[Predictive factors of seclusion duration in patients hospitalized in psychiatry settings. A prospective multisite study in the DTRF Paris-Sud]

In psychiatric inpatient settings seclusion is a last resort to ensure the safety of the patient, other patients, and staff from disturbed behaviors. Despite its major interest for patients, seclusion could negatively impact treatment adherence and patient/staff relationships. Indeed, some secluded patients report a feeling of guilt during the measure and do not consider seclusion to be a healthcare intervention. To be more beneficial and to reduce the feeling by patients of being forced, seclusions should be as short and rare as possible. In other words, measures to reduce seclusion are available and have been clearly identified. Such measures could be applied, in the first instance, in patients with longer duration. In this way, the aim of this study was to investigate predictive factors of a seclusion of long duration.

Our study was based on the dataset of the EPIC study, an observational prospective French multicenter study of seclusion and restraint. The EPIC study occurred in seven French psychiatric hospitals in the southern region of Paris. Inclusions were realized for 73days and allowed a data collection of 302 seclusion measures. Of these measures 236 were effectively a seclusion in a standardized room. Because the median duration was 7days, we defined two groups of patients: duration<7days and duration ≥ 7 days. Our variable to be explicated was duration ≥ 7 days. Explicative variables available in EPIC study were age, sex, forced hospitalization, autoagressivity, heteroagressivity, use of sedative treatment (oral or intramuscular), history of seclusion and patient diagnoses. We used bivariate and multivariate analyses to explore the association between a seclusion duration ≥ 7 days and explicative variables. Diagnoses were classified as psychotic disorders, mood disorders and others diagnoses. To be included in multivariate logistic regressions, diagnoses were treated as dummy variables (mood disorder vs psychotic disorders; psychotic disorders vs others; mood disorders vs others). Statistical analyses were performed using SPSS software 20.0 and R 3.4.0.

Of the 236 measures of seclusion the mean age was 38.2 (±12.8), 196 (83%) patients were forcibly hospitalized prior to their seclusion, 147 (62%) had a diagnosis of psychotic disorder, 43 (18%) a diagnosis of mood disorder and 33 (14%) an "other diagnosis". Mean duration was 10.2 (1.5) days and median was 7.1 days. One hundred and thirty-five (47%) patients were in the group of duration ≥ 7 days. In bivariate analyses, variables associated with a duration ≥ 7 days were: being in forced hospitalization prior to the seclusion (P=0.04), administration of a sedative treatment (P=0.01) and against the group of others diagnoses the diagnosis of mood disorders (P<0.0005) and psychotic disorders (P=0.001). Multivariate analyses showed that, against the group of other diagnoses, the group of psychotic disorders [OR=3.3, CI 95% (1.3-8.4), P=0.01], the group of mood disorder [OR=2.7, CI 95% (1.4-4.9), P=0.002] and administration of sedative treatment [OR=8.1, CI 95% (2.0-32.5), P=0.003] were significantly associated with a duration ≥ 7 days. These results were independent from other confusion variables. Considering the hospitalization status, psychotic disorders was the only diagnosis which showed an association between duration ≥ 7 days and forced hospitalization [OR=2.9 CI 95% (1.1-7.8), P=0.03].

Our study highlighted two profiles of higher risk to remain ≥ 7days in seclusion. The first one is patients with a diagnosis of mood disorder who needed sedative treatment. The second one is patients with a diagnosis of psychotic disorder who needed sedative treatment and forced hospitalized before seclusion. Thus, these two profiles could be a good target to practice, in the first instance, measures to reduce seclusion duration in psychiatry settings.