How Can We Improve Initial Public Response During Emergencies? Recommendations from a Systematic Review of Pre-Incident Information

Abstract

This systematic review examines the effect of pre-incident information on public preparedness prior to an emergency or disaster. Preparing members of the public for adverse events can improve self-sufficiency and improve health outcomes, particularly during periods when emergency responders are not immediately available. Twenty-three studies were identified, addressing both natural and human-influenced events. All the studies investigated pre-incident training targeting members of the public rather than specialist responders. The synthesis considered training content, delivery approaches and evaluation methods. The studies included preparation, personal safety, triage, first aid and evacuation in scenarios involving terrorism, fire, earthquake, flood and CBRN events. Pre-incident education generally improves knowledge and intention to act, with higher-intensity and interactive training yielding greater engagement and response. Due to the difficulty of simulating emergencies and disasters, several studies used self-reporting and hypothetical testing, while others attempted to create real-life scenarios. The immediate effects of pre-incident education were generally positive, although many studies tested outcomes theoretically or within a classroom environment. It was also noted that few studies considered retention over the medium to long term; this is a concern as temporal decay may reduce preparedness. This review provides a basis for continued development of public-facing pre-incident education to increase resilience to both terrorist attacks and natural disasters.

1. Introduction

At present, there is a high likelihood of malicious attacks on targets in publicly accessible locations across the world [1,2]. Any such event may result in an emergency incident (or series of incidents) requiring resources beyond those available from local specialised responders and healthcare services [3,4]. Natural hazards such as floods, earthquakes and fires are further potential causes of large-scale incidents, as is accidental or deliberate contamination by chemical, biological, radiological and nuclear (CBRN) agents [5]. Any major emergency may result in a mass casualty incident (MCI) in which multiple members of the public require assistance; urban examples include London Underground, UK (2005), Boston Marathon (USA 2015), Nice (France 2016), Manchester Arena (UK 2017), Colombo (Sri Lanka 2019) and Nova Scotia (Canada 2020).

During any major incident and subsequent potential MCI, there is a high degree of reliance on the actions of either trained responders who attend as members of the professional emergency services (police, ambulance, fire service) and security forces, or from other trained groups such as medical staff, local first-aiders and skilled volunteers. While emergency services are vital in major emergencies, in many cases nearby members of the public are also able to add critical and timely support to the response process. This is most likely to be the case when professional responders are either not immediately available or are unable to react to the scale of the incident and the number of possible casualties, creating a potential care gap between the available specialist response and the needs of those involved.

Despite the difficulties in dealing with these incidents, the provision of pre-incident public information explaining mitigation actions may allow members of the public who are at the scene, sometimes described as ‘zero responders’ [6], to take actions to reduce their own risk, mitigate losses, communicate information about the situation and help others who may be injured [7,8]. Previous studies suggest the existence of an inclination towards pro-social behaviour and spontaneous co-operation between individuals involved in or close to an incident and that this behaviour is likely to be helpful in mitigating negative consequences, e.g., [9,10,11]. Pro-social and cooperative behaviour may be advantageous in both urban and rural settings and may include shelter-in-place, evacuation, aid and emergency medical care. Following incidents involving contamination, practical measures may include removing contaminated clothing providing that it is not stuck to the skin, rinsing with water and, for certain contaminants, using available absorbent materials [12]. For other incidents, priority first aid may be applied when it is safe to do so including cardiopulmonary resuscitation (CPR) [13] and attending to heavy bleeding [14]. The above are examples of basic measures which, when applied appropriately and quickly, may reduce or prevent harm to those affected [15,16].

In areas prone to natural hazards such as flooding, earthquake and wildfire, there is often the provision of information intended to enable citizens to reduce their risk either in the absence of emergency response or in situations which may result in insufficient emergency response being immediately available. Such information, provided by governments and other organisations, may be available via diverse media including websites, printed material and video campaigns and may be targeted towards the general adult population, school-age children or other specific groups such as the elderly or those with young families. Evidence suggests that this pre-incident information may effectively encourage and empower people to be better able to protect themselves and their families in the event of a natural hazard; studies have further suggested that post-event recovery may be enhanced by perceived individual and community self-efficacy [17,18,19]. Pre-incident information may also increase the recipient’s understanding of and confidence in mitigation, creating more willingness to take action [7,20]. Additionally, pre-incident information and enhanced understanding may make it easier for members of the public to communicate and cooperate with emergency responders by creating shared behavioural norms [21].

Recent events and current threats have led to an increased understanding, amongst policy makers, of the need for members of the lay public to be involved in emergency planning. In the United States (US) there has been a long-standing campaign “The first 72 is on you” in several states; this communicates the expectation that, after an emergency, members of the public need to prepare to survive for 72 h before help arrives, e.g., [22,23]. Following recent extreme weather and other emergency events, the US Government has launched a dedicated preparedness website [24] and there has been a move to suggest that people should aim to be self-sufficient for as long as possible [25]. In the UK, as well as in other countries, there have been initiatives to involve the public in planning for both natural hazards [26,27] and deliberate attacks [28]. The UK Government has recently set out a strategic approach to resilience based on prevention, shared understanding of potential civil contingencies and the idea of resilience as a ‘whole of society’ concept requiring both the transparency of communication and the empowerment of the general population [29]. This framework recognises the importance of integrating emergency planners and practitioners with businesses, local governments, voluntary organisations, community groups and the public in order to develop preparedness and resilience. Current UK preparedness campaigns intended to improve the ability of members of the public to respond to an emergency have included ‘Run, Hide, Tell’ which provides information and a strategy for use in marauding terrorist attacks [30], ‘See it, Say it, Sorted’ to improve vigilance and safety on rail networks [31], and ‘Remove, Remove, Remove’ for hazardous materials incidents involving caustic chemicals [32]. Although limited testing of the efficacy of these initiatives has been conducted, encouraging results have been reported [33,34,35].

A review undertaken in 2019 [7] identified studies that examined the efficacy of pre-incident information in improving public preparedness for major incidents and emergencies. This review found that most studies focused on pre-incident preparedness for natural hazards, with little research relating to the effectiveness of pre-incident education and preparedness for terrorist attacks or other malicious incidents. While some of the included studies suggest that members of the public would be either unwilling to engage with pre-incident education or unable to retain the information for later use, e.g., [36], others found that providing information prior to an event may increase public knowledge and confidence, enabling faster deployment of protective actions and potentially saving lives [20,37]. Overall, Carter et al. [7] found evidence for the short-term effectiveness of pre-incident education campaigns but identified few studies that investigated the long-term retention of information. Their review concluded that useful components of pre-incident education include: information about the potential nature and consequences of emergencies, the actions that people can take to protect themselves and others, and the efficacy of these measures.

