Data analytic and predictive model for analysis of causes for a loss from diagnosis related groups (DRGs) payment of inpatient at Golden Jubilee Medical Center, Faculty of Medicine Siriraj Hospital, Mahidol University

Abstract

Golden jubilee Medical Center, Faculty of Medicine Siriraj Hospital, Mahidol university is a secondary care level general hospital. Hospital have in-patient service since 2008, most of the patients was reimbursed from the Comptroller General’s Department. Hospital used Diagnostic Related Group Method (DRGs), (DRG Version 6.2.1) to calculate for a reimbursement from Comptroller General’s Department. We observed a loss from DRGs reimbrusememt in some patients (cost from treatment more than money from reimbursement)

From hospital in-patient database, we analyse data since October 2016 to April 2019. Hospital have 6,871 admission number, we observed a loss from DRGs reimbursement in 1,441 admission number or 20.95%, and cost 13,489,521.75 bath.

We analyzed data by ranking a diseases by amount of loss, number of cases and sum of length of stay. We selected 3 diseases within top-10 ranking in each category, we selected 1) Urinary tract infection, site not specific
2) Lobar pneumonia, unspecified) and 3) Cerebral infarction, unspecified to run a predictive model (Decision tree) by program R.

After we run a predictive model, we found; Immobility condition, Length of stay, GI hemorrhage complication, E.coli organism, Age range and Diabetes in co-morbid is a predictive factors for a loss in interested diseases and hospital can bring  this data to improve hospiatal process and utilize medical treatment to reduce a loss in the future.