Impact of Tele-Intensive Care Units on the Clinical Outcomes of Critically Ill Patients With COVID-19: Retrospective Cohort Study

Background The COVID-19 pandemic imposed an unprecedented demand for intensive care unit (ICU) resources in Brazil, where shortages of trained intensivists prompted the implementation of telemedicine-based critical care support strategies. Objective This study aimed to evaluate the association between adherence to the Tele-ICU COVID-19 Brazil Program and clinical outcomes of ICU patients with COVID-19. Methods We conducted a retrospective cohort study including all ICUs participating in the Tele-ICU COVID-19 Brazil Program between April and December 2020. Program adherence was assessed at 2 levels: patient coverage, defined as the number of daily multidisciplinary rounds per patient divided by the patient’s total ICU length of stay (LOS), and ICU coverage, defined as the number of daily multidisciplinary round days in the ICU divided by the total number of patient-days in that ICU. We compared outcomes between groups categorized by an empirically defined 50% cutoff: low patient coverage (<50%) versus high patient coverage (≥50%) and low ICU coverage (<50%) versus high ICU coverage (≥50%). Multilevel mixed-effects models accounting for ICU-level clustering were used to assess outcomes: logistic regression for ICU mortality (adjusted odds ratios) and linear mixed-effects regression with log-transformed ICU LOS (exponentiated coefficients, expβ). Results A total of 1680 patients were included. Compared with the low patient coverage group (<50%), patients in the high patient coverage (≥50%) had lower Sequential Organ Failure Assessment scores (median 2, IQR 0-5 vs median 3, IQR 0-6; P=.007); shorter ICU LOS (median 6, IQR 3-11 days vs median 11, IQR 6-20 days; P<.001); and shorter hospital LOS (median 9, IQR 5-16 days vs median 14, IQR 8-26 days; P<.001). In unadjusted analyses, ICU mortality did not differ significantly between the low and high patient coverage groups (50.1% vs 46.3%; P=.16). In multilevel analysis, mechanical ventilation and vasopressor use were independently associated with higher ICU mortality. Higher patient coverage was independently associated with lower ICU mortality (adjusted odds ratio 0.52, 95% CI 0.27-0.99; P=.048). In the log-transformed mixed-effects model for ICU LOS, a higher Sequential Organ Failure Assessment score (expβ 1.037, 95% CI 1.02-1.05; P<.001) and use of mechanical ventilation (expβ 1.23, 95% CI 1.05-1.43; P=.01) were associated with longer ICU LOS, whereas higher patient coverage was independently associated with shorter ICU LOS (expβ 0.17, 95% CI 0.13-0.21; P<.001). ICU coverage was not independently associated with ICU mortality or ICU LOS. Conclusions Greater patient-level coverage by remote intensivist–led multidisciplinary rounds within the Tele-ICU program was independently associated with lower ICU mortality and shorter ICU LOS. These findings support the potential contribution of tele–critical care strategies to expanding specialist support during public health emergencies.