Since 2019, the COVID-19 pandemic has prompted reviews of emergency plans and brought a new public understanding of preparedness for and response to major incidents. There have also been substantial advances in technology, with an increasing use of immersive technology in education settings [38,39]. Therefore, this review identifies and appraises recent research that evaluates pre-incident information campaigns relating to public emergencies and disasters and evaluates the efficacy of such campaigns for improving public preparedness for mass emergencies. Specifically, this review aims both to identify the types of information that have been found to have practical value and hence should be included in public preparedness campaigns, as well as to identify the information provision methods that are most effective in conveying this information. Carter et al. [7] identified a lack of evidence relating to the longer-term efficacy of emergency preparedness information; therefore, this review considers whether any evidence has subsequently emerged to overcome the shortfall. In line with systematic review update guidance [40,41], this paper uses similar inclusion criteria to those of the previous review, while incorporating new data published since 2019.

2. Materials and Methods

The Ovid research platform was used to search three databases: Embase, PsychInfo and Medline. As studies prior to June 2019 were examined by [7], this update includes relevant papers published from June 2019 onwards. Initial paper selection included three focused searches. Search 1 focused on terms related to the timing of information provided, such as “pre-incident”, “pre-emergency”, and “prior” in order to identify studies in which information was provided prior to the event. Search 2 focused on terms related to the intervention intention, these include “educat*”, “inform*”, and “communicat*”. Search 3 focused on terms related to the type of emergency, such as “chemical”, “CBRN”, “disaster”, and “flood”. The results of these three searches were combined using ‘AND’ logic to obtain the list of papers for review. Supplementary Materials documents the full search strategy.

The criteria for inclusion in this review were that a paper must:

• Examine the effect of pre-incident information on levels of public preparedness prior to an emergency or disaster, whether natural or caused by human action;
• Examine the effect of pre-incident information on improving the public ability to take effective protective actions during mass emergencies and disasters;
• Examine factors that improve the efficacy of pre-incident information;
• Include original data, whether qualitative or quantitative;
• Be available in English;
• Be published in a peer-reviewed journal.

A paper was excluded from the review if it:

• Examined the effect of information provided during or after a disaster or emergency on public behaviour or in response, as the focus of this review is pre-incident information;
• Examined the effect of training or pre-incident information on responder (first responder, healthcare staff, government officials, etc.) disaster preparedness or response, as our concern is information directed at the general public;
• Focused on provision of preparedness information in a warning context rather than on practical preparedness;
• Was unavailable in English;
• Did not contain original data;
• Was not published in peer-reviewed journals (grey literature was excluded from this study).

The review focused exclusively on peer-reviewed journal articles to ensure a consistent level of methodological quality and reliability across the included studies. Grey literature, including government reports, policy papers, and industry publications, was excluded because these sources often lack standardised peer-review procedures and may vary in methodological reporting. Additionally, only English-language publications were included due to translation resource constraints and to maintain consistency in the screening and analysis process.

Following the initial search, paper screening and selection was carried out using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework [42], and documented according to the PRISMA protocol [43]. The first author carried out title, abstract, and full-text screening; a second researcher screened a selection of the papers for consistency. In addition, some papers were screened by a second researcher where there was uncertainty as to whether a particular paper met the inclusion criteria. Approximately 20% of papers underwent additional screening in this manner at either the abstract or final selection stage. Finally, the Mixed Methods Appraisal Tool [44] was used to critically appraise the methodological qualities of each of the selected papers. The MMAT was initially designed in 2006, the revised version used here was developed from further testing, input from methodological experts and a review of critical appraisal tools [45]. All papers were considered to be sufficiently methodologically robust for inclusion and analysed using narrative synthesis. This analytical approach yields more meaningful evaluation given the heterogeneity of methodology and results across the included studies which precludes quantitative comparison. Screening results are reported in Supplementary Materials. This systematic review has been registered during peer review with the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY, Middletown, DE, USA) [46], registration ID INPLASY202630085.

3. Results

3.1. Selection of Papers for Review

The database search generated a set of 13,484 article records. De-duplication reduced this to 9193 records and subsequent title screening produced a set of 712 papers, which were subject to abstract screening. This yielded 92 papers for full-text screening and each of these papers was reviewed in full, resulting in the selection of 22 papers for inclusion in the review. Forward and backward citation analysis of these papers yielded one further reference. The final 23 papers were then used for the subsequent data synthesis. Figure 1 shows a PRISMA flow diagram for this process.

Table 1. Studies included in the review.

