223 Evaluating the Clinical Respiratory Score for Initiating High Flow Nasal Cannula in the Pediatric Emergency Department

Abstract

There are no established guidelines for the initiation of oxygen therapy in pediatric patients. With the increased utilization of high flow nasal cannula (HFNC) for patients with respiratory failure, it is critical to develop a systematic approach for the initiation of this oxygen delivery modality. Currently at our institution, providers utilize the Clinical Respiratory Score (CRS) to assess all respiratory distress patients in the emergency department (ED). CRS is a 12-point assessment that includes respiratory rate, work of breathing, wheezing, hypoxia, dyspnea, and cough. While this score serves as a guidance for asthma treatment, it is not extended to HFNC therapy decisionmaking. We investigate the correlation between CRS and HFNC use in the ED. To determine the relationship between ED-assessed CRS and respiratory therapy, we completed a structured explicit retrospective chart review at a single institution between 2/1/15-12/13/16. We included all patients, 0-18 years old, admitted to the floor or Pediatric Intensive Care Unit (PICU) from the ED with an ED-assessed CRS. If a patient has multiple CRS during their course in the ED, only the first CRS was used in the analysis. Diagnosis code were used to include patients with asthma, pneumonia, bronchiolitis, viral upper respiratory illness, wheezing, and hypoxia. Patients who have tracheostomy tube, baseline oxygen dependence, cystic fibrosis, and croup were excluded from the study. We calculated the predicative values, sensitivity, and specificity for the CRS relative to the incidence of HFNC initiation in the ED. Additional stratification by diagnosis was conducted to understand patterns of assessment between patients with bronchiolitis, pneumonia, and asthma. A total of 1,093 patients were included in the assessment. The median age was 3.1 years (median 0.9, IQR 7.2). 63% of the cohort had a diagnosis of asthma; 35% of the cohort had a diagnosis of bronchiolitis; and 29.6% of the cohort had a diagnosis of pneumonia. With respect to ED-based supplemental oxygen treatment, 23% of the cohort received HFNC compared to 51% for NC. Patients receiving HFNC received a statistically significantly higher CRS compared to non-HFNC (HFNC mean 4.57, SD 1.87 versus non-HFNC mean 3.28, SD 1.98; p<0.001). Differences in distributions are slightly more pronounced for the bronchiolitis cohort (HFNC mean 4.26, SD = 1.77 versus non-HFNC mean 2.65, SD 1.76; p<0.001) compared to asthma (HFNC mean 5.05, SD = 1.73 versus non-HFNC mean 3.71, SD 1.92; p<0.001) and pneumonia (HFNC mean 4.63, SD = 2.17 versus non-HFNC mean 3.05, SD 2.07; p<0.001). Stratification by diagnosis showed a CRS of ≥ 5 in the asthma cohort has sensitivity of 65% and specificity of 67%, and CRS of ≥ 4 in the bronchiolitis cohort has sensitivity of 71% and specificity of 71%. Based on this cohort review, CRS was not a highly sensitive or specific tool to discern the need for HFNC. Initiation of HFNC within the ED has shown to be an effective rescue modality for some patients in respiratory distress, but also requires costlier equipment and higher likelihood of PICU observation. It is therefore critical to develop a clinical support tool to help guide providers in the usage of HFNC in the ED. Additional refinement and validation of components within CRS (or other respiratory scoring tools) should be explored in order to establish guidelines and best practices for initiating HFNC therapy.