral analysis, with interaction and feedback proposed as supports for engagement and learning.Examples include fire, aviation, and earthquake preparedness training, as well as recording evacuation paths and exit choices.Source: §2.1 · pp. 4–5
- IVR can provide credible virtual environments for realistic behavior, and prior studies report memory-recall and knowledge-retention benefits under IVR conditions.The cited work attributes these outcomes to task focus, engagement, and life-like experience.Source: §2.1 · p. 5
- Prior IVR serious-game studies report emergency-training benefits, but few address full earthquake behavioral responses and post-earthquake evacuation preparedness in buildings.Examples include children’s fire-safety knowledge, airplane evacuation time, and earthquake self-protection skills.Source: §2.1 · p. 6
2.2 National and hospital earthquake response procedures in New Zealand
- Auckland City Hospital was selected for a virtual drill because operational disruption and ethical risks restrict hospital evacuation drills.The paper notes that hospital plans can be assessed through alternatives such as computer simulations.Source: §2.2 · pp. 6, 8
- New Zealand guidance promotes Drop, Cover and Hold during shaking, followed by post-earthquake actions such as checking routes, helping others, and evacuating by stairs.Building occupants often have little chance to practice the full recommended response set because realistic hazards are too dangerous to reproduce in physical drills.Source: §2.2 · p. 7
- Auckland hospital procedures advise staff to shelter during earthquakes, remain inside immediately afterward, and take additional steps such as first aid and checking hazards.The manual also recommends advising visitors to remain until safety is assessed and turning off damaged utilities where practicable.Source: §2.2 · p. 7
3. The IVR SG training system
The system recreates part of Auckland City Hospital for a full indoor-earthquake and post-earthquake evacuation experience. Its scripted scenarios embed recommended actions and use immediate and post-game feedback to teach them.
3. The IVR SG training system
- The prototype trains Auckland City Hospital occupants through a full indoor-earthquake and post-earthquake evacuation experience focused on recommended practice.The training aims to improve preparedness and knowledge of post-earthquake evacuation practice.Source: §3 · p. 8
3.1 Virtual Environment
- A BIM-based workflow models a portion of ACH’s fifth floor, including public and administrative areas, then imports the model into Unity for IVR and game development.The model includes structural and non-structural components.Source: §3.1 · p. 8
3.2 Earthquake Simulation and Building Damage
- Earthquake shaking and damage are simulated qualitatively; the system depicts non-structural damage but excludes major structural failure or collapse.Examples include falling ceiling tiles, toppling partitions and furniture, and breaking glass.Source: §3.2 · p. 9
3.3 Non-playable Characters
- Scripted non-playable characters represent other evacuees and let participants practice actions such as helping a trapped person.One scenario has a female NPC stuck under furniture asking for help.Source: §3.3 · p. 10
3.4 Navigation and Interaction
- Waypoint navigation limits movement along predefined routes; at stopping points, action panels let participants choose actions tied to recommended behaviors.Participants move between waypoints and click a panel to choose an action.Source: §3.4 · pp. 10–11
3.5 Training Outcomes and Storyline Narrative
- The action-driven storyline progresses from taking cover and checking for aftershocks to helping others, finding alternatives, responding to hazards, and reaching an outdoor assembly point.Recommended behaviors include using stairs and unplugging damaged electrical equipment.Source: §3.5 · pp. 11–12
3.6 Teaching Method
- Immediate green or red lights indicate whether a choice is recommended, followed by a post-game report and audiovisual playback explaining the rationale for recommended actions.The post-game assessment recaps actions against the full list of recommended behaviors.Source: §3.6 · pp. 14–15
4. Research Methods
The study assessed training using pre- and post-tests of the outcome of interest. The IVR SG ran on a Unity workstation with an Oculus Rift headset and one-button remote control.
