信息分析与最大似然法在卵巢囊肿早期鉴别诊断上的联合应用及其软件开发  

Combining Information Analysis and Maximum Likelihood Method to Diagnose Early Ovarian Cyst,and Software Developing

作　　者：杨海涛[1] 王莹莹 席海峰[3] 闫丽娜[1] 刘荷一[4] 杨俊英[1] 

机构地区：[1]河北医科大学流行病学与卫生统计学教研室,河北石家庄050017 [2]石家庄市疾病与预防控制中心,河北石家庄050010 [3]河北省卫生厅,河北石家庄050057 [4]河北医科大学肿瘤医院,河北石家庄050011

出　　处：《数理统计与管理》2015年第5期792-798,共7页Journal of Applied Statistics and Management

基　　金：河北省卫生厅科研基金项目(20090318)

摘　　要：本文基于串联的联合诊断形式,联合应用信息分析与最大似然法对卵巢囊肿进行早期鉴别诊断,并在对两种方法的真实性和可靠性做过评价后,应用Delphi7.0程序开发软件对两种方法在卵巢囊肿鉴别诊断上的应用实施软件开发。结果显示两种方法的准确性中等,一致性较好;串联形式的联合诊断提高了诊断的特异度,弥补了单一诊断方法的不足;开发的软件使得临床医生无需掌握复杂的统计学理论和方法,只需通过简单的软件操作,即可方便、高效地完成门诊的辅助诊断。Based on serial testing,we combine information analysis and maximum likelihood method to discriminate ovarian cyst in early stage and develop the multivariate statistical application in the clinical into software by use of Delphi7.0,on the basis of the two method＇s authenticity and reliability evaluation.We find that the discriminant accuracy of both methods are moderate and they have good consistency.The serial testing improves diagnostic specificity and made up for the lack of single statistical method.The developed software can help clinician to improve disease diagnosis accuracy and objectivity,more realistically to obtain a diagnosis without mastering sophisticated statistical methods,the developed software can be used for clinical assistant diagnosis.

关 键 词：信息分析 最大似然法 联合诊断 ROC曲线 软件开发 

分 类 号：R195.1[医药卫生—卫生统计学] O212[医药卫生—卫生事业管理]