Mental health during and after protests,riots and revolutions: A systematic review

Abstract

Objectives:

Protests, riots and revolutions have long been a part of human history and are increasing globally, yet their impact on mental health remains largely unknown. We therefore systematically reviewed studies on collective actions and mental health.

Method:

We searched PubMed, Web of Science, PsycINFO and CINAHL Plus for published studies from their inception until 1 January 2018. Study quality was rated using the Newcastle–Ottawa Scale.

Results:

We identified 52 studies (n = 57,487 participants) from 20 countries/regions. The prevalence of post-traumatic stress disorder ranged from 4% to 41% in riot-affected areas. Following a major protest, the prevalence of probable major depression increased by 7%, regardless of personal involvement in the protests, suggestive of community spillover effects. Risk factors for poorer mental health included female sex, lower socioeconomic status, exposure to violence, interpersonal conflicts, frequent social media use and lower resilience and social support. Nevertheless, two studies suggested that collective actions may reduce depression and suicide, possibly due to a collective cathartic experience and greater social cohesion within subpopulations.

Conclusion:

We present the first systematic review of collective actions and mental health, showing compelling evidence that protests even when nonviolent can be associated with adverse mental health outcomes. Health care professionals therefore need to be vigilant to the mental and psychological sequelae of protests, riots and revolutions. Further research on this emerging sociopolitical determinant of mental health is warranted.

Keywords

Protest riot revolution collective action mental health mental disorders depression post-traumatic stress disorder anxiety psychological civil disorder demonstration campaign social movement political movement

Background

In the past few decades, collective actions such as protests, riots and revolutions (Tarrow, 2011) have taken place in over 180 countries, accounting for 99% of the world’s population (Banks and Wilson, 2017) (Figure 1). They continue to increase globally (Ortiz et al., 2013; Stephan and Chenoweth, 2008), yet the mental health consequences of collective actions remain largely unknown. By contrast, the mental health impact of other large-scale population events such as natural disasters, human-made disasters, terrorist attacks and armed conflicts have been much better documented (Charlson et al., 2019; Galea et al., 2005; Galea and Maxwell, 2009; Goldmann and Galea, 2014; Norris et al., 2002a; North, 2014; North and Pfefferbaum, 2002; Perlman et al., 2011). These studies indicate significant levels of both psychopathology and resiliency in the face of mass community trauma, but the relevance of collective actions to health care professionals remains unclear. As the nature of collective actions differs from other major population events, a separate field of inquiry is needed. We therefore systematically reviewed studies on collective actions and mental health around the world.

Figure 1.

Number of collective actions from 1967 to 2017 (upper panel), number of published studies of collective actions and mental health from 1967 to 2017 (lower panel) and correlation between the number of collective actions and the number of published studies (right panels). Data source for the number of collective actions: Banks and Wilson (2017).

Methods

Search strategy and selection criteria

We pre-specified our methods in a study protocol registered in PROSPERO (PROSPERO 2015:CRD42015029732), and followed the Preferred Reporting Items for a Systematic Review and Meta-analysis (PRISMA) guidelines (Liberati et al., 2009). We included studies reported in English and published in peer-reviewed journals. Studies of all designs, qualitative and quantitative, cross-sectional and longitudinal, were included.

The inclusion criterion for the outcome was that the study measured mental health, either by reporting psychiatric diagnoses, admissions, prescriptions, symptoms or distress of individuals who were directly or indirectly influenced by an identified collective action. We excluded studies based on labour disputes as these aimed to change the distribution of resources or roles within an organisation. Incidents that involved prolonged, paramilitary conflicts organised by states or state-like actors were also excluded.

PubMed, Web of Science, PsycINFO and CINAHL Plus were searched for studies published from the inception of each database until 1 January 2018. Reference lists of retrieved articles were reviewed to identify additional studies. Terms that capture diverse types of collective action, such as ‘protest’, ‘civil disorders’, ‘demonstration’, ‘revolution’ or ‘campaign’ were searched, along with social movements employing collective actions as a tool, using ‘social movement’, ‘political movement’ and ‘socio-political movement’. These were used in combination with terms such as ‘mental health’, ‘mental disorders’, ‘behavioral symptoms’ and ‘psychological’ to identify potential articles addressing collective action and mental health outcomes. An example of the detailed search strategy is presented in Table 1.

Table 1.

Electronic search strategy for PubMed.

Database	Search	Filters/limits
PubMed	(‘protest’ [All Fields] OR ‘revolution’ [All Fields] OR ‘Civil Disorders’ [Mesh] OR ‘civil disobedience’ [All Fields] OR ‘demonstration’ [All Fields]) OR ‘social movement’ [All Fields] OR ‘political movement’ [All Fields] OR ‘campaign’ [All Fields] AND (‘Mental Health’ [Mesh] OR ‘Mental Disorders’ [Mesh] OR ‘Behavioural Symptoms’ [Mesh] OR psychological [All Fields])	Humans English

Mesh: Medical Subject Headings.

The full search strategy for each database and the search results are available from the authors.

Once studies were retrieved, two authors independently screened the articles to be included and any disagreement was resolved by consensus. Titles were examined in the first round and abstracts in the second round of screening to determine whether the articles met the inclusion criteria. After the first two rounds of screening, full texts of articles were carefully assessed, and those that met the exclusion criteria were removed.

Two authors constructed a data extraction form, and the first and second author independently completed the form for all included studies. For each article, location, date and duration, type of the collective action, level of violence (death, looting, arson, gun and tear gas use), study design, study population, sampling methods, sample size, analytic methods, exposure and outcome measurement, social support, media exposure, resilience and funding information were extracted.

Evidence grading

Three authors independently rated the methodological quality of all included studies using the Newcastle–Ottawa Scale (NOS), which is recommended by the Cochrane collaboration for the appraisal of non-randomised studies (Higgins and Green, 2012a). The NOS was modified for the quality assessment of cross-sectional studies with reference to previous adaptations (Anglin et al., 2013; Patra et al., 2015).