Study	Population	Sample Size	Disaster Type	Intervention Type	Preparedness Outcomes	Study Type
Amini 2021 [47], Iran	Women	Intervention n = 60 Control n = 60	Earthquake	Group lecture, Q&A, discussion, written material using Health Belief Model constructs for experiment group. Control group received health care education, written material	Questionnaire on perceived susceptibility, severity, benefits and self-efficacy in earthquake preparedness	Intervention test
Banerski 2020 [48], Poland	General population 18–65 in flood-prone area	Intervention n = 286 Control n = 101	Flood	Video with three different graphics	Knowledge of risks, confidence in taking protective action	Intervention test
Bhandari 2023 [49], Japan	Nepalese immigrants 18+	n = 165	Earthquake	Group lecture, demonstration, discussion	Knowledge perception and practice of earthquake preparedness immediately, 2 weeks, 12 weeks after intervention	Intervention test, longitudinal study
Carter 2021 [35], UK	18+	n = 1000	CBRN	Given information sheet to read	Perception, understanding, preparedness, willingness and ability to take action	Intervention test uncontrolled study
Celik 2019 [50], USA	School personnel	n = 104	All	30 min lecture on SALT triage, brief lecture and group demonstration of haemorrhage control and tourniquet application	Ability to triage, ability to apply tourniquet	Intervention test uncontrolled study
Chittaro 2023 [51], Italy	Adult	Serious game n = 30 Print material n = 30	Any, aiding evacuation of people with disabilities	Comparison of serious game vs. training with printed material	Knowledge of and reported willingness to help people with disabilities during evacuation	Intervention test
Dennis 2023 [52], Germany	Adult	n = 18	CBRN	Written pre-incident information, pictograms, decontamination drill	Self-reported confidence and knowledge of actions to take, opinion on help to be given to/expected of others	Intervention test uncontrolled study
Elkbuli 2019 [53], USA	High school students	n = 232	Bleeding after school attack	Stop the Bleed training programme	Knowledge of and comfort/willingness to use a tourniquet in an emergency	Intervention test longitudinal study
Fatmah 2022 [54], Indonesia	Families with elderly	n = 30	Flood	One day group programme. Printed media (leaflets and flipcharts) and visual media (video)	Knowledge of flood risk, preparedness	Intervention test longitudinal study
Feng 2021 [55], New Zealand	11–15-year-old secondary students	n = 125	Earthquake	Leaflet, serious game	Earthquake drill knowledge and self-efficacy	Test of serious game intervention vs. studying leaflet (no baseline testing)
Gouda 2023 [56], Myanmar	Adults	Project n = 91 Control n = 91	Natural disaster, focus on cyclone	Community disaster awareness training programme including meetings, first aid training, drill, community activity, pamphlet/flyer, visit from expert	Self-reported disaster preparedness and knowledge	Cross-sectional survey after intervention
Heagele 2022 [57], USA	Parents of newborn children	n = 68	All	Booklet on emergency plans and preparedness, checked by nurse then kit provided	Household Emergency Preparedness score	Intervention test uncontrolled study
Hung 2021 [58], China	Adults in rural villages	n = 110	Earthquake, flood	Workshop, group discussion, pictorial presentation	Emergency preparedness knowledge and confidence	Uncontrolled study, intervention, some retest after 7 years
Liu 2023 [59], USA	Adult	Intervention n = 54 Control n = 54	Fire or active shooter	Training video	Perceived self-efficacy and optimal behaviour in emergency interactive video scenario	Intervention test
Liu 2023 [60], USA	Adult	Intervention n = 54 Control n = 54	Active shooter	Training using either video or VR	Perceived self-efficacy and optimal behaviour in emergency interactive video scenario	Intervention test
Mailizar 2023 [61], Indonesia	Year 7 (12–13 years)	Intervention n = 61 Control n = 61	Earthquake, tsunami	Mathematical comic with embedded information	Disaster awareness measure	Intervention test
Marcus 2019 [62], USA	Adult (student)	n = 97	All (tourniquet application)	Face to face training in small groups	Confidence in applying tourniquet	Intervention test uncontrolled study
Noviana 2020 [63], Indonesia	Primary school students grade 5 (10–11 years)	n = 72	Any natural disaster	Information contained in ‘KOASE’ comics given to children for one week	Test of disaster mitigation knowledge	Intervention test uncontrolled study
Salita 2019 [64], Philippines	School teachers	n = 45	All	Lectures, demonstrations	Questionnaire measuring knowledge, attitude, intentions, perceived efficacy	Intervention test uncontrolled study
Schildkraut 2020 [65], USA	School staff and pupils	n = 5562 baseline n = 4081 first test n = 4283	Lockdown drill	Group training with presentations, videos and discussion, Standard Response Protocol Extended (SRP-X) used	Lockdown behaviour in test, expressed perception of safety and preparedness	Intervention test uncontrolled study
Stephens 2023 [66], USA	Residents of rural community	n = 70	Wildfire	Pre-planned evacuation drill	Self-reported perceived susceptibility, preparedness, intent to share preparedness information	Case study (no pre-drill data)
Susmini 2022 [67], Indonesia	Community members in hurricane-prone area	Intervention n = 40 Control n = 40	Hurricane	Two weeks of disaster preparedness phone app (N = 40), information given in meetings (N = 40)	Self-reported intended response to a future hurricane	Intervention test comparing education methods
White-Lewis 2021 [68], USA	High school students	Intervention n = 31 Control n = 25	Any	Lecture, small group rotation through first aid practise and bioterrorism information stations	Emergency preparedness Information Questionnaire	Intervention test

documents the key attributes of each of the 23 studies selected for inclusion in this review. Of the 23 studies identified, 19 used quantitative methods and four employed a mix of qualitative and quantitative methods. Public preparedness for a range of threats and disasters was examined, encompassing both natural hazards and human acts. Types of threats and emergencies included: earthquake (n = 4), flood and other weather-related events (n = 4), wildfire (n = 1), terrorism (n = 3), accidental spillage of CBRN material (n = 3), and preparedness for emergencies in general (n = 5). Some studies considered more than one single type of threat and others focused on response across threat categories including first aid (n = 5), evacuation (n = 4) and lockdown (n = 1).

The studies were conducted in several parts of the world including the USA (n = 9), Indonesia (n = 4), Iran, Poland, Japan, the UK, Italy, Germany, New Zealand, Myanmar, China and the Philippines (all n = 1). While the most studied participant group was the general adult population (n = 14), other groups studied were children and teenagers (n = 6), families with elderly members (n = 1), parents of neonates (n = 1) and women (n = 1).

The majority of the studies (n = 22) tested the effectiveness of an educational intervention on the preparedness of participants. One study used a cross-sectional design to contrast preparedness in villages which had had pre-incident education with similar populations that had not been provided with pre-incident education.

3.2. Overview of Results

The results of this review follow four main themes:

• Evidence for the overall effectiveness of public pre-incident education;
• The type of information that is useful for this education;
• The nature and efficacy of the delivery methods used;
• The characteristics of pre-incident information relating to terrorism or human error emergencies compared with that relating to natural hazards.

Details of the preparedness outcomes, types of information and methods of providing information included within each study are documented in

Table 2. Study preparedness outcomes.

Preparedness Outcome	Studies
Knowledge	Carter, Gauntlett [35], Amini, Biglari [47], Banerski, Abramczuk [48], Bhandari, Rahman [49], Chittaro [51], Dennis, Weston [52], Elkbuli, McKenney [53], Fatmah [54], Feng, Gonzalez [55], Heagele and Nurse-Clarke [57], Hung, Macdermot [58], Mailizar, Johar [61], Marcus, Pontasch [62], Noviana, Kurniaman [63], Salita, Liwanag [64], Susmini, Feri [67], White-Lewis, Beach [68]
Taking action	Bhandari, Rahman [49], Feng, Gonzalez [55], Marcus, Pontasch [62], Celik, Mencl [50], Liu, Zhu [59], Liu, Becerik-Gerber [60], Schildkraut and Nickerson [65], Stephens, Powers [66]
Intention/willingness to take action	Carter [37], Banerski, Abramczuk [48], Dennis, Weston [52], Elkbuli, McKenney [53], Fatmah [54], Heagele and Nurse-Clarke [57], Susmini, Feri [67], Stephens, Powers [66]
Confidence/perceived self-efficacy	Carter [37], Amini, Biglari [47], Chittaro [51], Dennis, Weston [52], Elkbuli, McKenney [53], Feng, Gonzalez [55], Marcus, Pontasch [62], Salita, Liwanag [64], White-Lewis, Beach [68], Liu, Zhu [59], Liu, Becerik-Gerber [60], Schildkraut and Nickerson [65], Stephens, Powers [66], Gouda and Yang [56]
Engagement	Dennis, Weston [52]
Perceived usefulness	Carter [37], Dennis, Weston [52], Salita, Liwanag [64]

,

Table 3. Study information types.

