我国眼病抽样调查中的常见偏倚问题与对策

Several common biases and control measures during sampling survey of eye diseases in China

偏倚是眼病抽样调查中常见的人为误差，是影响调查结果真实性与可靠性的主要原因。本文分析评论了我国眼病抽样调查中常见的几种偏倚，包括抽样偏倚、无应答偏倚及诊断偏倚等的产生原因及其对策；强调控制偏倚是确保调查研究质量的关键；指出只要随机化抽样、样本量足够、认真检录和询问病史、提高受检率、准确检查诊断、严格培训及预试验，并做好质量控制工作，就可以将偏倚消除或减少到最低程度，从而提高我国眼病抽样调查质量。

Bias is a common artificial error during sampling survey in eye diseases, and is a major impact factor for validity and reliability of the survey. The causes and the control measures of several biases regarding current sampling survey of eye diseases in China were analyzed and discussed, including the sampling bias, non-respondent bias, and diagnostic bias. This review emphasizes that controlhng bias is the key to ensure quality of sampling survey. Random sampling, sufficient sample quantity, careful examination and taking history, improving examination rate, accurate diagnosis, strict training and preliminary study, as well as quality control can eliminate or minimize biases and improve the sampling survey quality of eye diseases in China.