Abdel-Rahman Susan M, Jacobsen Ryan, Watts Jennifer L, Doyle Stacy L, OʼMalley Donna M, Hefner Tiffany D, Dowd M Denise
From the *Division of Clinical Pharmacology, Children's Mercy Hospital; †Department of Pediatrics, University of Missouri-Kansas City, School of Medicine; ‡Division of Emergency Medicine, Children's Mercy Hospital; §Department of Emergency Medicine, Truman Medical Center; and ∥Department of Nursing, Children's Mercy Hospital, Kansas City, MO.
Pediatr Emerg Care. 2017 Aug;33(8):548-552. doi: 10.1097/PEC.0000000000000543.
We compared performance characteristics of 7 weight estimation methods examining predictive performance and human factors errors.
This was a prospective study of 80 emergency care providers (raters) and 80 children aged 2 months to 16 years. Raters estimated weights in 5 children with the following 7 strategies: visual estimation, Advanced Pediatric Life Support, Luscombe and Owens, Broselow tape, devised weight estimation method, 2D Mercy TAPE (2DT), and 3D Mercy TAPE (3DT). Quantitative errors were determined by checking rater values against values returned with optimal method use.
Four hundred rater-child pairings generated 2800 weight estimates. For all methods, rater-estimated weights were less accurate than weights derived by optimal application. Skill-based, perception, and judgment/decision error were observed. For visual estimation, weights were underestimated in most children. For Advanced Pediatric Life Support/Luscombe and Owens, order of operations markedly impacted errors with 23% of calculations requiring addition first performed incorrectly versus 9% of calculations requiring multiplication first. For Broselow tape, only 63% of cases were eligible for estimation with this device, yet raters assigned a weight in 96% of cases. For Devised Weight Estimation Method, 96% of overweight and 48% of obese children were classified as slim or average. For 2DT/3DT, the 2DT was prone to more errors most commonly use of the wrong side of the device (24%). The impact of rater characteristics on error was most pronounced for methods requiring calculation.
Skill-based, perception, or judgment errors were observed in more than 1 of 20 cases. No singular strategy was used with 100% accuracy.
我们比较了7种体重估计方法的性能特征，考察其预测性能和人为因素误差。
这是一项针对80名急救人员（评估者）和80名年龄在2个月至16岁之间儿童的前瞻性研究。评估者采用以下7种策略对5名儿童的体重进行估计：视觉估计、高级儿科生命支持、卢斯科姆比和欧文斯方法、布罗泽洛卷尺法、自行设计的体重估计方法、二维梅西卷尺（2DT）和三维梅西卷尺（3DT）。通过将评估者给出的值与使用最优方法得出的值进行核对来确定定量误差。
400次评估者与儿童的配对产生了2800个体重估计值。对于所有方法，评估者估计的体重都不如通过最优应用得出的体重准确。观察到了基于技能、感知和判断/决策方面的误差。对于视觉估计，大多数儿童的体重被低估。对于高级儿科生命支持/卢斯科姆比和欧文斯方法，运算顺序对误差有显著影响，23%需要先进行加法的计算被错误执行，而9%需要先进行乘法的计算被错误执行。对于布罗泽洛卷尺法，只有63%的病例适合用该工具进行估计，但评估者在96%的病例中都给出了体重估计值。对于自行设计的体重估计方法，96%的超重儿童和48%的肥胖儿童被归类为苗条或正常体重。对于2DT/3DT，2DT更容易出现错误，最常见的是使用了工具错误的一面（24%）。对于需要计算的方法，评估者特征对误差的影响最为明显。
在每20个病例中，超过1例观察到基于技能、感知或判断的误差。没有一种单一策略的准确率能达到100%。