Synthesis of results

Findings were summarised by mental health outcomes. In accordance with recommendations for systematic reviews, we focused first on findings from longitudinal studies and when none were available then cross-sectional studies with high NOS scores were reviewed (JAMA, 2018). When assessing the prevalence of mental health outcomes, we prioritised findings based on random samples drawn from the general population.

Results

Description of studies

The initial search yielded 18,181 articles. After applying the inclusion and exclusion criteria and searching the reference lists of included studies, we identified a total of 52 studies (Figure 2). These included 6 longitudinal, 8 time series, 3 serial cross-sectional, 22 cross-sectional, 8 case series and 5 qualitative studies, involving a total of 57,487 participants (excluding time series and qualitative studies that did not specify sample size). In total, our review identified reports from 20 countries/regions, including the United States (16 studies), Hong Kong SAR (six), India (five), Northern Ireland (five), Egypt (three), Czechoslovakia (two), Tunisia (two) and one each from Chile, China, Georgia, Iran, Israel, Italy, Kenya, Malaysia, Nigeria, Romania, South Africa, Turkey and the United Kingdom (Figure 1). The Spearman correlation coefficients between the number of collective actions and the number of published studies from 1967 to 2017 were 0.40 (p < 0.001) by region and 0.75 (p = 0.08) by continent (Figure 1). Half of the studies covered riots, 14 were on protests and 9 studied revolutions. The collective actions ranged in time from the 1965 Watts riots to the 2015 Baltimore Protests. Tables 1 –3 and Supplemental Tables S1–S4 summarise the study designs and main findings from five decades of research.

Figure 2.

PRISMA flow diagram for studies on collective actions and mental health.

Sample characteristics

Thirteen studies used random samples, some drawn from the general population (Hou et al., 2015; Lau et al., 2016, 2017; Ni et al., 2016, 2017), others by sex (Kubicka et al., 1995, 1998), from health facilities (Bamrah et al., 2013), schools (Castillo-Carniglia et al., 2017), or affected communities (Greenley et al., 1975; Hanson et al., 1995; Obilom and Thacher, 2008; Sasao and Chun, 1994). Sample sizes ranged from 5 to 37,319 (median 343). The majority of studies (n = 36) were conducted after the collective action, while seven started before the collective action (Brown et al., 1994; Fraser, 1971; Garbarino et al., 2012; Klee and Gorwitz, 1970; Kubicka et al., 1995, 1998; Lyons, 1972), and six studies included assessments before, during and after the collective action (Castillo-Carniglia et al., 2017; Fishbain et al., 1991; Greenley et al., 1975; Ni et al., 2016, 2017; Yimgang et al., 2017).

Assessment of mental health outcomes

Depressive sequelae were the most frequently assessed outcome, recorded in a third (n = 18) of all studies, followed by post-traumatic stress disorder (PTSD) (n = 11) and anxiety disorder/anxiety symptoms (n = 11). Other outcomes included psychiatric admissions and prescriptions, psychological distress, substance abuse and suicide (Table 2).

Table 2.

Published studies on collective actions and mental health.

	Number of studies (%)
Study design
Longitudinal	6 (12)
Time series	8 (15)
Serial cross-sectional	3 (6)
Cross-sectional	22 (42)
Case series	8 (15)
Qualitative	5 (10)
Study population
General population	6 (12)
Residents of affected areas	8 (15)
Children	9 (17)
Patients	17 (33)
Demographic group (e.g. ethnicity)	7 (13)
Occupational groups (e.g. police)	5 (10)
Sample size^a
1–499	26 (50)
500–999	10 (25)
1000–1999	7 (18)
⩾2000	1 (3)
Outcome^b
Depression	18 (35)
Post-traumatic stress disorder	11 (21)
Anxiety	11 (21)
Psychological distress	6 (12)
Substance abuse	6 (12)
Psychiatric admissions and prescriptions	5 (10)
General psychiatric diagnosis	5 (10)
Suicide	5 (10)

Time series studies and qualitative studies which do not specify the sample size have been excluded.

The number of studies is greater than 52 as some studies examined more than one outcome.

Depressive sequelae

Prevalence

Five studies (Hou et al., 2015; Lau et al., 2016, 2017; Ni et al., 2016, 2017) used probability samples in the general population to assess the prevalence of depression following a collective action. However, all of these studies focused on Hong Kong’s 2014 Occupy Central/Umbrella Movement. Using pre-event data from a population-representative cohort, the prevalence of probable major depression was 1.5% before the Occupy Central/Umbrella Movement. Following the Movement, the prevalence increased to 6.7% (first month), 6.3% (second month), 6.8% (third month) and 8.5% (sixth month) (Ni et al., 2016) (Table 3). Based on the same cohort, four depression trajectories were identified before, during and 1 year after the Occupy Central/Umbrella Movement: ‘resistant’ (22.6% of sample), ‘resilient’ (37.0%), ‘mild depressive symptoms’ (32.5%) and ‘persistent moderate depression’ (8.0%) (Ni et al., 2017). These findings contrast studies from Northern Ireland. Findings based on three districts with varying degrees of exposure to the 1969 Northern Ireland riots found no differences in anti-depressant prescriptions (Fraser, 1971), and indeed a decline in health care utilisation for depression (Lyons, 1972).

Table 3.

Longitudinal studies on collective actions and mental health.

