Confiabilidade inter e intra examinador na fotozoometria bidimensional digital de cavalos

Abstract

This study aimed to evaluate the intra and inter-rater reliability, in standardized fotos and in random photos, using the statistical software RStudio®, with the intra-class correlation coefficient (ICC) test with a significance level of 0 .05. The study was carried out at the Federal Technological University of Paraná (UTFPR-DV), and 4 horses from UNEPE Equinocultura da Fazenda Experimental were used, 2 castrated mixed breed males and 2 Crioula females, aged between 6 and 10 years and weight of these varying between 350 and 450 kg. It was necessary the presence of 8 trained evaluators to take the measurements. Digital photographs were taken of the horses, marking the reference joints for measurements on the skin with an adhesive label. Digital photos were captured using cell phone cameras, observing two ways to enable the determination of intra-rater (repeatability) and inter-rater (reproducibility) reliability, as well as whether the camera position would influence the determination of measurements. Therefore, they were taken in fixed and standardized positions with the use of a tripod (standardized photos) and in random positions (random photos). The standardized photos were taken with the camera of an Iphone 11 device (Apple®) positioned at a distance of 1.50 m and height equivalent to half the withers, duly leveled. For the analysis of repeatability and reproducibility, the digital photographs were sent to the main researcher, who organized the files in folders identified with reference codes for each evaluator. One week after the photo session, the evaluators were trained to perform the measurements using the free Image J® , version 1.51j and received instructions on how to record the data using a spreadsheet with the following data: folder name, animals, angles and measurements. Each evaluator recorded the measurements of the 8 angles in 36 photos, totaling 288 measurements. After 8 weeks, the evaluators were invited to perform the measurements again, using the same photographs and methodology. The analysis of the measures in the standardized photos demonstrates that this technique presents intra- and inter-evaluator reliability by presenting, respectively, repeatability and reproducibility in general (p<0.05). Considering that all other measurements of the angles showed a significant ICC, it is possible to state that the technique of using standardized digital photos in this study, with standardized positioning of height and distance between the digital camera and the horse, using a tripod in a flat and level place, can be used to measure the joint angles of horses. With regard to random photos, all data referring to intra- and interrater reliability underwent the ICC correlation with a significance level of 0.05, demonstrating that this technique presents intra- and inter-rater reliability. Therefore, the technique demonstrates positive evidence, as the same evaluators are able to evaluate both the standardized photos and the random photos.