率的标准化与率差分解法在行为医学研究中的应用

目的 将率的标准化和率差分解法引入行为医学研究，调整因研究总体内部构成不同对率的影响。方法 应用重复抽自助法估计成分效尖的标准,并进行显著性检验。结果 美国北部某静脉吸毒人群血清HIV抗体阳性率资料分析表明，由年龄构成不同造成的效应贡献率为68．9％，有统计学意义；由种族构成不同造成的效应和年龄－种族别不同造成的率效应贡献率分别为－0．71％和31．65％，无统计学意义。结论 率的标准化和率差分解法可广泛应用于行为医学研究。

Standardization and Decomposition of Rates and its Use in Behavior and Health Studies

Objective To use standardization and decomposition of rates in behavior and public health studies for adjusting the impact of compositional factors on rates.Methods Bootstrapping is used to estimate standard errors of the component effects and conduct significance tests for them.Results Empirical data on HIV seropositivity rates in two injection drug using populations in the northeastern United States is used to demonstrate the application of standardization and decomposition analysis in behavior and public health studies.The results show that the contribute rate of age effect is 68\^9% and it has the statistics significance,the contribute rate of gender effect and age gender effect are -0 71% and 31\^65% respectively,it is no statistics significance.Conclusion standardization and decomposition of rates can be used in the behavior and public health studies widely.