Suicide in Hong Kong during the COVID-19 pandemic: an observational study

Introduction

The global impact of the COVID-19 pandemic has had far-reaching consequences, including a concerning rise in suicide rates (SR). According to the World Bank, the pandemic has caused a contraction of economic activity in nearly 90% of countries, surpassing the declines witnessed during World Wars and the Great Depression.1 This economic downturn has significantly impacted households, with surveys indicating that over a third of respondents have ceased working due to the pandemic and 64% of households have experienced income reductions. The unprecedented changes and restrictions associated with the pandemic have contributed to heightened levels of stress, isolation and uncertainty, thereby elevating the risk of suicide. Studies examining the impact of COVID-19 on suicide mortality have yielded mixed findings. Some research has suggested an increase in SR during the pandemic, potentially attributed to factors such as economic stress, social isolation and disruptions to mental health.2 Conversely, other sources have indicated that there is no definitive evidence of a change in SR since the onset of the pandemic.3 4

Hong Kong, a prosperous metropolis located in the southeast of China, is an autonomous region with a heavily externally oriented economy. Hong Kong was greatly impacted by the COVID-19 pandemic. Specifically, the overall excess mortality per 100 000 population was 25 in 2020,5 and the unemployment rate hit 7.2% in December 2020, the highest value in 16 years.6 Daily arrivals into the city dramatically declined after the Chinese New Year holidays at the end of January 2020, and further plummeted following the implementation of home quarantine arrangements for all arrivals from mainland China on 8 February 2020.7 Furthermore, Hong Kong experienced a relatively high COVID-19 fatality rate, as evidenced by the case fatality rate during the Omicron outbreak, which was significantly higher than that of Singapore, standing at 0.53% compared with 0.06%.8 The prolonged stress and uncertainty caused by the pandemic can lead to profound risks of mental health issues in the city. Particularly, one research conducted in the aftermath of the 2003 SARS outbreak in Hong Kong revealed a notable increase in the SR among the elderly population.9 Considering Hong Kong’s ageing population and the more severe nature of COVID-19, the risk of suicide may be even higher than what was observed during the 2003 SARS outbreak.

However, reliable empirical evidence establishing a clear link between the COVID-19 pandemic and suicide mortality remains scarce, and conclusions drawn from previous research on whether the SR will rise as the pandemic spreads have been inconsistent. Some evidence suggests that deaths by suicide have increased, and mental health has deteriorated during the pandemic.2 10 Nevertheless, several previous reviews propose that although SR may sometimes increase following public health emergencies, these changes may not necessarily occur immediately and there may even be an initial reduction in risk.11 In contrast, a comprehensive international study encompassing data from 33 countries revealed that the majority of these nations did not experience an increase in suicide cases (SC).12 Other studies have also demonstrated no significant increases or even decreases in suicide deaths, especially during the initial months of the pandemic.13 The impact of the pandemic on suicide outcomes can vary depending on factors such as a country’s public health control measures, sociocultural and demographic structures, availability of digital alternatives to face-to-face consultation and existing support systems.11 Compared with other regions and countries examined in similar studies, Hong Kong may have suffered more severe economic shocks and stricter travel restrictions during the pandemic due to its unique economic structure and geographical location. To the best of our knowledge, this paper represents the first focused investigation into the suicide pattern in Hong Kong during the COVID-19 pandemic.

This study provides large-scale evidence examining the association between the COVID-19 pandemic and changes in the SR in Hong Kong from 2020 to 2022. To achieve this, we aggregated individual SC at the district-month level and tested two hypotheses in this paper. First, we hypothesised that there was a corresponding change in the SR following the evolving trends of COVID-19 situations and anti-COVID-19 policy events. Second, taking into account demographic and socioeconomic characteristics, we posited that certain population groups were more susceptible to the adverse effects of the pandemic.

Materials and methods

Results

Main results

Figure 1 depicts the yearly average of combined cases of suicide deaths and suicide attempts per 100 000 population in each district in Hong Kong from 2019 to 2022. We found the majority of districts witnessed an elevated SR during the COVID-19 period. Comparable trends were also observed in SDC (online supplemental figure S4). Furthermore, online supplemental table S3 demonstrates that districts with lower proportions of owner-occupiers exhibited higher SR. However, the income level, number of local households and average household size did not exhibit statistically significant associations with local SR.

