A Comparative Analysis of the Respiratory Subscore of the Sequential Organ Failure Assessment Scoring System

RATIONALE The Sequential Organ Failure Assessment (SOFA) is a commonly used measure of illness severity. Calculation of the respiratory subscore of SOFA is frequently limited by missing PaO2 data. Although missing PaO2 data is commonly replaced with normal values, performance of different methods of substituting PaO2 for SOFA calculation is unclear. OBJECTIVE To compare the performance of different substitution strategies for missing PaO2 data for SOFA score calculation. METHODS This retrospective cohort study was performed using the Weill Cornell-Critical carE Database for Advanced Research (WC-CEDAR), from a tertiary care hospital in the United States. All adult patients admitted to an intensive care unit from 2011-2019 with an available respiratory SOFA score were included. We analyzed the availability of PaO2/FiO2 on the first day of ICU admission. In those without a PaO2/FiO2, the SpO2/FiO2 was used to calculate a respiratory SOFA subscore according to four methods (linear substitution (Rice), non-linear substitution (Severinghaus), modified respiratory SOFA, and multiple imputation by chained equations (MICE)) as well as missing as normal. We then compared how well the different total SOFA scores discriminated in-hospital mortality. We performed several subgroup and sensitivity analyses. RESULTS We identified 35,260 unique visits, of which 9,172 had predominant respiratory failure. PaO2 was available for 14939 (47%). The AUC for each substitution technique discriminating inhospital mortality was higher than missing as normal in all analyses, modified 0.80 (0.79-0.81), Rice 0.80 (0.79-0.81), Severinghaus 0.80 (0.79-0.81), and MICE 0.80 (0.79-0.81) compared to missing as normal, 0.78 (0.77, 0.79), p<0.01. Each substitution method had a higher accuracy discriminating in-hospital mortality, MICE 0.67, Rice 0.67, modified 0.66, Severinghaus 0.66 compared to missing as normal. Model calibration was less precise for in-hospital mortality for missing as normal compared to other substitution techniques at the lower range of SOFA, and amongst the subgroups. CONCLUSIONS Using physiologic and statistical substitution methods improved the total SOFA score's ability to discriminate mortality compared to missing as normal. Treating missing as normal may under-report severity of illness compared to substitution. The simplicity of a direct SpO2/FiO2 modified SOFA technique make it an attractive choice for electronic health recordbased research. This knowledge can inform comparisons of severity of illness across studies that used different techniques.