The final search strategy yielded 3584 records. Following the removal of duplicates, 2191 records remained and were included in title and abstract screening. A PRISMA flow diagram is presented in Figure 1.
At the title and abstract screening stage, the process became more iterative as the inclusion criteria were developed and refined. For the first iteration of screening, CH and DW sorted all records into ‘include,’ ‘exclude,’ and ‘unsure’. All ‘unsure’ papers were re-assessed by CH, and a random selection of ~20% of these were also assessed by DW. Where there was disagreement between authors the records were retained, and full text screened. The remaining papers were reviewed by CH, and all records were categorised into ‘include’ and ‘exclude’. Following full-text screening, 26 papers were retained for use in the review.
This section of the review addresses study characteristics of those which met the inclusion criteria, which comprised: date of publication, country of origin, study design, disaster, and variables examined.
Date of publication
Publication date across the 26 papers spanned from 2008 to 2020 (see Figure 2). The number of papers published was relatively low and consistent across this timescale (i.e. 1-2 per year, except 2010 and 2013 when none were published) up until 2017 where the number of papers peaked at 5. From 2017-2020 there were 15 papers published in total. The amount of papers published in recent years suggests a shift in research and interest towards CR and SC in a disaster/ public health emergency context.
Country of origin
The locations of the first authors’ institutes at the time of publication were extracted to provide a geographical spread of the retained papers. The majority originated from the USA (34-40), followed by China (41-45), Japan (46-49), Australia (50-52), The Netherlands (53, 54), New Zealand (55), Peru (56), Iran (57), Austria (58), and Croatia (59).
There were multiple methodological approaches carried out across retained papers. The most common formats included surveys or questionnaires (35-37, 41, 45-49, 52-54, 56, 58), followed by interviews (38, 39, 42, 50, 51, 59). Four papers used both surveys and interviews (34, 40, 44, 57), and two papers conducted data analysis (one using open access data from a Social Survey (43) and one using a Primary Health Organisations Register (55)).
Multiple different types of disaster were researched across the retained papers. Earthquakes were the most common type of disaster examined (44, 46, 48, 49, 52, 55-57), followed by research which assessed the impact of two disastrous events which had happened in the same area (e.g. Hurricane Katrina and the Deepwater Horizon oil spill in Mississippi, and the Great East Japan earthquake and Tsunami; (35-37, 41, 43, 47)). Other disaster types included: flooding (50, 53, 54, 58, 59), hurricanes (34, 38, 40), infectious disease outbreaks (42, 45), oil spillage (39), and drought (51).
Variables of interest examined
Across the 26 retained papers: eight referred to examining the impact of SC (34, 36, 38, 40, 45, 48, 54, 59); eight examined the impact of cognitive and structural SC as separate entities (39, 41, 44, 47, 49, 53, 56, 58); one examined bridging and bonding SC as separate entities (57); two examined the impact of CR (37, 55); and two employed a qualitative methodology but drew findings in relation to bonding and bridging SC, and SC generally (50, 51). Additionally, five papers examined the impact of the following variables: ‘community social cohesion’ (35), ‘neighbourhood connectedness’ (43), ‘social support at the community level’ (46), ‘community connectedness’ (42) and ‘sense of community’ (52). Table 1 provides additional details on this.
How is CR and SC measured or quantified in research?
The measures used to examine CR and SC are presented Table 1. It is apparent that there is no uniformity in how SC or CR is measured across the research. Multiple measures are used throughout the retained studies, and nearly all are unique. Additionally, SC was examined at multiple different levels (e.g. cognitive and structural, bonding and bridging), and in multiple different forms (e.g. community connectedness, community cohesion).
[TABLE 1 INSERT]
What is the association between CR and SC on mental wellbeing?
To best compare research, the following section reports on CR, and facets of SC separately.
CR relates to the ability of a community to withstand, adapt and permit growth in adverse circumstances due to social structures, networks and interdependencies within the community (10).
