The relation of unrest-related distress with probable depression during and after widespread civil unrest

Background This study investigated whether subjective unrest-related distress was associated with probable depression during and after the 2019 anti-ELAB movement in Hong Kong. Methods Population-representative data were collected from 7157 Hong Kong Chinese in four cross-sectional surveys (July 2019–July 2020). Logistic regression examined the association between subjective unrest-related distress and probable depression (PHQ-9 ⩾ 10), stratified by the number of conflicts/protests across the four timepoints. Results Unrest-related distress was positively associated with probable depression across different numbers of conflicts/protests. Conclusion Unrest-related distress is a core indicator of probable depression. Public health interventions should target at resolving the distress during seemingly peaceful period after unrest.


Introduction
In 2019, a controversial bill was introduced by the Hong Kong SAR Government for extraditing individuals accused of crimes to countries/regions which do not have a legal mechanism for doing so. The resultant anti-extradition law amendment bill (anti-ELAB) movement gradually escalated from peaceful demonstrations into massive violence and vandalism since June 2019. Protests diminished in early 2020 due to the coronavirus disease 2019 (COVID-19) but mounted again in May-June 2020 due to the hasty introduction of the National Security Law. The movement involved approximately 2000000 participants lasting over months, leaving serious mental health consequences (Holbig, 2020;Turnbull et al., 2020). The prevalence of depression rose from 1.9% a decade ago to as high as 25% in the heat of the movement (Ni et al., 2020;Hou et al., 2021). Given that depression forms one of the heaviest public health burdens and elevates the suicidal risks (WHO, 2021), it is pressing to further elucidate its development and maintenance under potentially traumatic events such as social unrests.
Risk factors of psychopathology embed within the event level (e.g. exposure, proximity) or the individual level (e.g. a lack of personal coping resources) (Galovski et al., 2016;Wong et al., 2021a). The anti-ELAB movement incubated, developed, and faded over a dynamic process, and this synchronized with the trend of depression over time (Hou et al., 2022). Subjective experiences of support and relationship quality with close social partners were positively associated with better mental health among adolescents during the anti-ELAB movement (Wong et al., 2021b). An experience sampling study illustrated that individuals experienced significant negative changes in their everyday emotions during a social movement, which predicted subsequent higher psychiatric symptoms and lower subjective well-being (Hou and Bonanno, 2018).
It is intuitive that individuals experience higher levels of stress with greater intensity of conflicts/protests, with direct or indirect exposure to street protests and related social media posts positively associated with elevated psychological distress, probable depression, and even suicidal ideation (Galovski et al., 2016;Turnbull et al., 2020;Hou et al., 2021;Lam et al., 2021). However, currently limited research has investigated whether the association between subjective unrest-related distress and psychiatric symptoms changes as a function of the intensity of unrest over time. There is also a deficit of knowledge about the comparative centrality of subjective experience of traumatization and objective facets of event exposures in understanding mental health consequences (Maschi et al., 2011;Boals, 2018;Su, 2018;Danese and Widom, 2020). Existing evidence is mainly based on recalls of objective and/or subjective traumatic experiences over a fraction of one's life history (Maschi et al., 2011;Boals, 2018;Danese and Widom, 2020), but less so following instant measures of civil unrest as an acute traumatic exposure.
This study aims to examine the association of subjective unrest-related distress with probable depression in Hong Kong between July 2019 and July 2020, a period with varying intensity of widespread conflicts/protests (Holbig, 2020). We also tested whether the positive association between unrest-related distress and depression remained the same across different number of conflicts/protests at different timepoints. This study hypothesized that unrest-related distress would be positively associated with probable depression. The positive association between unrestrelated distress and depression would be consistent across different numbers of conflicts/protests over time.

Sample
This study was approved by the Human Research Ethics Committee of The Education University of Hong Kong. This study analyzed data from four population-representative samples of Hong Kong Chinese in serial cross-sectional surveys in July 2019 (T1), February 2020 (T2), April 2020 (T3), and July 2020 (T4) (n = 7157). The respondents were approached through random digit dialing by a Computer-Assisted Telephone Interview system, with contact numbers drawn from the Hong Kong Communication Authority databases. Inclusion criteria were: (1) Hong Kong Chinese, (2) aged ⩾15 years, (3) Cantonese speaking. Oral informed consent was obtained before each interview. The cooperation rates (i.e. completed over eligible) across the four surveys were 72.0%, 67.3%, 73.5% and 72.8%. More details were documented in prior studies (Lai et al., 2021;Hou et al., 2022). Demographics are summarized in Table 1.