Study	Collective action (duration)	Study population (sample size)	Exposure	Outcome	Main findings
Greenley, Arch Gen Psychiatry 1975	1967 New Haven riots (8 days)	Residents of suburban and urban areas (n = 938)	Before, during and after the riot	Psychological distress	Psychological impairment decreased by 9.7% and 6.7% among suburban men and women, respectively.
Kubicka, Addiction 1995	1989 Velvet Revolution (45 days)	Young and middle-aged women (n = 608)	Before and after the revolution	Alcohol use	Mean annual alcohol consumption increased from before (3.6 L) to after the revolution (4.8 L). The rise in alcohol consumption has been suggested to be due to an expansion of social contacts.
Kubicka, Addiction 1998	1989 Velvet Revolution (45 days)	Young men (n = 586)	Before and after the revolution	Alcohol use	The price of alcohol increased and mean annual alcohol consumption decreased by 26% before the revolution. Following the revolution, political control over drinking was abolished and alcohol consumption increased by 16%.
Ni, Am J Epidemiol 2016	2014 Occupy Central/Umbrella Movement (79 days)	General adult population (n = 909)	Protest participation; media exposure; social support	Probable major depression; depressive symptoms	The absolute prevalence of probable major depression increased by 7%, regardless of personal involvement in the protests. Depressive symptoms were associated with exposure to the protest via online/social media, frequent Facebook use and intra-familial sociopolitical conflicts.
Ni, Am J Public Health 2017	2014 Occupy Central/Umbrella Movement (79 days)	General adult population (n = 1170)	Protest participation; social support; resilience	Probable major depression; depressive symptoms	Four depression trajectories were identified: ‘resistant’ (22.6% of sample), ‘resilient’ (37.0%), ‘mild depressive symptoms’ (32.5%) and ‘persistent moderate depression’ (8.0%). Baseline predictors that appeared to protect against persistent depression included higher household income, resilience, family support and self-rated health.
Hou, J Couns Psychol 2018	2014 Occupy Central/Umbrella Movement (79 days)	General adult population (n = 108)	During and after the protests	Emotional reactivity; psychological distress	Lower positive emotional reactivity and higher negative emotional reactivity for positive events were detected during the protests.

Correlates of depressive sequelae

The negative impact of exposure to collective action appears to vary with the level of violence. For collective actions associated with fatalities, proximity to violence was an important predictor for depression. During the 2011 Arab Spring, the prevalence of depression among children in schools closer to Tahrir Square (a focal point of the revolution) was 62% (Moussa et al., 2015). Living in riot-affected areas was also associated with depression risk (Coudreaut et al., 1995; Fraser, 1971; Tan and Simons, 1973; Yimgang et al., 2017). By contrast, participation in nonviolent collective actions was not associated with depressive sequelae (Lau et al., 2017; Ni et al., 2016). Case series have identified other risk factors for depression following collective actions, including media coverage, unemployment and poor social cohesion (Brown et al., 1994; Coudreaut et al., 1995; Tan and Simons, 1973). These have been corroborated by subsequent studies where exposure to collective actions via television (Galovski et al., 2016; Lazarescu et al., 1993) or social networking platforms (Lau et al., 2016, 2017; Ni et al., 2016) was associated with more depressive symptoms.

Sociodemographic factors such as female sex, lower socioeconomic status and not having children were also associated with more depressive symptoms (Lau et al., 2017; Moussa et al., 2015; Ni et al., 2017). Other risk factors included poorer resilience, pre-existing health conditions, and the loss of personal and social resources (Brown et al., 1994; Hou et al., 2015; Ni et al., 2017). Lower baseline family support and family harmony were predictors of persistent depression following the Occupy Central/Umbrella Movement (Ni et al., 2017), while conflicts with family members (Lau et al., 2017; Ni et al., 2016) were also associated with more depressive symptoms. Qualitative studies have suggested that if the collective action is not able to achieve its aims, hopelessness could result in depression (Matthies-Boon, 2017).

PTSD

Prevalence

Four studies have assessed the prevalence of PTSD after collective actions (Galovski et al., 2016; Hanson et al., 1995; Obilom and Thacher, 2008; Wang et al., 2011). All the studies were cross-sectional, so the incidence and course of PTSD following collective actions remain unknown. The prevalence of PTSD drawn from probability samples ranged from 4.1% of residents after the 1992 Los Angeles riots (Hanson et al., 1995) to 41% of residents in Nigeria’s riot-affected areas (Obilom and Thacher, 2008). Studies have also examined the prevalence of PTSD among specific groups such as teachers (Stuber et al., 1997), children (Kumar and Fonagy, 2013; Moussa et al., 2015), patients (Coudreaut et al., 1995; Ouanes et al., 2014), medical and paramedical professionals (Ouanes et al., 2012), police officers (Harvey-Lintz and Tidwell, 1997) and victims of looting and arson (Kim-Goh et al., 1995). The prevalence of PTSD among these groups ranged from 6.7% in teachers (Stuber et al., 1997) to 75% in victims of looting and arson (Kim-Goh et al., 1995).

Correlates of PTSD

Exposure to violence was most consistently associated with increased risk of PTSD: direct victims of looting, arson and physical injury demonstrated the highest levels (Kim-Goh et al., 1995; Obilom and Thacher, 2008; Unuvar et al., 2017). Witnessing a personal attack (Kumar and Fonagy, 2013; Obilom and Thacher, 2008), being in close proximity to violence (Moussa et al., 2015) and the level of media exposure (Galovski et al., 2016) were also associated with PTSD risk. Demographic characteristics such as female sex and lower socioeconomic status were associated with increased risk of PTSD among school children (Moussa et al., 2015). In the context of racial (2014 Ferguson Protests in Missouri, USA) or ethnoreligious riots (2002 Gujarat Riots in India), social identity was associated with PTSD symptoms (Galovski et al., 2016; Patel and Nath, 2013).