The impact of CR on mental wellbeing was consistently positive. For example, research indicated that there was a positive association between CR and number of common mental health (i.e. anxiety and mood) treatments post-disaster (55). Similarly, other research suggests that CR is positively related to psychological resilience, which is inversely related to depressive symptoms) (36). The same research also concluded that CR is protective of psychological resilience and is therefore protective of depressive symptoms (36).
SC reflects the strength of a social network, community reciprocity, and trust in people and institutions (13). These aspects of community are usually conceptualised primarily as protective factors that enable communities to cope and adapt collectively to threats.
There were inconsistencies across research which examined the impact of abstract SC (i.e. not refined into bonding/bridging or structural/cognitive) on mental wellbeing. However, for the majority of cases, research deems SC to be beneficial. For example, research has concluded that, SC is protective against post-traumatic stress disorder (54), anxiety (45), psychological distress (49), and stress (45). Additionally, SC has been found to facilitate post-traumatic growth (37), and also to be useful to be drawn upon in times of stress (51), both of which could be protective of mental health. Similarly, research has also found that emotional recovery following a disaster is more difficult for those who report to have low levels of SC (50).
Conversely, however, research has also concluded that when other situational factors (e.g. personal resources) were controlled for, a positive relationship between community resources and life satisfaction was no longer significant (59). Furthermore, some research has concluded that a high level of SC can result in a community facing greater stress immediately post disaster. Indeed, one retained paper found that high levels of SC correlate with higher levels of post-traumatic stress immediately following a disaster (38). However, in the later stages following a disaster, this relationship can reverse, with SC subsequently providing an aid to recovery (40). By way of explanation, some researchers have suggested that communities with stronger SC carry the greatest load in terms of helping others (i.e. family, friends and neighbours) as well as themselves immediately following the disaster, but then as time passes the communities ‘bounce back’ at a faster rate as they are able to rely on their social networks for support (40).
Cognitive and Structural Social Capital
Cognitive SC refers to perceptions of community relations, such as trust, mutual help and attachment, and structural SC refers to what actually happens within the community, such as participation, socialising (15).
Cognitive SC has been found to be protective (48) against PTSD (53, 56), depression (39, 53)) (mild mood disorder; (47)) and anxiety (47, 53).
For structural SC, research is again inconsistent. On the one hand, structural SC has been found to: increase perceived self-efficacy, be protective of depression (39), buffer the impact of housing damage on cognitive decline (41) and provide support during disasters and over the recovery period (58). However, on the other hand, it has been found to have no association with PTSD (53, 56) or depression, and is also associated with a higher prevalence of anxiety (53). Similarly, it is also suggested by additional research that structural SC can harm women’s mental health, either due to the pressure of expectations to help and support others or feelings of isolation (48).
Bonding and Bridging Social Capital
Bonding SC refers to connections among individuals who are emotionally close, and result in bonds to a particular group (16), and bridging SC refers to acquaintances or individuals loosely connected that span different social groups (17).
One research study concluded that both bonding and bridging SC were protective against post-traumatic stress disorder symptoms (57). Bridging capital was deemed to be around twice as effective in buffering against post-traumatic stress disorder than bonding SC (57).
Other community variables
Community social cohesion was significantly associated with a lower risk of post-traumatic stress disorder symptom development (34), and this was apparent even whilst controlling for depressive symptoms at baseline and disaster impact variables (e.g. loss of family member or housing damage) (35). Similarly, sense of community, community connectedness, social support at the community level and neighbourhood connectedness all provided protective benefits for a range of mental health, wellbeing and recovery variables, including: depression (52), subjective wellbeing (in older adults only) (42), psychological distress (46), happiness (43) and life satisfaction (52).
Research has also concluded that community level social support is protective against mild mood and anxiety disorder, but only for individuals who have had no previous disaster experience (47). Additionally, a study which separated SC into social cohesion and social participation concluded that at a community level, social cohesion is protective against depression (48) whereas social participation at community level is associated with an increased risk of depression amongst women (48).
