Wright H J, Walker P C, Webster J
J R Coll Gen Pract. 1983 Aug;33(253):493-7.
A random sample of 617 white Caucasian primiparae was identified from notifications made to the Leeds Area Health Authority over a 12-month period. Of these, 534 were interviewed regarding personal characteristics, events during pregnancy, and their experience of infant feeding. Significant associations between these factors and the mode of infant feeding initially used are described.From discriminant analysis of two half-populations, weightings were derived for each significantly associated variable, and discriminant scores were calculated for each participant. The distributions of these scores were similar in both half-populations, suggesting that the weightings were stable and thus have potential predictive importance. Further analysis showed that, from all the significant variables, the choice of two (maternal age-group at confinement and age at leaving school) correctly predicted 79 per cent of mothers who will choose to bottle-feed. The potential clinical yield of using these two factors is discussed, and a simple predictive tool for use in everyday practice is presented.
从向利兹地区卫生局提交的为期12个月的通知中，随机抽取了617名白人初产妇样本。其中，534人就个人特征、孕期事件及其母乳喂养经历接受了访谈。描述了这些因素与最初采用的婴儿喂养方式之间的显著关联。通过对两个子样本的判别分析，得出了每个显著相关变量的权重，并为每个参与者计算了判别分数。这两个子样本中这些分数的分布相似，表明权重是稳定的，因此具有潜在的预测重要性。进一步分析表明，从所有显著变量中，选择两个变量（分娩时的母亲年龄组和离校年龄）能够正确预测79%选择奶瓶喂养的母亲。讨论了使用这两个因素的潜在临床价值，并提出了一种用于日常实践的简单预测工具。
