What Do We Mean by 'Community Resilience'? A Systematic Literature Review of How It Is Defined in the Literature.

1The ability to operationalize the concept of ‘community resilience’ against a disaster is highly sought after by disaster-response professionals, government officials, and academics. With the effects of climate change and demographic movements into large cities, disasters are occurring more frequently and in many cases with higher intensity than in previous years,, When disasters strike, governments and aid organizations are not always in a position to help communities immediately. For example, in Canada, the official emergency preparedness guide urges families to “be prepared to take care of yourself and your family for a minimum of 72 hours” if an emergency happens in their community because “it may take emergency workers some time to reach you (p. 3)”. In America, States such as New Hampshire and California have created public awareness campaigns and websites to increase disaster preparedness activities, such as assembling household emergency kits,. Even when governments and organizations can help, their resources are often stretched. Speaking to the press in 2014, Valerie Amos, Under-Secretary-General for Humanitarian Affairs and Emergency Relief Coordinator at the United Nations, noted that “the world’s collective response capacity and resources are being stretched to the limit” by the number of disasters that had occurred in 2013. Because of these issues, recent attention has focused on how best to help communities to help themselves, with a concomitant focus on understanding what factors contribute to making a community resilient to disasters.

2This focus has been accompanied by a change in rhetoric from government, industry and charitable organisations from discussing ‘disaster vulnerability’ to ‘disaster resilience,’ which is “viewed as a more proactive and positive expression of community engagement with natural hazard reduction (p. 598)”. The term ‘disaster’ is defined by UNISDR as “a serious disruption of the functioning of a community or a society involving widespread human, material, economic or environmental losses and impacts, which exceeds the ability of the affected community or society to cope using its own resources”, while a ‘community’ can be broadly defined as a constituent population such as a neighbourhood, town, or city. For example, the Rockefeller Foundation and Arup International Development recently created the City Resilience Framework to aid with evidence-based policy, to reduce the disaster risk in cities and to identify functions of what makes a resilient city. On an international scale, substantial improvements have been made with shifting policies and activities related to disaster risk reduction. Recent efforts by the United Nations have led to the creation of the Sendai Framework for Disaster Risk Reduction 2015-2030, which is built upon the Hyogo Framework for Action 2005-2015. This recent international treaty emphasises specific outcomes and priorities related to disaster risk reduction to be achieved by 2030, such as, “the substantial reduction of disaster risk and losses in lives, livelihoods and health in the economic, physical, social, cultural and environmental assess of persons, businesses, communities and countries (p. 12)".

3The concept of ‘community resilience’ is almost invariably viewed as positive, being associated with increasing local capacity,,, social support, and resources,, and decreasing risks,,, miscommunication,, and trauma,,. Yet consensus as to what community resilience is, how it should be defined and what its core characteristics are does not appear to have been reached, with mixed definitions appearing in the scientific literature, policies and practice,,. This confusion is troubling. The way we define community resilience affects how we attempt to measure and enhance it. For example, the Communities Advancing Resilience Toolkit (CART) describes a resilient community as one that “has the ability to transform the environment through deliberate, collective action” and “requires that the community as a whole must cope effectively with and learn from adversity (p. 1)" and as such suggests the measurement of community resilience requires the measurement of specific constructs such as ‘transformative potential,’ ‘connection and caring,’ ‘resources’ and ‘disaster management’ in order to identify areas of weakness. In contrast, the Conjoint Community Resiliency Assessment Measure (CCRAM) defines community resilience as “the community’s ability to withstand crises or disruptions (p. 1732)” and emphasises variables relating to leadership, collective efficacy, place attachment, preparedness, and social trust. Identical communities may score very differently on these two measures of what is supposed to be the same phenomenon.

4We conducted a systematic review of definitions of community resilience as it relates to disasters, in order to identify the range of definitions of community resilience present in the literature and to identify the range of constituent elements of community resilience that have been proposed.

