| Objective: To study general demographic characteristics and internal composition of insured acute myocardial infarction patients in Tianjin from 2008 to 2011. Making sure of the structure and distribution of hospitalization expenses patients to provides the basis in hospital cost trend analysis for patients with acute myocardial infarction. Moreover, we tried to find out the external factors influencing the hospitalization costs through the analysis of the data who with acute myocardial infarction, and Provide a reference for reasonable cost possible control. At last, Using data mining technology, to join in combination of case study in patients with acute myocardial infarction. We Calculated over-quota patients cases and surcharge under every possible combination, and tried to explore single diseases Diagnosis Related Group(Diagnosis Related Group System).Methods: Data rooted in Tianjin urban employee basic medical insurance database system from January 1, 2008 to December 31, 2011. 6467 qualified acute myocardial infarction patients were collected from 50% cases. Adopt new grey correlation and structural changes of analysis to quantify the hospitalization cost, and describe the hospitalization expense form, define the general changes in four years. Using artificial neural network model selection on the hospitalization expenses of patients with external influencing factors based on the single factor analysis results. Finally, using the method of decision tree in data mining, fitting regression tree model and build case mix model of single diseases based on the database system.Results: Male patients with acute myocardial infarction in Tianjin from 2008-2011 was 4716 people, average age was(65.08±11.78); Female patients with acute myocardial infarction was 1751 people, average age was(70.98±9.34), and focused on the elderly. Patients with acute myocardial infarction occurring Frequently in the autumn and winter seasons and few complications after onset. Furthermore, inpatients with acute myocardial infarction concentrate in the high-level medical institutions, the majority of patients in hospital days for 8~14 days. For four years, the total cost of insured patients hospitalized with the upward trend, the results of both methods have shown that the cost of medical materials is the main factor affectinghospitalization cost, accounted for 50%. Structural change analysis results show that the cumulative contribution is 82.44%. The neural network model and decision tree model indicate that major influential factors were surgery or not, hospital’s level and length of stay, there are 8 case mix and cost standard from decision tree. A total of 151 cases over each combination control standards as a suspected cases, total excess cost was 10579418.33 yuan, accounting for 6.87% of the total expenses occurred.Conclusion: Patients with acute myocardial infarction are concentrated in the high level medical institutions, and the composition of hospitalization medical expenses is not reasonable. The key for policymakers is taking effective measures to control medical cost of materials for AMI patients in hospitalization cost. The new grey correlation analysis and structure analysis can reflect overall situation and characteristics of AMI patients in hospital medical expenses in the form of numeric, to provide health management advice and decision making. In addition, data mining is a process of extracting effective information from huge amounts of data, applied in DRGs research has its own advantages. Two methods confirm that the influence factors and significance of AMI inpatient’s expense are the same, the uppermost splitting node is surgery or not in case mix model. The hospitalization expense is more apparently distinguished according to differences between surgical and non-surgical patients when carry out classification related disease diagnosis. The upper limit of medical charge can assist medical institutions in finding suspicious cases, early warning, and perfecting the social medical insurance system. |
PDF Full Text Request