Analysis of Mortality and Length of Stay Associated With Implementation of a Large Telecritical Care Program in an Integrated Healthcare System.

Abstract

Multiple studies of telecritical care (TCC) programs have demonstrated an association between TCC implementation and reduced mortality, as well as decreased hospital and ICU length of stay (LOS), but these effects have been heterogeneous. We evaluated the impact of implementing a TCC program on inpatient mortality and LOS and examined heterogeneity in outcomes by acute severity of illness and chronic comorbidity groups. Retrospective study of adult ICU patients with and without TCC program exposure between January 1, 2015, and April 30, 2024. Eighteen medical surgical units in an integrated healthcare system in Northern California before and after implementation of an overnight TCC program. A total of 172,878 patient encounters (137,457 pre-TCC and 35,421 post-TCC implementation). No intervention. We used generalized linear models to estimate adjusted relative risk (ARR) of inpatient mortality and differences in hospital and ICU LOS comparing TCC exposed and unexposed groups. We performed mortality-adjusted LOS analysis by setting decedents' LOS to the 99th percentile of non-TCC encounters' LOS. Unadjusted analyses showed that the TCC exposed group had higher inpatient mortality (17.8 vs. 14.3% TCC vs. non-TCC), mean acuity indices, rates of COVID-19, and mechanical ventilation despite lower mean Comorbidity Point Score, Version 2, at admission. After adjustment for patient demographics, acuity, and comorbidity, TCC exposure was associated with decreased inpatient mortality with an ARR of 0.93 (95% CI, 0.89-0.98). The mortality-adjusted ICU and hospital LOS were not statistically significant in the TCC exposed group, with a difference of -0.07 days ICU LOS (95% CI, -0.18 to 0.05 d) and -0.10 days hospital LOS (95% CI, -0.35 to 0.15 d). Multihospital implementation of night coverage by a TCC program was associated with a reduced adjusted risk of inpatient mortality with no significant differences in mortality-adjusted ICU or hospital LOS.