Probable depression
Respondents rated nine items on depressive symptoms over the past two weeks on a 4-point scale (0 = not at all, 1 = on several days, 2 = on more than half of the days, 3 = nearly every day) using the Chinese version 9-item Patient Health Questionnaire (PHQ-9) (Yeung et al., 2008). Higher scores (range = 0-27) indicated higher levels of depressive symptoms. Scores of 10 or above indicated probable depression (Levis et al., 2019). Cronbach's alphas were 0.842, 0.854, 0.850 and 0.827 across the four administrations respectively.

Unrest-related distress
Respondents rated to what extent they felt distress over (1) the government's handling of unrest, (2) confrontations between police and protestors, and the use of riot control measures such as physical assault, tear gas, and rubber bullets, and (3) widespread and ongoing demonstrations and protests on a 4-point scale (0 = not at all, 1 = some, 2 = quite a bit, 3 = a lot). Scores of each item were recoded into low (0) and high (1), with low indicating 'not at all'/'some' in all three items and high indicating 'quite a bit'/'a lot' in at least one item. Binary coding of distress relating to large-scale disasters demonstrated statistical sensitivity in detecting important mental health problems in previous studies (Bonanno et al., 2007;Hou et al., 2022).

Exposure to unrest
Exposure to unrest was quantified as the number of conflicts/protests occurred during the past month at each timepoint. Event count has been widely used to reflect the objective levels of traumatic stress across previous studies. For example, evidence showed good predictive utility of the number of traumatic events on PTSD among a postconflict population (Wilker et al., 2015) and first responders (Geronazzo-Alman et al., 2017). The number of events related to the anti-ELAB movement from June 2019 to July 2020 was extracted from the Armed Conflict Location and Events Dataset (ACLED) (Raleigh et al., 2010). ACLED is a database of real-time data on political violence and protest events across the globe. We adopted strict inclusion criteria for data extraction, and counted events as valid only if (1) the item description contained keywords like 'name of the unrest' and 'slogans' and 'event name' related to the anti-ELAB movement and (2) the item involved conflicts/protests as categorized in ACLED. Duplicates were excluded. Event count was the approximate of the potential level of exposure individuals could possibly encounter at that timepoint.

Demographics
Respondents' age, gender, education level, employment status, marital status, and monthly household income were recorded.

Analytic plan
The trends of the prevalence of probable depression, unrestrelated distress, and number of conflicts/protests across the four timepoints were described. Logistic regression models were used to examine the association of subjective unrest-related distress ('high' = 1) with probable depression (PHQ-9 scores:10-27 = 1), controlling for COVID-19 stress and socio-demographics. The association between unrest-related distress and probable depression was then stratified by the number of conflicts/protests across timepoints. Adjusted odds ratio (aOR) with 95% CI indicated the independent association of each correlate with the outcome. All analyses were performed using SPSS (Version 26).

Descriptive statistics
The prevalence of probable depression was 23.8% in July 2019, 25.9% in February 2020, 14.2% in April 2020, and 14.5% in July 2020. The prevalence of unrest-related distress was 66.8% in July 2019, 79.6% in February 2020, 78.4% in April 2020, and 77% in July 2020. We extracted a total of 1227 conflicts/protests relating to the anti-ELAB movement between 6th June 2019 and 31st July 2020. There were 98 conflicts/protests in July 2019, 48 in February 2020, 42 in April 2020, and 97 in July 2020.

Logistic regressions
Controlling for COVID-19 stress and socio-demographics and taking into consideration the potential association between number of conflicts/protests and probable depression, unrest-Global Mental Health related distress was associated with 211% increased odds of probable depression (aOR 3.11, 95% CI 2.19-4.43). The positive association between unrest-related distress and probable depression remained consistent and significant across the four timepoints with varying numbers of conflicts/protests ( p = 0.108-0.686) ( Table 2). The number of conflicts/protests (i.e. 'exposure to unrest' events) at each timepoints is presented. c Scores of 10 or above in the 9-item Patient Health Questionnaire (PHQ-9) were used to define probable depression.

