Breathing variability predicts the suggested need for corrective intervention due to the perceived severity of...

Patient-ventilator asynchrony is associated with intolerance to noninvasive ventilation (NIV) and worsened outcomes. Our goal was to develop a tool to determine a patient needs for  intervention by a practitioner due to the presence of patient-ventilator asynchrony. We postulated that a clinician can determine when a patient needs corrective intervention due to the perceived severity of patient-ventilator asynchrony. We hypothesized a new measure, patient breathing variability, would indicate when corrective intervention is suggested by a bedside practitioner due to the perceived severity of patient-ventilator asynchrony. With IRB approval data was collected on 78 NIV patients. A panel of experts reviewed retrospective data from a development set of 10 NIV patients to categorize them into one of the three categories. The three categories were; "No to mild asynchrony-no intervention needed", "moderate asynchrony-non-emergent corrective intervention required", and "severe asynchrony-immediate intervention required". A stepwise regression with a F-test forward selection criterion was used to develop a positive linear logic model predicting the expert panel's categorizations of the need for corrective intervention. The model was incorporated into a software tool for clinical implementation. The tool was implemented prospectively on 68 NIV patients simultaneous to a bedside practitioner scoring the need for corrective intervention due to the perceived severity of patient-ventilator asynchrony. The categories from the tool and the practitioner were compared with the rate of agreement, sensitivity, specificity, and receiver operator characteristic analyses. The rate of agreement in categorizing the suggested need for clinical intervention due to the perceived presence of patient-ventilator asynchrony between the tool and experienced bedside practitioners was 95% with a Kappa score of 0.85 (p < 0.001). Further analysis found a specificity of 84% and sensitivity of 99%. The tool appears to accurately match the suggested need for corrective intervention by a bedside practitioner. Application of the tool allows for continuous, real time, and non-invasive monitoring of patients receiving NIV, and may enable early corrective interventions to ameliorate potential patient-ventilator asynchrony.