4. Research Methods
- A pre-test before training and a post-test afterward measured the outcome of interest; participants received indoor-earthquake and post-earthquake evacuation training through the IVR SG.Source: §4 · p. 16
4.1 Apparatus
- The Unity system ran on a DELL workstation with an Oculus Rift headset and remote; its display was mirrored to an LED screen so researchers could observe and record in-game behavior.The remote let participants choose action panels with one button.Source: §4.1 · p. 16
4.2 Participants
The experiment recruited hospital staff and visitors, with 87 of 93 participants completing it. Participants reported more experience with fire drills than earthquake drills.
4.2 Participants
- 87 of 93 recruited participants completed the experiment; six stopped the IVR experience because of motion sickness.Source: §4.2 · p. 16
- Of 93 participants recruited (43 male, 50 female), 87 completed the experiment; completers comprised 25 ACH staff and 62 visitors, with ages ranging from less than 20 to over 70.One-third were aged 20 to 29.Source: §4.2 · p. 16
- Participants reported more experience with fire drills than earthquake drills.Source: §4.2 · p. 17
4.3 Measures
The study assessed preparedness through knowledge of recommended earthquake responses and self-efficacy, alongside participants’ perceived training efficacy and engagement. Knowledge was assessed with open-ended questions, while the other dimensions were measured with questionnaires.
4.3 Measures
- Preparedness measures covered knowledge and emergency self-efficacy, with perceived learning efficacy and game engagement assessed to inform the system’s applicability.Knowledge and self-efficacy questionnaires were administered before and immediately after training; training efficacy and engagement were measured after training.Source: §4.3 · p. 18
4.3.1 Knowledge acquisition
- Five open-ended questions assessed responses inside a building during and after an earthquake and outside afterward; answers were recorded, transcribed, cross-checked by three researchers, and scored from 1 to 4.The knowledge scale was based on recommended behaviors identified as learning outcomes.Source: §4.3.1 · p. 18, Table 5 · p. 18
- For inside-building responses, the highest score was 4 points for knowing to drop, cover, and hold under a table and pay attention to falling or unsteady objects and glass.The scale also assigned 3 points for knowing to drop, cover, and hold under a table, and 2 points for knowing to take cover or find a shelter.Source: §4.3.1 · p. 19
- After an earthquake, the knowledge scale credited actions such as checking surroundings, helping others, and attending to hazards; outside, it credited staying clear of buildings and falling objects.The outside-building scale awarded the highest score for staying in an open space away from buildings and not returning until safe.Source: §4.3.1 · p. 19
4.3.2 Self-efficacy
- Self-efficacy was defined as belief in one’s ability to accomplish difficult tasks and assessed before and after training with six emergency statements on a 7-point agreement scale.Statements covered confidence in managing an earthquake emergency, including when the building is damaged, fire is present, an exit is blocked, or hazards are encountered.Source: §4.3.2 · p. 20
4.3.3 Self-reported training efficacy
- Perceived training efficacy and engagement were each assessed after training using one statement on a 7-point agreement scale.The statements asked whether participants could easily learn the recommendations in the virtual game and whether the game was engaging or fun.Source: §4.3.3 · p. 20, §4.3.4 · p. 20
4.4 Procedure
The procedure moved participants from consent and baseline assessments through IVR orientation and a roughly 20-minute training session, then repeated assessments after training.
4.4 Procedure
- At Auckland City Hospital, participants received study information and consent forms covering questionnaires, audio recordings, and recorded in-game actions; they could stop at any time.The experiment was conducted in a meeting room at ACH.Source: §4.4 · p. 21
- Before training, participants completed demographic and experience questionnaires, including self-efficacy, then orally answered a five-question knowledge test.The questionnaires covered drill practice, video-game play, and IVR experience.Source: §4.4 · p. 21
- After IVR controls and safety instructions, participants used a tutorial to learn navigation and interaction, then completed the roughly 20-minute training while seated in a swivel chair.Participants adjusted a head-mounted display for a clear, comfortable view before the tutorial and training.Source: §4.4 · p. 21
- After training, participants repeated the oral knowledge test and completed questionnaires on self-efficacy, training efficacy, and engagement.Source: §4.4 · p. 21
5. Results
Staff and visitors showed significant post-training increases in knowledge and self-efficacy. Participants also reported that the game supported learning and was engaging.