Discussion
This study aims to investigate the association of subjective unrest-related distress with probable depression amid and after widespread civil unrest in Hong Kong between July 2019 and July 2020. We also examined whether the positive association between unrest-related distress and depression changed as a function of the number of conflicts/protests over time.
Unrest-related distress was consistently associated with higher odds of probable depression, and the association remained significant at different timepoints. Importantly, this association between unrest-related distress and probable depression held in spite of the slight difference in the patterns of distress (sustained prevalence) and probable depression (decreased prevalence), as this difference could have suggested an increase in the 'dosage' of unrest-related distress needed to sufficiently trigger probable depression overtime. The role of unrest-related distress in predicting probable depression during and after widespread civil unrest could be explained by the cognitive model of stress-related disorders (e.g. depression, PTSD), for the easily accessible traumatic memories and/or the negative appraisals of such memories could underlie the perpetuating psychiatric symptoms beyond the end of objective incidents (Ehlers and Clark, 2000;Creamer et al., 2005;Rubin et al., 2008;Duyser et al., 2020). Rumination has been identified as a prospective predictor of more severe depressive symptoms in a month's time amidst combined social unrest and COVID-19 (February-March 2020) in Hong Kong (Wong et al., 2021b).
Our current results are consistent with not only the conclusion that subjective perceptions of traumatic incidents explain depressive symptoms independent of objective lifetime exposure to these incidents (Boals, 2018) but also the notion that mental health consequences following exposure to potentially traumatic events might be fairly minimal in the absence of negative subjective appraisals (Maschi et al., 2011;Danese and Widom, 2020). There is evidence showing that only a fraction of objective traumatic events was subjectively experienced by individuals, suggesting a discrepancy between objective events and subjective experiences amid trauma (Creamer et al., 2005;Maschi et al., 2011;Boals, 2018). A previous study involving burn survivors found that individuals could experience intrusive memories two years after the accident exposure, and whether they cognitively  Abbreviations. aOR, adjusted odds ratio; CI, confidence interval. Note. p values are two-sided, * p < 0.05, ** p < 0.01, *** p < 0.001. a Scores of 10 or above in the 9-item Patient Health Questionnaire (PHQ-9) were used to define probable depression. b US$1 ≈ HK$7.80. c The number of conflicts/protests (i.e. 'exposure to unrest' events) at each timepoint is presented in bracket.
appraised these memories in a negative way was related to the severity of both depressive and PTSD symptoms, over and beyond the adverse consequences brought by burn-related disabilities (Su, 2018). Our current results are generally in line with this. While the trends of depression and number of conflicts/protests were relatively independent, the positive association between unrestrelated distress and probable depression remained significant irrespective of the exact numbers of conflicts/protests across time, providing further supportive evidence that mental health during and after political unrest is dependent upon subjective experiences, regardless of objective criteria defining traumatic incidents (e.g. intensity of protests).
Our study has some limitations. First, repeated cross-sectional data could reflect population-level but not intra-personal changes in unrest-related distress and probable depression as the civil unrest unfolded. Second, the current objective measure of conflicts/protests did not take into account heterogeneity in the nature and intensity across these incidents, or individual differences in the actual exposure to conflicts/protests. Finally, probable depression was based on self-report. Future studies would demonstrate a clearer picture of depression using clinical diagnoses/interviews. Notwithstanding the limitations, this study has built upon our own work and further investigated the subjective versus objective pillars of mental health impact following acute traumatic exposures. Accurate screening and effective interventions are needed to complement the existing already-overloaded mental health care system (Galovski et al., 2016;Hou and Hall, 2019;Hou et al., 2022). The main implication of the current study is that high unrest-related distress could be one core component of continuous mental health assessment and interventions irrespective of the presence/absence of actual incidents. Its utility could last from the acute phase of the unrest to the seemingly peaceful time after the unrest.
Financial support. This work was supported by the Policy Innovation and Co-ordination Office, Hong Kong SAR Government [grant number SR2020.A5.019]; and the Research Grants Council, University Grants Committee, Hong Kong SAR, China [grant numbers C7069-19 GF, 18600320]. The funding source had no role in any process of our study.

Conflict of interest. None.
Ethical standards. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.