世界卫生组织健康调查资料的结构方程模型  

A structural equation model for the WHO health survey data

作　　者：刘素芳[1] 林岳卿[1] 何泽慧[1] 付蓉[1] 谭文艳[1] 欧顺云[1] 方积乾[1] 

机构地区：[1]中山大学公共卫生学院医学统计与流行病学系,广州510080

出　　处：《中华预防医学杂志》2010年第7期631-635,共5页Chinese Journal of Preventive Medicine

基　　金：WHO资助项目（HQMH10802021）

摘　　要：目的 探索2002年世界卫生组织（WHO）健康调查数据中,人群自述健康水平、实际健康水平、社会人口学特征、危险因素之间的潜在关系,及各调查指标对自述健康和自述健康对实际健康水平的影响程度.方法 数据来源于WHO提供的2002年包含印度、巴西、布基纳法索、匈牙利、尼泊尔、俄罗斯、西班牙、突尼斯、越南等9个国家在内的健康调查数据（包括自我评价部分和健康情景部分）（29 971例）.本研究利用健康情景资料对自我评价资料进行校正和消除切点位移偏倚后,用自我评价资料构建自述健康水平、实际健康水平、社会人口学特征、危险因素之间关系的结构方程模型.结果 最终结构方程模型：实际健康水平=0.80×自述健康水平＋（-0.04）×社会人口学特征＋（-0.08）×危险因素（R2=0.66）,自述健康水平=（-0.70）×社会人口学特征＋0.10×危险因素（R2=0.55）;自述健康水平对实际健康水平的标准化总效应达0.80,人口学特征对自述健康水平和实际健康水平的标准化总效应分别为-0.70与-0.60;自述健康水平的16项指标含8个维度,按照对其影响程度由大到小依次为：活动能力、疼痛不适、睡眠情况、认知、情绪、自理能力、视觉能力、交际活动能力.结论 实际健康水平与自述健康水平之间、自述健康水平与社会人口学特征之间均有较强的线性关系,自述健康水平对实际健康水平有较大影响,2002年WHO健康调查所用16项指标在人群健康评价中占有重要的地位.Objective Based on the 2002 WHO health survey data, to explore the latent relationship among self-reported health level, the actual level of health, the social demographic characteristics and the risk factors, and to analyze the influence of the various surveillence indicators on self-reported health and the degree that the self-reported health explained the actual level of health.Methods Field tests for various components of the World health survey were conducted in nine countries during 2002, including India, Brazil, Burkina, Hungary, Nepal, Russia, Spain, Tunisia, and Vietnam （29 971 ）.The survey questionnaire included a self-assessment component and anchoring vignette component.The self-assessment component data was adjusted and eliminated the affect of ＂cut-point bias＂ by using the anchoring vignette component data,and then was used to build the structural equation model on the relationship among selfreported health level, actual health level, social demographic characteristics and the risk factors.Results In the final structural equation model, ＂the actual level of health＂ = 0.80 × ＂the self-reported health level＂ ＋（ - 0.04） × ＂the social demographic characteristics＂ ＋ （ - 0.08 ） × ＂the risk factors＂ （ R2 = 0.66 ）, and＂the self-reported health level＂ = （ -0.70） × ＂the social demographic characteristics＂ ＋0.10 × ＂the risk factors＂ （R2 = 0.55 ）.The standardized total effect of self-reported health to the actual level of health was 0.80 ,and that of the social demographic characteristics to the self-reported health and the actual level of health were - 0.70 and - 0.60, respectively.And the 16 items of self-reported health consisted of8 dimensions; and sorted by the power of impact to the actual health level, they were mobility, pain and discomfort, sleep, cognition, feelings, self-care ability, visual capacity and interpersonal activities.Conclusion There were significant linear correlation relationship between the actual level of health and th

关 键 词：健康调查 模型 统计学 健康状况 

分 类 号：R195[医药卫生—卫生统计学]