B104-24 Tele-ICU Implementation and Mortality Outcomes in Safety-net Hospitals: A Multi-site Pre-post Study

Rationale

Tele-Intensive Care Unit (Tele-ICU) programs have emerged as a strategy to extend intensivist expertise to hospitals facing critical care workforce shortages and high patient acuity. Safety-net hospitals, in particular, often care for medically complex populations with limited access to specialty staffing, making Tele-ICU a potentially high-impact intervention. However, prior studies have shown mixed effects on patient outcomes, and there remains limited evidence from U.S.-based, multi-site implementations in safety-net settings. Understanding the real-world impact of Tele-ICU programs in these environments is essential to inform scalable critical care delivery models.

Methods

We conducted a retrospective pre-post observational cohort study of adult ICU admissions across four Los Angeles safety-net hospitals during the one-year period before and one-year period after Tele-ICU implementation. Patient-level administrative and clinical data were used to characterize demographics, diagnostic categories derived from ICD-10 codes, mechanical ventilation utilization, ICU length of stay, and mortality outcomes. Baseline characteristics and outcomes were compared between pre- and post-implementation cohorts using univariate statistical tests appropriate to variable type. Multivariable logistic regression models were used to evaluate adjusted ICU mortality, accounting for study period and facility-level differences. All analyses were descriptive and comparative, without causal inference.

Results

A total of 5,737 ICU admissions were included in the analysis. Baseline patient characteristics were similar between pre- and post-Tele-ICU cohorts, with no significant differences in mean age (66.7 vs 67.6 years, p = 0.37), sex distribution (female 19.9% vs 19.8%, p = 0.98), or major diagnostic categories including respiratory failure, sepsis, cardiac disease, neurologic disorders, and renal disease. Unadjusted ICU mortality (19.3% pre vs 20.2% post, p = 0.41), hospital mortality (41.9% vs 44.4%, p = 0.22), ventilator days (5.73 vs 5.66 days, p = 0.80), and ICU length of stay (7.93 vs 8.10 days, p = 0.19) did not differ significantly between periods. In adjusted analyses, Tele-ICU implementation was associated with lower ICU mortality (odds ratio 0.72, p = 0.01), with evidence of variation by facility and study period.

Conclusions

In this large multi-site study of safety-net hospitals, Tele-ICU implementation was associated with improved adjusted ICU mortality, while length of stay and ventilator utilization remained unchanged. Observed heterogeneity by facility and time underscores the importance of contextual factors, including seasonal variability and implementation differences, in evaluating Tele-ICU effectiveness. These findings highlight both the potential benefits of Tele-ICU programs and the complexity of measuring their impact in real-world safety-net settings.

This abstract is funded by: None