Evaluation of emergency evacuation in school buildings: protocol for a systematic review

R. P. Martinsa, J. Duarteb, Mário Vazc aAssociated Laboratory for Energy, Transports and Aeronautics (PROA/LAETA), Faculty of Engineering, University of Porto, PT (rpmar@fe.up.pt) ORCID: 0000-0001-8438-0597. bAssociated Laboratory for Energy, Transports and Aeronautics (PROA/LAETA), Faculty of Engineering, University of Porto, PT (jasduarte@fe.up.pt) ORCID: 0000-00025856-5317. cAssociated Laboratory for Energy, Transports and Aeronautics (INEGI/LAETA), Faculty of Engineering of University of Porto, Porto, PT (gmavaz@fe.up.pt).


Background
Disasters profoundly affect society in a social, economic and psychological term (Mermer, Donmez, & Daghan, 2018). The outcomes of natural disasters include death casualties, destruction of health facilities, and health crises (Bustami & Baharuddin, 2018). Human reaction to threats and emergencies in crowded spaces is a crucial issue for disaster preparedness, emergency management and evacuation planning (Haghani & Sarvi, 2018). Understanding the social and dynamical behaviours of a crowd is essential for any decision process related to safety (Bellomo, Clarke, Gibelli, Townsend, & Vreugdenhil, 2016).
Children are amongst the most vulnerable affected groups by natural and human-made disasters (Codreanu, Celenza, & Jacobs, 2014;Fu & Underwood, 2015;Gibbs et al., 2014). In the 2008 earthquake in Sichuan, China, many children died, in one of the biggest disasters for children (Watts, 2008). The increasing number of affected children draws a concerning picture for their future. During an emergency, children are dependent on adults to communicate the risk and to instruct or assist them with the appropriate response (Codreanu et al., 2014;Fothergill, 2017 Nouchi, 2018). Schools need to have disaster plans that are uniquely designed for the school culture and interface with a broader community (Blum et al., 2008).
Either quantitative or qualitative methods are used to prepare a school community for responding safely to a disaster. The evacuation process is dynamic and incorporates human behaviour, which has a critical effect on evacuation efficiency (Chen, Tang, Song, Huang, & Guo, 2019). For an in-depth exploration and insight into the students' perceptions of the school evacuation plan, focus groups, interviews, and individual mapping exercises can also be performed (Vásquez et al. 2018). Evacuation drills allow to test safety practices and procedures of the school emergency plan and evaluate school personnel performance during a crisis (Allen, Lorek, & Mensia-Joseph, 2008). Evacuation time is essential in the evacuation process for recognising the required time by the occupants to move from the hazard area to a safe region when they acknowledge the danger and start to evacuate (Ng & Chow, 2006).
Understanding behavioural patterns and the corresponding complex phenomena during an evacuation is critical to enhancing evacuation safety and efficiency.
The purpose of the systematic review is to find evidence of emergency evacuation response in schools. In order to do so, data will be gathered concerning the evacuation dynamics and response towards disasters in school building: observed behaviour during evacuation drills, travel speeds and total evacuation time.

METHODOLOGY
This systematic review protocol follows the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) Statement Shamseer et al., 2015).

Eligibility criteria
Studies will be selected according to the criteria outlined below. To be considered eligible, a study must examine the disaster response behaviour of children and adults during evacuation drills in school buildings. Participants of the studies will, therefore, include children and adults that participate in drill exercises in a school building. Interventions addressing the performance of evacuation drill exercises or real drills are also studies of interest.
The comparators relevant to this study will be the time of evacuation and the type of evacuation procedures.
Articles only considering pre-disaster analysis or post-disaster analysis, without performing evacuation drills, will not be included. Similarly, the articles that only include computer simulation drills will not be considered eligible.
Literature reviews and conference papers will not be included. Articles from 2013 upward will be retrieved. The references list of the obtained articles will be checked in order to find and include more relevant studies. This way, older articles would be included.
Only publications written in English will be considered.

