Predicting the patients’ proportions classified by number of groups of diseases using the Markov chain

Abstract

The patients’ profile classified by the number of diseases per patient can be reflected the services level of a hospital, if the proportions by patients’ with higher number of diseases are reduced from the previous year.  Each patient can changed the number of diseases from one number to the other number in the consecutive years. This study uses the number of groups of diseases (NGD) instead of number of diseases to present appropriate size of statistical tables.  The patients’ record from the hospital data base can be processed NGD and the proportion by NGD and the transition from one NGD to another NGD.  Applying the Markov chain to the district hospital of about 3,900 patients per annum, if their operations are the same as in 2016 and 2017 (reflected by the transition from 2016 to 2017), then for the next 5 years (2018-2022); the group of less than two diseases will be increased from 0.714 to 0.764, 0.783, 0.792, 0.795, and 0.796, respectively, or year on year percentage increase at 7.0, 2.5, 1.1, 0.4, and 0.1, respectively.  In the next 6^th  year (2023), the proportion by number of groups of diseases will go to steady state,  the group of less than two diseases will be at 0.798 while the group of more than one disease will be at 0.202.  If the hospital aims to reduce the proportion of the group of more than one disease, the detail study is needed to implement the suitable service plan.