Validated Assessment Tools for Reconstructing Facial Defects in Simulation

Abstract

Objectives:

1) Create a 3D functional organosilicate model for assessing facial reconstruction that is valid and has high inter-rater reliability. 2) Perform a prospective blinded study using 3D and 2-dimensional (2D) models for evaluating facial reconstruction.

Methods:

A 3D organosilicate model of a human adult face was built with a corresponding 2D paper model. Surgical subspecialty trainees from six institutions prospectively drew a z-plasty on the paper model and the organosilicate model. Next, the subjects each performed a z-plasty on the organosilicate model with surgical instruments and suture. Three blinded reviewers rated two drawings and one scar revision per subject.

Results:

Forty-three trainees encompassing 3 PGY1, 10 PGY2, 13 PGY3, 11 PGY4, 3 PGY5, and 3 PGY6 (fellows) completed the study. Subjects were placed into three groups, novices (PGY1+PGY2), intermediates (PGY3+PGY4), and experts (PGY5+PGY6). Experts performed better than intermediates and intermediates better than novices for performing the z-plasty and drawing on the organosilicate model. This statistically significant difference between groups was determined by one-way ANOVA (p =<.05) and Tukey post-hoc test p=<.05 and showed substantial inter-rater reliability agreement among the three raters with an average Cohen’s Kappa=0.76. On the 2D paper model the experts performed better than novices (p=<.05); however, there was no difference between experts and intermediates (p=0.98).

Conclusions:

We have created a 3D organosilicate model for facial reconstruction that is valid and shows high inter-rater reliability. The 3D model was able to detect a difference between experts and intermediates where the 2D model could not.