5.1 Knowledge acquisition
- Knowledge of behavioral responses in all three assessed situations increased significantly after training for both staff and visitors.The comparisons are reported in Table 6 and illustrated with knowledge-score boxplots in Figure 10.Source: §5.1 · p. 21, Table 6 · p. 22, Figure 10 · p. 22
5.2 Self-efficacy
- Self-efficacy increased significantly after training for both staff and visitors.Factor scores were estimated from six statements; comparisons are reported in Table 7 and shown as boxplots in Figure 11.Source: §5.2 · p. 23, Figure 11 · p. 23, Table 7
5.3 Self-reported training efficacy
- Participants agreed that the IVR SG facilitated learning immediate earthquake responses and post-earthquake evacuation practice.Mean training-efficacy ratings were 2.52 (SD = 0.59) for staff and 2.65 (SD = 0.66) for visitors; Figure 12 presents the ratings.Source: §5.3 · p. 23, Figure 12 · p. 24
5.4 Self-reported engagement
- Participants perceived the IVR SG as engaging.Mean engagement ratings were 2.28 (SD = 0.79) for staff and 1.79 (SD = 1.26) for visitors; Figure 13 presents the ratings.Source: §5.4 · p. 24, Figure 13 · p. 24
6. Discussion
The discussion reports significant post-training gains in earthquake-response knowledge and self-efficacy, alongside training participants found engaging and applicable. It also considers uneven learning outcomes, shared staff and visitor performance, and motion sickness as a participation constraint.
6. Discussion
- Participants significantly increased knowledge of earthquake and evacuation responses and confidence in handling emergencies after training.The authors conclude that the system increased preparedness.Source: §6 · pp. 24–26
- The 13 behavioral responses span three phases, and participants significantly increased knowledge in all three phases.The phases are the indoor earthquake, pre-evacuation and indoor evacuation, and outdoor evacuation phases.Source: §6 · p. 25
- Knowledge scores were highest for indoor earthquake behaviors, but open-ended answers may be inaccurate because they depend on respondents’ effort.The authors also suggest limited awareness of what to do immediately after an earthquake as a possible explanation for uneven performance.Source: §6 · p. 25, Table 8 · p. 26
- The authors suggest multiple training environments and repetitive rehearsals as ways to improve learning and memory recall.They connect varied learning contexts with recall and cite prior work recommending repeated IVR SG rehearsals.Source: §6 · p. 26
- Staff and visitors performed similarly in knowledge acquisition; the training used shared evacuation behaviors rather than staff-specific behaviors.The authors note that the selected behaviors were also included in national earthquake response procedures.Source: §6 · p. 26
- Participants reported high perceived efficacy and engagement, while six of 93 participants quit because of motion sickness.The 87 final participants were mainly adults, including 11 aged 60–79; over half had never experienced IVR.Source: §6 · p. 27
7. Conclusions
The IVR SG training was associated with significant increases in participants’ knowledge and self-efficacy, and participants found it engaging and easy to use. The study did not compare learning outcomes with traditional approaches or assess knowledge retention over time; participant age and motion sickness also limit what can be concluded.
7. Conclusions
- Participants’ knowledge of earthquake behavioral responses and self-efficacy increased significantly after IVR SG training, which they found engaging and easy to use.The study presents the system as effective for preparedness, while noting that the results do not establish superiority over traditional training.Source: §7 · p. 27
- Learning outcomes were not compared with seminars, posters, or videos, so relative effectiveness versus those approaches remains unknown.Source: §7 · p. 27
- Knowledge retention over time was not assessed, leaving the durability of participants’ newly learned knowledge unmeasured.Source: §7 · p. 28
- Most participants were young adults, fit for children and seniors remains unclear, and about 6% stopped the experiment because of motion sickness.Source: §7 · p. 28