Reproducibility and Predictive Value of Objective Cough Measurement to predict extubation success - COBRE-US trial

<bold>Background:</bold> Utilizing clinical tests, such as objective cough measurement (OCM), can assist in predicting the success of the weaning and extubation process in critically ill patients. <bold>Methods:</bold> Intra- and inter-observer reproducibility of the cough test was evaluated using the intraclass correlation coefficient (ICC) and the Bland-Altman agreement plot. We employed receiver operating characteristic curves (ROC curves) to assess the predictive ability of OCM for both the success of the spontaneous breathing trial (SBT) and extubation. <bold>Results:</bold> We recruited 367 subjects who were receiving invasive mechanical ventilation. A total of 451 OCM and 456 SBTs were conducted. The predictive capability of the OCM for SBT success had a ROC-curve of 0.58 (95%CI:0.56-0.61). OCM to predict successful mechanical ventilation withdrawal had a ROC-curve of 0.61 (95%CI:0.56-0.66). The intraobserver reproducibility exhibited an ICC of 0.94 (95%CI:0.89-0.96;p<0.001), while the interobserver reproducibility demonstrated an ICC of 0.72 (95%CI:0.51-0.85;p< 0.001). The Bland-Altman intra-observer plot of 39 double observations showed that 10.26% were outside the agreement limits, with a mean difference of 0.05, and 95% agreement limits between -0.57 and 0.67. Finally, inter-observer agreement for 39 double observations had 95% agreement limits between -0.99 and 1.51. <bold>Conclusions:</bold> The OCM using the method employed in our study reveals almost perfect intra-observer reproducibility. However, its discriminative ability to predict success or failure in extubation is limited.