Excellence in ECMO Outcomes: Getting Back to Basics

Between 4/2020 and 1/2021 we identified 7 VV ECMO patients (4 IJ, 3 fem/fem) with bloodstream infections due to Enterococcus faecalis. Time from cannulation to first positive blood culture ranged from 8 to 63 days. There was no geographic clustering apparent. A culture of the heater/ cooler water reservoir of an actively infected patient was sterile. Based on our preliminary analysis, we felt that there were likely multiple contributing factors leading to our spike in infections rather than one simple cause. We hypothesized that our nursing focus had shifted during the COVID-19 pandemic to a culture of minimizing spread of COVID-19 to our staff, with less focus on basic patient care and standard patient infection prevention. Knowing that E faecalis can be spread from the patient (it is a common enteric bacteria) to the environment and then back to lines, we began intensive reeducation and focus on basics of ICU patient care emphasizing: (1) Basic hand hygiene/gloving, (2) General room sanitation including disinfecting surfaces daily, (2) Sterility during line and cannula dressing changes, (3) Weekly dressing changes as well as PRN blood or fluid under the dressing, and (5) Diarrhea containment. Since implementing these policies we have had no blood stream infections in our ECMO population. The importance of daily routine care is too easily forgotten. A strategy of teaching and emphasizing basics can produce large and sustained benefits.

On the effect of decomposition granularity on DeTraC for COVID-19 detection using chest X-ray images

COVID-19 is a growing issue in society and there is a need for resources to manage the disease. This paper looks at studying the effect of class decomposition in our previously proposed deep convolutional neural network, called DeTraC (Decompose, Transfer and Compose). DeTraC can robustly detect and predict COVID-19 from chest X-ray images. The experimental results showed that changing the number of clusters (decomposition granularity) affected the performance of DeTraC and influenced the accuracy of the model. As the number of clusters increased, the accuracy decreased for the shallow tuning mode but increased for the deep tuning mode. This shows the importance of using suitable hyperparameter settings to get the best results from the DeTraC deep learning model. The highest accuracy obtained, in this study, was 98.33% from the deep tuning model. © ECMS Khalid Al-Begain, Mauro Iacono, Lelio Campanile, Andrzej Bargiela (Editors)