5We initially completed a systematic literature search with no start date and a publication cut-off date of October 2013 for scientific peer-reviewed articles and of January 2014 for grey literature. Peer reviewed papers were found through searches of MEDLINE and PsycInfo, which were searched from inception. The search used keywords based on the stem words of resilience AND disaster AND definition (Table 1). We subsequently updated the search for peer reviewed papers using the same strategy for papers published up to May 2015.

6Grey literature searches using the keywords listed in Table 1 were applied using the search engine Google. The grey literature search was initially undertaken in October 2013, and updated in January 2014. Due to the volume of pages found in this search, we only reviewed the first 60 unique links. If two or more links came from the same main website, this was considered one unique link. Within the first 60 unique links, we extracted any publications that appeared relevant for further examination (e.g. annual reports or educational handouts). In addition to the electronic searches, we reviewed the references cited in all papers and reports.

7We included papers if they were written in English, were on the topic of disasters and had a description or discernible definition of community resilience. We defined ‘disasters’ as meeting the UNISDR definition of disaster and including acts of violence such as war and terrorism as well as natural disasters. Acts of violence such as war and terrorism were included in this review because there are strong similarities between aspects of community resilience to terrorism and war and aspects of community resilience to disasters. We excluded papers discussing epidemics, for example the HIV/AIDS epidemic in Southern Africa. Publications that included resilience definitions relating to individuals,,,,,, children,,,,, or hospital-based systems,, were not included, unless the authors also related the definition to a community as a whole. We acknowledged the ambiguity inherent in the definition of ‘community’ and accepted a publication if it attempted to describe resilience as a population-based concept. If publications suggested a wider theory, which included community resilience as part of the theory, we accepted it, but only if these publications included a description or definition of community resilience in respect to the theory (e.g. Zakour and Gillespie).

8We screened publications by reading the abstract or summary to remove duplicates, non-English reports, or papers that did not discuss resilience in a community setting in relation to a disaster. If an abstract or summary was not provided, we electronically searched the document for references where the term “resilience” was mentioned and read the relevant section to see if resilience was discussed in a community setting and in relation to a disaster. Potentially relevant publications were read in full.

9We created an evidence table, extracting the following from each selected publication: study description, direct quotation of definition, elements of definition as described or inferred by study and any measurable examples given by study. Using QSR International's NVivo 10 qualitative data analysis software, we carried out an inductive thematic analysis to compile a list of common elements within the concept of community resilience based on the definitions found in the literature,. This list was determined by the descriptions used in the definitions, as we categorized them into relevant overarching themes (elements) based on their similarities. The final number of elements was reached when no new theme could be uniquely supported or reduced into other existing elements. The same approach was used for the sub-elements within each element.

10Figure 1 shows the results of the search strategy. Overall, we identified 615 publications. Of these, 578 papers were found through MEDLINE and PsychInfo and 37 were from the grey literature.

11Definitions of community resilience

12From 62 publications found in the first search, we identified 57 unique definitions of community resilience as it applies to disasters. From the second search, 18 publications were added along with their respective definitions of community resilience. The unique definitions can be found in on-line supplementary Table 1.

13Three general types of definition were found: 1) ‘process’ definitions (i.e. an ongoing process of change and adaptation); 2) ‘absence of adverse effect’ definitions (i.e. an ability to maintain stable functioning); and 3) ‘range of attributes’ definitions (i.e. a broad collection of response-related abilities). More recent studies tended to adopt the first type of definition. For example, in an article discussing response enhancements to chemical, biological, radiological and nuclear terrorism, Lemyre and colleagues called the construct of resilience “a process or the attainment of positive outcomes at the individual, family, and community levels despite adversity (e.g., natural disaster, terrorist attack) (pg. 319).” In a review paper on community resilience, Norris and colleagues defined community resilience as “a process linking a set of networked adaptive capacities to a positive trajectory of functioning and adaptation in constituent populations after a disturbance (p. 131).” Citing publications by Paton and colleagues and Norris and colleagues, Cox and Perry defined community resilience as “a reflection of people’s shared and unique capacities to manage and adaptively respond to the extraordinary demands on resources and the losses associated with disasters” (p. 396). Furthermore, in a recent literature review on resilience, Castleden and colleagues defined community resilience as “a capability (or process) of a community adapting and functioning in the face of disturbance (p.370).”