Anxiety

Prevalence

Four studies assessed the prevalence of anxiety after a collective action in the general population (Hou and Bonanno, 2018; Hou et al., 2015; Lau et al., 2016, 2017) and seven in special groups (Bamrah et al., 2013; Dhavale et al., 2002; Garbarino et al., 2012; Lazarescu et al., 1993; Lyons, 1971; Matthies-Boon, 2017; Moussa et al., 2015). Two studies in the general population followed the 2014 Hong Kong Occupy Central/Umbrella Movement, although different anxiety scales were used. Immediately after the Movement, 10.5% of the general population had probable anxiety (7-item General Anxiety Disorder scale) (Lau et al., 2017). In a separate study 2 months later, 47.4% were found to have moderate-to-severe anxiety (State-Trait Anxiety Inventory) (Hou et al., 2015). A time series analysis of patients in a psychiatric clinic showed an increase in anxiety symptoms after the 1989 Romanian Revolution (Lazarescu et al., 1993). Symptoms of anxiety were common among subgroups, including 69% of general practice patients during the 1969 Northern Ireland riots (Lyons, 1971), 55% of children in primary schools near protest sites (Moussa et al., 2015) and nearly half of selected young political activists after the 2011 Arab Spring in Egypt (Matthies-Boon, 2017). Anxiety symptoms were reported for 34% of children attending schools in riot-affected areas in Bombay, India (Dhavale et al., 2002), and in 37% of internally displaced people and 44% of HIV patients in health facilities after the 2007–2008 Kenyan crisis (Bamrah et al., 2013).

Correlates of anxiety

As with depression, anxiety symptoms were correlated with younger age, female sex, lower socioeconomic status (Moussa et al., 2015), being single, and without children (Lau et al., 2017), personal and social resource loss (Hou et al., 2015), political uncertainty (Matthies-Boon, 2017), and exposure to media coverage (Lazarescu et al., 1993).

Other outcomes

Alcohol consumption increased among men and women after Czechoslovakia’s 1989 Velvet Revolution, although this has been attributed to a loosening of political control over drinking rather than the revolution itself (Kubicka et al., 1995, 1998). A school-level aggregated panel design showed that changes in marijuana use following Chile’s 2011 Occupy School movement were due to secular trends rather than the collective actions (Castillo-Carniglia et al., 2017). Findings related to reactive psychosis have been mixed. Reactive psychosis increased among patients during the 1989 Romanian Revolution (Lazarescu et al., 1993). In contrast, the 1969 Northern Ireland riots did not appear to impact reactive psychosis (Lyons, 1971) and was associated with fewer suicides (Lyons, 1972).

Appraisal of studies

Five of the six longitudinal studies received the maximum NOS score (Greenley et al., 1975; Kubicka et al., 1995, 1998; Ni et al., 2016, 2017); all six were assessed to be of ‘good quality’ (cutoff: NOS score ⩾ 7 out of maximum 9) (McPheeters et al., 2012; Patra et al., 2015) (Supplemental Table S5). Of the 22 cross-sectional and serial cross-sectional studies, 5 received the maximum score of 6 (Castillo-Carniglia et al., 2017; Hanson et al., 1995; Hou et al., 2015; Lau et al., 2017; Sasao and Chun, 1994) and 3 had a NOS score of 5 (Galovski et al., 2016; Moussa et al., 2015; Obilom and Thacher, 2008) (Supplemental Table S6).Studies with lower scores generally did not control for confounders such as sociodemographics.

Discussion

To our knowledge, this is the first systematic review of collective actions and mental health. Depression and PTSD are the two most commonly studied mental health outcomes for collective actions, disasters and armed conflicts (Charlson et al., 2019; Goldmann and Galea, 2014). Notably, the prevalence of depression and PTSD following collective actions both in the general population and among victims of violence appear comparable to those experiencing natural disasters, terrorist attacks and armed conflicts (Charlson et al., 2019; Goldmann and Galea, 2014; Hanson et al., 1995; Hou et al., 2015; Kim-Goh et al., 1995; Ni et al., 2016, 2017; Obilom and Thacher, 2008). Risk factors for psychopathology following collective actions included individual characteristics (e.g. female sex, low socioeconomic status, psychological resilience, health status, social support), the nature of the event itself (e.g. exposure to violence, media coverage, level of engagement) and post-event circumstances (e.g. loss of personal and social resources, unemployment) (Goldmann and Galea, 2014; Hanson et al., 1995; Kumar and Fonagy, 2013; Lau et al., 2017; Moussa et al., 2015; Ni et al., 2016, 2017; Norris et al., 2002a, 2002b; North and Pfefferbaum, 2013).

There is some evidence that collective actions may also be associated with better mental health (n = 2; 4% of all studies) such as reduced depression and suicides (Greenley et al., 1975; Lyons, 1972). This finding may be explained by collective actions serving as a cathartic experience when people collectively express grievances (Curran, 1988). Furthermore, greater social cohesion among subpopulations, either supporting or opposing the cause of the collective action, could strengthen social ties, which in turn could buffer the adverse impact of the stressful environment (Curran, 1988; Kawachi and Berkman, 2001; Lyons, 1973).

For natural disasters and terrorist attacks, the threat of harm or death is a key mediator of mental health consequences (Goldmann and Galea, 2014). Yet collective actions can still exact a substantial, pervasive and persistent toll on mental health even in the absence of large-scale violence such as fatalities, looting or arson (Hou et al., 2015; Ni et al., 2016, 2017). Nonviolent collective actions must therefore influence health via alternative mechanisms. Stressors and stress appraisal processes are potential explanation for the mental health impact of protests (Ni et al., 2017). Stressors are operationalised as environmental demands that challenge a person’s adaptive capacity, while stress reactivity refers to a greater risk of mental disorder following stress exposure (Hammen, 2015; Porta, 2014). The importance of stress reactivity in the development of psychiatric disorders has been recognised by its inclusion as the first domain of the Research Domain Criteria matrix (Insel et al., 2010). Thus, risk factors for mental health disorders following collective actions could be interpreted as the product of exposure and the person’s stress reactivity (Hammen, 2005; Ni et al., 2017). Collective actions can generate stressors via interpersonal conflicts and the disruption of health services (Ni et al., 2016; Yimgang et al., 2017). Social media may also act as a stressor due to interactions with other online users holding different ideological views (American Psychological Association, 2016; Bakshy et al., 2015). This could account for the association of social media use (Ni et al., 2016) and ‘unfriending’ (Lau et al., 2016) with depressive sequelae in the 2014 Occupy Central/Umbrella Movement. Protective factors may operate via stress appraisal processes, where higher baseline resilience (a measure of ability to cope with stressors) was associated with fewer depressive sequelae (Ni et al., 2017). Similarly, the protective role of social support might be explained by the stress buffering function of social ties (Goldmann and Galea, 2014; Kawachi and Berkman, 2001).