Information Type	Studies
Preparedness/protective actions that may be taken	Carter, Gauntlett [35], Amini, Biglari [47], Banerski, Abramczuk [48], Bhandari, Rahman [49], Chittaro [51], Elkbuli, McKenney [53], Fatmah [54], Feng, Gonzalez [55], Heagele and Nurse-Clarke [57], Hung, Macdermot [58], Mailizar, Johar [61], Marcus, Pontasch [62], Noviana, Kurniaman [63], Salita, Liwanag [64], Susmini, Feri [67], White-Lewis, Beach [68], Celik, Mencl [50], Liu, Zhu [59], Liu, Becerik-Gerber [60], Schildkraut and Nickerson [65], Stephens, Powers [66], Gouda and Yang [56]
Disaster effects	Amini, Biglari [47], Banerski, Abramczuk [48], Bhandari, Rahman [49], Fatmah [54], Heagele and Nurse-Clarke [57], Mailizar, Johar [61], Noviana, Kurniaman [63], Salita, Liwanag [64], Susmini, Feri [67], Celik, Mencl [50], Gouda and Yang [56]
Disaster severity	Amini, Biglari [47], Banerski, Abramczuk [48], Fatmah [54], Heagele and Nurse-Clarke [57], Mailizar, Johar [61], Salita, Liwanag [64], Susmini, Feri [67], Gouda and Yang [56]
Disaster likelihood	Banerski, Abramczuk [48], Fatmah [54]
Positive effect of taking action	Carter, Gauntlett [35], Amini, Biglari [47], Chittaro [51], Feng, Gonzalez [55], Mailizar, Johar [61], Salita, Liwanag [64], White-Lewis, Beach [68]
Self-efficacy	Feng, Gonzalez [55], Heagele and Nurse-Clarke [57], Hung, Macdermot [58], Noviana, Kurniaman [63], Salita, Liwanag [64], Liu, Becerik-Gerber [60]
Likely stress during/after incident	Mailizar, Johar [61], Salita, Liwanag [64], White-Lewis, Beach [68], Gouda and Yang [56]
Overcoming barriers to action	Amini, Biglari [47], Chittaro [51], Dennis, Weston [52], Salita, Liwanag [64]
Preparedness of others	Fatmah [54], Gouda and Yang [56]
Sources of further information	Bhandari, Rahman [49], Heagele and Nurse-Clarke [57]

and

Table 4. Study training methods.

Training Method	Studies
One-to-one	Heagele and Nurse-Clarke [57], Stephens, Powers [66]
Small group	Marcus, Pontasch [62], Salita, Liwanag [64], White-Lewis, Beach [68], Stephens, Powers [66]
Large group	Amini, Biglari [47], Banerski, Abramczuk [48], Bhandari, Rahman [49], Elkbuli, McKenney [53], Fatmah [54], Hung, Macdermot [58], Susmini, Feri [67], Schildkraut and Nickerson [65], Gouda and Yang [56]
Demonstration/presentation	Bhandari, Rahman [49], Elkbuli, McKenney [53], Hung, Macdermot [58], Marcus, Pontasch [62], Salita, Liwanag [64], White-Lewis, Beach [68], Celik, Mencl [50], Schildkraut and Nickerson [65], Stephens, Powers [66], Gouda and Yang [56]
Practise of technique	Elkbuli, McKenney [53], Hung, Macdermot [58], Marcus, Pontasch [62], White-Lewis, Beach [68], Celik, Mencl [62], Gouda and Yang [56]
Written material	Carter, Gauntlett [35], Amini, Biglari [47], Bhandari, Rahman [49], Chittaro [51], Dennis, Weston [52], Fatmah [54], Feng, Gonzalez [55], Heagele and Nurse-Clarke [57], Mailizar, Johar [61], Noviana, Kurniaman [63], Stephens, Powers [66]
Online app	Susmini, Feri [67]
Serious game	Chittaro [51], Liu, Zhu [59], Liu, Becerik-Gerber [60]
Virtual Reality	Feng, Gonzalez [55], Liu, Becerik-Gerber [60]
Discussion/Q&A	Amini, Biglari [47], Bhandari, Rahman [49], Heagele and Nurse-Clarke [57], Marcus, Pontasch [62], Salita, Liwanag [64], Schildkraut and Nickerson [65], Stephens, Powers [66]
Video	Banerski, Abramczuk [48], Fatmah [54], Salita, Liwanag [64], Liu, Zhu [59], Liu, Becerik-Gerber [60]

respectively.

3.3. Evidence for the Overall Effectiveness of Public Pre-Incident Education

Twenty-two studies examined the effect of pre-incident education interventions on public preparedness using a pre–post study design, while one [56] used a natural experiment approach to compare two participant groups (villages) participating in education initiatives with two that did not have this provided. All the studies included in this review described some degree of improved preparedness as a consequence of pre-disaster education. Many studies tested intervention effectiveness with questionnaires or self-reporting. Seventeen studies examined the effect of education on knowledge of potential emergencies and how to prepare for them, while fifteen considered actual (7) and/or intended (8) protective action. Fourteen studies measured the effect of the intervention on participants’ confidence in their ability to take protective action; all the studies found that participants felt more confident after pre-incident information had been delivered.

Three studies [35,52,64] performed detailed investigations of participants’ reflections on the pre-incident education process and its impact on their self-efficacy. Carter et al. [35] studied public perception of education in response to hazardous substance contamination (n = 1000). Their study found that members of the public considered the information useful, thought the recommended actions would be effective, and reported that they would be able and willing to implement the recommendations. In particular, this study collected data on the receptiveness of participants to the provision of pre-incident information, finding that only 4% of the participants indicated that they would prefer not to receive the information. Dennis et al. [52] used a small participant group (n = 18) to investigate responses to pre-incident information relating to a decontamination drill; this study also noted participant responses that they would feel comfortable in carrying out the recommendations and that this would be helpful. Salita et al. [64] delivered lay responder disaster training to participants (n = 45), finding that perceived knowledge and understanding of any likely threat increased while fear decreased. However, self-efficacy did not increase by the intervention employed in this study, possibly indicating that participants’ intention to act had not increased.

