Fig 1.
Triage algorithm for critical care initiation under resource scarcity due to COVID-19.
Summary of the first step (day 0, critical care initiation) of the SFAR/SSA critical care prioritization/triage protocol, adapted from [17] with proposed substitution of “age ≥ 85 & at least one comorbidity” to “age ≥ 85” alone.
Fig 2.
Triage algorithm for critical care continuation under resource scarcity due to COVID-19.
Summary of the second step (day 7–10 or typical disease turning point, critical care continuation) of the SFAR/SSA critical care prioritization/triage protocol, adapted from [17]. * Note: Initial criteria to withhold critical care may have been unknown due to missing information. They should be reassessed in view of the updated level of resource scarity.
Table 1.
Baseline characteristics and comparison of survivors and non-survivors.
Table 2.
Patient severity and outcome by priority levels (day 0, saturation).
Table 3.
Patient severity and outcome by priority levels (day 7–10, saturation).
Fig 3.
Outcome of COVID-19 ICU patients by priority level in saturation: P4 vs. others.
Cumulative incidence (c.i.) of alive discharge from ICU and survival (= 1 –c.i. of death in ICU) for COVID-19 patients: P4 compared with other priority levels at day 0 (A), and at reassessment on day 7 to 10 (B). Shaded area: initial prioritization no longer relevant due to reassessment.
Fig 4.
Outcome of COVID-19 ICU patients by priority level in saturation: All priority levels.
Cumulative incidence (c.i.) of alive discharge from ICU and survival (= 1 –c.i. of death in ICU) for COVID-19 patients: comparison between all priority levels at day 0 (A) and at reassessment on day 7 to 10 (B). Shaded area: initial prioritization no longer relevant due to reassessment.
Fig 5.
Raw length of ICU stay by priority level in saturation.
Length of ICU stay, irrespective of patient outcome, compared between all priority levels at day 0 (A, N = 225) and at reassessment on day 7 to 10 (B, N = 151). Boxes: median, 1st and 3rd quartiles; whiskers: Tukey’s convention (farthest points within 1.5 x IQR distance from box).
Fig 6.
Length of mechanical ventilation by priority level in saturation.
Length of mechanical ventilation, irrespective of patient outcome, compared between all priority levels at day 0 (A, N = 225) and at reassessment on day 7 to 10 (B, N = 151). Boxes: median, 1st and 3rd quartiles; whiskers: Tukey’s convention (farthest points within 1.5 x IQR distance from box).
Fig 7.
SAPS2 distribution by initial priority level in saturation.
SAPS2 compared between all priority levels at day 0 (N = 225). Boxes: median, 1st and 3rd quartiles; whiskers: Tukey’s convention (farthest points within 1.5 x IQR distance from box).
Fig 8.
Outcome of COVID-19 ICU patients by SAPS2 quartile.
Cumulative incidence (c.i.) of alive discharge from ICU and survival (= 1 –c.i. of death in ICU) for COVID-19 patients: comparison between SAPS2 quartiles.
Fig 9.
Age distribution by initial priority level in saturation.
Age compared between all priority levels at day 0 (N = 225). Boxes: median, 1st and 3rd quartiles; whiskers: Tukey’s convention (farthest points within 1.5 x IQR distance from box).
Fig 10.
Outcome of COVID-19 ICU patients by age quartile.
Cumulative incidence (c.i.) of alive discharge from ICU and survival (= 1 –c.i. of death in ICU) for COVID-19 patients: comparison between age quartiles.
Fig 11.
Initial severity, age and outcome compared between investigation centers.
Inter-center variations in SAPS2 (A), age (B), and ICU outcome (C). Boxplots: box–median, 1st and 3rd quartiles; whiskers–Tukey’s convention, farthest points within 1.5 x IQR distance from box. C: cumulative incidence (c.i.) of alive discharge from ICU and survival (= 1 –c.i. of death in ICU).