Researchers at one U.S. hospital used machine learning to develop a model that predicted which inpatients with COVID-19 would deteriorate. They validated it in 9,291 patients at 13 medical centers, with results published by The BMJ on Feb. 17. The model used nine variables (age, respiratory rate, oxygen saturation, oxygen flow rate, pulse oximetry type, head-of-bed position, position of patient during blood pressure measurement, venous blood gas pH, and partial pressure of carbon dioxide in arterial blood) to predict a composite outcome of in-hospital mortality or need for ventilation, heated high-flow nasal cannula, or IV vasopressors. The model had an area under the receiver operating characteristic curve of 0.80 (95% CI, 0.77 to 0.84), and the study authors calculated that using it to triage low-risk patients could potentially save up to 7.8 bed days per patient. The model could be integrated into the electronic health record as an indicator of the patient's condition or as a trigger to summon rapid response, they said. An accompanying editorial cautioned that it's uncertain how clinicians would react to such an alert and that while it's laudable that the algorithm is open sourced, its use is limited to hospitals that have Epic and are “resource rich enough to employ bioinformaticians.”
The organization CHEST updated its recommendations on the optimal approach to thromboprophylaxis in hospitalized patients with COVID-19. According to the society's expert panel, critically ill patients should receive standard thromboprophylaxis for venous thromboembolism, while moderately ill patients with a low bleeding risk might benefit from therapeutic heparin. The experts saw no role for intermediate-dose thromboprophylaxis in either setting. The update, which includes discussion of the evidence and the experts' thoughts on it, was published by CHEST on Feb. 12. The Infectious Diseases Society of America (IDSA) also updated its COVID-19 guidelines on Feb. 16. For patients (ambulatory or hospitalized) with mild-to-moderate COVID-19 at high risk for progression to severe disease, the IDSA guideline panel suggests remdesivir initiated within seven days of symptom onset rather than no remdesivir.
In other COVID-19 research, a randomized trial in Malaysia of high-risk patients with mild-moderate COVID-19 found no effect of ivermectin on risk of progression to severe disease, mechanical ventilation, ICU admission, or 28-day in-hospital mortality, according to results published by JAMA Internal Medicine on Feb. 18. A Canadian study, published by JAMA on Feb. 17, matched 9,087 omicron cases with an equal number of delta cases and found omicron to be associated with reduced risk for hospitalization or death, ICU admission or death, or death (hazard ratios of 0.41, 0.19, and 0.12, respectively). However, residence in a socially vulnerable ZIP code was associated with higher risk of ICU admission, mechanical ventilation, organ dysfunction or failure, and in-hospital death, according to a Michigan analysis of hospitalized COVID-19 patients published by Annals of Internal Medicine on Feb. 22. Finally, MMWR published two studies on Feb. 18 that investigated risk of COVID-19 spread during conventions, looking at a cluster of infections and the effectiveness of preventive measures.