Background Coercive measures such as involuntary psychiatric admission are considered a last resort in the treatment of people with psychiatric disorders. So far, numerous factors have been identified that influence their use. However, the link between a pandemic – in particular, restrictions such as lockdowns – and the use of involuntary psychiatric admission is unclear. Aim To examine the association between COVID-19 lockdowns and involuntary psychiatric admissions in Austria. Method This retrospective exploratory study assessed all involuntary psychiatric admissions and use of mechanical restraint in Austria, except for the federal state of Vorarlberg, between 1 January 2018 and 31 December 2020. Descriptive statistics and regression models were used. Results During the 3-year study period, 40 012 individuals (45.9% females, mean age 51.3 years) had 66 124 involuntary psychiatric admissions for an average of 10.9 days. Mechanical restraint was used during 33.9% of these admissions. In weeks of nationwide COVID-19 lockdowns (2020 v. 2018/2019), involuntary psychiatric admissions were significantly fewer (odds ratio = 0.93, P = 0.0001) but longer (11.6 (s.d.: 16) v. 10.9 (s.d.: 15.8) days). The likelihood of involuntary admission during lockdowns was associated with year (2020 v. 2018–2019; adjusted odds ratio = 0.92; P = 0.0002) but not with sex (P = 0.814), age (P = 0.310), use of mechanical restraint (P = 0.653) or type of ward (P = 0.843). Conclusions Restrictions such as lockdowns affect coercive measures and resulted in fewer but longer involuntary psychiatric admissions during weeks of lockdown in Austria. These results strengthen previous findings that showed the dependence of coercive measures on external factors, highlighting the need to further clarify causality and desired prevention effects when using coercive measures.