IntroductionThe Roth score is an alternative strategy to estimate oxygen saturation by using a simple verbal test. We designed this clinical study to assess the validity of the Roth score (Spanish version) as a screening test for hypoxemia. A secondary objective was to evaluate the agreement and consistency between the oxygen saturation obtained via pulse oximetry (SpO2) and arterial blood gas test (SaO2). MethodsAn observational study was conducted in two hospital emergency departments. Adult patients who underwent arterial blood gas tests were included in the analysis. Pulse oximetry values were determined, and the Roth score was applied in the Spanish language. The validity of the Roth score was assessed in terms of sensitivity and specificity by creating ROC curves and by calculating the area under the curve (AUC) for SpO2, SaO2, and oxygen pressure in the arterial blood (PaO2). Agreement between SpO2 and SaO2 values was assessed by using the intraclass correlation coefficient (ICC), and consistency between both measures was calculated by following the method of Bland and Altman. ResultsThe ROC curve models of the Roth score results that were obtained for SaO2 < 90%, ≤92%, and < 95% had AUCs of 0.574, 0.462, and 0.543, respectively, for the highest number that was obtained in the test, as well as AUCs of 0.403, 0.376, and 0.495, respectively, for the maximum time that was used. The AUCs for PaO2 ≤ 60, ≤70, and ≤80 mmHg were 0.534, 0.568, and 0.512, respectively, for the maximum number that was obtained in the test, as well as AUCs of 0.521, 0.515, and 0.519, respectively, for the maximum time that was spent. The ICC between SaO2 and SpO2 was 0.817 (p < 0.001); additionally, the mean difference between the two measurements was −0.55. ConclusionThe Roth score performed in Spanish is not a valid test for hypoxemia screening. There is sufficient agreement and consistency between SaO2 and SpO2 measurements.