Font Size: a A A

Study Of Severity Scale Of Inpatients With Various Diseases

Posted on:2004-03-07Degree:DoctorType:Dissertation
Country:ChinaCandidate:D H LiuFull Text:PDF
GTID:1104360092491741Subject:Epidemiology and Health Statistics
Abstract/Summary:PDF Full Text Request
It is of great practical significance to describe how severe of a patient is, which can help evaluating the efficacy and efficiency of medical performances. In the process of medical intervene, such factors differentiating patients from each other, which are beyond of doctors' control, as the severity degree of a patient also have certain influences to the outcome of the patient together with therapies he or she received. In order to make an "Apples to Apples" comparison of health outcome among different health providers, we should take the severity of the patients being treated into account. For the same reason, severity should also be considered when the efficiency of health providers is analyzed and consequently how much of each hospital should be paid for certain kind of patients is determined. Only by this kind of solutions can health resources be used reasonably and achieve better health result.Although severity is hard to describe and assess, to classify and scale the relative severity of patients' on certain background and for certain purpose is possible. Clinically, many diseases were staged or categorized on the basis of their severity such as stages of malignant tumor and high blood pressure. This kind of severity description were developed by summarizing clinical investigation data in more detail and has been used in estimating patients' prognosis and making clinic decisions. Because of its disease-specific characteristics, this kind of severity description cannot be shared across various kinds of diseases and cannot be used as a common language when evaluating the severity of patients whose diseases are completely different. In order to produce such a common language, many patient classification systems according to severity have been developed science 1980' such as Comprehensive Severity Index (CSI) and Disease Staging. The most famous casemix model, Diagnosis Related Groups (DRGs) is also a kind of patient classification system based on resources consumption originally and is used as the foundation of Prospect Payment System (PPS) in the U.S. For the purpose of making hospitalreimbursement meet the patients' actual needs, DRGs was modified into several versions and each of them can express patients' severity to certain degrees. All these systems mentioned above have been played important roles in improving healthcare performance and curbing health cost rising. Although there were similar studies in China, we have not got an objective severity classification system, which is developed with large discharge data and statistical techniques. With the increasing demand of health service and reformation of healthcare systems in our country, evaluation methods of healthcare performance are urgently needed and thus severity assessment system must be put forward in advance.â—† The objects of the study: to develop a patient classification system according to patient' severity using large discharge data and statistical techniques, which can be used as severity standardization method in management of healthcare quality and cost.â—† The data used in our study: all discharge abstract data of 248 hospitals around the whole country of 1998 provided by health unit of rear-service department of PLA, which include 1,420,692 cases. After deleting illogic records and cases with M, E and V ICD-9 codes as first diagnosis, there are 1,244,887 records available in the database.â—† The start point of patient classification and the method of database splitting: we take disease categories (labeled by ICD-9 three-digit categories) as the beginning of further classification. Thus, the whole database was divided into two kinds of smaller databases according to how frequent the diseases present. One is composed of 120 frequent diseases and the other includes remained non-frequent diseases. The first database was further spitted into 128 smaller databases on the basis of disease categories and weather operations were performed on the majority of patients, and the latter was div...
Keywords/Search Tags:Severity, Casemix, Evaluation, Data mining, Decision tree
PDF Full Text Request
Related items