Results from the selected studies suggest that any type of pre-incident disaster education is likely to improve preparedness compared with no intervention. Knowledge and intention were frequently found to be increased by education, e.g., [37,52,53], although there is less data available on the extent to which engagement in preparedness behaviours is increased. While many of the studies in this review used questionnaires, knowledge tests or self-reported confidence to test the efficacy of training, some (n = 8) conducted practical testing of participants’ ability to perform actions after intervention. For example, Celik et al. [50] and Marcus et al. [62] supplied training in haemorrhage control, both reported that the majority of previously untrained participants were able to demonstrate the appropriate application of a tourniquet to a mannequin in a training room. Celik also trained volunteers to perform triage on casualties and found during theoretical scenario testing that a level of accuracy similar to that of emergency responders could be attained.

Virtual Reality (VR) or immersive video applications may offer the possibility of more realistic testing. Feng et al. [55] tested earthquake response within a VR environment designed to emulate a realistic situation, while Liu, Zhu et al. [59] and Liu, Becerik-Gerber et al. [60] used an immersive video with embedded questions to test response to a fire or active shooter. These studies measured video and training effectiveness by developing an immersive online environment in which participants’ virtual actions during an emergency could be analysed, all reported positive results from training.

Two studies tested pre-incident information by attempting to create a real-world scenario that mimicked an emergency with the aim of eliciting realistic responses from participants. Stephens et al. [66] documented a wildfire drill case study with evacuation, noting high satisfaction with the process and confidence after the drill. A realistic simulation was also devised by Schildkraut and Nickerson [65] who provided lockdown training then simulated school active shooter situations by triggering full-scale lockdown drills involving over 3000 pupils and 1000 staff. This test mimicked an actual attack as closely as possible within the school setting and indicated effectiveness of the pre-incident education, reporting a large increase in both drill compliance and participant confidence after training.

In considering the efficacy of pre-incident education, the time between delivering training and testing may also be relevant. The majority of intervention studies tested participants immediately after the intervention (n = 18) or within 2 weeks (n = 2). The lockdown drill study by Schildkraut and Nickerson [65] tested up to several weeks post-training and reported high levels of knowledge retention although the study did not compare these results with knowledge immediately after training. Bhandari et al. [49] tested immediately after training and again after two and 12 weeks, finding an immediate increase in both knowledge and earthquake preparedness which then increased further after two weeks before reducing but remaining above baseline after 12 weeks. A much longer-term study by Hung et al. [58] retested participants after 7–8 years and found little increase in disaster preparedness from baseline at this point; the drop is likely to indicate time-decay of the training effect although the researchers commented that results may also be affected by life events and experience during the intervening time.

3.4. Type of Pre-Incident Information Used in the Studies

The nature of pre-incident education in the 22 intervention studies varied widely. The studies included: information about the preparedness and protective actions that people can take prior to and during disasters (n = 22); information about the effects of different types of disasters (n = 11); information about the efficacy of different prevention or protection measures, and the importance of taking such measures (n = 7); information designed to increase perceptions of the severity of different types of disasters (n = 8); information designed to increase perceptions of likelihood of different types of disasters (n = 2); information designed to increase self-efficacy in relation to taking preventative or protective actions (n = 7); information about likely stress reactions to disasters (n = 3); and information about how to overcome different barriers to taking preparedness actions (n = 3).

Most studies communicated multiple information types during the intervention, all included information about both the effects of disasters and the actions that can be taken to mitigate these effects. Three studies included information on preparedness and protective actions, information about the effects and severity of different types of disasters and the positive effect of taking action [47,61,64]. These three studies all reported positive intervention results, although Mailizar [61] identified the potential influence of demographic differences between urban and rural populations.

Several of the papers in this review included information about how to help others as well as oneself to survive an emergency, with three specifically focusing on meeting the needs of vulnerable populations and reporting successful intervention in carers of these groups. Fatmah [54] offered a one-day training meeting on flood preparedness to the families of elderly people, applying a self-reported questionnaire before and after the session and finding that knowledge of flooding and how to look after the family, including the elderly, was increased after the session. Heagele and Nurse-Clarke [57] worked with parents of neonates using a nurse-led training intervention and with supplied disaster kits, finding that household emergency preparedness scores increased by the end of the process. Chittaro [51] trained volunteers (n = 60) in how to assist those with sight or mobility limitations to evacuate in an emergency; this study reported a significant increase in participants’ knowledge of how to help people with these disabilities.

One study [56] reported a cross-sectional natural experimental study involving four villages, two of which received preparedness training interventions during an aid project over a period of two years, and two of which did not. The pre-incident information provided included the areas of preparedness for an event, and the effects and severity of a disaster, and was offered through a number of projects over a two-year period to two of the four villages studied. Residents of the villages that received training reported greater knowledge and preparedness than residents of the villages that did not receive training.

3.5. Methods of Providing Information

A variety of training methods were used in the studies including paper-based training (e.g., leaflets, pamphlets, posters), information films, group sessions, practical demonstrations, serious games, interactive video and VR. As previously noted, most of the studies found that it is better to provide information in any form than to not provide it at all. Many of the studies in this review used more than one method of delivery for pre-incident training materials, but most did not directly compare methods.

Studies that included hands-on elements in the training, e.g., [62,65,68] reported results suggesting that these interventions successfully taught practical skills in the areas of bleed control, lockdown behaviour and first aid. However, these studies did not compare this practical pre-education with purely theoretical training in the same skills.

Some recent research has included novel delivery methods for pre-disaster education with the goal of increasing engagement and immersion and hence potentially enhancing learning outcomes. Susmini et al. [67] asked participants to use an interactive app to play disaster simulation games related to hurricane risk for two weeks, while a control group received home visits and conventionally presented information. Disaster knowledge and reported intention to act in the event of a hurricane before and after the intervention were then compared. The researchers found that both groups achieved improvements post-intervention, but that the app group scored significantly higher on knowledge and intention to act.

Two of the studies selected used comic book formats to educate children in disaster awareness and safety, both focused on disasters arising from natural hazards. Noviana et al. [63] conducted a longitudinal study with primary-school-aged participants in Indonesia, finding that students’ knowledge of disaster mitigation improved significantly after exposure to comics. Another Indonesian study of disaster education using comics among urban and rural schoolchildren [61] found no significant difference in disaster awareness after using comics versus classroom-based learning, although the authors commented that there were disruptions in learning delivery due to the study taking place during the COVID-19 pandemic. Banerski et al. [48] observed responses to three different flood awareness videos with increasing levels of visual enhancement and found results that suggest viewing a 3D visual representation of a flood results in greater motivation to prepare for the threat than viewing a video of somebody talking about flood threat. This study also considered whether the effect of greater cognitive load in the more immersive 3D representation might have a negative impact on the ability to respond to the message but found that this was not the case.