Limitations

This review is subject to several limitations. First, all identified studies were observational, so causality between the collective action and mental health outcomes cannot be directly inferred (Amenta et al., 2010; Giugni, 1998). It is difficult to ascertain whether changes in mental health were due to the circumstances that led to the collective action or to the collective action itself. Nevertheless, we have highlighted encouraging examples of longitudinal studies (Table 3) with before, during and after data for collective actions (Castillo-Carniglia et al., 2017; Fishbain et al., 1991; Greenley et al., 1975; Ni et al., 2016, 2017; Yimgang et al., 2017), which represent an advance from the preponderance of cross-sectional studies mostly conducted after the event in the literature thus far. Second, collective actions are disparate and our review may have omitted relevant studies. However, we used a combination of terms to identify different types of collective actions and accessed multiple bibliographic databases to broaden our search strategy. Third, there could also be publication bias as the majority of published studies were on riots and revolutions (Earl et al., 2004). As riots and revolutions commonly involve violence and deaths, they are more likely to reveal significant associations with adverse mental health. Fuller reporting of studies that find null or positive associations is needed, as well as research on nonviolent forms of collective actions given emerging evidence that they can also influence mental health (Hou et al., 2015; Lau et al., 2016, 2017; Ni et al., 2016, 2017). Fourth, comparisons between studies are limited by the heterogeneity of collective actions and the samples, a limitation that also applies to the disaster literature (Nandi et al., 2009). Due to the heterogeneity of the identified studies, a meta-analysis may not be appropriate (Higgins and Green, 2012b; Page et al., 2016). Fifth, this heterogeneity and the limited numbers of studies reduce the generalisability of our findings. This could be addressed by developing a standardised exposure assessment for collective actions, analogous to the Traumatic Exposure Severity Scale for disaster-related trauma (Elal and Slade, 2005). Sixth, the studies employed different measures of mental health, which made cross-study comparisons difficult. Future studies should assess mental health using diagnostic interviews or validated standardised scales (Galea and Maxwell, 2009). Finally, while the Cochrane collaboration recommends the NOS for the quality assessment of non-randomised studies (Higgins and Green, 2012a), the NOS was designed for cohort and case–control studies only. Nevertheless, the NOS has been modified with reference to previous reviews for the appraisal of cross-sectional studies (Anglin et al., 2013; Patra et al., 2015).

Implications for future research

Despite the extensive history and prevalence of collective actions around the world, our systematic review only identified 52 relevant studies of mental health outcomes (Figure 1). By comparison, there are more than 280 studies of PTSD following disasters (Nandi et al., 2009), 129 on conflicts (Charlson et al., 2019) and more than 150 studies have been conducted on the mental health consequences of the 11 September terrorist attacks (Perlman et al., 2011). The sparse documentation on collective actions has been attributed to their rapidly evolving and sporadic nature (Slone, 2003). Yet these methodological challenges also apply to disasters and armed conflicts (Galea and Maxwell, 2009).

As the majority of the population is resilient to a traumatic event, it would be more informative to examine trajectories of mental health rather than the population average (Bonanno et al., 2011). Longitudinal studies show that individuals exhibit remarkable resiliency following collective actions or disasters (Ni et al., 2017; Norris et al., 2009). We identify a lack of studies that assess potential mechanisms using theoretical frameworks relevant to mental health outcomes. One relevant framework is the social support deterioration model, which posits that declines in perceived social support after a population event mediate the immediate and delayed psychopathology (Kaniasty and Norris, 1993). The social support deterioration model is most applicable to human-induced events and thus could be relevant to collective actions (Kaniasty and Norris, 2009; Rochford and Blocker, 1991). Indeed, collective actions have the potential to exacerbate ideological divisions (Fraser, 1971; Ni et al., 2016) where disputes with family members over politics, rather than actual participation in protests, were associated with depressive sequelae during the 2014 Occupy Central/Umbrella Movement (Lau et al., 2017; Ni et al., 2016). In order to test the social support deterioration model, future studies should document social dynamics, together with other potential mediators, before, during and after an event as these could be as important in the long-term course of mental health as the event itself (Ajdukovic, 2004; Ni et al., 2017). Emerging methods could also be incorporated, for example, biological markers of stress (Hoyt et al., 2018), harnessing the social media as a form of mental health surveillance (Gruebner et al., 2016) and applying time-varying analytic approaches to disentangle underlying mechanisms (Galea and Maxwell, 2009; Slone et al., 2000).

None of the identified studies examined interventions to improve mental health following a collective action. Future research on interventions targeting risk factors for psychopathology and enhancing resilience is warranted. Psychological first aid, cognitive-behavioural therapy (CBT) and exposure-based therapy have been recommended in the disaster setting (Neria et al., 2009; North and Pfefferbaum, 2013). Interpersonal psychotherapy may also be promising since it could specifically address the social disruptions that are frequently observed in collective actions (Markowitz and Weissman, 2004; Ni et al., 2016, 2017).

