Abstract Rationale Over the past two decades, Telecritical Care (TCC) programs have expanded in number and scope. While studies have shown that TCC programs reduce mortality and length of stay (LOS), these effects have been inconsistent. This heterogeneity may be due to differences in program design, recipient ICU characteristics, or interactions between TCC programs and recipient ICUs. Methods We examined the effect of specific TCC program workflows on ICU patient outcomes in a large integrated health care network in Northern California, including 14 ICUs and 331 ICU beds between July 2022 and December 2024. Data for all calls from recipient ICUs to the TCC program were recorded in a database. Additionally, we used a custom system of Best Practice Advisories (BPAs) to proactively monitor patients at risk of deterioration. We fit generalized linear models to estimate whether the 30-day moving average of calls per patient or BPAs per 10 patients by facility is associated with differences in mortality, LOS, readmission, or discharge disposition. The models were adjusted for age, secular time, patient comorbidity (COPS2), clinical acuity (LAPS2), IV vasopressor use, COVID status, mechanical ventilation, nighttime ICU admission, and number of ICU stays. We used a binomial distribution with a logit link for binary outcomes and a normal distribution with identity link function for continuous outcomes. Results Data from 37,641 ICU encounters at 14 TCC-covered facilities was examined. A facility-level increase of one call per patient per day facility average was associated with a 0.99 day decrease in hospital LOS (CI -1.31, -0.67, p.001) and a 0.54 day decrease in ICU LOS (CI -0.67, -0.40, p 0.001). Similarly, a facility-level increase of one BPA per 10 patients was associated with a 0.25 day decrease in ICU LOS (CI -0.32, -0.18, p 0.001) and a 0.18 day decrease in Hospital LOS (CI -0.35, -0.02, p = 0.03). Facility-level call volume and frequency of BPAs average did not have a significant effect on mortality, readmission, or discharge disposition. In a combined model of both call volume and BPAs, the associated decrease in LOS generally persisted with a diminished effect size. Conclusion In this study of ICUs covered by TCC programs in a large integrated health network, an increased frequency of interaction by facilities with the TCC program in the form of increased calls or proactive monitoring BPAs was associated with a reduction in LOS. This abstract is funded by: None
Tom et al. (Fri,) studied this question.