Objective To compare the correlations among different measurement methods for living standards of household and to further evaluate the association between living standards of household and antenatal care services utilization in rural areas .Methods Data were collected from the Perinatal Health Care Survey in Rural Areas of Shaanxi Province in 2010.The data of families of 4 719 pregnant women were classified by self -rated economic condition and wealth index constructed by the first principal component and latent class model , and antenatal examination indicators among families with different living standards were compared . Unconditional multivariable logistic regression analyses and receiver operating characteristic ( ROC) were considered in the analysis to evaluate which measurement of living standards of household was more predictive over the antenatal care services utilization .Results It was significantly coherent between wealth indexes constructed by the first principal component and the latent class model , but not for the self-rated economic condition ( Spearman value was -0.175 and -0.111 respectively, and the corresponding P value was less than 0.001 and 0.003, respectively).Rates of antenatal care services utilization in the richer household were higher than those in the poor household according to classifies of wealth indexes (χ2 value between family economy index based on principal component classification and frequency of prenatal examination more than or equal to 5 times, the first prenatal examination earlier than gestational 12 weeks, completing recommend examination contents , and delivering in hospitals at or beyond the township level was 21.8, 20.5, 13.1 and 13.7, respectively; χ2 value between family economy index based on latent class classification and frequency of prenatal examination more than or equal to 5 times, the first prenatal examination earlier than gestational 12 weeks, completing recommend examination contents , and delivering in hospitals at or beyond the township level was 64.2, 35.1, 17.6 and 20.7, respectively, all P<0.001).The area under ROC curve(AUC) of the living standards of household classified by the latent class model over the antenatal care services was higher than that by first principal component , but the difference was not significant (P>0.05), so was predictive probabilities of Logistic analysis .Conclusion The household living standards measured by wealth index constructed by the latent class model and first principal component are suitable for evaluating the associations between living standards of household and antenatal care services utilization in rural areas , and the latent class model seems to be more predictive .%目的：比较对家庭生活水平的不同测量方法间的相关性，在此基础上，进一步研究家庭生活水平与产前检查相关指标的关系。方法数据来源于本课题组在2010年所完成的《陕西省农村围产期保健状况调查》项目。对4719名孕妇家庭的资料分别用主成分分析法和潜在类别模型构造的财富指数及自感经济对家庭生活水平进行分类，比较不同家庭生活水平群体间产前检查相关指标的差异。不同测量方法下的家庭生活水平预测产前检查相关指标的能力比较采用受试者操作特征曲线（ ROC）和多因素Logistic回归分析。结果主成分分析法和潜在类别模型构造的财富指数表现了较好的相关性和一致性，而与自感经济呈负相关（Spearman值分别为－0．175和－0．111，相应的P值分别为＜0．001和＜0．003）；以前两种方法对家庭生活水平进行分类，富裕家庭的产前检查相关指标均显著高于贫困家庭（以主成分法分类的家庭经济指标－产前检查次数≥5次、首次产前检查＜12孕周、完成推荐产检内容和产检地点在乡及以上医院的比例有差异的χ2值分别为21．8、20．5、13．1和13．7；以潜在类别法分类的家庭经济指标－产前检查次数≥5次、首次产前检查＜12孕周、完成推荐产检内容和产检地点在乡及以上医院的比例有差异的χ2值分别为64．2、35．1、17．6和20．7，均P＜0．001）。以潜在类别模型构造的财富指数分类的家庭生活水平对产前检查相关指标的ROC曲线下面积大于主成分分析法，但差异不具有统计学意义（ P＞0．05）；对产前检查相关指标的多因素Logistic回归分析的预测概率略大于主成分分析法。结论用潜在类别模型和主成分分析法构造的财富指数衡量的农村地区家庭生活水平，均适用于研究其与产前检查的关系，且潜在类别模型法更具优势。
展开▼
