Speaker: Tamas Szakmany, MBE, MD, PhD, EDIC, DESA, FRCA, FFICM, FCCM
Tamas Szakmany, MBE, MD, PhD, EDIC, DESA, FRCA, FFICM, FCCM: Hello, my name is Tamas Szakmany, the Associate Editor for for social media at Critical Care Explorations. Here are my top reads from Critical Care Explorations this month. The use of generative large language models in scientific discourse has been at the forefront of many journal editors, reviewers and authors. While the capability is certainly present, it is not clear how well a generative pre-trained transformer or GPT for short would compare to established medical writers. In fresh research from Huespe and colleagues, we got some indication of the reliability of using GPT-3.5 for writing the background of an article for a critical care topic. Their work confirmed that the GPT generated text was highly rated by blinded reviewers. However, they noted the lack of referencing and that the model didn't provide reliable sources. Telecritical care has been embedded in clinical practice for some time now, but the exact benefits of such practice have been difficult to quantify. In a before-after study in three academic centers, Boyle and colleagues demonstrated a striking reduction of mortality in those patient groups where the standardized mortality ratio was above 1.5. Interestingly, they found no association between outcomes and the telecritical care support in the higher performing patient groups. They argue that the focused deployment of telecritical care to those groups who were at higher risk enabled more focused nursing and physician involvement from the remote team. The use of prone positioning has become more standard practice in severe hypoxemia and ARDS during the COVID-19 pandemic. In general, it is associated with better outcomes. However, the mechanism by prone positioning can exert its beneficial effects is unclear. Many practitioners also noted little change in oxygenation or ventilators following proning, indicating a possibility of a heterogeneous treatment effect. In an elegant study by Morais and colleagues, 22 ARDS patients with COVID-19 were investigated using electric impedance tomography and an esophageal manometry to see if the changes of global respiratory compliance and regional ventilation during proning and at various levels of PEEP can be established. Their results indicate that proning does not homogenize the ventilation distribution in the lung; rather, ventilation is regionally redistributed according to lung geometry along a gravitational gradient. They have also found significant individual heterogeneity in the response to proning, which can only be monitored with more advanced techniques such as electric impedance tomography. Please take time to read this article and more by visiting ccejournal.org ccejournal.org Also, follow us on social media at @critcareexplore. Stay safe, everybody. Goodbye.