What is the impact of infectious disease outbreaks / disasters and emergencies on community resilience?
From a cross-sectional perspective, research has indicated that disasters and emergencies can have a negative effect on certain types of SC. Specifically, cognitive SC has been found to be impacted by disaster impact, whereas structural SC has gone unaffected (44). Disaster impact has also been shown to have a negative effect on community relationships more generally (51).
Additionally, of the eight studies which collected data at multiple time points (34, 35, 40, 41, 46, 48, 55, 59), three reported the effect of a disaster on the level of SC within a community (39, 41, 48). All three of these studies concluded that disasters may have a negative impact on the levels of SC within a community. The first study found that the Deepwater Horizon oil spill had a negative effect on SC and social support, and this in turn explained an overall increase in the levels of depression within the community (39). Similarly, the second study found that SC (in the form of social cohesion, informal socialising and social participation) decreased after the 2011 earthquake and tsunami in Japan; it was suggested that this change correlated with incidence of cognitive decline (41).However, the third study reported more mixed effects based on physical circumstances of the communities’ natural environment: Following an earthquake, those who lived in mountainous areas with an initial high level of pre-community SC saw a decrease in SC post disaster (48). However, communities in flat areas (which were home to younger residents and had a higher population density) saw an increase in SC (48). It was proposed that this difference could be due to the need for those who lived in mountainous areas to seek prolonged refuge due to subsequent landslides (48).
What types of intervention enhance CR and SC and protect survivors?
There were mixed effects across the 26 retained papers when examining the effect of CR and SC on mental wellbeing. However, there is evidence that an increase in SC (55, 56), with a focus on cognitive SC (56), namely by: building social networks (44, 50, 52), enhancing feelings of social cohesion (34, 35) and promoting a sense of community (52), can result in an increase in CR and potentially protect survivors’ wellbeing and mental health following a disaster. An increase in SC may also aid in decreasing the need for individual psychological interventions in the aftermath of a disaster (54). As a result, recommendations and suggested methods to bolster CR and SC from the retained papers have been extracted and separated into general methods, preparedness and policy level implementation.
Suggested methods to build SC included organising recreational activity-based groups (43) to broaden (50, 52) and preserve current social networks (41), introducing initiatives to increase social cohesion and trust (50), and volunteering to increase the number of social ties between residents (58). Research also notes that it is important to take a ‘no one left behind approach’ when organising recreational and social community events, as failure to do so could induce feelings of isolation for some members of the community (48). Furthermore, gender differences should also be considered as research indicates that males and females may react differently to community level SC (as evidence suggests males are instead more impacted by individual level SC; in comparison to women who have larger and more diverse social networks (48)). Therefore, interventions which aim to raise community level social participation, with the aim of expanding social connections and gaining support, may be beneficial (41, 46).
In order to prepare for disasters, it may be beneficial to introduce community-targeted methods or interventions to increase levels of SC and CR as these may aid in ameliorating the consequences of a public health emergency or disaster (56). To indicate which communities have low levels of SC, one study suggests implementing a 3-item scale of social cohesion to map areas and target interventions (41).
It is important to consider that communities with a high level of SC may have a lower level of risk perception, due to the established connections and supportive network they have with those around them (60). However, for the purpose of preparedness, this is not ideal as perception of risk is a key factor when seeking to encourage behavioural adherence. This could be overcome by introducing communication strategies which emphasise the necessity of social support, but also highlights the need for additional measures to reduce residual risk (58). Furthermore, support in the form of financial assistance to foster current community initiatives may prove beneficial to rural areas, for example through the use of an asset-based community development framework (51).
At a policy level, the included papers suggest a range of ways that CR and SC could be bolstered and used. These include: providing financial support for community initiatives and collective coping strategies, (e.g. using asset-based community development (51)); ensuring policies for long-term recovery focus on community sustainable development (e.g. community festival and community centre activities) (43); and development of a network amongst cooperative corporations formed for reconstruction and to organise self-help recovery sessions among residents of adjacent areas (57).