Request permission

Figure 1

Yearly suicide incident cases per 100 000 population in Hong Kong from 2019 to 2022. (A) Suicide incident cases per 100 000 population in each district of Hong Kong in 2019. (B) Suicide incident cases per 100 000 population in each district of Hong Kong in 2020. (C) Suicide incident cases per 100 000 population in each district of Hong Kong in 2021. (D) Suicide incident cases per 100 000 population in each district of Hong Kong in 2022.

Figure 2 presents the monthly SC in Hong Kong from 2019 to 2022, highlighting the pandemic situation. The y-axis demonstrates the number of monthly SC and SDC. The implementations of key COVID-19-related policies are marked in dotted lines. A steadily increasing trend in both SC and SDC was identified in waves 1 and 2 (figure 2). After that, SC and SDC declined substantially between waves 2 and 3. However, SC increased again significantly during wave 3 in May 2020, coinciding with the implementation of quarantine requirements for all travellers. Following a slight decrease in SC during wave 4, SC and SDC dropped remarkably in October and December 2021, attributed to the CVS 2021 by the government. Wave 5 exhibited the highest increase in SC (March and May 2022), although the impact was mitigated by the introduction of the CVS in 2022.

Request permission

Figure 2

Monthly suicide cases in Hong Kong from 2019 to 2022. The dotted lines indicate the implementation of important antipandemic policies. CVS, Consumption Voucher Scheme.

Figure 3 reports the estimated IRR of SC for each month from 2020 to 2022 relative to the corresponding months in the year 2019 (see online supplemental methods and table S1). The estimates, accounting for district-specific effects, revealed a 30% decline in the overall SR (IRR=0.70, 95% CI 0.56 to 0.87) during the initial COVID-19 outbreak compared with previous years. In contrast, the SR increased by 33% (IRR=1.33, 95% CI 1.02 to 1.74) during wave 2 and 51% (IRR=1.51, 95% CI 1.09 to 2.09) during wave 5 (online supplemental table S1 and figure 3). A declining suicide trend was apparent during the third wave, with a 37% decrease in October 2021 and a 23% decrease in November 2021. In online supplemental figures S4 and S5, we test the robustness of these results. Online supplemental figure S5 examined the change in IRR of SDC. Online supplemental figure S6 used ordinary least squares regression instead of Poisson regression to test the fluctuation of suicide incidence patterns with local COVID-19 developments. All the results remained similar to the baseline pattern.

Request permission

Figure 3

Changes in suicide incidence rate ratio (IRR) for each month from 2020 to 2022, relative to the corresponding months in the year 2019.

Heterogeneity across subgroups

Figure 4 describes the cumulative SC in Hong Kong from 2019 to 2020 in the entire population and across different gender and age groups. Notably, a significant rise in cumulative SC was observed among the elderly population (65+ years old), while the total number of cases remained relatively stable among children and adolescents. In 2019, the cumulative SCs among elders aged over 65 years were 318, while this number reached 410 in 2022, bringing about 92 excess SC during the pandemic. Similar patterns were observed for SDC, as depicted in online supplemental figure S7.

Request permission

Figure 4

Cumulative suicide cases in the entire population and in different gender and age groups in Hong Kong from 2019 to 2022. (A) Results of the cumulative number of suicide cases among the whole population from 2019 to 2022. (B, C) Results of comparing cumulative number of suicide cases among males (B) and females (C). (D–F) Results for different age groups: under 25 years (D), 25 - 65 years (E) and ≥65 years (F).

We then formally and separately investigated the IRR across different gender groups (figure 5A,B), age groups (figure 5C–F) and economic status groups (figure 5G,H). These analyses revealed notable heterogeneous effects of the pandemic among these population groups. Figure 5A,B investigate the impacts across different gender groups. We did not observe a significant difference in the SR between females and males during waves 1–4. However, the SR increased dramatically among the female population in March 2022 (IRR=1.77, 95% CI 1.04 to 3.02) and May 2022 (IRR=1.67, 95% CI 1.13 to 2.46) during wave 5 (online supplemental table S5). By contrast, the SR for males only increased by 35% in March 2022 (IRR=1.35, 95% CI 0.95 to 1.92).

