Modeling patient arrivals in a radiology department

We develop improved statistical methods for modeling the arrival process of patients in radiology department of West China Hospital of Sichuan University. Two-year data from 2013/1/1 to 2014/12/31 are used in this study. Based on the massive historical records in the database, data are screened, revised, analyzed and fitted by SPSS and MATLAB. A comprehensive analysis of patient arrival times shows that patient sources and time factors (day-of-week and time-of-day) should be taken into consideration. Also patients' unpunctuality show different pattern among patient sources. We propose a non-stationary Poisson process to model emergency patient arrivals, and a normal distribution for patients' unpunctuality. Results are of great value to optimize the capacity allocation and appointment scheduling practices.