趋势推测数学模型(MMI)的构建及其在大肠埃希菌头孢他啶耐药趋势与规律的应用  

A mathematical trend-inferring model's development and its application in prediction of ceftazidime-resistance of E.coli

作　　者：黄晓明[1] 丁凡[1] 王和[1] 

机构地区：[1]贵阳医学院微生物学教研室,贵阳550004

出　　处：《中国抗生素杂志》2010年第5期388-390,共3页Chinese Journal of Antibiotics

基　　金：贵州省卫生厅科研计划基金立项(黔卫发2006)

摘　　要：目的构建耐药趋势推测的数学模型(MMI)并用于对大肠埃希菌头孢他啶耐药性规律推测应用的价值。方法收集1996年至2006年期间发表的国内期刊全文数据库(CNKI)中大肠埃希菌头孢他啶耐药性数据,分别用趋势χ2检验和MMI对耐药数据进行推测并与以误差平方和(SSE)、误差标准差(RSE)、平均绝对误差(MAD)衡量推测结果的准确性。结果根据MMI数学模型建立的曲线,1996年至2006年期间大肠埃希菌头孢他啶耐药率显示逐渐增高的趋势(P<0.01),与实际数据趋势吻合。在构建的MMI中,当参数取(M=2,K=2/3)时的SSE、RSE、MAD值均为最小,则推测效果最好。结论 MMI可用于细菌耐药性数据的分析与推测,对大肠埃希菌头孢他啶耐药性数据的分析与推测显示较高的准确度,其中M和K参数的取值可对推测结果产生较大影响。Objective To establish a mathematical model for trend-inference （MMI） and to estimate its application value by using it to analyze the developing trend and regularity of ceftazidime-resistance of E.coli. Methods Data about drug-resistance of E.coli to cefiazidime during 1996-2006 were collected from China National Knowledge Internet （CNKI） and analyzed statistically. The annual means of drug-resistance were analyzed by using x^2 trend test and presumed with MMI. Sum of squares of error （SSE）, relative standard error （RSE）, and mean absolute deviation （MAD） were simultaneously calculated and compared in order to estimate the accuracy of MMI. Results According to the curves produced by MMI, the annual mean drug-resistance showed a escalating trend （P〈0.01）, which was accordant with the practical status. Being analyzed with MMI, the smallest values of SSE, RSE and MAD, and best presumption were obtained when parameters of M=2, and K=2/3 were used. Conclusion （1） MMI is suitable for analyzing and inferring of developing trend of drug-resistance of bacteria; （2） In this study, a satisfied presumption accuracy of ceftazidime-resistance of E.coli was obtained; （3） M and K are important parameters affecting the presumption results.

关 键 词：数学模型 时间序列 耐药性 应用 评价 

分 类 号：R978.11[医药卫生—药品]