| Objective: To learn about the status quo of insured osteoporosis hospitaliz ed patients in Wuhan, analyze the composition of the costs, and figure out the relationship between individual costs and hospitalization costs. To analyze the hospital costs and actual compensation rate of factors by univariate analysis. T o explore the influencing factors of hospitalization costs of patients with osteop orosis. To explore the adaptation of K-means clustering algorithm and Support Vector Machine in analyzing the factors that impact on hospitalization costs, so as to illustrate its future application in medical fields. To provide advice on c ontrolling medical costs for related departments according to the findings in thi s paper.Methods: Datasets was from medical insurance system in Wuhan, and it was pre-treated via R language, then a total of 1069 cases of patients were sel ected. Descriptive analysis was made to the conditions and cost structures of i npatients with osteoporosis, univariate analysis was employed to explore of th e factors influencing hospitalization costs and actual compensation ratio such as gender, hospital grades and the length of stay in hospital inpatients; using the R language and K-means clustering to discretize hospitalized costs, then we performed support vector machine fitting using SPSS modeler14.1, and exploredfactors that influence the cost of hospitalization patients with osteoporosis.Results: The analysis of status quo of osteoporosis patients showed that o steoporosis occurs mainly to the females and the elders; the cost structure anal ysis showed that drug costs accounted for 39.74% of all hospital costs which i s the highest, followed by cost of tests(28%) and treatment expenses(12%), t he sum ofwhich accounted for more than 80% of the cost. Univariate analysis showed that the category of employees, hospital grade, surgical dummy, compli cations, the length of stay and the way of settling cost are statistically signific ant(P <0.05) different in the cost of inpatient test; K-means cluster analysis s uggested that hospitalization costs be divided into three categories: the first cat egory was 736.02-22401.08 yuan, and the cluster center is 6807.62 yuan, with 929 cases; the second category is 23034.47-57042.72 yuan, and the cluster cent er is 38,719.18 yuan, with 110 cases, the third category is 60083.40-110672.56 yuan, and the cluster center is 7056.99 yuan, with 30 cases. Then we chose discretized cost as the target variable, and apply support vector machine to ana lyze factors that influence the hospitalization expenses of patients with osteopor osis, and it showed that surgery dummy, length of stay, transfusion, hospital gr ade and cost settlement patterns are major factors influencing hospitalization co sts in patients with osteoporosis.Conclusion: By analyzing the cost of hospitalization in 1069 patients with osteoporosis in Wuhan,this study verified the elderly and women are high r isk victims of osteoporosis, and found that the constitutes of hospital costs of osteoporosis inpatients are unreasonable, with a high proportion of drug use, an d the analysis of the influence factors found that surgery dummy, length of st ay, transfusion therapy, hospital grade and cost settlement patterns are major fa ctors for osteoporosis patients in hospital costs, and proposed targeted strategiesto control hospital costs. This study used a combination of K-means clusterin g and support vector machines to analyze osteoporosis hospitalization costs, wh ich is a good model for evaluation and explanation, and worthy of promotion i n research and analysis. |
PDF Full Text Request