Conclusion

Our review shows that the mental health outcomes following collective actions can be comparable to natural disasters, terrorist attacks or armed conflicts (Charlson et al., 2019; Goldmann and Galea, 2014; Hanson et al., 1995; Hou et al., 2015; Kim-Goh et al., 1995; Ni et al., 2016, 2017; Obilom and Thacher, 2008). Health care professionals therefore need to be vigilant to both the short- and medium-term psychiatric sequelae of collective actions. They should be particularly aware of potential community spillover effects, where even those who did not take part in the collective action can experience psychiatric sequelae (Galovski et al., 2016; Lau et al., 2017; Lazarescu et al., 1993; Ni et al., 2016). The Constitution of the World Health Organization (1995) states that ‘health is one of the fundamental rights of every human being without distinction of race, religion, political belief, economic or social condition’. This underlines the importance for all sides to find a peaceful resolution. On the other hand, studies have also suggested positive mental health influences of collective actions (Greenley et al., 1975; Lyons, 1972). Given the extensive research on disasters and conflicts, studies of collective actions may benefit from drawing upon the disaster- and conflict-related literatures, which have been summarised in classic reviews and texts (Charlson et al., 2019; Galea et al., 2005; Goldmann and Galea, 2014; Norris et al., 2002a, 2002b; North and Pfefferbaum, 2002). However, given the qualitative differences between collective actions, disasters and armed conflicts, tailored studies for collective actions are needed. Despite the increase in collective actions globally and their documented mental health consequences (Ortiz et al., 2013), collective actions have often been overlooked as a sociopolitical determinant of mental health. The literature has just begun to assess the mental health impact of collective actions and the underlying mechanisms. Research in this emerging field of inquiry would therefore be timely (Ni et al., 2020).

Supplemental Material

Appendix_clean_Jan9_ver2 – Supplemental material for Mental health during and after protests, riots and revolutions: A systematic review

Supplemental material, Appendix_clean_Jan9_ver2 for Mental health during and after protests, riots and revolutions: A systematic review by Michael Y Ni, Yoona Kim, Ian McDowell, Suki Wong, Qiu Hong, Irene OL Wong, Sandro Galea and Gabriel M Leung in Australian & New Zealand Journal of Psychiatry

Footnotes

Acknowledgements

We thank Irene Wong, Candi Leung, and Cynthia Yau for technical assistance.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: General Research Fund (RGC GRF Grant #17609818), Research Grants Council, University Grants Committee of Hong Kong and Seed Fund for Basic Research for Resubmission of GRF/ECS Proposals 2017/18, University Research Committee, The University of Hong Kong.

ORCID iD

Michael Y Ni

Supplemental material

Supplemental material for this article is available online.

References

Ajdukovic

(2004) Social contexts of trauma and healing. Medicine, Conflict and Survival 20: 120–135.

Amenta

Caren

Chiarello

, et al. (2010) The political consequences of social movements. Annual Review of Sociology 36: 287–307.

American Psychological Association (2016) Stress in America: U.S. Presidential Election 2016. American Psychological Association. Available at: https://www.apa.org/news/press/releases/stress/2016/presidential-election.pdf

Anglin

Samaan

Walter

, et al. (2013) Vitamin D deficiency and depression in adults: Systematic review and meta-analysis. British Journal of Psychiatry 202: 100–107.

Bakshy

Messing

Adamic

(2015) Exposure to ideologically diverse news and opinion on Facebook. Science 348: 1130–1132.

Bamrah

Mbithi

Mermin

, et al. (2013) The impact of post-election violence on HIV and other clinical services and on mental health – Kenya, 2008. Prehospital and Disaster Medicine 28: 43–51.

Banks

Wilson

(2017) Cross-National Time-Series Data Archive. Jerusalem: Databanks International.

Bonanno

Westphal

Mancini

(2011) Resilience to loss and potential trauma. Annual Review of Clinical Psychology 7: 511–535.

Brown

Melchior

Reback

, et al. (1994) Psychological functioning and substance abuse before and after the 1992 Los Angeles riot in a community sample of women. Journal of Psychoactive Drugs 26: 431–437.

10.

Castillo-Carniglia

Kaufman

Pizarro

, et al. (2017) School collective occupation movements and substance use among adolescents: A school-level panel design. Drug and Alcohol Dependence 176: 21–27.

11.

Charlson

Van Ommeren

Flaxman

, et al. (2019) New WHO prevalence estimates of mental disorders in conflict settings: A systematic review and meta-analysis. The Lancet 394: 240–248.

12.

Coudreaut

Frye

Pynoos

(1995) Civil unrest and exacerbation of psychiatric symptoms: A case series. Journal of Nervous and Mental Disease 183: 109–111.

13.

Curran

(1988) Psychiatric aspects of terrorist violence: Northern Ireland 1969-1987. British Journal of Psychiatry 153: 470–475.

14.

Dhavale

Damani

Kedare

, et al. (2002) Impact of riots on children. Indian Journal of Social Work 63: 194–205.

15.

Earl

Martin

Mccarthy

, et al. (2004) The use of newspaper data in the study of collective action. Annual Review of Sociology 30: 65–80.

16.

Elal

Slade

(2005) Traumatic Exposure Severity Scale (TESS): A measure of exposure to major disasters. Journal of Traumatic Stress 18: 213–220.

17.

Fishbain

Aldrich

Goldberg

, et al. (1991) Impact of a humanmade disaster on the utilization pattern of a psychiatric emergency service. Journal of Nervous and Mental Disease 179: 162–166.

18.

Fraser

(1971) The cost of commotion: An analysis of the psychiatric sequelae of the 1969 Belfast riots. British Journal of Psychiatry 118: 257–264.

19.

Galea

Maxwell

(2009) Methodological challenges in studying the mental health consequences of disasters. In: Neria

Galea

Norris

(eds) Mental Health and Disasters. Cambridge: Cambridge University Press. pp. 579–593.

20.

Galea

Nandi

Vlahov

(2005) The epidemiology of post-traumatic stress disorder after disasters. Epidemiologic Reviews 27: 78–91.

21.

Galovski

Peterson

Beagley

, et al. (2016) Exposure to violence during Ferguson protests: Mental health effects for law enforcement and community members. Journal of Traumatic Stress 29: 283–292.