14The ‘absence of adverse effect’ definitions used the desired outcome of ‘maintaining stable functioning’ as their basis. Bonanno examined the evidence on resilience based on loss and trauma among adults and defined it as an ability of adults to “maintain relatively stable, healthy levels of psychological and physical functioning (p.20).” Contrasting to the first type of definition, Gibson stated in a paper exploring the 2009 Victorian Bushfire in Australia that “…resilience is not a process, it is not a management system standard, nor is it a consulting product. Resilience is a demonstrable outcome of an organization’s capability to cope with uncertainty and change in an often volatile environment. Resilience is thus a product of an organization’s capabilities interacting with its environment (p. 246).”

15This notion of community resilience as an outcome was adapted by others who noted the importance of specifically identifying and strengthening abilities in a community, creating a third type of definition: the ‘range of positive attributes’ definition. An example of these definitions can be found in a publication by the UK Cabinet Office which defined community resilience, as “communities and individuals harnessing local resources and expertise to help themselves in an emergency, in a way that complements the response of the emergency services (p. 11)”. This report suggests that primarily, community resilience has to do with having a responsive and collective action of local support to help the community after an incident. Research carried out by Coles and Buckle into community resilience in Australia and the United Kingdom, led them to see resilience as “a multi-dimensional attribute that in its different forms contributes in various but equally important ways to disaster recovery (p. 6)." Their definition of community resilience is inferred from their publication in 2004 and summarized in a review article by Norris and colleagues as “a community’s capacities, skills and knowledge that allows the community to participate fully in the recovery from disasters (p. 129).” Moreover, by carrying out an epidemiological study of violence, injury, and resilience among the low-income communities in Western Cape of South Africa, Ahmed and colleagues defined community resilience by the features of a community. They found the following as key defining dimensions of community resilience specific to their study: “household relationships, levels of education and literacy, employment-seeking behaviours, social support networks, ability to seek support services, sense of communal safety and hope, and physical security measures (p. 393)".

16Additionally, definitions exist that blend one or more of these general definition types. In a recent review on assessment models and tools of community disaster resilience, Ostadtaghizadeh and colleagues produced a definition of community resilience now used by the United Nations International Strategy for Disaster Reduction: the “ability of a system, community, or society exposed to hazards to resist, absorb, accommodate to and recover from the effects of a hazard in a timely and efficient manner including through the preservation and restoration of its essential basic structures and functions (p. 3).” This definition blended the general types of ‘absence of adverse effect’ definition and ‘range of positive attributes’ definition. Even more broadly, Pfefferbaum and colleagues generally defined resilience “as an attribute (e.g., ability, capacity), a process, and/or an outcome associated with successful adaption to, and recovery from adversity” and that it “differs depending on context and purpose (p. 241-242).”

17Community resilience was therefore found to be an amorphous concept that was understood and applied differently by different research groups. In essence, depending on one’s stance, community resilience can either be seen as an ongoing process of adaptation, the simple absence of negative effects, the presence of a range of positive attributes, or a mixture of all three. However, common elements of community resilience were found among the literature based on these various definitions.

18Elements of community resilience

19Within the definitions, we identified nine main elements and 19 sub-elements that have been proposed as important within the concept of community resilience, shown in supplementary Table 2. These are described below with the main element listed in bold and sub-elements in italics. Additionally, some of the sub-elements could feasibly have been placed within other elements. We decided where to place each sub-element based on the emphasis given by the original authors.

20A comparison of the elements found in widely cited reviews, models and measurements of community resilience is given in supplementary Table 3. No existing models or measurements incorporated all of the elements and sub-elements identified.

21Local knowledge: The effects of a disaster, whether short-term or long-term, could be mitigated if a community understands its existing vulnerabilities. These vulnerabilities, if addressed prior to a disaster, are believed to build resilience within a community. For example, Kennedy and colleagues emphasised the importance of having a community assess and understand their own vulnerabilities. Three sub-elements were found within this. The first was the factual knowledge base of the community. Defined as the information, education and experience acquired in relation to a disaster, factual knowledge included specific learned information related to a disaster or disaster preparedness, such as knowledge about first aid, and other issues translatable to disaster preparedness, mitigation, response, and recovery.

