| This study investigated social and adaptive functioning as a predictor of inpatient psychiatric hospital length of stay (LOS). LOS is an important index of the utilization of inpatient mental health treatment resources. The accurate prediction of LOS has become an important aspect of managing resources allocated to the mental health care system. Social and adaptive functioning was hypothesized to be a unique predictor of LOS, after controlling for variables such as diagnosis, past inpatient treatment history, and medical comorbidity. The current study used a naturalistic archival design with a large sample of individuals (n = 1003) admitted into an inpatient psychiatric setting. The Scale of Functioning (SOF), a multidimensional measure of social and adaptive functioning, was assessed at the time of admission and discharge. Additionally, demographics (age, gender, education), employment status, clinical information (treatment history, psychiatric and medical diagnoses), legal status, and insurance eligibility status were collected through interviews, and reviews of clinical records.; Hierarchical multiple regression analyses were used to simultaneously assess the contributions of the SOF and other specified variables as predictors of LOS. As hypothesized, the SOF total score significantly predicted LOS after controlling for demographic variables (DeltaR2 = 4.0%, p < .001). The SOF total score also demonstrated a unique incremental contribution to the prediction of LOS after controlling for all other predictors (DeltaR2 = 1.2%, p = .004). This complete model significantly accounted for approximately 19% of LOS variance, corresponding to a medium magnitude of effect, while the SOF total score demonstrated the largest correlation with LOS ( r = -.23, p < .001) when compared to all other specified predictors. DSM-IV Axis I diagnoses of schizophrenia and psychosis NOS were the strongest unique predictors of LOS and SOF intake total scores accounted for the next largest amount of unique variance in LOS. Principal components analysis (PCA) of the individual SOF items was used to extract two orthogonal SOF Intake Factor scores. As hypothesized, these factors were significant predictors of LOS while controlling for demographic predictors in a hierarchical MRA (DeltaR2 = 4.0%, p < .001). Individual admission SOF item scores, entered as a block in a hierarchical MRA following demographic variables, significantly predicted LOS, jointly accounting for 8.4% of the observed variance in LOS (p < .001). A model which specified the full set of predictors, including SOF item scores, significantly accounted for 21% of the variance observed in LOS (p < .001), improving upon the results of the majority of prior studies which investigated predictors available at the time of admission.; Overall, social and adaptive functioning significantly contributed to the prediction of inpatient psychiatric LOS. These findings are interpreted as further evidence for the validity of social and adaptive functioning in the management of clinical services in inpatient psychiatric settings. |