Serious games are games that are designed to be used for education, simulation or another purpose beyond simple entertainment [69,70]. Training via serious games was compared with more conventional training in two studies [51,55], both found that information imparted through serious games appeared to be more effectively utilised and better retained. Feng et al. [55] used virtual reality for the serious game design and found VR participants to have greater feelings of self-efficacy than those who were given information in written form. The study went on to examine the timing of feedback during and after the game and found that post-game assessment may be more effective, possibly because this allows more time for performance reflection than in-game feedback does. Similarly, Chittaro [51] found evidence that participants using a serious game displayed a greater level of knowledge retention than those who used printed materials. Liu, Zhu et al. [59] used video training for active shooter and fire evacuation scenarios and found this to be more effective than paper-based materials in improving confidence and reducing following behaviour, although there was no significant difference in outcome as measured by pre-evacuation and evacuation time or decision making. A further study [60] extended this work by comparing video training with two levels of VR training, concluding that all methods improved perceived self-efficacy but that VR training produced a better survival outcome during testing as indicated by faster decision making and a greater tendency to run from an attacker. This result suggests that immersion and interactivity may produce better learning outcomes.

3.6. Consideration of Emergencies Related to Natural Hazards Versus Those Resulting from Human Action

Table 1 indicates the emergency type and intervention outcome measured for each of the selected studies. Twelve of the studies evaluated pre-incident information about general preparedness for response to mainly natural hazards such as floods, tsunami, earthquakes and hurricanes. These studies recruited participants from geographic areas more likely to be subject to such hazards and applied interventions that included risk recognition, preparation of an emergency plan, preparation of supplies, evacuation from the area and pre-event organisation of essential information including official documents and telephone numbers. Studies of this type focus on clear local risks and emphasise behaviour likely to enhance safety during, immediately after, and in the days following a natural hazard-related disaster. These events may not be predicted with precision, but can be planned for in order to increase individual, family and community resilience [71].

The other eleven studies considered specific actions which may mitigate the immediate effects of an unforeseen potential emergency likely to be related to non-natural hazards including terrorism, crime (school shooter) and accidents. While these events are likely to be hard to predict in terms of impact and who would be affected, lives may be saved by prompt action in the form of emergency first aid, evacuation and situation control.

Five studies [50,53,62,64,68] considered first aid and triage training. The studies undertaken by Elkbuli et al. [53] and Marcus et al. [62] trained members of the public in tourniquet application and used self-reported measures of knowledge and confidence to confirm the success of these interventions. Both studies reported positive results after training. Marcus et al. [62] found that 94.8% of participants reported that they would be comfortable applying a tourniquet. Elkbuli et al. [53] recorded a 95.2% level of confidence among trainees with 74.7% of those trained in tourniquet use expressing an intention to obtain their own tourniquet kit, although only 16.4% had actually done so when followed up one month later, suggesting a potential intention–behaviour gap. Celik et al. [50] taught both direct pressure and tourniquet techniques and checked the ability of participants to apply these to a mannequin; the pass rate for this test was 96% after training. These studies suggested that haemorrhage control techniques may be successfully taught to non-medical participants and potentially utilised in the event of an MCI. Triage training [50] is a potential extension of this as it enables zero responders to identify the casualties most in need of urgent first aid. Salita et al. [64] trained participants in basic triage and first aid, reporting an increase in confidence after the training, while White-Lewis et al. [72] delivered training including general first aid, triage and bioterrorism response to a group of students from under-resourced high schools and found that self-reported knowledge of the topics was significantly increased.

Evacuation from the site of an emergency is frequently an urgent requirement in case of an attack or accident; this was considered in four studies. Liu, Zhu et al. [59] and Liu, Becerik-Gerber et al. [60] examined video-based and VR training as preparation for active shooter and fire emergencies, reporting results that suggest that immersive training can facilitate self-efficacy and potentially improve participants’ ability to escape an active shooter attack. The focus of Chittaro [51] was on evacuation from an unsafe space including the provision of help to people with disabilities, statistically significantly greater knowledge was recorded after training. Feng et al. [73] also found knowledge and self-efficacy to be generally increased by the provision of pre-incident information, although, as outlined in the previous section, results were also influenced by the training method. Lockdown of all or part of a site may also be required, this was studied by Schildkraut and Nickerson [65] who found that training had a strongly positive effect on performance in lockdown drills.

Specific, prompt actions are required in the case of a CBRN incident to mitigate damage to both those affected and anybody helping them. Two studies examined preparedness in relation to CBRN threats. Carter et al. [35] tested responses to the ‘Remove’ campaign describing action to be taken during a hazardous chemical incident, while Dennis et al. [52] provided information to members of the public before a live hazardous chemical incident field exercise. Both studies found that CBRN-related information is welcomed by members of the public and appears to significantly increase knowledge and confidence caused by terrorism or human error.

4. Discussion

4.1. Is Public Pre-Incident Education Effective?

The studies included in this review are heterogeneous, considering a wide range of emergency scenarios and mitigating actions. However, consistent results and common features across studies allow the emergence of several themes which are documented in this section.

Short-term, pre-incident training prior to emergencies appears to be effective in improving:

• Knowledge;
• Intention to take mitigating action;
• Likelihood of actual mitigating action being taken;
• Confidence in taking action;
• Belief that taking action is helpful.

Education is particularly effective for improving self-reported knowledge and behavioural intentions, although self-reported outcomes may not always correlate with objective behaviour [74] and there may be a gap between intention to act and actual behaviour [75]. An example of this intention–action gap was noted by Elkbuli et al. [53] who studied first aid training and found that only a small proportion of trained participants who expressed an intention to obtain a tourniquet had done so within one month of receiving training; lack of a tourniquet or suitable alternative would render this training far less useful even if the participant was confident of their skills. Within the studies included in this review, the positive effect of intervention on theoretical measures such as intent and perceived self-efficacy is more substantiated than the effect on practical measures such as success in a test task and actual emergency preparedness; this is a similar finding to that of a previous systematic review [7]. While objective measurement of actual behaviour in an emergency is challenging, it is this practical outcome that ultimately determines whether training has succeeded in preparing participants to act when required.

Incidents occur after delays of varying length post-training and the effects of training tend to diminish over time for both emergency [76] and more general education [77,78]. The sixteen studies that evaluated the effectiveness of pre-incident education immediately or shortly after delivery reported that the intervention was effective. However, training decay effects over time were not generally examined with only three of the studies including analysis of retention levels beyond a two-week window. One long-term study reported substantial decay over 7–8 years, suggesting the desirability of maintaining knowledge and beliefs through continuing education [58]. The short-term follow-up earthquake-related study by Bhandari et al. [49] reported an unexpected increase in both knowledge and preparedness after two weeks compared with immediately after training; the researchers comment that this may be due to participants being able to process and practise the training. The possible consolidation of learning during a time when participants have time to process the information is also suggested by Feng et al. [55] who considered training effects when feedback was given during or after training and suggested that the superior knowledge found in a delayed feedback group may be a result of additional reflection time. However, both Hung et al. [58] and Bhandari et al. [49] suggest environmental factors as parameters that may affect learning consolidation in their studies.

