Determining the Length of Stay and Duration of Illness for Psychiatric Inpatients Using Multivariate Modelling

Abstract

Mental and behavioural disorders is a significant contributor of global burden of disease. As per WHO estimates, this burden is likely to increase by 15 percent by 2020, significantly impacting health and major social, human rights and economic consequences in all countries of the world. This paper provides a procedure for estimation of length of stay (LOS) in the hospital, total duration of illness (TDI) and recent duration of illness (RDI) using multivariate normal (MVN) and bivariate (BVN) normal distributions. To accomplish this, a retrospective data of 146 patients with complete record history, diagnosed with mental and behavioural disorders (as per APA’s DSM-V as well as the WHO’s ICD-10) is collected from the Department of Psychiatry, Lady Hardinge Medical College & Smt. S.K, Hospital, New Delhi, India for the calendar year 2013-2014. The estimated values of the above mentioned variables are found to be consistent with the observed values. Finally MVN distribution is applied to estimate the variables LOS and TDI for the patients for whom the information on these variables is not known. The model derived in this paper will facilitate the medical fraternity to not only guide the patients about their approximate length of stay in the hospital at the time of admission, but also assist them in development and management of appropriate interventions for patients.