Special Section on the GAF: Determinants of Indicated Versus Actual Level of Care in Psychiatric Emergency Services

OBJECTIVE: This study was undertaken to improve understanding of the admission decision process by distinguishing between the clinically indicated level of care and actual level-of-care decisions in emergency psychiatry. METHODS: Clinicians in emergency psychiatric services in Rotterdam, the Netherlands, prospectively rated 720 patients by using the Severity of Psychiatric Illness Scale and collected information on demographic, clinical, and contextual parameters. The clinically indicated level of care and actual level-of-care decisions were studied independently, by using multivariate logistic regression analyses. The decision-making process was divided into three consecutive steps: evaluation of clinically indicated inpatient or outpatient level of care (step 1), voluntary or involuntary admission (step 2), and actual admission of patients for whom voluntary admission was indicated (step 3). RESULTS: Each step was determined by separate factors. Specifically, clinically indicated admission (step 1) was associated with family or friends' desire for admission (odds ratio [OR]=3.7), previous admissions (OR=2.9), symptom severity (OR=2.7), and personality disorder (OR=.4). Involuntary admission (step 2) was associated with lack of motivation (OR=5.7), symptom severity (OR=3.7), time of referral (OR=3.5) and danger to self or others (OR=2.7). Actual voluntary admission (step 3) was associated mainly with bed availability (OR=8.7). The overall percentage of correctly predicted cases was 82 percent for all steps in the decision process. CONCLUSIONS: This study showed that each step in the admission decision process is determined by a unique set of variables and provided evidence that contextual factors influence decision making. Guidelines for voluntary admission and civil commitment need to be based on the results of studies that distinguish between the clinical needs of patients and contextual factors.