22.

Garbarino

Magnavita

Chiorri

, et al. (2012) Evaluation of operational stress in riot and crowd control police units: A global challenge for prevention and management of police task-related stress. Journal of Police and Criminal Psychology 27: 111–122.

23.

Giugni

(1998) Was it worth the effort? The outcomes and consequences of social movements. Annual Review of Sociology 24: 371–393.

24.

Goldmann

Galea

(2014) Mental health consequences of disasters. Annual Review of Public Health 35: 169–183.

25.

Greenley

Gillespie

Lindenthal

(1975) A race riot’s effect on psychological symptoms. Archives of General Psychiatry 32: 1189–1195.

26.

Gruebner

Sykora

Lowe

, et al. (2016) Mental health surveillance after the terrorist attacks in Paris. The Lancet 387: 2195–2196.

27.

Hammen

(2005) Stress and depression. Annual Review of Clinical Psychology 1: 293–319.

28.

Hammen

(2015) Stress sensitivity in psychopathology: Mechanisms and consequences. Journal of Abnormal Psychology 124: 152–154.

29.

Hanson

Kilpatrick

Freedy

, et al. (1995) Los Angeles County after the 1992 civil disturbances: Degree of exposure and impact on mental health. Journal of Consulting and Clinical Psychology 63: 987–996.

30.

Harvey-Lintz

Tidwell

(1997) Effects of the 1992 Los Angeles civil unrest: Post traumatic stress disorder symptomatology among law enforcement officers. Social Science Journal 34: 171–183.

31.

Higgins

JPT

Green

(2012a) Tools for assessing methodological quality or risk of bias in non-randomized studies [Online]. Available at: http://handbook-5-1.cochrane.org/chapter_13/13_5_2_3_tools_for_assessing_methodological_quality_or_risk_of.htm (accessed 2016).

32.

Higgins

JPT

Green

(2012b) What is heterogeneity? [Online]. Available at: http://handbook-5-1.cochrane.org/chapter_9/9_5_1_what_is_heterogeneity.htm (accessed 2016).

33.

Hou

Bonanno

(2018) Emotions in everyday life during social movements: Prospective predictions of mental health. Journal of Counseling Psychology 65: 120–131.

34.

Hou

Hall

Canetti

, et al. (2015) Threat to democracy: Physical and mental health impact of democracy movement in Hong Kong. Journal of Affect Disorders 186: 74–82.

35.

Hoyt

Zeiders

Chaku

, et al. (2018) Young adults’ psychological and physiological reactions to the 2016 US presidential election. Psychoneuroendocrinology 92: 162–169.

36.

Insel

Cuthbert

Garvey

, et al. (2010) Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychiatry 167: 748–751.

37.

JAMA (2018) Instructions for authors [Online]. Available at: https://jamanetwork.com/journals/jama/pages/instructions-for-authors#SecSystematicReviews (accessed 14 June 2018).

38.

Kaniasty

Norris

(1993) A test of the social support deterioration model in the context of natural disaster. J Pers Soc Psychol 64: 395–408.

39.

Kaniasty

Norris

(2009) Distinctions that matter: Received social support, perceived social support, and social embeddedness after disasters. In: Neria

Galea

Norris

(eds) Mental Health and Disasters. Cambridge: Cambridge University Press. pp. 175–200.

40.

Kawachi

Berkman

(2001) Social ties and mental health. Journal of Urban Health 78: 458–467.

41.

Kim-Goh

Suh

Blake

, et al. (1995) Psychological impact of the Los Angeles riots on Korean-American victims: Implications for treatment. American Journal of Orthopsychiatry 65: 138–146.

42.

Klee

Gorwitz

(1970) Effects of the Baltimore riots on psychiatric hospital admissions. Mental Hygiene 54: 447–449.

43.

Kubicka

Csémy

Kozeny

(1995) Prague women’s drinking before and after the ‘velvet revolution’ of 1989: A longitudinal study. Addiction 90: 1471–1478.

44.

Kubicka

Csémy

Duplinsky

, et al. (1998) Czech men’s drinking in changing political climates 1983-93: A three-wave longitudinal study. Addiction 93: 1219–1230.

45.

Kumar

Fonagy

(2013) Differential effects of exposure to social violence and natural disaster on children’s mental health. Journal of Traumatic Stress 26: 695–702.

46.

Lau

Kim

, et al. (2017) The Occupy Central (Umbrella) movement and mental health distress in the Hong Kong general public: Political movements and concerns as potential structural risk factors of population mental health. Social Psychiatry and Psychiatric Epidemiology 52: 525–536.

47.

Lau

Hou

Hall

, et al. (2016) Social media and mental health in democracy movement in Hong Kong: A population-based study. Computers in Human Behavior 64: 656–662.

48.

Lazarescu

Ardeleanu

Jianu

(1993) Study on the reactive psychosis occurred during the revolution of December 1989: Timisoara, Rumania. European Journal of Psychiatry 7: 219–228.

49.

Liberati

Altman

Tetzlaff

, et al. (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: Explanation and elaboration. British Medical Journal 339: b2700.

50.

Lyons

(1971) Psychiatric sequelae of the Belfast riots. British Journal of Psychiatry 118: 265–273.

51.

Lyons

(1972) Depressive illness and aggression in Belfast. British Medical Journal 1: 342–344.

52.

Lyons

(1973) Violence in Belfast: A review of the psychological effects. Public Health 5: 163–168.

53.

McPheeters

Kripalani

Peterson

, et al. (2012) Closing the quality gap: Revisiting the state of the science (Vol. 3: Quality improvement interventions to address health disparities). Evidence Report/Technology Assessment 3: 1–475.

54.

Markowitz

Weissman

(2004) Interpersonal psychotherapy: Principles and applications. World Psychiatry 3: 136.

55.