22Training and education was a second sub-element. For example, Moore and colleagues found exemplary practices in community education, such as including public disaster education within routine education curricula, having early warning and public communications, partnering with the media for public education and risk communication, and communicating with affected populations via newsletters. Additionally, Moore and colleagues proposed practice to be an element of community resilience with activities such as community training and exercises proposed to build local knowledge and capacity. Effective training and education should lead to learning. For example, Cutter et al, emphasise the importance of learning how to respond effectively to an emergency.

23The third sub-element found was collective efficacy and empowerment. This was defined as a community’s shared belief of its ability to overcome potential hardships caused by a disaster, for example through self-reliance. What the community knows and understands about their own processes to endure and respond to a disaster can be crucial in relief efforts especially if a community is dependent upon its own resources. This was described by Chandra and colleagues, who suggested the importance of strengthening: 1) personal and community preparedness; 2) civic responsibility; 3) effective bystander responses; and 4) self-and community-reliance.

24Community networks and relationships: Positive effects on a community and its members can occur during a crisis when its members are well connected and form a cohesive whole. The connectedness of a community, sometimes called its ‘social network’, was defined by the linkages within a community. Creating links among community members based on social relationships,, and/or between communities were examples of connectedness. The cohesion of a community is based on the nature of these links, typically described as weak or strong ties. Several factors which determine the strength of a tie, including trust,,, and shared values,, might be relevant to enhanced community resilience. The connectedness of the networks and their cohesion were also discussed as important aspects of social capital, which conceptually focuses on bonding, bridging, and linking.

25Communication: Effective communication was seen as an important by most authors. Different ways to achieve effective communication were highlighted by different authors, however. For example, Norris and colleagues defined communication as “the creation of common meanings and understandings and the provision of opportunities for members to articulate needs, views, and attitudes (p. 140).” The authors interpreted effective communication as having occurred if the community used common meanings for all to understand and if the community provided opportunities for open dialogue. Another aspect of effective communication was the establishment of appropriate communication infrastructure that could be coordinated in a pre- or post-disaster setting. Chandra and colleagues noted that “strong communication networks are critical for resilience (p. 20)” and suggested that networks should have “diversity of mode and content,” for example through the use of social media to support and promote emergency messages, preferably using a trusted and established source of information.

26Both before and after a disaster, risk communication should provide accurate information about possible threats. This was mentioned in several papers, especially by Chandra and colleagues where the authors suggested training “partners and lay health advisors in proper risk communication techniques (p. 20).” Another suggestion was that government officials “should consider community norms and the range of individual beliefs (p. 21)” when crafting risk messages to ensure that the messages address the expectations of community members and are placed in an appropriate social context to help the public understand them,. Castleden and colleagues proposed integrating such steps into a detailed vulnerability analysis and subsequent community awareness initiatives.

27During a disaster, crisis communication should provide up-to-date information to community members about the ongoing impact and relief efforts. For example, Ganor and Ben-Lavy found that flow of information in real time was important for relief efforts while Dawes and colleagues identified open communication during an incident as important to crisis communication. Both issues can also be seen in a community and media approach to community resilience by Houston and colleagues. Their approach focused on the interactions between the strategic communication processes, such as community competence and community narratives; community relationships, such as social capital and media relations; community attributes, such as equality and diversity; and community systems and resources, such as communication infrastructure and traditional and social media. Effective response coordination within a community was found to rely on the communication between agencies, organizations, and community members. Furthermore, Cox and Hamlen suggested the importance of communications during the crisis, specifically infrastructure and technology.

