Fig 1.
The PAWPER XL-MAC tape system.
There are three steps to the use of the PAWPER XL-MAC tape. The first step is to measure the length of the child from his/her head to his/her heel (Panels A and B). The user thus identifies into which length segment the child falls. The second step is to measure the child’s mid-arm circumference using the tape (Panel C). The final step is to read off the estimated weight in the appropriate length segment based on the measured mid-arm circumference (Panel D). The user should be familiar and well-practised with the tape before using it in a resuscitation situation.
Table 1.
Basic demographic information of the children in the dataset.
Table 2.
Children classified according to normal and abnormal growth or body weight.
Fig 2.
Analyses of the overall accuracy of the weight estimation systems.
The chart shows the proportion of weight estimations falling within 10% and 20% of actual weight (p10 and p20 respectively) as well as the proportion of critical weight estimation errors (>20% error). The data for the Ralston method was obtained from the published study [9]. The McNemar test was significant at the p<0.001 level for every comparison of p10 and p20. The PAWPER XL-MAC method’s p10 beached the 10% improvement criterion when compared with the other methods. The p10 of the Broselow 2011A was clinically inferior to all other methods.
Fig 3.
Analyses of the bias and precision of the weight estimation systems.
The forest plot shows the overall bias (the black circles indicate the mean percentage error), as well as the precision (the whiskers indicate the Bland & Altman 95% limits of agreement) for each system. The green shaded area denotes an acceptable MPE (within ±5%), while the dashed lines indicate an acceptable range for the 95% LOA (within ±20%). The paired t-test was significant at the p<0.001 level for every comparison of MPE and RMSPE. The MPE of the PAWPER XL-MAC and Ralston methods were clinically superior to the Broselow tape methods. The precision (quantified using the RMSPE) of the PAWPER XL-MAC method was clinically superior to all other methods.
Fig 4.
Accuracy of the three systems in “normal weight” children.
The chart shows the proportion of weight estimations falling within 10% and 20% of actual weight (p10 and p20 respectively) as well as the proportion of critical weight estimation errors (>20% error) for children with a BMI-for-age Z-score between -1.4 and 1.4. This subgroup data was not available for the Ralston method as it information was not presented in the original publication. The McNemar test was significant at the p<0.001 level for every comparison of p10 and p20. However, the differences were not clinically important.
Fig 5.
Accuracy of the three systems in underweight children.
The chart shows the proportion of weight estimations falling within 10% and 20% of actual weight (p10 and p20 respectively) as well as the proportion of critical weight estimation errors (>20% error) for children with a BMI-for-age Z-score less than -1.4. This subgroup data was not available for the Ralston method as it information was not presented in the original publication. The McNemar test was significant at the p<0.001 level for every comparison of p10 and p20. The PAWPER XL-MAC method p10 and p20 were clinically superior to both versions of the Broselow tape and the Broselow 2007B was clinically superior to the Broselow 2011A.
Fig 6.
Accuracy of the three systems in overweight and obese children.
The chart shows the proportion of weight estimations falling within 10% and 20% of actual weight (p10 and p20 respectively) as well as the proportion of critical weight estimation errors (>20% error) for children with a BMI-for-age Z-score greater than 1.4. This subgroup data was not available for the Ralston method as it information was not presented in the original publication. The McNemar test was significant at the p<0.001 level for every comparison of p10 and p20. The PAWPER XL-MAC method p10 and p20 were clinically superior to both versions of the Broselow tape. The p10 of the Broselow 2011A was clinically superior to the Broselow 2007B, but the p20 of the Broselow 2007B was clinically superior to the Broselow 2011A.
Table 3.
Weight estimation performance by subgroups of sex.
Table 4.
Weight estimation performance by subgroups of age.
Table 5.
Weight estimation performance by subgroups of weight.
Table 6.
Weight estimation performance by subgroups of weight status.
Fig 7.
Accuracy outcomes by subgroups of weight-for-height.
The p10 data is shown in the upper panel and the p20 data in the lower panel. Children with a weight-for-height Z-score of ≥-2.0 were categorised as “normal”, between -2.0 and -3.0 as “moderately wasted” and below -3.0 as “severely wasted”. The p10 data for the Ralston method was imputed from the MPE data. The McNemar test was significant at the p<0.001 level for every comparison of p10 and p20. We considered the PAWPER XL-MAC method p10 to be clinically superior to all the other methods in all subgroups. The Ralston method p10 was clinically superior to the Broselow tapes in wasted children. The PAWPER XL-MAC method p20 was clinically superior to all other methods in all wasted children. The Ralston method p20 was also clinically superior to the Broselow tapes in all wasted children.