Matthies-Boon

(2017) Shattered worlds: Political trauma amongst young activists in post-revolutionary Egypt. Journal of North African Studies 22: 620–644.

56.

Moussa

El Kholy

Enaba

, et al. (2015) Impact of political violence on the mental health of school children in Egypt. Journal of Mental Health 24: 289–293.

57.

Nandi

Tracy

Beard

, et al. (2009) Patterns and predictors of trajectories of depression after an urban disaster. Annals of Epidemiology 19: 761–770.

58.

Neria

Galea

Norris

(eds) (2009) Mental Health and Disasters. Cambridge: Cambridge University Press.

59.

Pang

, et al. (2016) Direct participation in and indirect exposure to the Occupy Central Movement and depressive symptoms: A longitudinal study of Hong Kong adults. American Journal of Epidemiology 184: 636–643.

60.

Yao

Leung

KSM

, et al. (2020) Depression and post-traumatic stress during major social unrest in Hong Kong: a 10-year prospective cohort study. The Lancet. Epub ahead of print 9 Jan. DOI: 10.1016/S0140-6736(19)33160-5.

61.

Pang

, et al. (2017) Longitudinal patterns and predictors of depression trajectories related to the 2014 Occupy Central/Umbrella Movement in Hong Kong. American Journal of Public Health 107: 593–600.

62.

Norris

Friedman

Watson

(2002a) 60,000 disaster victims speak: Part II. Summary and implications of the disaster mental health research. Psychiatry 65: 240–260.

63.

Norris

Friedman

Watson

, et al. (2002b) 60,000 disaster victims speak: Part I. An empirical review of the empirical literature, 1981–2001. Psychiatry 65: 207–239.

64.

Norris

Tracy

Galea

(2009) Looking for resilience: Understanding the longitudinal trajectories of responses to stress. Social Science & Medicine 68: 2190–2198.

65.

North

(2014) Current research and recent breakthroughs on the mental health effects of disasters. Current Psychiatry Reports 16: 481.

66.

North

Pfefferbaum

(2002) Research on the mental health effects of terrorism. The Journal of the American Medical Association 288: 633–636.

67.

North

Pfefferbaum

(2013) Mental health response to community disasters: A systematic review. The Journal of the American Medical Association 310: 507–518.

68.

Obilom

Thacher

(2008) Posttraumatic stress disorder following ethnoreligious conflict in Jos, Nigeria. Journal of Interpersonal Violence 23: 1108–1119.

69.

Ortiz

Burke

Berrada

, et al. (2013) World Protests 2006-2013. New York: Initiative for Policy Dialogue and Friedrich-Ebert-Stiftung.

70.

Ouanes

Bouasker

Ghachem

(2014) Psychiatric disorders following the Tunisian revolution. Journal of Mental Health 23: 303–306.

71.

Ouanes

Bouasker

Zgueb

, et al. (2012) Tunisian revolution events: Psychological impact on emergency personnel. European Psychiatry 27: 1.

72.

Page

Shamseer

Altman

, et al. (2016) Epidemiology and reporting characteristics of systematic reviews of biomedical research: A cross-sectional study. PLOS Medicine 13: e1002028.

73.

Patel

Nath

(2013) Where identity and trauma converge: Hindu-Muslim perceptions of the 2002 Gujarati riots. Journal of Muslim Mental Health 7. DOI: 10.3998/jmmh.10381607.0007.202.

74.

Patra

Bhatia

Suraweera

, et al. (2015) Exposure to second-hand smoke and the risk of tuberculosis in children and adults: A systematic review and meta-analysis of 18 observational studies. PLOS Medicine 12: e1001835.

75.

Perlman

Friedman

Galea

, et al. (2011) Short-term and medium-term health effects of 9/11. The Lancet 378: 925–934.

76.

Porta

(2014) A Dictionary of Epidemiology. New York: Oxford University Press.

77.

Rochford

Blocker

(1991) Coping with ‘natural’ hazards as stressors: The predictors of activism in a flood disaster. Environment and Behavior 23: 171–194.

78.

Sasao

Chun

(1994) After the Sa-i-gu (April 29) Los Angeles riots: Correlates of subjective well-being in the Korean-American community. Journal of Community Psychology 22: 136–152.

79.

Slone

(2003) The Nazareth riots – Arab and Jewish Israeli adolescents pay a different psychological price for participation. Journal of Conflict Resolution 47: 817–836.

80.

Slone

Kaminer

Durrheim

(2000) The contributions of political life events to psychological distress among South African. Political Psychology 21: 465–487.

81.

Stephan

Chenoweth

(2008) Why civil resistance works: The strategic logic of nonviolent conflict. International Security 33: 7–44.

82.

Stuber

Nader

Pynoos

(1997) The violence of despair: Consultation to a HeadStart program following the Los Angeles uprising of 1992. Community Mental Health Journal 33: 235–241.

83.

Tan

E-S

Simons

(1973) Psychiatric sequelae to a civil disturbance. British Journal of Psychiatry 122: 57–63.

84.

Tarrow

(2011) Power in Movement: Social Movements and Contentious Politics. New York: Cambridge University Press.

85.

Unuvar

Yilmaz

Ozyildirim

, et al. (2017) Usage of Riot Control Agents and other methods resulting in physical and psychological injuries sustained during civil unrest in Turkey in 2013. Journal of Forensic and Legal Medicine 45: 47–52.

86.

Wang

Zhang

Shi

, et al. (2011) Comparing alternative factor models of PTSD symptoms across earthquake victims and violent riot witnesses in China: Evidence for a five-factor model proposed by Elhai et al. (2011). Journal of Anxiety Disorders 25: 771–776.

87.

World Health Organization (1995) Constitution of the World Health Organization. Geneva: World Health Organization.

88.

Yimgang

Wang

Paik

, et al. (2017) Civil unrest in the context of chronic community violence: Impact on maternal depressive symptoms. American Journal of Public Health 107: 1455–1462.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB