Introduction: Trust in government to take care of its citizens may affect mental health outcomes such as anxiety and depression, particularly as measures become more stringent. The study aimed to investigate the associations between stringency of COVID-19 social distancing policies and mental health outcomes, and the moderating effects of trust in government and gender. Methods: The study consisted of secondary analysis of publicly available cross sectional data from a global online survey COVID-19 survey conducted between 20 March and 7 April 2020. There were 106,497 adult participants (18 years of age and over) from 58 countries. The main outcome measures were indices for depression and worries. The exposure measure was the stringency index. The effect modifier measures were gender and trust in government. Multivariable regression was conducted to determine the three-way interaction between the exposure, modifier and outcome measures, adjusting for age, income and education. Results: The median age of participants (56.4% women) was 37 years. Women had higher worries and depression than men. The proportion of people trusting (44%) and distrusting (45%) the government was almost the same. Among those who strongly trusted the government, an increase in policy stringency was associated with an increase in worries. Among men who distrusted the government, an increase in policy stringency was associated with an increase in depression, but in distrusting women there was an inversed Ushaped association between policy stringency and both worries and depression. Once policies exceeded the 50-point mark on the stringency index, women benefited from the most stringent policies, yet men did not, particularly men who strongly trust or distrust the government. Conclusion: As the stringency of public health measures increases, so too do depression and worries. For safe and effective public health measures, governments should develop strategies to increase trust in their actions.