Abstract: Objective　To build a discriminant function of clonorchiasis sinensis using stepwise discriminant analysis, so as to investigate the feasibility of discriminant analysis for clonorchiasis sinensis screening. Methods　Ten villages in Mulan County of Harbin City were sampled as the study sites using a cluster random sampling method. The fecal samples were collected from the permanent residents in the study sites and detected for Clonorchis sinensis eggs using the modified Kato?Katz smear technique. A questionnaire of clonorchiasis sinensis was designed. All data were entered into Epidata 3.1, and a discriminant analysis was performed using the software SPSS version 15.0. Variable were screened using the stepwise discriminant analysis, and the discriminant function was built using the Fisher’s discriminant analysis method. The effectiveness of the discriminant function for clonorchiasis sinensis screening was evaluated by comparison with the modified Kato?Katz smear method. Results　Eight variables with statistical significance were included to build the discriminant function, including chronic cholecystitis, cholangitis, gender, eating raw fish, abdominal distension, ethnicity, abdominal pain and age, and the correction rate of the discriminant function was 88.75% to identify clonorchiasis sinensis. Conclusions　The discriminant function of clonorchiasis sinensis can be rapidly and simply built based on the strong data processing and analysis capability of the SPSS software, which is rapid to screen clonorchiasis sinensis in Harbin City. Such a function has a high discriminant analysis capability, and provides insights into the establishment of rapid screening of clonorchiasis sinensis in other endemic areas.

Key words: Clonorchiasis sinensis, Screening, Stepwise discriminant analysis, Fisher’s discriminant analysis

摘要： 目的　采用逐步判别分析法建立华支睾吸虫病判别函数，探讨其在华支睾吸虫病筛查中应用的可行性。方法　在黑龙江省哈尔滨市木兰县采取整群随机抽样法，抽取10个沿江村屯常住人口作为调查对象，采用改良加藤厚涂片法检测受检者粪便中的华支睾吸虫虫卵。设计华支睾吸虫病调查表，用EpiData 3.1软件建立数据库并录入数据，用SPSS 15.0软件对数据进行判别分析。采用逐步判别分析法筛选变量，然后采用Fisher判别法建立判别函数，通过与改良加藤厚涂片法结果进行对比，评价判别函数的判别能力。结果　判别函数中慢性胆囊炎、胆管炎、性别、吃生鱼、腹胀、民族、腹痛、年龄等8个变量有统计学意义，华支睾吸虫病判别正确率为88.75%。结论　利用SPSS软件的强大数据处理和分析能力，可以快速简便地建立华支睾吸虫病判别分析函数。该函数具有较高判别分析能力，能够快速筛检本地区华支睾吸虫病患者，同时为其他华支睾吸虫病流行区建立快速筛检模式提供借鉴。

关键词: 华支睾吸虫病, 筛查, 逐步判别分析, Fisher判别分析