The focus on immediate evaluation within most of these studies results in little information being currently available on the effects of pre-incident information over time. This lack of data creates evaluation challenges as emergency incidents may occur some time after training has been received. Determining the effect of time and other factors on the retention of training would aid the development of effective training delivery.

4.2. What Information Should Be Included?

Emergencies and disasters leading to mass casualty incidents are often unpredictable, difficult to manage and therefore extremely frightening. Nevertheless, there is almost always some action that can be taken by those involved which will mitigate the harm to themselves, their families and others. The emergency preparedness measures in the studies selected for this review included having a disaster kit at home (to allow survival for several days without outside help) [47,57], having an evacuation plan [51,59,60], and being aware of sources of information should an incident occur [66]. In an emergency situation, members of the public can take appropriate protective action and ideally should be encouraged to make dynamic decisions based on their knowledge of the type of disaster and the current situation; actions may include escape, physically protecting themselves and their family [54,57], first aid, e.g., [64,68], and communication with authorities and responders. The studies analysed in this review underline the necessity to educate people about protective actions that they can take before and during an incident.

Given the need for prompt action, the public should be informed before any incident about what the dangers are, what actions they should take during and after the incident, and the efficacy of taking such actions [35]. The results of previous reviews also emphasise this, finding that people want to have information about what they can do and how it will help in order to improve their confidence and willingness to act [7,79,80]. The current study supports the findings of other reviews, confirming that public preparedness and willingness to act are likely to be enhanced by increased knowledge of the effects of different incident types, their attributes and effective approaches to them.

Provision of information about the likelihood and severity of a particular situation allows members of the public to understand what may happen and its impact; evidence suggests that people want to receive this type of information, and that the provision of this type of information may motivate people to learn more about what they can do to mitigate damage in an emergency. Several studies in this review noted a significant increase in preparedness after education, including information on likelihood and severity of natural hazards, was provided [47,48,49,54,56,58,66]. Carter et al. [35] asked participants specifically about their preference, finding that a large majority of participants expressed a wish to receive information about a potential hazardous material threat. It may also be that certain groups are particularly interested in disaster preparation and that this inclination can be encouraged to ensure that vulnerable members of the population are protected. For example, the studies by Heagele and Nurse-Clarke [57] and Fatmah [54] worked with new parents, and families with elderly members, and both reported high knowledge and preparedness scores after intervention. These two groups are likely to be motivated to protect their vulnerable family members which may lead to a high degree of engagement.

Training participants and equipping them to potentially mitigate the negative effects of an emergency is important to the process of helping them to protect themselves. All studies in this review focus on how members of the public can help themselves and others, all studies report broadly positive results for this training. Self-efficacy and confidence appear to be particularly boosted by training, although care is required in the interpretation of this finding as some studies suggest that confidence and intent are not always fully correlated with practical skills or actions, e.g., [53,55,60]. The success of studies that deliver and measure practical skills such as first aid, evacuation and lockdown drills suggest that these interventions do appear to be successful at least in the short term, e.g., [50,62,65,66,72].

Members of the public also need to be confident that their actions will be appropriate and that barriers to action can be overcome. This may be achieved by discussing actions and barriers [47], ensuring that the intervention explicitly covers these areas [35] and providing specific instructions, such as how to prepare an emergency kit [57].

Socioeconomic and cultural circumstances may also shape the design and implementation of pre-incident education. Studies included in this review used designs focused on both natural disaster awareness and non-natural hazards such as terrorism and other malicious acts. Design and outcome differences between these studies may reflect acceptance of educational interventions by local communities and potentially influence delivery approaches. Most of the included studies provided limited information on cultural and socioeconomic context, restricting the ability of this review to compare intervention effectiveness across different settings in a systematic manner. However, contextual differences should be considered during the design of scalable and culturally appropriate education programmes that can be deployed across diverse regions.

4.3. How Should Information Be Provided?

As noted above, presentation styles that employ greater immersion in the form of videos, VR, comics or games appear to improve interest, concern and intention to act although it is important to ensure that this translates to behaviour in real-life emergencies [51,55,59,60,67]. The research of Banerski, Abramczuk [48] illustrates that details within a particular presentation medium may also affect the outcome; the researchers found that providing dynamic visual information about flooding risk, showing what it would look like and suggesting ways to reduce harm, appeared to elicit more concern and resulted in greater motivation than showing the same information in static form.

Active involvement by participants also appears to consolidate knowledge and lead to a better learning outcome [50,57,62,65,66,72]. A stronger intervention effect may also be created by presentations involving small groups and question and answer sessions that encourage participants to contribute and hence learn more, feel more confident and be more motivated to act [47]. Confirmation that higher-intensity active involvement education elicits a stronger learning response is in accord with the findings of [7] and of other previous studies [81]. However, the studies in this review have generally shown that any pre-incident information is better than none, so practical constraints of audience, availability and resources should also be considered in planning an educational intervention. To better understand the effects of different types of presentation over time, further research is required with longer periods before re-testing. In addition to economic considerations, demographic differences between populations may also inform the most effective method of communicating pre-incident training, for example, one study suggested that students in some rural populations experienced learning barriers due to lower engagement with education, while participation in urban areas was more accessible [61].

4.4. Considerations When Developing Information Campaigns for Terrorism and Other Human-Influenced Mass Casualty Incidents

Terrorist and other malicious incidents are extremely frightening and usually dynamic in nature requiring a fast response and awareness of the dangers and potential mitigating actions. It is therefore particularly important for members of the public to be prepared and know what to do to protect themselves and others in the case of an active shooter, attack with a toxic or corrosive substance, and other types of incidents with multiple potential casualties. Information supplied needs to be concise, accurate and easy to remember and put into practice, this may include immediate personal safety, evacuation, triage, first aid and decontamination. The studies in this review consider the delivery of pre-incident information across different potential scenarios, both natural and human-created, and it is likely that their findings are transferrable to MCI emergency situations, with common factors evident in the efficacy of training delivery and content. Two of the studies in this review [35,52] target CBRN agents, their conclusions concur with those found by previous researchers [7], confirming that pre-incident information campaigns appear effective for these situations, providing that clear information is provided about the unique nature of these agents, the differences between them, mitigating actions and the need for speed in deploying help. Other studies which focused on terrorism and other human-created potential MCI events similarly stressed the need for speed and attention to procedures, e.g., [50,65,68].

