Predictors of readmission to a psychiatry inpatient unit

To determine predictors of time to readmission to a general psychiatry inpatient unit.

Data from the Minimum Data Set-Mental Health (MDS-MH), a standardized assessment used to collect demographic and clinical information, were retrospectively reviewed from April 2006 through October 2008. A total of 758 patients were eligible for the study. A set of clinically relevant predictors was generated based on a literature review. A Cox regression model was applied to determine which variables were most predictive of shorter time to readmission, and their respective hazard ratios (HR).

Covariates that were significantly associated with readmission (HR [95% CI]) included receiving a pass (3.48 [2.33, 5.17], p ≤ 0.0005), 1-2 psychiatric admissions in the past two years (15.63 [7.50, 32.55], p ≤ 0.0005), and more than 3 psychiatric admissions in the past two years (24.15 [11.58, 50.36], p ≤ 0.0005). Post hoc analysis indicated that those issued passes were more commonly male (57.1% vs. 43.9%, p=0.03), with a longer length of stay (25.4 ± 21.2 days vs. 18.7 ± 21.1 days, p=0.008), and higher GAF score (62.8 ± 11.1 vs. 57.8 ± 13.9, p=0.003), but were otherwise similar.

The factors that were associated with reduced time to readmission were a history of previous admissions and receipt of a pass prior to discharge. These results suggest that while physicians may be able to identify patients at high risk of early readmission, issuing a pass may not fully mitigate this risk. There is a need for critical research evaluating the potential benefits of passes.