[Crisis unit at the general hospital: Determinants of further hospitalization]

The availability of short-stay beds for brief admission (less than 72hours) of crisis patients presenting to the emergency room is a model that has gained a growing interest because it allows time for developing alternatives to psychiatric hospitalization and favors a maintained functioning in the community. Still, the determinants influencing the disposition decision at discharge after crisis intervention remain largely unexplored.

The primary objective of this study was to determine the factors predicting aftercare dispositions at crisis unit discharge: transfer for further hospitalization or return to the community. Secondary objectives included the description of clinical and socio-demographic characteristics of patients admitted to the crisis unit upon presentation to the emergency room.

All patients (n=255) admitted to the short-stay unit of the emergency department of Rambouillet General Hospital during a one-year period were included in the study. Patient characteristics were collected in a retrospective manner from medical records: patterns of referral, acute stressors, presenting symptoms, initial patient demand, Diagnostic and Statistical Manual, 5th edition (DSM-5) disorders, psychiatric history, and socio-demographic characteristics were inferred. Logistic regression analysis was used to determine the factors associated with hospitalization decision upon crisis intervention at discharge.

Following crisis intervention at the short-stay unit, 100 patients (39.2%) required further hospitalization and were transferred. Statistically significant factors associated with a higher probability of hospitalization (P<0.05) included the patient's initial wish to be hospitalized (OR=4.28), the presence of a comorbid disorder (OR=3.43), a referral by family or friends (OR=2.89), a history of psychiatric hospitalization (OR=2.71) and suicidal ideation on arrival in the emergency room (OR=2.26). Conversely, significant factors associated with a lower probability of hospitalization were the presence of a personality disorder (OR=0.31), a precipitating conflict situation (OR=0.41), age between 20 and 39 years (OR=0.42), being employed (OR=0.49).

Our study confirms that clinical factors such as the presence of a personality disorder or the context of a precipitating conflict situation are predictive of a community return. Interestingly, it points out the importance of the patient's initial wish in the hospitalization decision.