28Health: The pre-existing health of a community and delivery of health services after a disaster are important for community resilience. Understanding and addressing health vulnerabilities can build resilience before a disaster and mitigate long-term issues after a disaster. One main sub-element identified within this was health services. Health services can be disrupted during a disaster. For example, mass casualties and blackouts could cause problems for a small community healthcare facility. Rego and Mehta suggested “building hospitals to higher standards of disaster resilience (p. 34).” This approach was adopted by the Asian Disaster Preparedness Center’s (ADPC), improvement to healthcare services through training and capacity-building at the hospital and facility level to handle mass casualties. For example, ADPC assisted in integrating these standards to a hospital in Kachchh, Gujarat by replacing the hospital building, which collapsed during the Gujarat earthquake of 2001 and claimed 176 lives, to a building with earthquake-resistant technology.

29The delivery and quality of care for physical and mental health issues were also identified as important sub-elements in health. When disasters hit, casualties with physical injuries must be treated quickly and with a high quality of care. A suggestion by Chandra and colleagues overlapped with other elements, such as resources, economic investment and preparedness. They suggested a plan for having post-health incident housing, especially for low-income individuals, in order to help restore physical health and livelihoods. Having such a capability would ensure a high level of quality and delivery of care post-disaster.

30The mental health of a community can be affected long after a disaster, with the immediate trauma and more chronic secondary stressors resulting in a range of disorders including post-traumatic stress disorder, anxiety and depression in a significant minority of individuals,. Providing “‘psychological first aid’ or other early psychological or behavioral health interventions after a disaster (p. 15)” were examples of ways in which a community might reduce psychological distress. Additionally, bolstering psychological wellness through public health communication campaigns was another example of ways to bolster resilience.

31Governance/ leadership: Governance and leadership shape how communities handle crises. We found two sub-elements within governance and leadership: infrastructure and services, and public involvement and support. For a community’s infrastructure and services, their effectiveness,,,,, efficiency,,,,, and capability to respond quickly were all noted as important. A specific concern was that infrastructure must have the capacity to deal with disaster,, and provide capable responses while in the midst of the crisis. For example, the infrastructure must have processes in place to handle incoming information about a disaster and send instructions and implement a response during and after the disaster,,.

32In terms of public involvement and support, having local participation and representation in strategic planning, response, and recovery were described as important by multiple publications,,,,,. Additionally, public involvement may involve having local leaders who understand and represent a community’s uniqueness and aspirations. A sense of community empowerment can be an additional output of public involvement in governance and leadership, as can increased trust in risk and crisis communication stemming from local leaders.

33Resources: Numerous resources have been hypothesised to be connected with community resilience. From tangible supplies, such as food, water and first aid kits, to technical resources such as shelter, automobiles and essential machinery, a higher level of resources is generally supposed to lead to higher levels of resilience. Some publications have described ‘resources’ more generally as including intangible aspects such as “natural, physical, human, financial, and social resources (p. 2)" and suggested the importance of having these resources widely available and distributed in the community,,. Ensuring the fairness of resource allocation is also known as distributive justice. Additionally, for physical resources such as food or water, it has been suggested that merely possessing the resource is insufficient; a resilient community must be able to harness these resources and allocate them appropriately within the community.

34Economic investment: If not addressed, the direct and indirect economic costs of a disaster can plague an affected community long after it has occurred. Addressing the post-disaster economic situation can involve: (i) distribution of financial resources, (ii) economic programming and ensuring that interventions are cost-effective, and (iii) the economic development of the post-disaster infrastructure and increasing the diversity of economic resources. This can be achieved through proactive investments to rebuild the economy. Assessing a community’s current economy and developing its ability to sustain economic growth were also noted as important areas of concentration after a disaster. The papers fell short in identifying a consensus for a useful post-disaster economic policy framework but many examples did overlap with suggestions provided by Smith in his book Planning for Post-Disaster Recovery: A Review of the United States Disaster Assistance Framework. The publications found in this review varied in terms of what economic planning is needed for a community in the pre-disaster setting, but all focused on the specific needs for post-disaster setting, whether through a specific programme to revitalise the job market, distribute economic aid or stimulate economic growth. A community’s post-disaster economy may be vital not just for recovery, but also for mitigating future disaster risks.