Information sources
The research will be conducted in the following electronic databases: Scopus, Web of Science, ScienceDirect, Inspec, PubMed, PsycArticles, ERIC, Business Source Complete, SpringerLink, Wiley Online Library, Oxford Journals, SciELO -Scientific Electronic Library Online, Informaworld (Taylor and Francis), MetaPress, Directory of Open Access Journals (DOAJ), IEEE Xplore.
The research will be conducted on journal articles from 2013 to 2019 in order to obtain the most relevant information about emergency evacuation behaviour in schools. Additional studies will be screened in the reference chapter of the included articles identified through the first stage of research. For this step, the date will not be an exclusion criterion. The literature search will be limited to the English language.

Search strategy
The following keyword expressions will be used during the database search: [(emergency OR disaster) AND ((management AND school) OR (response AND school) OR (planning AND school) OR (preparedness AND school))], [("Reaction time" OR "Response time" OR "Evacuation time" OR "Human response" OR "Human behaviour" OR "Response preparedness") AND ((emergency AND school) OR (disaster AND school))], [("Risk assessment" OR "Hazard awareness" OR "Risk perception" OR Safety) AND ((emergency AND school) OR (disaster AND school))].
The search in the several databases will be set for the "Article title, Abstract, Keywords", whenever possible.
The selected articles will be analysed, with the identification of potential new keywords. A new search will then be managed if new keywords are found. For the search of any additional study within the scope of the review, the references of the collected articles will also be analysed. This process will be performed until no more related outcomes can be found.

Data management
An Excel table will be used to help register the number of articles obtained from the search. This process will allow registering all identified studies through the selection phases. After completing the search and recording the number of collected articles in Table 1 (Appendix 1), the selected articles from each database will be exported for screening and removing duplicates.
Mendeley software will be used for references management.

Selection process
As each keyword expression is entered in the databases, the following criteria will be considered with search filters: Date: Articles published between January 2013 and May 2019.

Article type: Articles and Articles in Press.
Source type: Journals.
Removal of duplicated articles.
Articles will not be included if any of the following conditions are found: • Studies that do not perform drill experiments; • Studies that only make computer simulation drills; • Studies that only consider pre-disaster emergency analysis (educational phase); • Studies that only consider post-disaster emergency analysis (recovery and rehabilitation phases); • Studies that only consider subjective methods of assessment, without execution of evacuation drills.
Potential eligible records will be identified through the screening phase. Two reviewers will independently analyse title and abstract; a third reviewer will solve any disagreement. All the information will be registered and justified.
For selecting the studies to be included in the systematic review, the full-text from the resultant studies will be retrieved and assessed to determine if the following conditions are satisfied: • Studies include the performance of evacuation drill experiments or real drills; • Studies include the preparedness response phase.
Any article removal, after examining the full-text, will be justified and recorded.

Data collection process
From the final selected studies, some information will be extracted, including general study information, sample characteristics, context, study characteristics and significant limitations. One reviewer will conduct the data extraction with the verification by another reviewer. This will allow to reduce bias and reduce data extraction errors.

Data items
Summary and descriptive tables will be built with information on the following topics: 1. Study general information: authors, publication year, country.
3. Context: type of emergency, type of school, number of schools, speed, density, flow, order and time of arrivals, pre-evacuation time, evacuation time; type of intervention (qualitative study). 4. Study characteristics: aims, procedures/methods, conclusions, equipment, and software. 5. Major study limitations.

Outcomes and prioritisation
Data will be gathered concerning the evacuation dynamics and response to disasters in the school building.
From the proposed research, the following primary outcomes are expected: 1. To identify the characteristics related to evacuation time; 2. To identify the adopted behaviour responses during the evacuation drills.
The most frequently evaluated parameters will be assessed, and an outcomes comparison from different studies will be presented. Therefore, the expected outcomes of this systematic review are the resulting time of evacuation and the comparison results between evacuation procedures.

Risk of bias in individual studies
The bias assessment in included articles will be considered, using the taxonomy suggested by Higgins et al. (2011).

Data synthesis
This parameter does not apply to the proposed systematic review.

Meta-biases(es)
This parameter does not apply to the proposed systematic review.

Confidence in cumulative evidence
This parameter does not apply to the proposed systematic review.

Protocol registration
The protocol is under registration.