Human Factor Errors in the use of the PAWPER Tape Systems: An Analysis of Inter-Rater Reliability

Background International guidelines have promoted the use of length-based tapes for emergency weight estimation in children. This is primarily because of a perception that more modern methods might require more training than can practically be achieved. This study aimed to evaluate the inter-rater reliability of novice users of the PAWPER XL (paediatric advanced weight-prediction in the emergency room) tape as an indicator of the ability to use the device effectively with limited training. Methods This was a secondary analysis of unpublished data from three previous studies. Inter-rater reliability analyses were performed for each study for the accuracy of weight estimations and for the assignment of body habitus score. Inter-rater reliability was analysed using percentage agreement and Cohen's Kappa for Study 1 and intraclass correlation (ICC) for Study 2 and Study 3. A value of <0.7 was considered to indicate inadequate agreement, 0.7 to 0.89 was considered to indicate good agreement and ≥0.9 was considered to indicate excellent agreement. Main Results A total of 7034 data points were available for analysis in the three studies. In Study 1, the percentage agreement of an accurate weight estimation in 50 children, with two raters, was 47/50 (94%) with a Cohen's Kappa of 0.93 (0.81 - 1.0). In Study 2, the ICC for 6720 habitus score assessments, with 112 raters, was 0.73 (0.68 - 0.80). In Study 3, the ICC for 264 weight estimations (in terms of an accurate weight estimation, with 33 raters) was 0.88 (0.72 - 0.97). Conclusions There was good inter-rater reliability in the assessment of habitus and the accuracy of weight estimation for the PAWPER XL tape in this secondary analysis. The findings suggest that reasonable proficiency with the system can be achieved with minimal training. It is therefore unlikely that systems such as this require too much training to be practical in emergency care.