A Study Of Scheduling Policies In CT Department With Simulation

CT examination is one of the bottlenecks in the hospital, patients’appointment times c ould be reasonably determined by a well-designed scheduling policy so that patients’waitin g times and throughput times in the hospital could be reduced, the hospital could provide m edical service to more patients. There has been many studies on the "single-stage, single-serv er" outpatient scheduling while little studies on the "multi-stage, multi-service" medical exami nation scheduling. Therefore, based on the business processes and historical data of the CT department of a3A-grade domestic large general hospital, a simulation model was utilized to analyze the effects of different scheduling policies on both hospitals and patients in a "mul ti-stage, multi-server" queuing system so that a reference for designing a scheduling policy f or the department and other similar institutions is provided.Firstly, based on the business process and historical data from April2011to March2012of the CT department of a domestic3A-grade hospital, a "two-stage, two-queue, two-serv er" queuing model was built, using the queuing theory and the mathematic statistics theory, data was processed by Excel2007and SPSS19.0, the probability distributions of patient inter-arrival times and service times were fitted by Stat::Fit2.0. It was concluded as below:(1) For the hospital management, the extent of information sharing among all departmen ts in the hospital should be improved and doctors should be reminded of not providing a se cond application forms so that repeating information input is avoided and time of both patie nts and hospitals are saved; encourage the staff to delete invalid records or improve the wor k environment to reduce mistakes, data analysis would be much easier with these improvem ents.(2) For the probability distributions, the patient inter-arrival times obey a non-stationary e xponential distribution whose parameter changes with time, and the service times of examina tion rooms obey the Weibull distribution whose parameter varies from different types of pati ents. The result differs from assumption of the former studies that the probability distribution s of patient inter-arrival times and service times of all patients are identical, but it is close t o the reality.Secondly, using the simulation theory, a discrete event simulation model was built by E xtendSim8and results outputted from the steady-state simulation. The average waiting time o f enhanced scan patients and non-enhanced scan patients, the average overtime and idle time of examination rooms, total cost which is the linear weighted sum of the above four value s were used as the perfonnance measures;22scheduling policies (grouped by three factors named appointment rule, length of slots and time length enhanced scan patients arrive ahead) were evaluated by ANOVA and efficient frontier charts; the sensitivity analysis on the we ighted values of total cost, the appointment rules and the number of queues was carried out. It was concluded as below:(1) For the appointment rule, patients with shorter average service time should be sched uled before those with longer average service time within a session, the length of slots of a ny type of opatients is equal to their average service time, total number of scheduled patient s is equal to the result of total time of a session divided by the overall average service tim e, the number of both types of patients are calculated through relations between their averag e service times, total number of scheduled patients and total time of a session.(2) For the length of slots, that of patients with shorter and longer average service time is determined by the appointment rule and performance measure adopted; for time length th e enhanced scan patients arrive ahead, the time length enhanced scan patients should arrive ahead to reach the CT examination punctually is also determined by he appointment rule an d performance measure adopted.