35Preparedness: Almost all publications mentioned the importance of preparedness across a number of levels, including the individual, family and government. In spite of this, only a few publications identified specific preparedness activities that can be used to mitigate the effects of a disaster. For example, Tierney and Bruneau suggested that emergency management systems should create plans before a disaster on how the disaster-response processes would work. Similarly, risk assessment was believed to help with preparedness. Actively involving community stakeholders in planning before a disaster, and running practice drills or exercises with a focus on risk management were cited as contributing to community resilience,. Carlson and colleagues suggested mitigation measures such as relocating buildings and infrastructure from flood-prone areas and/or flood-proofing structures, prior to a disaster. Altogether, the outputs of the planning, mitigation measures, and overall preparedness were intended to enable a sustainable response and recovery by the community, and to reduce the likelihood of harm to community members.

36Mental Outlook: Mental outlook was defined as attitudes, feelings and views when facing the uncertainty that typically occurs after a disaster or when contemplating a future one. This term was conceptually different to mental health, as the latter dealt with well-being while the former dealt with attitudes towards uncertainty. After a disaster, uncertainty is a common feeling among the affected population. This uncertainty can manifest itself in different ways; from anxiety about what the future holds for families, to concerns about the long-term impacts on the community, uncertainty reaches across individual and group boundaries. The search for meaning and the quality of the meaning attached to the disaster can also affect a community’s outlook. The mental outlook of a community is therefore important in shaping the willingness and ability of community members to continue on in the face of uncertainty. For example, hope, the expectation that things will improve, can help people cope with the uncertainty caused by a disaster. Ganor and Ben-Lavy described hope as a vision of community that depicts a better future after a disaster. In addition to hope, adaptability can be defined as the ability and willingness to change after a disaster while accepting that things will be different. Many publications noted various aspects of adaptability as an inherent aspect of resilience (e.g.,,,,,,,,,,,,,,). Bahadur and colleagues argued that one of ten main characteristics of a resilient system is the “acceptance of uncertainty and change (p. 15).”

37Within the field of disaster preparedness and response, it has been suggested that “a required paradigm shift and a new national ‘culture of disaster resilience (p. 2)" needs to occur. Unfortunately, our review suggests that we currently have no consensus on what such a culture would look like within our communities. What, exactly, do we mean by community resilience? Until we resolve this basic question, attempts to measure or enhance resilience will remain discordant and inefficient, while the academic literature will continue to be confused by papers assessing different concepts but using the same terminology.

38At present, definitions within this field tend to either focus on specific aspects of the concept that may lead to overconfidence in the resilience of a given community that is deficient in elements that were not considered, or else tend towards all-encompassing definitions that may be too complex to apply at the local level. How can we advance from this apparent Catch-22? One option is to abandon the search for a single, precise definition of community resilience. Instead, it may be more appropriate to consider community resilience as a catch-all term for the range of elements which may be important for a community facing or recovering from a disaster. We are not alone in suggesting this. For example, Usher-Pines and colleagues point out, “these discussions [about the definition of community resilience], while important, distract stakeholders from the actual task at hand: to better prepare communities to respond and recover from incidents (p. 604).” They argue that despite the advantages of the term community resilience, such as its ability to inspire people and re-invigorate the field, the pitfalls of community resilience are that no entity is clearly accountable for it and it is difficult to measure. A recent publication on subjective understandings of resilience-oriented interventions suggests similar ambiguity exists with the concept of ‘resilience'. The authors argue that this ambiguity does not make the term meaningless but that researchers and professionals should understand that a diverse range of meanings exist, especially when considering interventions. Rather than use the term community resilience, we therefore propose that it may be easier, clearer and more useful for academics, policy-makers and responders to be explicit as to the particular elements of resilience they are focussing on in their research or interventions; the phrase community resilience is not precise enough to be useful in any detailed discussion of the issue.

