The use of coercive practices, i.e., interventions against a person’s will, is controversial. Recent observational studies highlighted their potential detrimental effects on patients’ mental health, but this topic remains understudied. This study investigated the effect of a common coercive practice, seclusion (i.e., being locked in a closed room), on mental health using a trial emulation of observational data to allow causal inference. We used data from 1200 psychiatric inpatients, classified as being either secluded or non-secluded during their hospital stay. Inverse probability of treatment weighting was used to emulate the random assignment to the intervention. The primary outcome was the Health of the Nations Outcome Scales (HoNOS). The secondary outcome was the first item of the HoNOS, which focuses on overactive, aggressive, disruptive, or agitated behavior. Both outcomes were assessed at hospital discharge. There was a significant effect of seclusion with increases in both total HoNOS score ( p = .002) and item 1 of the HoNOS ( p = .01). Seclusion may have a negative causal effect of patients’ mental health status and should therefore be avoided in mental health care settings. Training efforts should raise the awareness of the medical staff about potential adverse effects instead of therapeutic benefits.