Investigators from more than 65 sites across the country, led by a team from Brigham and Women’s Hospital, conducted a multicenter cohort examination of COVID-19 patients admitted to intensive care units, finding that 35 percent of patients died in the 28 days after ICU admission. They also found that treatment and outcomes varied greatly between hospitals, according to a press release from Brigham and Women’s Hospital. The results of their work are published in JAMA Internal Medicine.
The multicenter cohort study included 65 sites from the Northeast, South, Midwest and West regions of the U.S., including parts of the U.S. that were heavily affected by COVID-19. The study included 2,215 adults with laboratory-confirmed COVID-19 who were admitted to ICUs between March 4 and April 4, 2020.
Overall, 784 patients (35 percent) died within 28 days, with wide variation among hospitals. Factors independently associated with death included older age, male sex, higher body mass index, coronary artery disease, active cancer, and the presence of low oxygen levels, liver dysfunction, and kidney dysfunction at the time of ICU admission.
Even after adjusting for a variety of risk factors, death rates varied widely across hospitals, from 6 percent to 80 percent. The number of pre-COVID ICU beds in the hospital was strongly associated with death rate. Patients admitted to hospitals with less than 50 ICU beds had a more than three-fold higher risk of death than patients admitted to hospitals with 100 or more ICU beds.
In addition, hospitals varied widely in the proportion of patients who received medications and supportive therapy for COVID-19. During the time period studied, hydroxychloroquine, azithromycin, and anticoagulants were commonly prescribed, and interventions such as prone positioning were also being implemented. But the proportion of patients receiving these measures varied considerably – for instance, the use of prone positioning ranged from 4 percent of patients at one hospital to 80 percent at another.
While the team adjusted for a large number of demographic and severity of illness characteristics, its estimates of differences in death rates across hospitals may be impacted by other confounders, such as the socioeconomic status of patients — a risk factor increasingly recognized as important in health outcomes for COVID-19 patients. The team’s models also do not account for varying degrees of strain across hospitals.
