空间流行病学中的偏倚与混杂  

Study on bias and confounding in ‘Spatial Epidemiology’

作　　者：周艺彪[1] 姜庆五[1] 赵根明[1] 

机构地区：[1]复旦大学公共卫生学院流行病学教研室,上海200032

出　　处：《中华流行病学杂志》2005年第2期135-139,共5页Chinese Journal of Epidemiology

摘　　要：目的　探讨空间流行病学中的偏倚与混杂。方法　结合实例分析空间流行病学中可能存在偏倚与混杂及其对研究结果的可能影响。结果　空间流行病学研究中存在选择性偏倚 ,确证、分子和分母偏倚 ,由疾病诱导期 潜隐期的选择和暴露 疾病模式的错误载明所致的偏倚 ,暴露不准确偏倚 ,空间相关性 ,显著性检验 ,生态学偏倚和社会 经济混杂等 8种偏倚与混杂。结论　空间流行病学研究中的偏倚来源众多且较为复杂 ,由此可以夸大或掩盖研究结果 ,故对研究结果的解释应慎重。Objective To explore the biases and confoundings in Spatial Epidemiological studies. Methods Possible bias and confounding and their impact on study results in Spatial Epidemiology were analyzed in given examples. Results In Spatial Epidemiology,biases related to ascertainment/numerator/denominator induced by the choice of the disease induction/latency period and mis-specification of exposure-disease model,exposure inaccuracy,spatial dependency,significance tests etc. were involved,as well as to ecological,socio-economic confoundings factors. Conclusion The sources of bias in ‘Spatial Epidemiology’ were both numerous and complex, that might be overestimated or underestimated on the study results. Hence,careful interpretation of such studies was needed.

关 键 词：偏倚 流行病学研究 研究结果 暴露 疾病模式 混杂 影响 分母 结论 空间 

分 类 号：R181.3[医药卫生—流行病学]