IntroductionIn France, the use of seclusion in psychiatric hospitals is regulated by the Act of January 26th 2016 which enforces a reduction of the use of coercive measures and limiting their duration. Criteria that are unrelated to the patient's symptoms might affect this decision and extend the duration of seclusion. The goal of the current observational study is, firstly, to determine which factors - unrelated to the patient's symptoms - influence the length of stay in seclusion. Secondly, it is to compare the composition of the medical and nursing teams at two times, the beginning and the end of the seclusion time period. MethodWe conducted this study in a La Rochelle regional hospital from October 2017 to July 2018. There were one hundred and twenty-four episodes of seclusion that occurred in the different psychiatric inpatient wards during this time. The episodes were divided into two groups: short-term and long-term seclusion, defined by the median duration of seclusion. Data were collected using a survey, completed by the nursing teams, based on the recommendations of good clinical practice published in 1998 and revised in 2017 by the French health authority. For each episode of seclusion, we collected the following data: socio-demographic information, history of psychiatric care of the patient and of violent acting-out, admission status, medical indication for seclusion, day of the week and time at the beginning and the end of seclusion, prior administration of a sedative before seclusion, exceptional events that might justify the end of seclusion, transfer to a protected room after seclusion, as well as the composition of the medical and nursing team on duty at the beginning and the end of the seclusion period. We compared the anamnesis between the short-term and the long-term seclusion groups, and we analyzed the composition of the medical and nursing teams at the two time points of seclusion. Statistical analyses were performed using R software (v. 3.5.1). ResultsThe mean duration of seclusion was 4.7 days and the median was 1.9 days. The average age was 37-years-old, with a ratio of 0.6 for females to males. Variables associated with a long-duration of seclusion were: the male gender (P=0.005), Compulsory Admission at the Request of a State Representative (P=0.008), a prevention measure of any hetero-aggressive action (P=0.007), the lack of psychiatric care (P=0.004), previous medico-legal issues (P=0.006), violent behavior during a previous hospitalization (P=0.022) and the use of seclusion on the weekend (P=0.01). The composition of the medical and nursing teams related to the period of the end of seclusion were: the presence of the physician in charge of the patient (P<0.001), a male caregiver in the team (P<0.001), a specialized psychiatric nurse (P<0.001) and the training of caregivers in the management of violence (P<0.001). The presence of nurses who do not usually work in the psychiatric ward was associated with the period of the beginning of seclusion (P<0.001). ConclusionOur findings showed a strong relationship between several anamnestic factors and the duration of seclusion. Caregivers lacking information about patients, potential violent behaviors and the beginnings of seclusion on weekends are associated with a long-duration of seclusion. Our study also highlights the roles played by the caregivers according to their composition and level of training to determine the ending of patient's seclusion.