LAUSR

Length of stay (LOS) is one of the most common indicators of effective inpatient service; it is an objective measurement of the balance between hospital resource utilization and quality of care. Prolonged hospitalization may be counter-therapeutic due to the weakening of social support and threat to employment. Reducing LOS not only reduces hospital costs but also decreases the occurrence of complications in elderly patients, such as falls, infections, and delirium. This study reviewed administrative records from the database at Siriraj Hospital, a 2221-bed academic hospital in Bangkok, Thailand. Predictive variables included demographic data, and diagnosticand treatment-related variables. Diagnostic-related variables comprised principal diagnoses that were categorized into four groups (i.e., (i) neurocognitive disorder, (ii) psychotic disorder, (iii) mood disorder, and (iv) other disorders), underlying medical comorbidities, and psychiatric comorbidities. All were classified according to the ICD-10. Other variables included psychosocial/environmental problems diagnosed by axis IV of the DSM-IV-TR (presence/absence), duration of illness, prior psychiatric admission, readmission status, and the Global Assessment of Functioning (GAF) score. Treatment-related variables included the number of medications prescribed at admission; antipsychotic polypharmacy, which referred to using ≥2 antipsychotics concurrently; the number of electroconvulsive therapy (ECT) sessions; and unfavourable treatment results (i.e., medical complications). Inclusion criteria were patients aged 60 years or above who were admitted between January 2009 and December 2015. Those who were transferred to or from another ward or hospitalized for maintenance ECT were excluded. This study was approved by the Siriraj Institutional Review Board of the Faculty of Medicine Siriraj Hospital, Mahidol University. To determine factors associated with LOS, multiple linear regression (stepwise selection) was employed. Subgroup analyses were undertaken according to two principal diagnostic groups (neurocognitive disorder and mood disorder). A total of 254 inpatients (mean age 70.23 years, SD = 7.01) with a female predominance (66.5%) were included in the analysis. The median LOS was 16.5 days (range: 1–76). Of those, 37 patients (14.6%) received acute ECT treatment. Antipsychotic polypharmacy was reported in 35 patients (13.8%), and 130 patients (51.2%) had medical complications during hospitalization. Factors associated with LOS included medical complications, antipsychotic polypharmacy, number of ECT sessions, and baseline GAF score (F (3,249) = 40.76, P < 0.001, R = 0.39) (see Table 1). Although we found an insignificant result from the principal diagnosis, a lower baseline GAF score was significantly associated with longer LOS. This finding suggests the focus in effectiveness of care shifts from ‘what illness the patient has’ to ‘how functionally impaired the patient is’. Patients with severe impairment require higher-intensity care from inpatient unit staff compared to those with less impairment. Our results indicated a relatively lower rate of complications from ECT in comparison to pharmacotherapy (3.9% vs 20.7%, respectively), which was in line with previous findings. However, receiving ECT may lead to longer LOS due to the incidence of its adverse effects. Thus, the preand post-ECT evaluations and outcome assessments are beneficial approaches in preventing complications as well as determining the minimum effective number of ECT sessions. Patients hospitalized on a regimen of antipsychotic polypharmacy had longer LOS. A possible explanation for this result may be the increased use of antipsychotic polypharmacy to treat severe and difficult-to-treat patients, which ultimately leads to more side effects.