非消除型选择性偏倚的评估  

EVALUATION ON SELECTION BIAS OF NON-ELIMINATION PATTERN.

作　　者：陈彬[1,2,3,4] 张思霖[1,2,3,4] 俞红吉 左惠娟[1,2,3,4] 傅梅菁 黄强[1,2,3,4] 

机构地区：[1]华西医科大学公共卫生学院 [2]四川省计划生育科学研究所 [3]华西医科大学临床流行病教研室 [4]四川省运动技术学院数学教研室

出　　处：《现代预防医学》1998年第1期15-18,共4页Modern Preventive Medicine

基　　金：四川省计划生育委员会资助

摘　　要：选择性偏倚是医学实践中客观存在的，文中从特定因素远期效应考虑，将其分为消除型选择性偏倚与非消除型选择性偏倚。从盐亭县资料提示，非消除型选择性偏倚可分为平行型、递增型和递减型。对非消除型选择性偏倚评诂必须有适宜的观察时间。对资料可根据配对或成组设计进行分析，亦可作相关分析或趋势检验。做结论时，必须有充分观察时期。 Selection bias,which was sepatated into elimination and nonelimination pattern according to forward effect of specially designated factor,are objective existence in medical practice.According to forward effectiveness of a specific factor,and results to analysis data of vasectomized and nonvasecmized men for Yanting County,the nonelimination pattern was separated into three pattern:paralled pattern,progressive extension pattern and progressive decrease pattern.A observation period must be suitable for evaluation on selection bias of nonelimination pattern.A statistical methods for this indicatrix are selected according to paried sign or others.When observing period is full,the conclusion of this data can be obtained,but it is the best that situation would be observed in different area,and synthetic analysis would be executed for this data.

关 键 词：偏倚 选择性偏倚 非消除型 评估 

分 类 号：R181.32[医药卫生—流行病学]