Method Of Measuring The Complexity Of The Disease And Its Application In Predicting Hospitalization Expenses

With the aging of the population and the rapid development of social economy,the shortage of medical and health resources is also increasing.The Diagnosis Related Groups(DRGs)led by the National Medical Security Center and piloted since 2017 are considered to promote hospitals.It is an important measure to carry out cost control,optimize hospital management,curb long-term hospitalization,excessive medical treatment,and reduce hospitalization costs.The reform of the payment system centered on DRGs is no longer based on the current billing method used by most hospitals in my country according to the service items received by patients,but according to the personalized characteristics of the patient’s medical record homepage,patients are classified into the corresponding diagnosis-related groups,medical insurance institutions The funds are allocated according to the group of patients.Therefore,quantifying the complexity of the patient’s condition and predicting the patient’s DRGs hospitalization expenses based on the complexity of the condition are the key to the successful implementation of the DRGs payment system reform.There are currently three types of methods for measuring the complexity of the disease: the United States has built a complication information database for quantification,which relies excessively on expert knowledge and has problems with accuracy;and the patient clinical complexity level(PCCL)The quantification complexity of the method is not highly correlated with the cost;in the quantification of the diagnostic complexity level of the case clinical complexity model(Episode Clinical Complexity,ECC),the impact of the number of diagnoses on the hospitalization cost is mainly considered,and the number of identical diagnoses is ignored But the cost difference between patients with different diagnosis types.Therefore,the first problem to be solved in this article is to improve the disease complexity measurement method and establish a disease complexity measurement method that considers the number of diagnoses and the type of diagnosis.In terms of cost prediction for DRGs,the existing DRGs grouper mainly uses statistical analysis and combines the complexity of the patient’s condition to predict the patient’s hospitalization expenses.The inaccurate quantification of the complexity of the patient’s condition and the insufficient utilization of other features restrict DRGs Accuracy in the field of cost forecasting.Therefore,the second problem to be solved in this paper is to realize the improvement of the DRGs cost prediction model in combination with the quantified complexity of the disease.Aiming at the difficulty in quantifying the complexity of patients’ condition in clinical hospitalization data and the design of DRGs grouping device,this paper aims to improve the accuracy of quantifying the complexity of patients’ condition and to build a more reasonable DRGs grouping tool.The contribution of this article has the following two aspects:(1)The study proposed an improved ECC model(Iterative Correction-based ECC,IC-ECC)through iterative correction of errors,using high-dimensional sparse nominal complication diagnosis data to achieve a more accurate diagnosis of the complexity of the disease Quantify.(2)Study the introduction of the integrated learning model SAMME algorithm,design the cost prediction model,and improve the effect of cost prediction.The research is based on the data set of the respiratory system and circulatory system of the First Affiliated Hospital of Chongqing Medical University and the data set of the medical records of some third-class hospitals in Meishan City.The simulation cost of the case based on the IC-ECC model and the original ECC model is compared with the actual The cost correlation is used to evaluate the quality of the new feature of diagnostic complexity level quantified by the IC-ECC model.The results show that the complexity level of cases obtained by the IC-ECC method can better reflect the complexity of diagnosis and treatment of cases from the perspective of resource consumption.By comparing the IC-ECC-SAMME forecasting model and the ECC-SAMME forecasting model on the accuracy,macro P,macro R,and macro-F1 metrics of cost forecasting,the results show that the IC-ECC-SAMME forecasting model has more advantages in cost forecasting.The high performance further illustrates that the quantification of the complexity of the disease by the IC-ECC model is more accurate.The IC-ECC-SAMME prediction model is compared with the CHS-DRGs grouping results of the coefficient of variation of the hospitalization expenses of the patients in the group.The results show that the hospitalization expenses of the patients within the group predicted by the IC-ECC-SAMME have lower intra-group differences,indicating this This cost prediction model can improve the effect of DRGs cost prediction.