One notable recent addition is the use of ‘Stop the Bleed’ training [14] and similar programmes that are designed to enable lay people to carry out emergency first aid when a casualty has a major haemorrhage. Haemorrhage is a major cause of death after trauma and requires urgent attention [82], so the successful introduction of ‘Stop the Bleed’ and other similar procedures has the potential to save lives with fast treatment [50,53,62]. These procedures are examples of ways in which trained members of the public can provide timely and potentially life-saving help before emergency responders arrive at the scene of an incident. Furthermore, identifying these patients and treating them as fast as possible is a priority for saving lives in an MCI, suggesting that triage skills are also important for those on the scene [50,72].

4.5. Implications for Further Study

The findings of this review indicate that effective pre-incident education can be planned and delivered for natural hazards, terrorist threats and other mass casualty incidents. These encouraging findings are in line with the conclusions of previous research, although it is noted that many of the studies rely on self-reported intention and confidence which are not necessarily reliable indicators of actual behaviour in an emergency.

A key requirement of pre-incident education is that it prepares recipients to act effectively in the event of an actual emergency. As shown by [53], confidence and knowledge are neither necessarily robust nor objective indicators of behaviour. This is an area in which further research is recommended, such research should include the consideration of personal, cultural and economic factors that may influence inclination and ability to act. Research involving large scale ‘real life’ testing of behaviour in emergency scenarios is limited because it is complex, expensive and resource-intensive, although a number of studies did achieve this to some extent. Hands-on exercises [50] or live-action simulation [52,65,66] may add realism and enable practical skills to be evaluated. A potential improvement on self-reported data that does not incur the same logistical burden as live-action simulation, is to use immersive virtual environment technology (IVET). Some of the studies within this review required participants to work within an immersive video, game or VR environment, e.g., [55,59,60], and this may provide a workable proxy for real-life testing [83,84,85]. Testing within an immersive environment is likely to yield more reliable results if it is designed and calibrated to induce a feeling of being in a real-life emergency scenario, requiring participants to behave accordingly and stimulating realistic cognitive responses [85,86]. An IVET-based evaluation offers the possibility of allowing members of the public to interact with different virtual reality scenarios while experimental parameters are closely controlled and actual behaviour can be observed, providing a reliable proxy for behaviour during an emergency and overcoming the inherent subjectivity of self-reports. This suggests IVET as a potentially useful assessment tool, the use of which becomes more practical as the technology develops, and more realistic environments are available. Future applications should consider the cost–benefit ratio of these new technologies as developmental expense may be offset by economies of scale during deployment and higher quality, replicable results.

Few studies have been identified that consider the effect of pre-incident education beyond the immediate or very short term. In the real-world it is unlikely that members of the public would be trained immediately prior to encountering an emergency situation, so it would be preferable to test a training intervention after some weeks or months have elapsed. This delayed testing approach would contribute to clarifying the likely retention of information after education has been delivered. Such studies may also generate evidence-led recommendations for optimum timing of training refreshment in order to retain sufficient levels of readiness. Findings have previously indicated that more active interventions in which participants practise what they have learned may lead to more reliable retention, so future studies could also compare delivery methods and identify those most likely to lead to long-term improvements in preparedness.

Most of the studies in this review consider preparation for natural hazards. Although it is likely that many of the results of these studies may be transferrable to pre-incident education for terrorist and other mass casualty incidents, the nature of these events is sufficiently different to indicate that further research is required to examine effects of preparedness for these types of emergencies more specifically.

4.6. Limitations

Best efforts were made by the authors of this paper to ensure that this review was as systematic and complete as possible. However, the nature of this area of study, together with resource and timing constraints, introduce a potential element of subjectivity in the screening of papers. Procedures employed to minimise this include the re-screening of 20% of papers by the second and fourth authors who were also available for consultation in unclear cases.

The decision to only include papers published in English may introduce an element of cultural bias and it is noted that, similar to the first review, many of the studies included were conducted in the USA (8), Europe (4) or New Zealand (1). The studies included from other parts of the world mainly focused on the preparation for natural hazards which may complicate comparisons, but it is clear that many of the findings of the papers included here are replicated across cultures. It is important for any future research and intervention development in this area to consider the cultural, social and economic attributes of the population in order to achieve maximum public engagement [87,88].

This review included only published studies, suggesting that the possibility of publication bias cannot be completely excluded as studies with non-significant findings may not enter the published literature [89]. However, robustness of the synthesis findings is supported by the broad search strategy used and the consistency and methodological quality of the included studies.

5. Conclusions

This systematic literature review builds upon previous research by reviewing studies that evaluate the development and delivery of pre-incident information campaigns for critical incidents and emergencies, with the specific goal of generating recommendations for the development of pre-incident information for mass emergencies. While the review highlights the need for further research to establish the optimum method of providing pre-incident information, there are some common results across the studies. Based on the findings, we therefore make the following recommendations:

5.1. Recommendations for the Provision of Pre-Incident Information

Pre-incident public information should include:

• Information about the need to prepare and the importance of acting if an incident occurs;
• The need to act quickly if involved or a close bystander to the incident;
• Actions that people can take to mitigate the effects of an incident for themselves and others while maintaining safety precautions;
• A clear message that taking prompt action can save lives and minimise injury.

Interventions should be high intensity where possible, as demonstration, practise, small group discussion and interactive video are all likely to elicit greater engagement and learning. However, this must be balanced against the cost and difficulty of training a large population using high-intensity methods.

Consideration should be given to targeting people in particular demographic groups who may be most motivated to receive and use pre-incident education.

5.2. Recommendations for Future Research

Finally, we recommend the following future research agenda:

• Further research should be conducted specifically on the provision of pre-incident information relating to emergencies involving terrorism and other human-influenced events. This will complement the existing larger body of work relating to natural hazard-related disasters.
• Triage of casualties and emergency first aid are immediate actions that, if taken by those on the scene of an emergency, can save lives. Future research should include investigation of pre-incident education in these areas to ascertain whether trained members of the public can be effective zero responders.
• Research should establish the most effective method of delivery of pre-incident information to ensure that members of the public are able to respond to a sudden emergency, including consideration of the information requirements of different demographic groups.
• Further research should be conducted to investigate the retention of information and maintenance of preparedness in the medium term (3+ months) and beyond following an intervention.
• Given the difficulty of testing in a real incident and the resources required for a real-life test scenario, studies reviewing the usefulness of pre-incident information should consider the most accurate testing methodology. Future research should incorporate test scenarios that are as realistic as possible in order to elicit ecologically valid responses from participants, potentially including virtual reality scenarios to elicit a high sense of presence in the scene.