上海市全民控盐干预效果评价——大样本人群抽样效度分析

Evaluation on a Mass Campaign for Salt Control in Shanghai: Sampling Validity of a Large Population-based Study

[目的]探讨上海市全民控盐干预效果评价课题调查样本的代表性。[方法]通过玛叶指数（MI）、拟合度检验、相异指数（DI）以及人口金字塔图等方法和途径,检验样本与总体人口构成的符合程度;并对随访样本人群社会人口学特征进行前后比较。[结果]本次调查共3 000户家庭和7 000名居民,实际调查2 960户、6 886人,剔除不合格问卷,最终纳入分析的基线数据为2 941份家庭问卷和6 748份个人问卷;而在发放控盐勺后1、6、12个月家庭问卷和个人问卷的失访率在0.51%~4.49%和1.38%~4.70%之间。经检验,1、6、12个月时随访对象的基本人口学特征（地区、年龄、性别、文化程度、就业状态、是否固定掌勺者）与基线时比较,差异均无统计学意义（均P〉0.05）。计算结果显示,样本人群MI为7.47,其中男性为8.70,女性为6.35,均〈10。金字塔图可见,样本人群与2008年上海市总体人口（全人群、男性、女性）年龄构成左右基本对称。以2008年上海市年中人口构成数据作为理论数作拟合度检验,得到样本人群合计、男性和女性的χ2值分别为：5.30、5.88、4.84,均P〉0.05。同时,计算得到样本人群合计及男性、女性人口的DI分别为10.34和10.72、9.96。[结论]本调查样本失访率较低,随访对象的人口学特征与基线无差异,即前后调查样本的一致性较好;样本人群的年龄分布质量较好,样本的分布与总体分布无差异,对总体人群有较好的代表性,样本的调查结果可外推到全市的情况。

[ Objective ] To examine the representativeness of samples from Shanghai Mass Campaign for Salt Control. [ Methods ] Myer＇s index （MT）, test of goodness for fit, dissimilarity index （Dl）, and population pyramid were used to examine the distribution of samples versus the total population in Shanghai by various demographic characteristics. The demographic characteristics of subjects in follow-up visits were compared before and after the intervention. [ Results ] This intervention project invited 3 000 households and 7 000 residents in Shanghai, and finally enrolled 2 960 households and 6 886 residents. After unqualified questionnaire sheets were excluded, data from 2 941 households and 6 748 residents were valid and analyzed. The attrition rates of household and individual questionnaires at the lth, 6th, and 12th months after delivery of salt measuring spoons ranged 0.51%-4.49% and 1.38%-4.70% respectively. No differences were found in the demographic characteristics （such as region, age, gender, education, employment, fixed cook） between the follow-up participants at the lth, 6th, and 12th months and baseline subjects （all P〉0.05）. The M/was 7.47 for the total sampled population, 8.70 for the male, and 6.35 for the female, all less than 10. The population pyramid diagrams （total, male, and female） showed a symmetrical distributions of various age groups of the sampled population and the total population of Shanghai in 2008. Using the 2008 Shanghai population data as reference for test of goodness for fit, the ehi-square values for total, male, and female sampling population were 5.30, 5.88, and 4.84, respectively （all P〉0.05）. In addition, the Dls were 10.34, 10.72, and 9.96, respectively. [ Conclusion ] The intervention project is featured by low attrition rates and no demographic differences between participants in follow-up visits and at baseline, indicating high consistency. Fair age distributions between the sampled population and the total Shanghai population indicate a good representativeness of the sampling subieets. It is therefore valid to conclude that the intervention results can be extrapolated to the general Shanghai population.