Fitting aggregated phase-type distributions to the length-of-stay in intra-hospital patient transfers

The patient transfer, as a common seen and necessary healthcare procedure, plays an important role in maintaining efficient treatment and improving the quality of healthcare. Among various factors impacting and indicating the safety and efficiency of the patient transfer, the length-of-stay (LOS), which is not often studied in this field, is worth investigating. Phase-type (PH) distributions, as one of the popular methods of modeling LOS, will be integrated in an aggregated Markov chain to construct a model to describe the sequences of LOS in hospital units. In this paper, we model the intra-hospital transfer flow routes by fitting aggregated PH distribution and using Maximum Likelihood Estimation to estimate the parameters. Following the results of distribution fitting, the patients can be divided into different groups according to their LOS in the same unit. By analyzing each group to find out its common characteristics, intra-hospital transfer routes, admission and discharge situations, the associations among significant factors, the LOS and the treatment efficiency are evaluated.