ABSTRACT
Background. Infectious respiratory-track pathogens are a common trigger of healthcare capacity strain, e.g. the COVID19 pandemic. Patient risk stratification models to identify low-risk patients can help improve patient care processes and allocate limited resources. Many existing deterioration indices are based entirely on structured data from the Electronic Health Record (EHR) and ignore important information from other data sources. However, chest radiographs have been demonstrated to be helpful in predicting the progress of respiratory diseases. We developed a joint EHR and chest x-ray (CXR) model method and applied it to identify low-risk COVID19+ patients within the first 48 hours of hospital admission. Methods. All COVID19+ patients admitted to a large urban hospital between March 2020 and February 2021 were included. We trained an image model using large public chest radiograph datasets and fine-tuned this model to predict acute dyspnea using a cohort from the same hospital. We then combined this image model with two existing EHR deterioration indices to predict the risk of a COVID19+ patient being intubated, receiving a nasal cannula, or being treated with a vasopressor. We evaluated models' ability to identify low-risk patients by using the positive predictive value (PPV). Results. The image-augmented deterioration index was able to identify 12% of 716 COVID-19+ patients as low risk with 0.95 positive predictive value in the first 48 hours of admission. In contrast, when used individually, the EHR and CXR models each identified roughly 3% of the patients with a PPV of 0.95. Predicting Low Risk Patients Aggregated predictions for COVID19 positive patients within the first 48 hours of admission, shown with exponential weight moving average and 95% CIs. Each plot shows the number of patients flagged as low-risk by lowest aggregated prediction and the resulting accuracy for that fraction of patients. The bottom plot compares the MCURES fused model to the MCURES model. The top plot compares the EDI fused model to the EDI model. Conclusion. Our multi-modal models were able to identify far more patients at low-risk of COVID19 deterioration than models trained on either modality alone. This indicates the importance of combining structured data with chest X-rays when creating a deterioration index performance for infectious respiratory-track diseases.
ABSTRACT
RATIONALE Multiple case reports and case series have described pneumothorax and pneumomediastinum as a complication of patients hospitalized with COVID-19, particularly among those receiving invasive mechanical ventilation. However, it is not known whether patients with COVID-19 have a uniquely higher incidence of these events compared to historical ARDS (non-COVID-19 ARDS) patients. METHODS We compared barotrauma rates in patients hospitalized with COVID-19 who received invasive mechanical ventilation between March-July 2020 to patients with non-COVID-19 ARDS who received mechanical ventilation in 2016-2018. We defined barotrauma as pneumothorax or pneumomediastinum during mechanical ventilation. RESULTS We analyzed 222 patients with COVID-19 who received invasive mechanical ventilation and 421 patients with ARDS. Barotrauma events occurred in 13.1% of patients with COVID-19 and 9.3% of historical ARDS patients (p = 0.136). Mean tidal volumes were 5.7 and 6.4 mL/kg of predicted body weight, plateau pressures were 25.6 and 23.6, PEEP was 11.2 and 8.8, and driving pressures were 14.4 and 14.8 cmH2O, respectively, in COVID-19 and non-COVID-19 ARDS. There were 42 pneumothoraces among COVID-19 patients and 50 among historical ARDS patients (p = 0.144). Incidence rates were 1.7 and 2.7 per 100 ventilator days in COVID-19 and historical ARDS respectively (p=0.808). There were 14 cases of pneumomediastinum among patients with COVID-19 compared to 16 among patients with ARDS (p = 0.152). Overall, pneumothoraces were identified within 24 hours of ipsilateral internal jugular or subclavian line placement in 5.4% (5/92) of events. In both groups, barotrauma was associated with fewer vent-free days at 28 days (3.0 vs 9.2 in COVID-19, p < 0.001 and 7.6 vs 11.5 in historical ARDS, p = 0.0214). Barotrauma was not associated with an increased mortality at discharge for either cohort. For COVID-19 patients only, mean plateau pressure and driving pressure were associated with barotrauma events (28 vs 25 cmH2O, p = 0.0015;16.7 vs 14.0 cmH2O, p ≤ 0.01). Administered tidal volume, PEEP, age, sex, tobacco use, obesity, number of comorbidities, and the presence of lung comorbidities were not associated with barotrauma in either cohort. CONCLUSIONS Both COVID-19 and non-COVID-19 ARDS patients who are mechanically ventilated are at high risk of barotrauma;this was not unique to patients with COVID-19. Barotrauma is associated with prolonged ventilation and fewer ventilator-free days. Despite advances in lung-protective ventilation, barotrauma continues to be a significant source of morbidity in patients mechanically ventilated for respiratory failure.