Snapshot of a social movement: Mental health and protest classes in Hong Kong

•High rates of severe depression and anxiety symptoms in large samples of pro-democracy supporters.•Rates of severe symptoms worsened over two months of continued civil unrest and crackdown.•Latent class analysis suggested more severe mental health risk for youth and frontline protesters.•Over half of frontline protesters reported severe traumatic stress symptoms. Popular protests have broken out worldwide, particularly in the last few years. In 2019, numerous demonstrations against an extradition bill occurred in Hong Kong until pandemic restrictions were imposed. The policing response relied heavily on methods such as batons, tear gas and rubber bullets. Given the relevance for other geographical contexts, the current study investigated the mental health impacts on protest participants and spillover to community members. Surveys were disseminated on social media in August and October 2019 to collect demographics, political views, protest participation, exposure to (protest-related) potentially traumatic events (PTEs) and mental health symptoms. A latent class analysis (LCA) was conducted using demographic data and inter-class differences in PTEs and mental health symptoms were examined. There were 37,541 (59.8% female) and 40,703 (50.0% female) responses in August and October. Respondents, even those with low participation, reported significant levels of depression, anxiety, and symptoms of traumatic stress (STS). The LCA suggested a 5-class solution (youth, allies, supporters, sympathizers, and frontliners). Mental health symptoms and PTEs varied with class membership, with 50.8% of frontliners reporting severe STS. The non-random sampling and self-reported measures may over-estimate the prevalence of mental distress in the wider population. Large numbers of pro-democracy supporters in Hong Kong reported high rates of depression, anxiety and STS during mass protests. Younger and more heavily involved respondents faced the highest mental health risks, however elevated rates were also observed for respondents with low participation.