(1216) Variation in Tracheostomy Practices Before and After Pandemic: Impact on Outcomes and Resource Use

Introduction: The timing of tracheostomy remains controversial and thereby highly variable. Recent experience during the Covid-19 pandemic exacerbated these differences. To inform these discussions, we report on our single-center experience during the years before, during, and after the pandemic. We hypothesize that early tracheostomy ( < 10 days after intubation) is associated with better outcomes

Methods: With IRB approval, we performed a retrospective, observational study of patients who underwent tracheostomy at a single community hospital. We analyzed EMR data that included ICD billing codes for percutaneous tracheostomy performed by an intensivist, hospital and ICU length-of-stay (LOS), ventilator days, hospital charges and final discharge status. Three time periods were identified for analysis relative to the pandemic: before (1/18-2/20), during (3/20-12/21), and after (1/22-4/23). The primary outcome was hospital mortality. Data are expressed as the mean +/- SEM. P-values refer to t-statistics. Statistical analysis was performed using R.

Results: 210 adult patients underwent the procedure across all periods. The average ICU and hospital LOS, mortality rates (20.4 vs 34.9%), and ventilator days were each lower in the “early trach” group compared to the “later trach” group (p < 0.05). Similarly, the percentage of patients discharged home was higher with early tracheostomy (p < 0.05). The average days from intubation to tracheostomy varied by period: 12.6 (+/- 1.5), 13.5 (+/- 0.7) and 6.6 (+/- 0.6) days, respectively (p < 0.01 for the difference between “before + during” vs “after”). The ICU LOS for patients who underwent early (vs late) tracheostomy were significantly lower in each period: 14.3 vs 21.1 days, 15.6 vs 20.5 days, and 9.6 vs 16.7 days, respectively (p < 0.01 in each period).

Conclusions: Compared to later tracheostomy, early tracheostomy ( < 10 days) was associated with better results for several important value-metrics, irrespective of the pandemic. These metrics include decreased ventilator days, decreased LOS and mortality rates (ICU and hospital), and increased home discharge rate. The feasibility of performing these analyses real-time using a contemporary data platform should be pursued