摘要
目的 应用统计分析方法建立覆盖所有病种的住院患者病情分级方案。方法 用来自全国 2 48家各级各类医院的 12 44 887例住院患者的病案首页数据 ,用是否有一级护理为反应变量 ,患者年龄、入院方式、有无伴随疾病等为预测变量 ,采用决策树分析法 ,寻找每种疾病患者的病情分类规则 ,用危重度等级和危重度分值表示各类病例的病情危重度。结果 建立了 163个基于病种的病情分类模型、产生了 63 6个危重度类别和危重度分值。经新样本考核 ,该病情指标对治疗结果、住院费用有一定预测能力 ,且与医院规模相关。结论 病情分级结果能反映住院患者病情的相对危重度 ,可作为病情测量的有效工具。
Objective To classify and evaluate the severity of inpatients by statistical method.Methods By application of decision tree with abstract data of 1?244?887 inpatients, determining the target and predict variables, find out the classification pattern of severity.Results 163 models were established based on the categories of diseases. The total number of severity groups is 636, with each one a numeric severity score. The validation result of applying the model to new data showed that the score could predict clinical outcome and distinguish hospitals of different scales to a certain extent.Conclusion The models can indicate the severity of inpatients and thus can be used as a tool in severity measurement of inpatients.
出处
《中国卫生统计》
CSCD
北大核心
2004年第2期90-92,共3页
Chinese Journal of Health Statistics
基金
国家自然科学基金资助项目 (79870 0 50 )
军队医药卫生"十五"重点课题 (0 1G0 0 6 1 )
