| Department of Health Statistics, the Fourth Military Medical University, Xi'an, Shaanzi, 710032, ChinaWith the changes of disease spectrum and mortal spectrum, the health cost will be certain to be continuing rise at a surprising speed ,due to the fact that people pay more attention to treatment than prevention and the hospital services is not efficient sufficiently. Computing with the invariance price in 1994, the health cost is estimated to be as high as 1081.6 billions RMB in the year of 2010, which is 10% of GDP. The whole country will be suffered from such a heavy burden if this tendency could not be controlled by means of taking some stratygies into actions.Case mix, which is considered as one of the methods of restraining and controlling the increasing of health cost, has been focused and studied widely in the world. So far, it has been applied in many countries and been generally acknowledged as the most effective method in controlling healthcare cost. Case mix has applying value in health policy making, health administration and health economic research. Case mix was began to studied in 1990s in our country, but limited in studies on the feasibility of the case mix to be applied in China. No research has been found in studying the statistical classification of the case mix in China so far.The data which was used in this paper is extracted from 140 millions abstracts of the inpatient medical records in 1998 at 246 army hospitals distributed all over the country. The database was provided by the Health Statistical Information Center of the PLA General Logistic Department.This study was conducted as follows;1 .Preparing data, omitting the cases with some items absent or cases which was verified not to be logic from the raw database. In addition, some items were added to the database;2.Dividing the raw database into two main groups called surgical and medical database according to weather a patient was experienced operation or not;3.Describing the cases respectively, taking patients' social characteristics,the situation of diseases and other connecting factors into account;4.Quantifying the variables, giving co-morbidity, the degree of care, rescuing, infection inter-hospital, co-morbidity during in-hospital and subordinate operation a number with "1 "indicating "no" and "2" indicating "yes". Co-morbidity means that the patient has subsidiary diagnosis. A patient with the care degree "2"means the days that first intensive care was given was not absent. A patient with subordinate operation means the patient had a second diagnosis code;5.Choosing 100 kinds of diseases from the two database respectively, making up two frequent diseases databases;6.Establishing grouping database manually with the remaining data in the two main databases. The medical database was divided into 18 groups and 111 subgroups, while the surgical database was classified as 16 groups and 86 subgroups;7.Classifying the raw database with the manually grouped database as described above;8.Determining the classifying nodes;9.According to the principle of case mix, took the inpatient cost as the axis and other 11 variables carrying the information about the patients' characteristics as grouping nodes. The automatically interactive detectionæ¡ID method was used in the procedure of classification. The procedure was not stopped until the grouping principle was satisfied;10.Computing the mean and the median of the cost and longs of staying in hospitals;11 .Assessing the result of the grouping;12.Analyzing the re-computing data and calculating the standard cost of the re-computing hospitals;13.Calculating the case mix index of each hospital;H.Describing and analyzing the patients whose costs were much high than that of ordinary patients.All the patients involved were divided into 1208 groups, hi the grouping sample of internal disease database, the frequent diseases were classified into 313 groups by one grouping node, 216 groups by two grouping nodes and 22 groups by three grouping node...
|
PDF Full Text Request