Reed D R, Price R A
Center for Neurobiology and Behavior, Department of Psychiatry, University of Pennsylvania, Philadelphia 19104, USA.
Int J Obes Relat Metab Disord. 1998 Sep;22(9):827-35. doi: 10.1038/sj.ijo.0800666.
Information about the accuracy of family informant estimates of height and weight should assist investigators in evaluating the costs and benefits of using this type of data in genetic study designs.
To assess the accuracy of family informant estimates, 374 first-degree relatives from 94 Caucasian families, gave estimates about the heights and weights of their first degree relatives. These estimates were compared with measured heights and weights to determine their accuracy.
Informant estimates were highly predictive of measured heights (r=0.95), and weights (r=0.94), but informants systematically overestimated heights (mean=1.4 cm) and underestimated weights of their family members (mean=4.1 kg).
On average, height estimates were generally within 1% of the measured height and weight estimates were within 3-5% of the measured weight. Therefore, these proxy measures can provide useful data, when measured or self-reported heights and weights are not available.
关于家族 informant 对身高和体重估计准确性的信息，应有助于研究人员评估在基因研究设计中使用此类数据的成本和收益。
为评估家族 informant 估计的准确性，来自94个白种人家族的374名一级亲属对其一级亲属的身高和体重进行了估计。将这些估计值与测量的身高和体重进行比较，以确定其准确性。
informant 的估计对测量身高（r = 0.95）和体重（r = 0.94）具有高度预测性，但 informant 系统性地高估了身高（平均 = 1.4厘米），并低估了其家庭成员的体重（平均 = 4.1千克）。
平均而言，身高估计值通常在测量身高的1%以内，体重估计值在测量体重的3 - 5%以内。因此，当无法获得测量或自我报告的身高和体重时，这些替代测量方法可以提供有用的数据。