Request permission

Figure 5

Changes in suicide incidence rate ratio (IRR) in heterogeneous population groups from 2020 to 2022. (A-B) Results of comparing cumulative number of suicide cases among males (A) and females (B). C–F) Results for different age groups: below 25 years (C), 25–45 years (D), 45–65 years (E) and ≥65 years (F). (G, H) Results among population groups residing in different housing conditions: cases from private houses (G), cases from public houses (H).

Figure 5C–F explore the impacts across different age groups. We found that SR declined consistently among younger adults (individuals aged 25–45 years). For adults below 25 years old, the most significant decrease occurred in August 2020 (IRR=0.38, 95% CI 0.16 to 0.91) and February 2022 (IRR=0.22, 95% CI 0.05 to 0.95). For adults aged 25–45 years, a more pronounced decline in SR was observed in February 2020 (IRR=0.39, 95% CI 0.15 to 1.05) and December 2021 (IRR=0.53, 95% CI 0.29 to 0.96). In contrast, SR increased markedly among older adults (aged over 65 years) during wave 5, with the most pronounced elevation observed in March 2022 by 127% (IRR=2.27, 95% CI 1.50 to 3.42) and in May 2022 by 47% (IRR=1.47, 95% CI 1.03 to 2.09).

Figure 5G,H examine the heterogeneous pattern of SC among population groups residing in different housing conditions. In Hong Kong, living in public houses typically represents a relatively lower household socioeconomic status (SES). Due to the lack of specific population data for these two population groups in each district, the outcome variable in figure 5G,H presents the number of SC at the district-year-month level. We observed a substantial increase in the SR among the population with lower SES in March 2022 (IRR=1.19, 95% CI 0.10 to 2.27), while no significant increase in SR was observed among the population living in private houses.

Discussion

This study investigated the suicide patterns in Hong Kong during the COVID-19 pandemic from 2020 to 2022. There was a noticeable corresponding change in the SR with the trends of COVID-19 situations and the anti-COVID-19 policy events, particularly the policies of quarantine requirements for all travellers and the government’s CVS. Moreover, we highlighted that the impacts of the pandemic were not evenly distributed among different population groups, with distinct differences observed based on demographic and socioeconomic characteristics. Specifically, the SRs were more likely to increase among the elderly groups and the low-income population.

Consistent with previous research findings, our study observed a trend of increasing suicides during and after pandemics.18 19 During the COVID-19 pandemic, the number of SCs in Hong Kong increased steadily from waves 1–2 to wave 4, followed by a dramatic surge during wave 5, primarily attributed to the Omicron outbreak. Several factors may have influenced these SRs, including pre-existing health inequalities related to gender, age, economic status and underlying health conditions. Additionally, the implementation of strict measures such as isolating infected individuals, mandatory quarantine requirements, travel restrictions, school closures, mask mandates and compulsory social distancing would also burden the population’s mental health, leading to heightened emotional distress and an increased risk of subsequent psychiatric symptoms.20

Considering the discrepancy in gender identity and social norms, the suicide patterns between males and females might be different. In this study, we observed that the SR of the female population increased dramatically in March 2022 (IRR=1.77, 95% CI 1.04 to 3.02) and May 2022 (IRR=1.67, 95% CI 1.13 to 2.46). Comparatively, SC among males only increased by 35% in March 2022 (IRR=1.35, 95% CI 0.95 to 1.92) and did not show a noticeable growth in May 2022. Females are known to be more likely to bear the brunt of the social and economic consequences of the pandemic,21 and can consequently develop various forms of mental disorders, including depression, anxiety, post-traumatic stress disorder and stress.22 During the studied pandemic period, females experienced greater psychological distress due to their over-representation in industries negatively impacted by COVID-19, such as retail, service and healthcare. In addition to the disproportionate effects of employment disruptions on females, several research studies also indicate that females may have a higher prevalence of certain mental disorders due to differential neurobiological responses to stressors.23