39Our review identified an array of elements that have been proposed within the general notion of community resilience and that may be usefully explored further. Factual knowledge base, collective efficacy and empowerment, and training and education have been proposed as useful within the element of local knowledge in order to mitigate vulnerabilities caused by how a community understands its risks. The positive effects of connectedness and cohesion within the element of community networks and relationships have been seen, especially in recent times, to help people deal with uncertainty after a disaster. Effective communication, whether risk or crisis communication, was proposed as important in helping a community to articulate, coordinate and understand the risk and impact of disasters. Health services were clearly relevant for a disaster-affected community, though a lack of knowledge of a community’s pre-existing issues among its residents and/or difficulty in delivery of quick, high-quality care were identified as key areas of difficulty to guard against. Within the element of governance and leadership, ensuring that roles, participation/engagement, and front-line leadership during a crisis are clear at the local level appears to be the main emphasis in the current literature. Similarly, the fair distribution of resources may help communities in the short term, while economic investment was generally seen as a longer-term intervention to promote resilience. Preparedness overlapped with the elements of local knowledge and communication but was typified by an emphasis on specific actionable activities. Lastly, mental outlook arguably has the most potential to build resilience within a community through a focus on sub-elements such as hope and adaptability.

40Most of these elements are already well-known within the disaster preparedness and crisis management fields outside of the specific rhetoric of community resilience; for example, risk and crisis communication has been extensively studied in respect to its role in disaster preparedness,,. However, many of these elements are broad, overlapping in practice and need further clarification. For example, what specific economic processes are important in making a community resilient? What forms of social networks help in mitigating the effects of a disaster? What types of preparedness activities are most effective? Further progress on these and other questions might be best met by disaggregating the issues from the global concept of community resilience.

41Limitations: Our review has several possible limitations. First, confirmation bias could have occurred when identifying publications for review given that a single researcher chose the accepted papers. To guard against this, we established an explicit set of inclusion criteria to use.

42Second, it is possible to question the reliability of the thematic analysis of the review, given that only one researcher worked on it. Had someone else analysed the same data, they may have come to a different set of conclusions.

43Third, selection bias based on language could have occurred, as there could have been additional useful studies available in languages other than English. Whether other elements are relevant in the resilience of non English-speaking communities is unknown. Whether community resilience is conceptualised differently in other cultures is an interesting question that may benefit from further investigation.

44Fourth, it is unlikely that we identified every relevant study in the literature, especially the grey literature. Despite this, updating the literature search did not alter the fundamental structure of our results although it did add more examples for the main elements found in the review. This provides some reassurance that the inclusion of missing studies would not radically alter the nature of the elements that we identified.

45Fifth, not all authors included in our review set out to write an original or specific definition of community resilience. For some of the included publications, community resilience was briefly described as part of another theory or concept. Had the authors of these papers been asked to construct their own formal definition, they may have produced a more detailed or nuanced interpretation. In many ways, however, the definitions given in these papers are of more interest, as they represent the interpretations of the concept that are being used in practice in the literature.

46Sixth, the results of this review are based on the original authors’ definitions which were broken down and grouped by similar concepts through an inductive thematic analysis. This led us to identify nine elements. Further investigation is needed in order to determine whether these identified elements are attributes or processes which make a community resilient to a disaster. Possible determinants in classifying the elements might include the type of disaster, a community’s culture, and whether the element is measured before, during or after a disaster.

47The concept of community resilience is widely used in the academic and policy literature. Yet the meanings of the term differ from team to team. Nine core elements have been consistently suggested as constituting community resilience as it applies to disasters: local knowledge, community networks and relationships, communication, health, governance and leadership, resources, economic investment, preparedness, and mental outlook. Further exploration of these individual elements may lead to a greater understanding of what community resilience is and how it can be measured and enhanced. In the meantime, the use of the phrase community resilience, and attempts to define the concept, may be unhelpful if it obscures the importance of these individual elements.

48The authors have no competing interests.

49Sonny S. Patel (spatel@hsph.harvard.edu)

50G. James Rubin (gideon.rubin@kcl.ac.uk)

51All relevant data are within the article.

52CART: Communities Advancing Resilience Toolkit

53CCRAM: Conjoint Community Resiliency Assessment Measure

54PTSD: Post-Traumatic Stress Disorder

55PTSS: Post-traumatic Stress Syndrome

56CR: Community Resilience

57CDR: Community Disaster Resilience