Danielle Cristina Gomes

Agreement between MATLAB® 2D and AutoCAD® methods for the evaluation of thoracoabdominal mobility in newborns

Background: The available methods that assess thoracoabdominal complex in neonatology are costly and difficult to manipulate. Biophotogrammetry emerged as an accessible method, which can be analyzed through AutoCAD®, but it is time-consuming and evaluator-dependent. Then, an algorithmic sequence using MATLAB® software was created by a group of researchers, aiming to evaluate the thoracoabdominal movement in newborns bedside. Aim: to evaluate the agreement between MATLAB® and AutoCAD® methods for thoracoabdominal mobility evaluation of neonates. Method: This is an observational cross-sectional study, with healthy term newborns and up to 72 hours of life. The newborns were placed in supine to collect the videos assessing their thoracoabdominal mobility, which were later analyzed through MATLAB and AutoCAD®. The agreement between these methods was analyzed trough Bland Altman, evaluating the difference of means (BIAS) between AutoCAD® and MATLAB®, and considering a confidence interval of 95%. Results: Videos of 26 newborns were collected, totaling 156 frames, but 17 of these were excluded because showed alterations that could harm the analysis. Then, a total of 139 frames were evaluated. The concordance analysis between the mobility measures showed a value difference of -1.68cm2 (BIAS) with a limit of agreement from +3.22 to -6.59 cm2 (Figure 1). Conclusions: In the population studied, there was an agreement between MATLAB® and AutoCAD® for thoracoabdominal mobility evaluation, suggesting that the methods are interchangeable. Thus, MATLAB® emerges as a simple, fast, inexpensive, and reliable alternative software that can be used for bio-photogrammetric analysis.