The downsizing of the economy and the overwhelming focus of the medical system on the COVID-19 pandemic can potentially lead to unintended long-term consequences for vulnerable groups on the fringes of society. Individuals with chronic diseases and a history of medical or psychiatric illnesses showed more symptoms of anxiety and stress, consequently having an increased SR under the strict quarantine measures, travel restrictions, physical distancing protocols and lockdown policies implemented during the pandemic.24 The pre-existing conditions rendered them susceptible to infection and heightened the risk of mortality, with certain illnesses even compromising their immune systems, such as systemic lupus erythematosus.25 In Hong Kong, where over 18% of the population is aged over 65 years, only 3.6% of them reside in residential institutions or nursing homes. Compared with children and adolescents, older adults may possess inadequate resources to effectively cope with the stress brought about by the COVID-19 crisis. This may include material (eg, lack of access to digital devices and the internet), social (eg, limited family or social support), cognitive (eg, limited exposure to health-related information and knowledge) and biological (eg, pre-existing chronic diseases or the inability to engage in physical exercise) resources.26 Several reports also reveal substantially higher mortality rates among patients with diabetes, hypertension and other coronary heart diseases, yet the exact causes remain unknown,27 leaving those with these common chronic conditions in a state of fear and uncertainty.

In response to the crisis, the Hong Kong government initiated the CVS, which aimed to alleviate economic distress and potentially contribute to the decline in SC during the pandemic waves 4 and 5. The impact of this policy might be particularly noteworthy among the population with low household SES. In our subgroup analysis, we indeed observed a significant increase in the SR exclusively among the population residing in public houses (IRR=1.19, 95% CI 0.10 to 2.27) during wave 5. Conversely, we consistently observed a decline in SR among the population living in private houses from wave 1 to wave 5. It is important to note that in Hong Kong, only people with lower SES are eligible to apply for public housing. The pandemic has exacerbated the financial pressure for low SES families, making it more challenging to maintain a quality life and increasing job insecurity. Consequently, the low SES population may need to work longer hours to secure employment and meet daily necessities for their families,28 resulting in feelings of helplessness, fear and anxiety. Our findings were generally consistent with previous large-scale local surveys that have demonstrated a few significant differences in negative impacts on mental health between populations at the higher and lower ends of the SES spectrum, as well as between populations with higher and lower levels of depressive symptoms prior to the onset of the COVID-19 pandemic.29

Finally, it is important to acknowledge the limitations of this study. First, the accuracy of the media-reported suicide data may be compromised, as there is a possibility that some suicide deaths or attempts were not discovered or were reported with some delay. Previous studies have highlighted the inadequate precision of suicide risk factor reporting by the mass media in Hong Kong, potentially resulting in under-reporting of SR city-wide and throughout the pandemic.30 To address this concern, we performed a sensitivity analysis comparing our data with the aggregated statistics from the Coroner’s Court in Hong Kong. The findings from both datasets were consistent, providing reassurance regarding the reliability of our main results. Second, we used 1 year of prepandemic data from 2019 due to availability. This approach may introduce some instability when estimating the impact of COVID-19 on SR by juxtaposing 2019 data with postpandemic data spanning 2020–2022. Future studies incorporating a longer prepandemic period would yield more robust estimates of the pandemic’s influence on SR. Third, given that this study was conducted within the specific context of Hong Kong, caution must be exercised when extrapolating our conclusions to other regions with differing socioeconomic, cultural and healthcare landscapes, particularly in light of varying anti-COVID-19 policy measures. Prior investigations have underscored the presence of heterogeneous pandemic effects on SR across diverse demographic contexts.4 31 Consequently, future research endeavours should strive to encompass a more representative sample spanning multiple countries or regions to enhance the generalisability of findings.

The COVID-19 pandemic has presented numerous unprecedented challenges for health research, service provision and public health policies. While the impact of the pandemic and government responses on suicide-related consequences is not entirely new, pre-existing inequalities have the potential to deepen, making it even more challenging to address these issues, particularly in a relatively isolated city during this period. In Hong Kong, suicides increased during the pandemic period, especially among the elderly population and those with low SES. Some policy responses in Hong Kong have demonstrated their effectiveness in controlling the spread of COVID-19 and in mitigating some of the adverse effects of the pandemic, such as the alleviation of economic distress and the decline in SC. Nevertheless, it is crucial for governments and other agencies to remain attentive to mental well-being and target the determinants of poor mental health that have been exacerbated by the pandemic, especially among specific subgroups of the population during and after the pandemic. Since the infectious disease-induced epidemic is still happening worldwide, continued vigilance and close monitoring of the mental health of vulnerable populations remain a priority.